Site-specific modelling of turbulent fluxes on the Tibetan Plateau

A dissertation submitted to the Faculty of Biology, Chemistry and Geosciences University of Bayreuth to attain the academic degree of Dr. rer. nat. Site-specific modelling of turbulent fluxes on the Tibetan Plateau Wolfgang Babel M.Sc. Global Change Ecology born 3 November, 1974 in Pfronten, Germany Bayreuth, March 2013 supervised by Prof. Dr. Thomas Foken Site-specific modelling of turbulent fluxes on the Tibetan Plateau supervised by Prof. Dr. Thomas Foken i Die vorliegende Arbeit wurde in der Zeit von April 2009 bis M¨arz 2013 in Bayreuth an der Abteilung Mikrometeorologie unter Betreuung von Herrn Prof. Dr. Thomas Foken angefertigt. Vollst¨andiger Abdruck der von der Fakult¨at f¨ur Biologie, Chemie und Geowissenschaften der Universit¨at Bayreuth genehmigten Dissertation zur Erlangung des akademischen Grades eines Doktor der Naturwissenschaften (Dr. rer. nat.). Dissertation eingereicht am: 6. M¨arz 2013 Zulassung durch die Pr¨ufungskommission: 13. M¨arz 2013 Wissenschaftliches Kolloquium: 15. Mai 2013 Amtierende Dekanin: Prof. Dr. Beate Lohnert Pr¨ufungsausschuss: Prof. Dr. Thomas Foken (Erstgutachter) Dr. habil. Eva Falge (Zweitgutachter) Prof. Dr. Andreas Held (Vorsitz) Prof. Dr. Michael Hauhs Prof. Dr. Bernd Huwe ii Contents List of manuscripts v List of additional publications vi Acknowledgements viii Summary ix Zusammenfassung xi 1. Introduction 1 1.1. Motivation: Regional estimates of turbulent fluxes on the Tibetan Plateau 1 1.2. My contribution to the project objectives . . . . . . . . . . . . . . . . . 2 1.3. Objectives of the thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 2. Background 7 2.1. Observed energy balance and closure . . . . . . . . . . . . . . . . . . . 7 2.2. Land surface modelling on the Tibetan Plateau . . . . . . . . . . . . . 8 2.3. Lake surface modelling on the Tibetan Plateau . . . . . . . . . . . . . . 9 3. Methods 11 3.1. The Nam Co 2009 experiment . . . . . . . . . . . . . . . . . . . . . . . 11 3.1.1. Sitedescription........................... 11 3.1.2. Measurements............................ 11 3.1.3. Data post-processing . . . . . . . . . . . . . . . . . . . . . . . . 12 3.1.4. Energy balance closure and correction . . . . . . . . . . . . . . 14 3.2. Land surface modelling for Nam Co 2009 . . . . . . . . . . . . . . . . . 16 3.2.1. Modelversions ........................... 16 3.2.2. Model parameters . . . . . . . . . . . . . . . . . . . . . . . . . . 19 3.3. Lake surface modelling for Nam Co 2009 . . . . . . . . . . . . . . . . . 21 4. Results 23 4.1. Data quality on the Tibetan Plateau . . . . . . . . . . . . . . . . . . . 23 4.2. Flux measurements at Nam Co . . . . . . . . . . . . . . . . . . . . . . 24 4.3. Land surface modelling at Nam Co . . . . . . . . . . . . . . . . . . . . 25 iii 4.4. Influence of the energy balance correction method . . . . . . . . . . . . 27 4.5. Lake surface modelling at Nam Co . . . . . . . . . . . . . . . . . . . . 30 4.6. Flux heterogeneity at Nam Co . . . . . . . . . . . . . . . . . . . . . . . 32 5. Conclusions 37 References 41 A. Individual contributions to the joint publications 54 B. Gerken et al. (2012) 58 C. Babel et al. (2013) 75 D. Biermann et al. (2013) 88 E. Charuchittipan et al. (2013) 117 F. Li et al. (2013) 146 Erkl¨arung 156 iv List of manuscripts This dissertation is presented in a cumulative form. It is based on three peer-reviewed publications and two submitted manuscripts as listed below. Peer-reviewed publications Biermann, T., Babel, W., Ma, W., Chen, X., Thiem, E., Ma, Y., and Foken, T.: Turbulent flux observations and modelling over a shallow lake and a wet grassland in the Nam Co basin, Tibetan Plateau, Theor. Appl. Climatol., accepted Gerken, T., Babel, W., Hoffmann, A., Biermann, T., Herzog, M., Friend, A. D., Li, M., Ma, Y., Foken, T., and Graf, H.-F.: Turbulent flux modelling with a simple 2-layer soil model and extrapolated surface temperature applied at Nam Co Lake basin on the Tibetan Plateau, Hydrol. Earth Syst. Sci., 16, 1095–1110, doi:10.5194/hess-16-1095-2012, 2012. Li, M., Babel, W., Tanaka, K., and Foken, T.: Note on the application of planar-fit rotation for non-omnidirectional sonic anemometers, Atmos. Meas. Tech., 6, 221–229, doi:10.5194/amt-6-221-2013, 2013. Manuscripts submitted Babel, W., Chen, Y., Biermann, T., Yang, K., Ma, Y., and Foken, T.: Adaptation of a land surface scheme for modeling turbulent fluxes on the Tibetan Plateau under different soil moisture conditions, submitted to J. Geophys. Res. Charuchittipan, D., Babel, W., Mauder, M., Leps, J.-P., and Foken, T.: Extension of the averaging time of the eddy-covariance measurement and its effect on the energy balance closure, submitted to Bound.-Lay. Meteorol. v List of additional publications The following list summarises other publications of mine which are not included in the dissertation. They were split up in two parts: Publications which have reference to the thesis and other publications. The first part includes two master theses as well, with reference to the dissertation, initiated and supervised by myself. Publications with reference to this thesis Peer-reviewed publications Zhou, D., Eigenmann, R., Babel, W., Foken, T., and Ma, Y.: The study of nearground free convection conditions at Nam Co station on the Tibetan Plateau, Theor. Appl. Climatol., 105, 217–228, doi:10.1007/s00704-010-0393-5, 2011. Non peer-reviewed publications Babel, W., Eigenmann, R., Ma, Y., and Foken, T.: Analysis of turbulent fluxes and their representativeness for the interaction between the atmospheric boundary layer and the underlying surface on Tibetan Plateau, CEOP-AEGIS Deliverable report De1.2, Ed. University of Strasbourg, France, ISSN 2118-7843: 35 p., 2011a. Babel, W., Li, M., Sun, F., Ma, W., Chen, X., Colin, J., Ma, Y., and Foken, T.: Aerodynamic and thermodynamic variables for four stations on Tibetan Plateau – Introduction on the CEOP-AEGIS database in NetCDF, CEOP-AEGIS Deliverable report De1.3, Ed. University of Strasbourg, France, ISSN 2118-7843: 90 p., 2011b. Babel, W.: An R routine for the simplified usage of TERRAFEX (Characterisation of a complex measuring site for flux measurements), Work Report University of Bayreuth, Dept. of Micrometeorology, ISSN 1614-8916, in preparation Biermann, T., Babel, W., Olesch, J., and Foken, T.: Documentation of the Micrometeorological Experiment, Nam Tso, Tibet, 25th of June – 8th of August 2009, Work Report University of Bayreuth, Dept. of Micrometeorology, ISSN 1614-8916, 41, 38 pp., URL http://opus.ub.uni-bayreuth.de/opus4-ubbayreuth/frontdoor/ index/index/docId/626, 2009. Foken, T. and Babel, W.: Auf dem “Dach der Welt” – Die Rolle Tibets bei der Wasserversorgung S¨udostasiens, in: Spektrum-Magazin der Universit¨at Bayreuth, Ausgabe 1/2012, 46–48, 2012 vi Gerken, T., Babel, W., Hoffmann, A., Biermann, T., Herzog, M., Friend, A. D., Li, M., Ma, Y., Foken, T., and Graf, H.-F.: Turbulent flux modelling with a simple 2-layer soil model and extrapolated surface temperature applied at Nam Co Lake basin on the Tibetan Plateau, Hydrol. Earth Syst. Sci. Discuss., 8, 10 275–10 309, doi:10.5194/hessd-8-10275-2011, 2011. Gerken, T., Fuchs, K., and Babel, W.: Documentation of the Atmospheric Boundary Layer Experiment, Nam Tso, Tibet, 08th of July – 08th of August 2012, Work Report University of Bayreuth, Dept. of Micrometeorology, ISSN 1614-8916, 53, 48pp., 2013 Li, M., Babel, W., Tanaka, K., and Foken, T.: Note on the application of planarfit rotation for non-omnidirectional sonic anemometers, Atmos. Meas. Tech. Discuss., 5, 7323–7340, doi:10.5194/amtd-5-7323-2012, 2012. Master theses supervised by myself Baumer, M.: Vergleich zweier Lagrange’scher Modelle zur Bestimmung des Footprints ¨uber heterogenem Gel¨ande, Master thesis, University of Bayreuth, 66pp., 2012. Thiem, E.: Modelling of the energy exchange above lake and land surfaces, Master thesis, University of Bayreuth, 73pp., 2011. Other publications not included in this thesis Irl, S. D. H., Steinbauer, M. J., Babel, W., Beierkuhnlein, C., Blume-Werry, G., Messinger, J., Palomares Mart´ınez, ´ A., Strohmeier, S., and Jentsch, A.: An 11yr exclosure experiment in a high-elevation island ecosystem: introduced herbivore impact on shrub species richness, seedling recruitment and population dynamics, J. Veg. Sci., 23, 1114–1125, doi:10.1111/j.1654-1103.2012.01425.x, 2012. Babel, W., Huneke, S., and Foken, T.: A framework to utilize turbulent flux measurements for mesoscale models and remote sensing applications, Hydrol. Earth Syst. Sci. Discuss., 8, 5165–5225, doi:10.5194/hessd-8-5165-2011, 2011. Babel, W. and Foken, T.: Preliminary footprint analysis of LAS (Large aperture scintillometer) measurements at Qomolangma station, Tibetan Plateau, CEOPAEGIS technical report, 2009, available at the CEOP-AEGIS project office, but not published yet vii 1. Introduction Table 1.1. Permanent measurement sites on the Tibetan Plateau as selected for the CEOP-AEGIS project Site Coordinates Altitude Land cover Naqu (BJ) 31◦220700N 91◦5305500E 4502 m Alpine steppe Nam Co 30◦4602200N 90◦5704700E 4745 m Alpine steppe Linzhi 29◦4505600N 94◦4401800E 3327 m Alpine grassland Qomolangma 28◦2102900N 86◦5604700E 4293 m Gravel of such regional flux maps should be strengthened , but an adequate processing and quality control, as summarised by Rebmann et al. (2012), Foken et al. (2012) and demonstrated on the TP e.g. by Metzger et al. (2006), is a prerequisite. 1.2. My contribution to the project objectives My work of the last years is mainly related to the project: “Coordinated Asia-European long-term Observing system of Qinghai–Tibet Plateau hydrometeorological processes and the Asian-monsoon systEm with Ground satellite Image data and numerical Simulations (CEOP-AEGIS)”. It is a collaborative project / Small or medium-scale focused research project – Specific International Co-operation Action financed by the European Commission under FP7 topic ENV.2007.4.1.4.2 ”Improving observing systems for water resource management”, and is coordinated by the University of Strasbourg, France (www.ceop-aegis.org). Amongst other goals it aims at constructing an observing system to monitor the plateau’s water yield by a combination of ground measurements and satellite based observations. This is of immediate interest for water resources management in South-East Asia. The anticipated outcome is providing infrastructure, i.e. an interactive data portal will be delivered, featuring a three-year data set with observations of the water balance terms. Together with higher level products and distributed hydrological modelling a prototype monitoring system is formed including early warning on floods and droughts which is intended to remain in operation beyond project completion. Within this infrastructural project I was responsible for delivering quality checked ground based observations of surface energy fluxes. The data set incorporates fluxes over three years at four stations on the Tibetan Plateau (see Table 1.1), the turbulent fluxes (sensible heat and latent heat/evapotranspiration) were derived by the eddy-covariance method. Based on previous studies about quality assurance of eddycovariance measurements in the community and especially in the Department of Mi2 1.2. My contribution to the project objectives crometeorology, University of Bayreuth, I myself compiled a work flow for the processing of eddy-covariance data in three levels. These are: Level I Turbulent fluxes, being checked for data quality, footprint, and potential obstacles in the vicinity of the sensor. Level II Level I data, corrected with respect to the energy balance closure. Level III Gap-filled level II data. In the course of this I adapted respective tools for footprint analysis developed by Mathias G¨ockede (G¨ockede et al., 2005, 2006, 2008) and merged them into a single, user-optimised R routine (Work report in prep.). The scheme and its implementation up to the creation of NetCDF data sets is described in CEOP-AEGIS technical reports (Babel et al., 2011a,b). As it was not possible to get access to the Chinese original data, I introduced Chinese colleagues to the methods and supervised the processing at all stages. In general, the presented workload could be straightforwardly handled consulting the relevant literature and executing the necessary steps. Nevertheless, the special conditions on the Tibetan Plateau mentioned in Sect. 1.1 raise the need for investigations beyond existing studies. Gap-filling of turbulent fluxes for Level III aims at representing the system dynamics adequately for this unique environment rather than delivering correct annual sums of evapotranspiration. This cannot be achieved with empirical or statistical approaches as presented by Falge et al. (2001a,b), but with Soil - Vegetation - Atmosphere - Transfer (SVAT) models. Therefore own turbulent flux measurements over different types of surfaces have been conducted in order to derive the essential input parameters as well as turbulent flux observations for validation (Biermann et al., 2009; Gerken et al., in prep.). This was necessary, because the access to Chinese data gathered on the Tibetan Plateau was very limited. The measurements, however, show a significant non-closure of the energy balance, which is distinctive of such heterogeneous environments (Foken, 2008a). Different options exist to close this gap (e.g. Foken et al., 2011) and their application clearly has an impact on the outcome of model validation. The aim of these efforts is to provide surface model solutions for significant land surface types on the Tibetan Plateau, which serve as a high-standard gap-filling tool and estimation of representativeness of the permanent flux stations within CEOP-AEGIS. Furthermore, inconsistencies in the turbulence measurements caused by specific sensor types had to be investigated in order to assess the effect of such problems on sensible and latent heat flux measurements. From these problems I defined my research tasks further described in the next section. 3 1. Introduction 1.3. Objectives of the thesis The main focus of this thesis is the application of process based modelling to estimate surface sensible heat fluxes and latent heat fluxes (evapotranspiration) in the specific environment of the Tibetan Plateau. Although several surface model studies already exist on the Tibetan Plateau (see Sect. 2.2), a thorough analysis including eddy-covariance measurements of turbulent fluxes is still missing. Moreover, there is a lack of evaporation measurements above lake surfaces. To my best knowledge, Tobias Biermann and me conducted the first eddy-covariance measurements over a lake surface on the Tibetan Plateau. This made it possible to validate a lake surface model. Based on these preconditions the following research questions have been elaborated. •Land surface modelling under the specific conditions of the Tibetan Plateau using the land surface scheme SEWAB (Mengelkamp et al., 1999) and validation with eddy-covariance measurements. •Elaboration of the necessary model parameters: required efforts and impact on model performance. •Investigation of different methods to correct for the energy balance closure gap and their influence on model performance assessment. •Potential application of the elaborated model version. •Specific problems of eddy-covariance data quality on the Tibetan Plateau and mitigation of specific measurement problems occurring with some sonic anemometers. The publications and manuscripts listed on page v contribute to these research questions as follows: Babel et al. (2013, Appendix C) present an adaptation of the land surface scheme SEWAB (Mengelkamp et al., 1999) to the Tibetan Plateau. The adaptation is designed to consider specific issues of land surface modelling as mentioned in Sect. 2.2. The model performance with respect to turbulent fluxes is investigated by using eddy-covariance measurements above alpine steppe from two nearby sites at Nam Co lake. Parameter sets originating from both standard values and laboratory and in situ measurements are tested in this manuscript. Special emphasis is put on the energy balance closure of turbulent flux measurements and its correction. Besides of using the well known method of distributing the residual according to the Bowen ratio (Twine et al., 2000), a new correction method, suggested by Charuchittipan et al. (2013, Appendix E), has been applied: The residual is distributed according to the relative contribution of the turbulent fluxes to the buoyancy flux. Charuchittipan et al. (2013, Appendix E) analyse a comprehensive data set from the LITFASS-2003 campaign, Lindenberg, Germany. The influence of the averaging time for eddy-covariance fluxes is intensely studied utilising Ogive analysis, block ensemble averaging, wavelet 4 1.3. Objectives of the thesis and quadrant analysis. The manuscript suggests secondary circulations to be responsible for the gap in energy balance closure and attributes a dominant role to the sensible heat flux. The proposed new closure correction method is based on these experimental findings. As a novel approach, both two correction methods have been applied to the data and the consequences on model performance evaluation is discussed by Babel et al. (2013, Appendix C). The SEWAB model successfully simulates turbulent fluxes on the Tibetan Plateau, creating several benefits: It can be as a reference for a more simplistic parametrisations as done by Gerken et al. (2012, Appendix B). Therein the simplistic land surface scheme Hybrid is updated with a new soil model, aiming to eliminate a delayed surface response to atmospheric forcing in the original version. SEWAB is successfully utilised for comparisons in example daily cycles. As SEWAB showed no delay in surface response, its simulations have been used to evaluate Hybrid’s responsiveness with cross correlation (Gerken et al., 2012, Appendix B). Another application is the usage of the simulated timeseries in order to assess landscape heterogeneity at Nam Co (Biermann et al., 2013, Appendix D). The gappy observations of a wet alpine steppe and a shallow lake could be effectively described by modelled timeseries of SEWAB and a hydrodynamic multilayer model (Foken, 1984) with an extension to shallow water exchange (Panin and Foken, 2005). Turbulent fluxes for both surface types are then compared with the SEWAB simulations of the “standard” land surface at Nam Co, dry alpine steppe. It could be shown that the differences among land surface types likely exceed the model uncertainty, so the differences are considerable. The effect of this heterogeneity is discussed in terms of using the eddy-covariance data as ground truth for remote sensing. In order to use eddy-covariance data for model evaluation some issues about data quality should be clarified in advance. Although not included in the thesis Zhou et al. (2011)1provide a basis for further usage of eddy-covariance data at the Nam Co Monitoring and Research Station for Multisphere Interactions: In addition to standard evaluation of footprint and data quality the occurrence of near-ground free convection events is investigated. It is shown that such events can be created on the Tibetan Plateau already due to changing cloudiness, and their influence on data quality is assessed. Furthermore, at one of the CEOP-AEGIS sites the sonic anemometer DAT 600 TR61A probe from Kaijo-Denki is in use. The sensor is not omnidirectional, i.e. in a certain sector the wind field is disturbed by the sensor structure and irregular friction velocities occur as a consequence. The study by Li et al. (2013, Appendix F) highlights this problem and investigates its influence on scalar fluxes and whether such problems occur also with the commonly used CSAT3, Campbell Scientific Ltd. A sector-wise planar-fit is suggested as an appropriate coordinate rotation to mitigate 1Together with Rafael Eigenmann I introduced Zhou Degang into the post-processing of the turbulence data, including the usage of TK2 and footprint analysis tools, and into the investigation and relevance of near ground free convection conditions at Nam Co station. I personally supported him in the usage and interpretation of the data quality tools. 5 1. Introduction such problems. 6 2. Background 2.1. Observed energy balance and closure The eddy-covariance method is the only direct method to measure turbulent exchange of heat and scalars between the atmosphere and the underlying surface and is therefore preferred in the community to quantify long-term fluxes of water vapour and carbon dioxide (Foken and Wichura, 1996; Baldocchi et al., 2001). Despite the general trust placed in this method, it became apparent that the energy balance cannot be closed at most experimental sites (e.g. Foken, 2008a). The surface energy balance at the surface is given by −Rnet =QH+QE+QG+ ∆QS(2.1) with the net radiation Rnet, the sensible heat flux QH, the latent heat flux QE, the ground heat flux QG, and the change in energy storage ∆QS. The signs follow the convention that fluxes directed towards the surface are negative and vice versa. Although this theoretical balance should be reproduced with measurements as well, according to Foken (2008a) most studies report that the sum of turbulent energy (QHand QE) only yields 70%–100% of the available energy (−Rnet −QG−∆QS). This residual is typically larger in complex landscapes (e.g. Aubinet et al., 2000; Wilson et al., 2002), while in homogeneous areas or deserts the energy balance can be closed (e.g. Heusinkveld et al., 2004; Mauder et al., 2007). It is recognized that in the measured energy balance turbulent energy is missing rather than available energy being overestimated so long as the energy storage and the ground heat flux has been addressed adequately (Twine et al., 2000; Foken, 2008a; Foken et al., 2011; Leuning et al., 2012). Especially when comparing eddy-covariance derived turbulent fluxes with those land surface model simulations, which are actually constrained by the energy balance equation, a systematic mismatch can be expected as pointed out by Falge et al. (2005). Therefore a correction of the turbulent fluxes should be considered. A widely used correction method proposed by Twine et al. (2000) distributes the residual according to the Bowen ratio, assuming scalar similarity between latent and sensible heat with respect to the missing flux. In contrast some studies indicate that the missing energy stems from sensible heat only (Mauder and Foken, 2006; Ingwersen et al., 2011). The recent discussion hypothesizes an influence of secondary circulations in complex landscapes, triggering near-ground advective and low-frequency flux components (Steinfeld et al., 2007; Foken et al., 2010, 2011; Stoy et al., 2013; Br¨otz et al., 2013). Such circulation systems are mainly driven by buoyancy 7 2. Background suggesting the buoyancy flux to play a dominant role for the energy balance closure. 2.2. Land surface modelling on the Tibetan Plateau Due to its remoteness, regional flux estimation on the Tibetan Plateau does not have a long history, see also Babel et al. (2013, Appendix C). First multi-year estimates across a variety of sites on the Tibetan Plateau have been presented by Xu and Haginoya (2001), calculating fluxes from standard meteorological measurements. Within the GEWEX Asian Monsoon Experiment 1998 Takayabu et al. (2001) utilized four land surface models for a comparison study, reported large differences in turbulent flux partitioning, but could not validate the models due to a lack of soil moisture and flux measurements. Only very few studies exist using eddy-covariance flux measurements for validation (e.g. Yang et al., 2009; Hong and Kim, 2010). These studies find an overestimation of the sensible heat flux as a typical feature of land surface modelling on the Tibetan Plateau. They blame too high turbulent diffusion coefficients for this problem an draw a relationship to the special conditions on the Tibetan Plateau (see Sect. 1.1), in particular the strong diurnal cycle of surface temperature over dry and sparsely vegetated surfaces. Indeed, Yang et al. (2003) and Ma et al. (2002) observed a diurnal variation of the sublayer-Stanton number B, describing the logarithm of the ratio between aerodynamic and thermal roughness lengths κB−1= ln(z0m z0h ). Formulations of κB−1as a fixed fraction, or depending on the friction velocity (Zilitinkevich, 1995) cannot resolve this diurnal variation. Therefore (Yang et al., 2008) propose a new formulation with an additional dependence on the temperature scale T∗and empirical adaptation to Tibetan Plateau observations (see Sect. 3.2.1). Leading to the determination of a variable thermal roughness length in land surface models, this parametrisation has been successfully rated in some recent studies in Asian arid regions Yang et al. (2008); Chen et al. (2010, 2011); Liu et al. (2012); Zhang (2012). Other developments include the influence of soil vertical heterogeneity which is found to be significant in case of remarkable stratification (Yang et al., 2005; van der Velde et al., 2009). For the soil heat flux on the Tibetan Plateau Yang et al. (2005) adapted a parametrisation for the soil thermal conductivity, proposed by Johansen (1975) and recommended by Peters-Lidard et al. (1998). In this form it can be easily transferred to any conditions when dry and saturated thermal conductivities are known. Furthermore, latent heat fluxes can be observed on the Tibetan Plateau even if soil moisture drops below wilting point, a typical feature in deserts or arid landscapes (Agam et al., 2004; Balsamo et al., 2011; Wallace et al., 1991). Different parametrisations of bare soil evaporation in dependence on soil moisture are compared by Mihailovi´c et al. (1995), but these are not tested on the Tibetan Plateau yet. Another challenge on the Tibetan Plateau is a small-scale heterogeneity of soil moisture (Su et al., 2011). Cold semiarid conditions characterise the landscape consisting of 8 2.3. Lake surface modelling on the Tibetan Plateau dry grasslands (Kobresia pastures, alpine steppe, partly non-vegetated) and wetlands or grasslands with shallow groundwater. The role of the soil moisture interacting with climate is highlighted in a review by Seneviratne et al. (2010). The land–atmosphere coupling strength (influence of soil moisture on precipitation) on the Tibetan Plateau, however, is rated low by Koster et al. (2004), but the local patterns of soil moisture and precipitation cannot be resolved by such studies using ensembles of global circulation models. Land surface models are in principle able to take small-scale heterogeneity of soil moisture into account, but this has to be tested. Obtaining flux measurements in such a remote environment as the Tibetan Plateau is challenging (Ma et al., 2009b) and land surface modelling can support land surface flux estimation on local and regional scale. A wide range of land surface schemes being suitable for such a task have been evolved in the last decades: Starting with simple schemes (e.g. Manabe, 1969), second-generation land surface models evolved, adding e.g. detailed resistance schemes for evapotranspiration and more complexity into the description of soil processes (Pitman, 2003). More recent developments for third-generation models mainly include a dynamic vegetation, the “greening” of land surface models (Pitman, 2003). In the context of land surface modelling this thesis lays the focus on how the model description of the special Tibetan Plateau conditions affect land surface fluxes (see Sect. 1.1). Thereby a crucial feature is the comparison with site-specific eddy-covariance measurements. As feedback mechanisms of the land surface to the atmosphere are outof-scope, the most suitable approach is an offline forced land surface model with prescribed, site-specific state of vegetation and soil properties. The used model SEWAB is a representative of the second-generation land surface models and participated in the Project for Intercomparison of Land-surface Parametrization Schemes (PILPS: Henderson-Sellers et al., 1996; Chen et al., 1997). From this experience it has been improved regarding the description of soil and surface processes (Mengelkamp et al., 1999, 2001; Warrach et al., 2001) and is therefore adequately structured for the intended purpose. Initially developed for humid conditions SEWAB’s performance under dry conditions has to be tested yet. 2.3. Lake surface modelling on the Tibetan Plateau Lake surfaces should be taken into account for regional flux estimation on the Tibetan Plateau as approximately 45 000 km2is covered by lakes (Xu et al., 2009). The importance of lake surfaces for the regional energy balance and water cycle has been pointed out by Rouse et al. (2005) and Nordbo et al. (2011). Some evaporation estimates already exist for lake surfaces, modelled with simple bulk approaches based on daily or monthly forcing data from remote sensing or surface observations (Haginoya et al., 2009; Xu et al., 2009; Krause et al., 2010; Yu et al., 2011), in some cases validated with pan evaporation measurements. To my best knowledge no studies are reported 9 2. Background yet, utilising or conducting eddy-covariance measurements above lake surfaces on the Tibetan Plateau. There is a huge amount of lakes on the Tibetan Plateau (≈1090 lakes larger than 10 km2, Yu et al., 2011), therefore a variety of lake extent and depth can be expected to occur. These factors greatly influence surface temperature (and thereby atmospheric stability) and surface roughness (Rouse et al., 2005; Panin et al., 2006a; Nordbo et al., 2011). These variables in turn typically exhibit a diurnal variation on the one hand and their relationship to the surface fluxes is non-linear. Therefore the processes can only be represented by resolving the diurnal cycle, which is not possible with the methods mentioned above. For such a purpose a hydrodynamic multilayer (HM) model (Foken, 1979, 1984) is a suitable candidate. It is originally designed for energy exchange above the ocean. Shallow water conditions, however, increase the wave height, depending on wind velocity, and therefore enhance the turbulent exchange (Panin and Foken, 2005). The HM model and the shallow water approach has been validated with eddy-covariance data from a lake in Germany and the impact on exchange over the Caspian Sea has been discussed (Panin et al., 2006b,a). Nevertheless, it has to be tested whether these parametrisations work under the conditions of the Tibetan Plateau as well. 10 3. Methods 3.1. The Nam Co 2009 experiment As pointed out in Sect. 1.2 own experiments were inevitable to derive necessary data for input, parametrisation and validation of surface models. The Nam Co site (see description in the following) has been chosen, as it is an area where typical land surfaces of the Tibetan Plateau (dry alpine steppe, more wet and dense grassland, and lakes) occur closely together. Located at the intersection of the Westerlies with the Asian Monsoon circulation systems the Nam Co basin has been considered as a key area of interest (Haginoya et al., 2009; Keil et al., 2010). 3.1.1. Site description The Nam Co 2009 experiment was carried out from 26 June to 8 August within the 2009 summer monsoon season. The site is located 220 km north of Lhasa in the Nam Co Basin, Tibetan Plateau, with its lake surface at an elevation of 4730 km a.s.l. The basin is dominated by the lake itself and the Nyainqentanglha mountain range, stretched along its SE side and reaching up to 7270 m a.s.l. with an average height of 5230 m (Liu et al., 2010). The Institute of Tibetan Plateau Research (ITP), Chinese Academy of Sciences is operating the Nam Co Monitoring and Research Station for Multisphere Interactions near a small lake in 1 km distance SE to the Nam Co Lake (see Figure 3.1). The vegetation around Nam Co reflects the prevailing arid, high-altitude climate with alpine meadows and steppe grasses (M¨ugler et al., 2010). Near the Nam Co Station the vegetation coverage and composition are highly variable according to the soil moisture conditions determined by topographic features: Grass genera typical for Alpine steppe (Stipa, Carex, Helictotrichon, Elymus, Festuca, Kobresia, Poa, see Biermann et al., 2009; Miehe et al., 2011) have been observed on the hillocks, with a total vegetation coverage of 60 % or less (grass−), while more wet areas are densely covered (>90 %) with alpine meadows dominated by Kobresia species (grass+, see Figure 3.1). 3.1.2. Measurements Turbulent fluxes were obtained by two energy balance systems. One set-up is located directly at the Nam Co station over dry alpine steppe (grass−), further called NamITP, operated by the Institute of Tibetan Plateau Research. The NamITP complex is settled 11 3. Methods •hydrological modules containing tunable parameters, which cannot be determined, are disabled (ponding, variable infiltration capacity, ARNO concept for subsurface runoff and baseflow, depth dependency parametrisation of saturated hydraulic conductivity, see Mengelkamp et al., 1999, 2001) •offline forcing with measured precipitation, air temperature, wind velocity, air pressure, relative humidity, downwelling short-wave and long-wave radiation using the same data for both grass+and grass−. •internal model time step of 10 min, interpolation of 30-min forcing data and aggregation of output to 30 min. •initialisation of soil moisture and soil temperature profiles with a 3-year forcing data set extracted from the ITPCAS (Institute of Tibetan Plateau Research, Chinese Academy of Sciences) gridded forcing data set (Chen et al., 2011). Test simulations showed reasonable simulations of soil moisture for the grass−surface, but could not be used for the grass+surface, as the shallow ground water table at NamUBT could not be reproduced with a single column realisation. •initialisation of soil moisture and soil temperature profiles with observed profiles. This initialisation showed good agreement with the 3-year spin-up at grass−, therefore it has been solely used for all analysis for both surface types. The adaptation to the Tibetan Plateau (TP version) aims at addressing the issues mentioned in Sect. 2.2. The changes include: 1. A new calculation of the soil thermal conductivity λsfollowing Yang et al. (2005) λs(Θ) = λdry + (λsat −λdry) exp [0.36 ·(1 −Θsat/Θ)] (3.6) with the volumetric soil water content Θ and Θsat as the porosity. The dry and saturated thermal conductivity limits were estimated from field observations as λdry = 0.15 W m−1K−1and λsat = 0.8 and 1.3 W m−1K−1for grass+and grass−, respectively. This parametrisation replaced the original formulation featuring a weighted sum of individual thermal conductivities of dry clay/sand, water, ice and air according to the actual state. 2. To account for diurnal and seasonal variations of the thermal roughness length observed on the Tibetan Plateau (Yang et al., 2003), a formulation according to Yang et al. (2008) has been implemented z0h =70ν u∗ exp −βu0.5 ∗|T∗|0.25(3.7) with the kinematic viscosity of air ν, the friction velocity u∗, the dynamic temperature scale T∗=−w0T0/u∗and an empirical constant β= 7.2 s0.5m−0.5K−0.25. As 18 3.2. Land surface modelling for Nam Co 2009 T∗depends on z0h, the equation has to be solved iteratively (Yang et al., 2010). The original formulation estimates z0h as a fixed fraction of the aerodynamic roughness length z0h = 0.1z0m. 3. Like observed in desert landscapes (Agam et al., 2004; Balsamo et al., 2011; Wallace et al., 1991), latent heat fluxes occur on the Tibetan Plateau even when soil moisture drops below wilting point. The soil air humidity controlling bare soil evaporation is adjusted in SEWAB with a soil moisture dependent factor α (see Table 3.2). To account for dry conditions a formulation by Mihailovi´c et al. (1993) has been implemented α=   1−1−Θ ΘFC n ,Θ≤ΘFC 1,Θ>ΘFC (3.8) with the volumetric water content at field capacity ΘFC, the actual water content of the topsoil Θ and using n= 2 as exponent. The original parametrisation α= 0.5h1−cos Θ ΘFC πi for Θ ≤ΘFC (Noilhan and Planton, 1989) is very prohibitive for low Θ as pointed out by Mihailovi´c et al. (1995). 3.2.2. Model parameters In order to focus on the impact of the model versions on the performance, no optimisation algorithms were applied to constrain the parameter space. Instead, two ways of deriving “reasonable” parameter sets were explored, defining a “measured” parameter set and a “default” parameter set. While the latter can be obtained with a standard knowledge of the surface and soil types involved, the measured parameters represent detailed in situ and laboratory observations of the relevant site-specific properties. A summary of the most important parameters gives Table 3.3. The leaf area index, emissivity, minimum stomatal resistance and maximum stomatal resistance were not measured and therefore uniformly taken for both surface types and parameter sets (Hu et al., 2009; Yang et al., 2009; Alapaty et al., 1997). Default parameters differ most considerably between both surfaces in the description of the soil. The soil texture was classified as “sand” and “sandy loam”, USDA textural classes, for grass−and grass+, respectively. The corresponding parameters have been collected for SEWAB by Mengelkamp et al. (1997), originating from Clapp and Hornberger (1978) in case of the hydraulic properties. As both land use types are classified as short grassland, the surface parameters differ solely in the fraction of vegetated area and therefore in the over-all albedo (albedo for grassland and dry bare soil from Foken, 2008b). Measured parameters use meteorological observations for albedo and roughness length for momentum, fraction of vegetated area, canopy height and rooting depth were 19 3. Methods Table 3.3. Most important parameters for the model simulations: albedo a, emissivity ε, fraction of vegetated area fveg, leaf area index of vegetated area LAIveg, canopy height hc, rooting depth zr, roughness length z0m, minimum stomatal resistance Rs,min, maximum stomatal resistance Rs,max, thermal diffusivity νT, soil heat capacity CG·%G, porosity Θsat, matrix potential at saturation Ψsat, saturated hydraulic conductivity Ksat, volumetric water content at field capacity ΘFC, volumetric water content at wilting point ΘWP, and exponent bfor relationships after Clapp and Hornberger (1978). Default parameter Measured parameter Parameter Unit NamITP NamUBT NamITP NamUBT Surface and vegetation parameter a0.22 0.205 0.196 0.196 ε0.97 0.97 0.97 0.97 fveg - 0.6 0.9 0.6 0.9 LAIveg - 1.0 1.0 1.0 1.0 hcm 0.15 0.15 0.15 0.07 zrm 0.3 0.3 0.3 0.5 z0m m 0.005 0.005 0.005 0.005 Rs,min s m−160.0 60.0 60.0 60.0 Rs,max s m−12500 2500 2500 2500 Soil parameter νTm2s−10.84 ·10−60.84 ·10−61.5 ·10−72.5 ·10−7 CG·%GJ m−3K−12.10 ·1062.10 ·1062.10 ·1062.10 ·106 Θsat m3m−30.395 0.435 0.396 0.63 Ψsat m -0.121 -0.218 -0.51 -0.14 Ksat m s−11.76 ·10−43.47 ·10−52.018 ·10−51.38 ·10−5 ΘFC m3m−30.135 0.150 0.21 0.38 ΘWP m3m−30.068 0.114 0.06 0.19 b4.05 4.90 3.61 6.79 20 3.3. Lake surface modelling for Nam Co 2009 estimated in the field. Soil physical parameters were deduced from laboratory investigation of soil samples taken nearby the measurement set-up (Chen et al., 2012) assuming the samples to be representative on the scale of the EC footprint. Directly measured are soil texture, thermal conductivity, hydraulic conductivity at saturation and the soil water retention curve, providing matrix potential at saturation and exponent b(Clapp and Hornberger, 1978). Backward calculation of the last two yield the volumetric water content at field capacity (pF=2.5 assumed) and at wilting point (pF=4.5 assumed). The latter pF value differs from the standard 4.2, a reasonable assumption for mesophytic grass species (Larcher, 2001, p208). 3.3. Lake surface modelling for Nam Co 2009 Turbulent fluxes over the shallow lake surface near Nam Co were modelled with a hydrodynamic multilayer model (HM) (Foken, 1979, 1984). As the governing principle, surface – atmosphere exchange is parametrised based on a bulk approach, but resolving the molecular boundary layer, the viscous buffer layer and turbulent layer by an integrated profile coefficient Γ. It accounts for stratification by using Monin-Obukhov similarity theory. The model is forced by measurements, using the same data set as utilised for SEWAB (see Sect. 3.2.1). Lake surface temperature is approximated by the measured lake temperature, hence there is no need for radiation measurements and energy balance closure within the model. The lake surface temperature probe was shielded against direct radiation, a radiation error due to diffuse radiation in the water body has been estimated as approximately 0.2 K, see Biermann et al. (2013, Appendix D). Wendisch and Foken (1989) investigated which forcing variables are most influential to the model error and identified water temperature (50 %) and wind velocity, air temperature and air humidity (10 % to 20 % each). The HM model is designed for turbulent exchange over the ocean. Shallow water, however, induces larger waves leading to higher roughness and an enhanced exchange depending on wind velocity and lake depth H(Panin et al., 2006b). Therefore the shallow water correction proposed by Panin and Foken (2005) has been implemented in the HM code within a master thesis (Thiem, 2011) QSW H,E=Qocean H,E1 + kSW H,E·h·H−1(3.9) with the coefficient kSW H,E≈2. As postulated by the theoretical consideration, the shallow water turbulent fluxes QSW H,Eare always larger than the corresponding deep water fluxes Qocean H,E. The mean square wave height is parametrised with the empirical expression h≈0.07u2 10m ·g−1gHu−2 10m0.6, formulated by Davidan et al. (1985). Therefore the influencial parameters for the shallow water extension are wind velocity and lake depth, their impact on the relative increase in fluxes is displayed in Fig. 3.4. 21 3. Methods 0 5 10 15 0 2 4 6 8 10 12 u10m in ms−1 lake depth in m 1.01 1.05 1.1 1.15 1.2 1.3 1.4 1.6 QSW ⋅Qocean −1 Figure 3.4. Sensitivity of the shallow water term on lake depth Hand wind velocity in 10 m height u10m: Isolines show the relative increase of deep water fluxes QSW ·Q−1 ocean depending on u10m and H From the equations local sensitivities can be derived. Using the mean wind velocity of 4 m s−1and a water depth of 1.5 m and assuming corresponding typical errors of 0.3 m s−1and 1 m would lead to flux uncertainties of 1 % and 4 %, respectively. 22 4. Results 4.1. Data quality on the Tibetan Plateau The quality of turbulent fluxes from eddy-covariance data has been analysed on four stations on the Tibetan Plateau (Table 1.1) with respect to fulfilment of eddy-covariance requirements, energy balance closure, footprint, as well as obstacles in the vicinity of the sensor and potentially resulting internal boundary layers (Babel et al., 2011a,b). Despite some site-specific sources of disturbance not discussed here, two, more general features can be highlighted. For one thing near-ground free convective conditions have been found very frequently at Nam Co due to changes in the diurnal land-lake circulation system and due to changing cloud cover inducing sharp contrasts in the surface energy budget especially on the Tibetan Plateau (Zhou et al., 2011). This comes along with a degradation of data quality caused by both instationarity and mismatch to theoretical integral turbulence characteristics. Zhou et al. (2011) argue that data from these situations should not be routinely rejected, as they describe a typical daytime phenomenon within a convective boundary layer. Secondly, irregular friction velocities have been frequently found in the data from the BJ site (now Naqu station), related to the used sonic anemometer DAT 600 TR61A probe from Kaijo-Denki. Irregular friction means that momentum flux has been frequently observed with the wrong direction, rating the surface erroneously as a source of momentum rather than a sink. The DAT 600 is a non-omnidirectional sensor with a relatively small open sector of 120◦. It is shown by Li et al. (2013, Appendix F) that the problem can be reduced for data of the undisturbed (open) sector by applying a sector-wise planar-fit. Basically such partitions in disturbed and undisturbed sectors are relevant to all non-omnidirectional sensors, therefore the impact of using a sector-wise planar-fit is investigated for the CSAT3, Campbell Scientific Ltd. as well. The friction velocity of the sector-wise planar-fit deviates up to 10 % (DAT 600) from the “usual” planar-fit applied for the whole sector of all wind directions. Due to its large open sector of 340◦no such differences could be found for CSAT3, but irregular friction could be slightly reduced, especially when the front sector (i.e. the disturbing probe elements occur straight behind the measuring path) is rotated separately. This discrepancy between CSAT3 and DAT 600 is reflected by systematic differences found between the friction velocities derived by both instruments when applying the planarfit in the usual way. In contrast, for sector-wise planar-fit no differences between both 23 4. Results −Res Rnet QE QH QG grass− (a) −800 −600 −400 −200 0 200 Energy fluxes in Wm−2 −Res Rnet QE QH QG grass+ (b) Rnet QE QH lake(c) Tsfc−land Tair 0000 0600 1200 1800 5 10 15 20 25 30 Temp in °C Tsfc−land Tair 0000 0600 1200 1800 Tsfc−lake Tair 0000 0600 1200 1800 2400 Figure 4.1. Mean diurnal energy fluxes for the whole measurement period, separated for land (a: grass−at NamITP, b: grass+at NamUBT) and lake (c: NamUBT); all components are measured for land fluxes (a,b); for lake fluxes, the net radiation is calculated from measured downwelling radiation and using an albedo of 0.06 and the lake surface temperature with an emissivity of 0.96; the lower panel shows diurnal surface and air temperature. The time axis is displayed in Beijing standard time (CST), mean local solar noon during the observation period is at 1400 CST. From Biermann et al. (2013, Appendix D) instruments can be seen. It is also important to mention that scalar fluxes were not affected by the different planar-fit rotations. 4.2. Flux measurements at Nam Co During the monsoon season, the measured energy fluxes at Nam Co exhibit a distinct spatial heterogeneity corresponding to different surface types, see Babel et al. (2013, Appendix C) and Biermann et al. (2013, Appendix D). Mean diurnal energy fluxes for a dry (grass−) and a wet (grass+) alpine steppe and a shallow lake surface can be seen in Fig. 4.1. The measurements at NamUBT correspond to either grass+or lake surface, depending on wind direction (Fig. 3.2). The land surface fluxes (Fig. 4.1a, b) show a similar diurnal cycle in general, with latent heat fluxes dominating over sensible heat fluxes, a typical feature for the monsoon season on the Tibetan Plateau (e.g. Gu et al., 2005; Ma and Ma, 2006). Nevertheless, evaporation is higher at grass+on average due to soil moisture availability. While grass+is constantly supplied by a shallow 24 4.3. Land surface modelling at Nam Co ground water table, water availability is highly variable at NamITP with volumetric soil moistures below 5 % most of the time, but with saturated soils shortly after rain events. In such short periods, latent and sensible heat fluxes are approximately equal, some example days are shown by Babel et al. (2013, Appendix C). The NamITP site exhibits the typical features for dry surfaces on the TP with a huge diurnal cycle of the surface temperature (peaks reach up to 50 ◦C on dry days), and only moderate heat fluxes are not able to redistribute this surface heat content effectively (Yang et al., 2009). The premature change of the sign of the ground heat flux in the early afternoon indicates a strongly heated shallow soil layer, thermally decoupled from the deeper soil and supplying energy to the surface well before surface temperatures drop to the same magnitude as air temperature. Over the lake surface, the turbulent energy does not show a diurnal cycle, but were constant over the day (Biermann et al., 2013, Appendix D). The lake body is able to release energy at any time, so evaporation is mainly limited by wind velocity and vapour pressure deficit. The shallow lake in particular shows comparably large evaporation due to high wind velocities of 4 m s−1on average and enhanced turbulent exchange caused by unstable stratification even during daytime (Fig. 4.1c). In contrast, stable stratification typically prevails over lakes in daytime (e.g. Beyrich et al., 2006; Nordbo et al., 2011). 4.3. Land surface modelling at Nam Co In order to assess model performance, model runs for grass−and grass+were conducted for measured and default parameters, using both the original version and the adaptation to the Tibetan Plateau (TP version). The simulations were compared with energy balance corrected observations using both correction methods according to the Bowen ratio (EBC-Bo, Twine et al., 2000) and according to the Buoyancy flux (EBC-HB), see Sect. 3.1.4 and Babel et al. (2013, Appendix C). In general observed patterns (EBC-Bo corrected) were adequately reproduced showing correlation coefficients of 0.9 for both sites, parameter sets and model versions (Babel et al., 2013, Appendix C). This is a notable feature as no optimisation algorithm has been applied so far. Differences between model runs, however, can be found in model bias B=ξsim −ξobs and the Nash-Sutcliffe coefficient NS = 1 −PN(ξsim−ξobs)2 PN(ξobs−ξobs)2 (Nash and Sutcliffe, 1970), which can be interpreted similarly to the common coefficient of determination R2, but is sensitive to bias as well. A large and positive bias for turbulent fluxes is found at grass−, which is not apparent at grass+(Fig. 4.2a). This is in parts connected to the estimation of the ground heat flux and a high sensitivity of the new thermal conductivity formulation to changes in soil moisture in the dry range (Babel et al., 2013, Appendix C). Nevertheless, bias in ground heat flux could be reduced at grass−compared to the original version. The TP version reduces the sensible heat flux via the new implementation of thermal roughness and therefore its bias as 25 4. Results (a) (b) BQturb Bias B in Wm−2 BQH −40 −20 0 20 40 60 BQE NamITP NamUBT od om Td Tm od om Td Tm NSQH Nash−Sutcliffe coefficient NS NSQE 0.0 0.2 0.4 0.6 0.8 1.0 NamITP NamITP NamUBT od om Td Tm od om Td Tm Figure 4.2. Bias (a) and Nash-Sutcliffe coefficient NS (b) of turbulent fluxes (simulated vs. EBC-Bo corrected observations), NamITP corresponds to grass−, NamUBT to grass+; the individual blocks show od:original SEWAB version, default parameters; om:original SEWAB version, measured parameters; Td:TP version, default parameters; Tm:TP version, measured parameters. Modified from Babel et al. (2013, Appendix C) well, the bias of latent heat is also slightly reduced. This reduction in bias for the TP version induces a better performance with the NS coefficient (Fig. 4.2b). Simulations with default parameters yield predictions closer to the EBC-Bo corrected observations. This can be mainly attributed to larger field capacities and wilting points in the measured parameter set (Table 3.3), suppressing evapotranspiration on both sites. The new formulation for bare soil evaporation partly compensates this effect at NamITP, at NamUBT bare soil evaporation takes no effect as the fractional area of bare soil is too low. As expected, the simulations perform better for grass+than for grass−in general, as SEWAB formulations have not been validated for such dry conditions before, for example the stomatal resistance by Noilhan and Planton (1989). For the grass−site the TP version shows better performance without substantially compromising latent heat fluxes. Similar results are found for the grass+side. Therefore this study not only agrees with previous work over dry surfaces (Yang et al., 2008, 2009; Chen et al., 2010), but shows that the implemented scheme to calculate thermal roughness is not limited to dry surfaces. Furthermore, the new TP version seems to be less sensitive to soil parameters as its performance shows smaller differences between parameter sets than the original version. The results have been cross-checked with the ground heat flux and important state variables as soil moisture and surface temperature. It could be shown that the new TP version predicts the surface temperature more accurately and is able to reduce bias for the ground heat flux, even if the scatter has been increased. The soil moisture is 26 4.4. Influence of the energy balance correction method reasonably resembled for grass−with both parameter set. For details see Babel et al. (2013, Appendix C). As SEWAB performed well in general when being forced with measured data, it has been deployed as a reference time series to evaluate a new soil model incorporated in a simplistic land surface model called “Hybrid” (Gerken et al., 2012, Appendix B). The new soil model has been invented to enhance the responsiveness of the surface in hybrid showing a distinct time lag to the observations in its old version. As SEWAB did not show such a time lag, its simulations were ideally suited for a cross correlation analysis with hybrid, where the large gaps in the observed data complicate the interpretation of the results, see Gerken et al. (2012, Appendix B) for detail. 4.4. Influence of the energy balance correction method In the previous section model evaluations were carried out with EBC-Bo corrected observations only. In order to highlight the role of the energy balance closure for the evaluation of land surface models the new correction method according to the buoyancy flux (EBC-HB) has been considered as well. Table 4.1 summarises the change in performance with respect to (a) model parameters, (b) model version, and (c) method of energy balance closure correction. Table 4.1a and b confirm the results given in the previous section for EBC-Bo corrected observations. In contrast, EBC-HB corrected observations indicate that measured parameters perform now substantially better at the grass+site. The same happens there, to a less extent, with respect to model version in case of sensible heat although the positive effect of the TP version prevails in general. The reason for this behaviour is a shift in bias for both turbulent fluxes, as EBC-HB attributes a larger fraction of the residual to the sensible heat flux. Therefore the choice for the method to close the energy balance has a strong influence on the decision on the “right” model parameter set or version. It should be noted that a large bias remains for the sum of turbulent fluxes (Fig. 4.2) which is in fact independent of the method for energy balance correction and must be attributed to other reasons. Switching between correction methods (Table 4.1c) yields ambiguous results in bias and NS with respect to model version and parameters, but shows advantage for EBCBo at grass+and for EBC-HB at grass−. The pattern statisitics, however, offer another perspective: EBC-Bo yields substantially higher R2for the sensible heat flux in any case and lower R2for the latent heat flux at grass+. Therefore the SEWAB model is more compatible with EBC-Bo, as exemplarily shown in Fig. 4.3. The simulations of sensible heat flux show more scatter with the EBC-HB corrected observations than for EBC-Bo. This might be caused by an intrinsic model incompatibility to EBC-HB or by problems in estimation of the energy balance closure, incorporating additional uncertainty into the turbulent fluxes. On the other hand, uncorrected observations do not lead to higher R2for sensible heat than EBC-Bo corrected (not shown). The red 27 4. Results −−−−−−−−−−−−−−−−−−−−−−−−− −−−−−−−−−−−−−−−−−−−−−−−−− QH−EB in Wm−2 grass− 0 100 200 300 400 −−−−−−−−−−−−−−−−−−−−−− −−− −−−−−−−−−−−−−−−−−−−−−−−−− QE−EB in Wm−2 grass− −−−−−−−−−−−−− − −− −−−−−− −−−−−−− −−−− −−−−−−−−−−− QH−EB in Wm−2 grass+ 0 100 200 300 400 −−− −−−− −−− − −− − −− −−−− − − −−−−−−− −−− −− − −−− −−− −− − QE−EB in Wm−2 grass+ −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−− −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−− QH in Wm−2 lake 0000 0600 1200 1800 0 100 200 300 400 obs sim −−−−−−− −−−−− − −− −−−−−−−−−− −−− −−−−−−− − −−−− − − − −− −−−−− −−−−−−−−−−−−−−−−−−−−−−−−−−−− − −−−− − −− QE in Wm−2 lake 0000 0600 1200 1800 2400 Figure 4.6. Diurnal cycle of observations and simulations at Nam Co 2009. Land surface observations have been corrected according to the Bowen ratio (EBC-Bo), respective simulations have been run in the TP version using the measured parameters. Solid lines represent mean fluxes while horizontal bars and grey shaded areas denote respective standard deviations for observations and simulations. From Biermann et al. (2013, Appendix D). 34 4.6. Flux heterogeneity at Nam Co grass− grass+ lake Mean fluxes in Wm−2 0 50 100 150 200 250 −Rsw Rlw QGQHQE Figure 4.7. Mean fluxes from the whole period for the three surface types from observation-based model runs as used in Figs. 4.5 and 4.6. Net shortwave radiation (Rsw) and net longwave radiation (Rlw) are parametrised as explained in Fig. 4.1. For the land surface types, QGrepresents the ground heat flux. In case of lake surface, QGdenotes the residual of the energy balance, shown as a hatched bar, including the energy fluxes not accounted for, e.g. storage change in the water body and flux into the sediment. Error bars indicate 1.96 times the standard error of the mean, corresponding to the 95 % confidence interval. From Biermann et al. (2013, Appendix D). 35 4. Results in comparatively high surface temperatures. Mean differences of sensible and latent heat flux between grass+ and grassare 24.0 W m−2and −33.5 W m−2, respectively, and between grass+ and lake are −27.3 W m−2and 22.3 W m−2, respectively. 36 5. Conclusions The overarching goal of this thesis is to progress the estimation of turbulent fluxes over typical surface types on the Tibetan Plateau. Achievements have been made related to eddy-covariance observations on the Tibetan Plateau and site-specific land surface modelling. An eddy-covariance data processing scheme has been developed (Babel et al., 2011a,b), major features have been already applied on data of the Tibetan Plateau (Zhou et al., 2011; Biermann et al., 2013, Appendix D). Specific sensor problems with the Kaijo-Denki DAT600 TR61A probe, installed at one of the sites on the Tibetan Plateau, have been encountered with a sector-wise planar-fit rotation (Li et al., 2013, Appendix F). Furthermore, the land surface scheme SEWAB has been successfully adapted to a dry and a wet grassland site for the monsoon season at Nam Co, Tibetan Plateau (Babel et al., 2013, Appendix C). First eddy-covariance measurements over a lake surface have been gathered and a hydrodynamical multilayer model could resemble these observations even resolving the diurnal cycle (Biermann et al., 2013, Appendix D). From these achievements the following conclusions can be drawn: •Coordinate rotation for non-omnidirectional sonic anemometer need a careful handling, which applies to the Kaijo Denki DAT 600 TR61A probe in particular and, to a less extent, to the CSAT3 (Campbell Scientific Ltd.). Both instrument show frequent occurrences of physical implausible upward momentum fluxes, caused by sensor distortion. It is recommended to apply the planar-fit rotation for only the undisturbed wind sector, and discard momentum fluxes measured in other wind sectors. This can mitigate such problems, and the friction velocity may deviate up to 10 % from those derived by a conventional planar-fit, likely explaining the differences between DAT 600 and CSAT3 found in previous studies as well (Hong et al., 2004). In contrast, no influence of this problem on scalar fluxes could be shown. •The adaptation of SEWAB to the Tibetan Plateau (TP version) shows a slight, but overall better performance than the original version. The TP version leads to bias reduction especially for sensible heat on the dry surface, which even holds when using energy balance corrected observations according to the buoyancy flux. It performs reasonable also for wet surfaces although the results are ambiguous with respect to the used method to correct the observation for the energy balance closure gap. Nevertheless this demonstrates that the implemented modifications, including the z0h-scheme by Yang et al. (2008) are not limited to dry or bare 37 5. Conclusions soil surfaces and the small scale heterogeneity in soil moisture, anticipated by Su et al. (2011), can be resolved with this SEWAB version. •When comparing simulations with eddy-covariance derived turbulent fluxes, the performance is strongly affected by the measured energy balance closure gap. Thus the selection of the surface model and the choice of the EBC correction method are inter-related problems, which should never be overlooked when validating a model with eddy-covariance data. Kracher et al. (2009) could show, that SEWAB reproduces the observed Bowen ratio quite well. This dissertation supports their findings with data from a total different environment. On the other hand, the EBC-Bo correction may not be appropriate due to poor scalar similarity between sensible and latent heat flux for low frequency contributions, as pointed out by Ruppert et al. (2006). This is confirmed by Charuchittipan et al. (2013, Appendix E), and the new correction according to the buoyancy flux, EBC-HB, is proposed in agreement with findings by e.g. Mauder and Foken (2006); Foken (2008a); Foken et al. (2011); Ingwersen et al. (2011). Therefore future model development of turbulent flux parametrisation should recognize advective fluxes supplying secondary circulations as recent hypotheses concerning the energy balance closure rather than trying to get the best fit to the uncorrected eddy-covariance data. This is a challenging task, and a meaningful future study of model structure at least requires an individual parametrisation of all energy flux components and no item should simply serve as residual of the energy balance. •The integration of the energy balance closure and its correction demands a high quality measurements of all energy flux components. This may not always be given or, in case of the lake surface, only possible with large efforts. Especially the ground heat flux estimation (observed and simulated) is prone to uncertainties. In the particular case of the grass−site at Nam Co problems arise due to a large gravel content in the soil. More general, under dry conditions on the Tibetan Plateau, a shallow dry upper soil layer might thermally decouple from the deeper soil. Such a feature is usually not considered in heat flux parametrisations. •Site-specific land surface model simulations based on eddy-covariance observations has been rarely conducted on the Tibetan Plateau. From this study follows that some field investigations (additional to turbulent flux measurements and forcing data) are inevitable to derive a high quality parameter set (measured or default parameters), which are at least soil moisture measurements, soil temperature measurements and field based knowledge of the soil type, derived at least by a conventional pedological description. The land surface model SEWAB, constrained by such a data set, is reliable enough to be used for validating more simplistic models as done by Gerken et al. (2012, Appendix B). 38 •Turbulent flux observations and simulations with the HM model pose a unique data set on the Tibetan Plateau, as only studies conducting bulk approaches on daily or monthly basis were reported up to now. However, high quality estimates are necessary since the Tibetan Plateau is covered with a significant lake fraction of various sizes and therefore different characteristics. The HM model with the shallow water term proved its suitability to estimate lake evaporation on a high standard even resolving the diurnal course. Appropriate forcing can be achieved with land derived standard meteorological measurements, a representative surface temperature is the only measurement required directly from the lake. Given these requirements, the model can be transferred to lakes with different size and depth, for example the large Nam Co lake, although more efforts have to be put in estimating a representative surface temperature. •At the lake shore site (NamUBT) the simulations for land and lake have been integrated by their relative contribution to the measured flux according to the footprint of each time step. Thus measurements and simulations become comparable even under conditions with mixed footprints. It could be shown, that the performance does not deteriorate in such situations and the tile approach is valid in this terrain for spatial integration. Therefore representative flux simulations on a grid cell with edge lengths of 1 km to 5 km can be given for each time step in order to compare with remote sensing data. The simulations of turbulent fluxes for grass−, grass+and lake differ in mean values and temporal characteristics beyond model uncertainty, with deviations occasionally exceeding 200 W m−2on daytime. In contrast, the measurements over dry grassland at the Nam Co Monitoring and Research Station for Multisphere Interactions (grass−) are considered to be a reference for the land surface exchange in the Nam Co region. This study shows, that the land use distribution within the respective remote sensing pixel or grid cell for mesoscale modelling has to be carefully determined before validating with the dry grassland station. A potential representation error can be reduced by integrating the simulated fluxes of adjacent land use types according to their contribution to the respective grid cell. The essential findings in this thesis concerning modelling under the conditions of the Tibetan Plateau, energy balance closure and footprint applications provide the basis to process the eddy-covariance data in three levels as described in Sect. 1.2. This is exemplarily shown with the sensible heat flux from Nam Co station in 2009 (Fig. 5.1). Due to data quality filtering, the levels I and II cannot provide complete diurnal cycles. SEWAB simulations from grass−fill most of the gaps in level II data with high reliability concerning magnitude and dynamics (Fig. 5.1c). Whenever forcing data is missing for simulations, some gaps are still left. In order to obtain no more than seasonal or annual averages, remaining gaps can be eliminated with methods listed by Falge et al. (2001a,b). 39 5. Conclusions (a) Level I: eddy−covariance observations Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 0 4 8 12 16 20 24 n.V. −200 −100 −50 0 50 100 200 300 400 550 (b) Level II: EBC−Bo corrected observations Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 0 4 8 12 16 20 24 n.V. −200 −100 −50 0 50 100 200 300 400 550 (c) Level III: (b) gapfilled with simulations Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 0 4 8 12 16 20 24 n.V. −200 −100 −50 0 50 100 200 300 400 550 Figure 5.1. Data example: Sensible heat flux at Nam Co, 2009, on different levels of processing as defined for the CEOP-AEGIS project (Sect. 1.2); (a), level I: quality checked eddy-covariance data; (b), level II: energy balance closure corrected observations according to the Bowen ratio, for nighttime values, when EBC correction is not applicable (see Sect. 3.1.4), level I data is accepted unchanged; note that daytime level II data cannot be provided whenever an observed energy balance component is missing (c), level III: data from level II, gap-filled with SEWAB simulations for grass−. Remaining gaps indicate missing forcing data. 40 References Agam, N., Berliner, P. R., Zangvil, A., and Ben-Dor, E.: Soil water evaporation during the dry season in an arid zone, J. Geophys. Res., 109, D16 103–, doi: 10.1029/2004JD004802, 2004. Alapaty, K., Pleim, J. E., Raman, S., Niyogi, D. S., and Byun, D. W.: Simulation of Atmospheric Boundary Layer Processes Using Localand Nonlocal-Closure Schemes, J. Appl. Meteorol., 36, 214–233, doi:10.1175/1520-0450(1997)036h0214:SOABLPi2. 0.CO;2, 1997. Aubinet, M., Grelle, A., Ibrom, A., Rannik, U., Moncrieff, J., Foken, T., Kowalski, A., Martin, P., Berbigier, P., Bernhofer, C., Clement, R., Elbers, J., Granier, A., Gr¨unwald, T., Morgenstern, K., Pilegaard, K., Rebmann, C., Snijders, W., Valentini, R., and Vesala, T.: Estimates of the Annual Net Carbon and Water Exchange of Forests: The EUROFLUX Methodology, Adv. Ecol. Res., 30, 113–175, doi:10.1016/ S0065-2504(08)60018-5, 2000. Babel, W., Eigenmann, R., Ma, Y., and Foken, T.: Analysis of turbulent fluxes and their representativeness for the interaction between the atmospheric boundary layer and the underlying surface on Tibetan Plateau, CEOP-AEGIS Deliverable report De1.2, Ed. University of Strasbourg, France, ISSN 2118-7843: 35 p., 2011a. Babel, W., Li, M., Sun, F., Ma, W., Chen, X., Colin, J., Ma, Y., and Foken, T.: Aerodynamic and thermodynamic variables for four stations on Tibetan Plateau – Introduction on the CEOP-AEGIS database in NetCDF, CEOP-AEGIS Deliverable report De1.3, Ed. University of Strasbourg, France, ISSN 2118-7843: 90 p., 2011b. Babel, W., Chen, Y., Biermann, T., Yang, K., Ma, Y., and Foken, T.: Adaptation of a land surface scheme for modeling turbulent fluxes on the Tibetan Plateau under different soil moisture conditions, submitted to J. Geophys. Res., see Appendix C, 2013. Baldocchi, D., Falge, E., Gu, L., Olson, R., Hollinger, D., Running, S., Anthoni, P., Bernhofer, C., Davis, K., Evans, R., Fuentes, J., Goldstein, A., Katul, G., Law, B., Lee, X., Malhi, Y., Meyers, T., Munger, W., Oechel, W., Paw, K. T., Pilegaard, K., Schmid, H. P., Valentini, R., Verma, S., Vesala, T., Wilson, K., and Wofsy, S.: 41 References FLUXNET: A New Tool to Study the Temporal and Spatial Variability of EcosystemScale Carbon Dioxide, Water Vapor, and Energy Flux Densities, B. Am. Meteorol. Soc., 82, 2415–2434, doi:10.1175/1520-0477(2001)082h2415:FANTTSi2.3.CO;2, 2001. Balsamo, G., Boussetta, S., Dutra, E., Beljaars, A., Viterbo, P., and van den Hurk, B.: Evolution of land surface processes in the IFS, ECMWF Newslett., 127, 17–22, 2011. Beyrich, F., Leps, J.-P., Mauder, M., Bange, J., Foken, T., Huneke, S., Lohse, H., L¨udi, A., Meijninger, W., Mironov, D., Weisensee, U., and Zittel, P.: Area-Averaged Surface Fluxes Over the Litfass Region Based on Eddy-Covariance Measurements, Bound.-Lay. Meteorol., 121, 33–65, doi:10.1007/s10546-006-9052-x, 10.1007/s10546006-9052-x, 2006. Biermann, T., Babel, W., Olesch, J., and Foken, T.: Documentation of the Micrometeorological Experiment, Nam Tso, Tibet, 25th of June – 8th of August 2009, Work Report University of Bayreuth, Dept. of Micrometeorology, ISSN 1614-8916, 41, 38 pp., URL http://opus.ub.uni-bayreuth.de/opus4-ubbayreuth/frontdoor/ index/index/docId/626, 2009. Biermann, T., Babel, W., Ma, W., Chen, X., Thiem, E., Ma, Y., and Foken, T.: Turbulent flux observations and modelling over a shallow lake and a wet grassland in the Nam Co basin, Tibetan Plateau, Theor. Appl. Climatol., accepted, see Appendix D, 2013. Br¨otz, B., Eigenmann, R., D¨ornbrack, A., Foken, T., and Wirth, V.: Early-morning flow transition in a valley in low-mountain terrain, submitted to Bound.-Lay. Meteorol., 2013. Charuchittipan, D., Babel, W., Mauder, M., Leps, J.-P., and Foken, T.: Extension of the averaging time of the eddy-covariance measurement and its effect on the energy balance closure, submitted to Bound.-Lay. Meteorol., see Appendix E, 2013. Chen, T. H., Henderson-Sellers, A., Milly, P. C. D., Pitman, A. J., Beljaars, A. C. M., Polcher, J., Abramopoulos, F., Boone, A., Chang, S., Chen, F., Dai, Y., Desborough, C. E., Dickinson, R. E., D¨umenil, L., Ek, M., Garratt, J. R., Gedney, N., Gusev, Y. M., Kim, J., Koster, R., Kowalczyk, E. A., Laval, K., Lean, J., Lettenmaier, D., Liang, X., Mahfouf, J.-F., Mengelkamp, H.-T., Mitchell, K., Nasonova, O. N., Noilhan, J., Robock, A., Rosenzweig, C., Schaake, J., Schlosser, C. A., Schulz, J.- P., Shao, Y., Shmakin, A. B., Verseghy, D. L., Wetzel, P., Wood, E. F., Xue, Y., Yang, Z.-L., and Zeng, Q.: Cabauw Experimental Results from the Project for Intercomparison of Land-Surface Parameterization Schemes, J. Climate, 10, 1194–1215, doi:10.1175/1520-0442(1997)010h1194:CERFTPi2.0.CO;2, 1997. 42 References Chen, Y., Yang, K., Zhou, D., Qin, J., and Guo, X.: Improving the Noah Land Surface Model in Arid Regions with an Appropriate Parameterization of the Thermal Roughness Length, J. Hydrometeorol., 11, 995–1006, doi:10.1175/2010JHM1185.1, 2010. Chen, Y., Yang, K., He, J., Qin, J., Shi, J., Du, J., and He, Q.: Improving land surface temperature modeling for dry land of China, J. Geophys. Res., 116, D20 104–, doi: 10.1029/2011JD015921, 2011. Chen, Y., Yang, K., Tang, W., Qin, J., and Zhao, L.: Parameterizing soil organic carbon’s impacts on soil porosity and thermal parameters for Eastern Tibet grasslands, Sci. China Earth Sci., 55, 1001–1011, doi:10.1007/s11430-012-4433-0, 2012. Clapp, R. B. and Hornberger, G. M.: Empirical Equations for some Soil Hydraulic Properties, Water Resour. Res., 14, 601–604, 1978. Dai, Y., Zeng, X., Dickinson, R. E., Baker, I., Bonan, G. B., Bosilovich, M. G., Denning, A. S., Dirmeyer, P. A., Houser, P. R., Niu, G., Oleson, K. W., Schlosser, C. A., and Yang, Z.-L.: The Common Land Model, B. Am. Meteorol. Soc., 84, 1013–1023, doi:10.1175/BAMS-84-8-1013, 2003. Davidan, I., Lopatuhin, L., and Rogkov, V.: Volny v okeane (waves in the ocean), Gidrometeoizdat, Leningrad, 256pp., 1985. Duan, A. and Wu, G.: Role of the Tibetan Plateau thermal forcing in the summer climate patterns over subtropical Asia, Clim. Dynam., 24, 793–807, doi:10.1007/ s00382-004-0488-8, 2005. Falge, E., Baldocchi, D., Olson, R., Anthoni, P., Aubinet, M., Bernhofer, C., Burba, G., Ceulemans, R., Clement, R., Dolman, H., Granier, A., Gross, P., Gr¨unwald, T., Hollinger, D., Jensen, N.-O., Katul, G., Keronen, P., Kowalski, A., Lai, C. T., Law, B. E., Meyers, T., Moncrieff, J., Moors, E., Munger, J. W., Pilegaard, K., Rannik, U., Rebmann, C., Suyker, A., Tenhunen, J., Tu, K., Verma, S., Vesala, T., Wilson, K., and Wofsy, S.: Gap filling strategies for defensible annual sums of net ecosystem exchange, Agr. Forest Meteorol., 107, 43–69, doi:10.1016/S0168-1923(00)00225-2, 2001a. Falge, E., Baldocchi, D., Olson, R., Anthoni, P., Aubinet, M., Bernhofer, C., Burba, G., Ceulemans, R., Clement, R., Dolman, H., Granier, A., Gross, P., Gr¨unwald, T., Hollinger, D., Jensen, N.-O., Katul, G., Keronen, P., Kowalski, A., Lai, C. T., Law, B. E., Meyers, T., Moncrieff, J., Moors, E., Munger, J. W., Pilegaard, K., Rannik, U., Rebmann, C., Suyker, A., Tenhunen, J., Tu, K., Verma, S., Vesala, T., Wilson, K., and Wofsy, S.: Gap filling strategies for long term energy flux data sets, Agr. Forest Meteorol., 107, 71–77, doi:10.1016/S0168-1923(00)00235-5, 2001b. 43 References Ni, J.: Impacts of climate change on Chinese ecosystems: key vulnerable regions and potential thresholds, Reg. Environ. Change, 11, 49–64, doi:10.1007/ s10113-010-0170-0, 2011. Noilhan, J. and Planton, S.: A Simple Parameterization of Land Surface Processes for Meteorological Models, Mon. Weather Rev., 117, 536–549, doi:10.1175/ 1520-0493(1989)117h0536:ASPOLSi2.0.CO;2, 1989. Nordbo, A., Launiainen, S., Mammarella, I., Lepp¨aranta, M., Huotari, J., Ojala, A., and Vesala, T.: Long-term energy flux measurements and energy balance over a small boreal lake using eddy covariance technique, J. Geophys. Res., 116, D02 119, doi:10.1029/2010JD014542, 2011. Panin, G., Nasonov, A., and Foken, T.: Evaporation and heat exchange of a body of water with the atmosphere in a shallow zone, Izv. Atmos. Ocean. Phys., 42, 337–352, 2006a. Panin, G. N. and Foken, T.: Air–sea interaction including a shallow and coastal zone, J. Atmos. Ocean Sci., 10, 289–305, doi:10.1080/17417530600787227, 2005. Panin, G. N., Nasonov, A. E., Foken, T., and Lohse, H.: On the parameterisation of evaporation and sensible heat exchange for shallow lakes, Theor. Appl. Climatol., 85, 123–129, doi:10.1007/s00704-005-0185-5, 2006b. Peters-Lidard, C. D., Blackburn, E., Liang, X., and Wood, E. F.: The Effect of Soil Thermal Conductivity Parameterization on Surface Energy Fluxes and Temperatures, J. Atmos. Sci., 55, 1209–1224, doi:10.1175/1520-0469(1998)055h1209: TEOSTCi2.0.CO;2, 1998. Pitman, A. J.: The evolution of, and revolution in, land surface schemes designed for climate models, Int. J. Climatol., 23, 479–510, doi:10.1002/joc.893, 2003. Rannik, U., Aubinet, M., Kurbanmuradov, O., Sabelfeld, K. K., Markkanen, T., and Vesala, T.: Footprint analysis for measurements over a heterogeneous forest, Bound.- Lay. Meteorol., 97, 137–166, doi:10.1023/A:1002702810929, 2000. Rebmann, C., Kolle, O., Heinesch, B., Queck, R., Ibrom, A., and Aubinet, M.: Data Acquisition and Flux Calculations, in: Eddy Covariance: A Practical Guide to Measurement and Data Analysis, edited by Aubinet, M., Vesala, T., and Papale, D., Springer Atmospheric Sciences, pp. 59–83, Springer Netherlands, doi: 10.1007/978-94-007-2351-1 3, 2012. Rouse, W. R., Oswald, C. J., Binyamin, J., Spence, C., Schertzer, W. M., Blanken, P. D., Bussi`eres, N., and Duguay, C. R.: The Role of Northern Lakes in a Regional Energy Balance, J. Hydrometeorol., 6, 291–305, doi:doi:10.1175/JHM421.1, 2005. 50 References Ruppert, J., Thomas, C., and Foken, T.: Scalar Similarity for Relaxed Eddy Accumulation Methods, Bound.-Lay. Meteorol., 120, 39–63, doi:10.1007/s10546-005-9043-3, 2006. Schotanus, P., Nieuwstadt, F., and De Bruin, H.: Temperature measurement with a sonic anemometer and its application to heat and moisture fluxes, Bound.-Lay. Meteorol., 26, 81–93, doi:10.1007/BF00164332, 1983. Sellers, P., Randall, D., Collatz, G., Berry, J., Field, C., Dazlich, D., Zhang, C., Collelo, G., and Bounoua, L.: A Revised Land Surface Parameterization (SiB2) for Atmospheric GCMs. Part I: Model Formulation, J. Climate, 9, 676–705, doi: 10.1175/1520-0442(1996)009h0676:ARLSPFi2.0.CO;2, 1996. Seneviratne, S. I., Corti, T., Davin, E. L., Hirschi, M., Jaeger, E. B., Lehner, I., Orlowsky, B., and Teuling, A. J.: Investigating soil moisture–climate interactions in a changing climate: A review, Earth-sci. Rev., 99, 125–161, doi:10.1016/j.earscirev. 2010.02.004, 2010. Steinfeld, G., Letzel, M., Raasch, S., Kanda, M., and Inagaki, A.: Spatial representativeness of single tower measurements and the imbalance problem with eddycovariance fluxes: results of a large-eddy simulation study, Bound.-Lay. Meteorol., 123, 77–98, doi:10.1007/s10546-006-9133-x, 2007. Steppeler, J., Doms, G., Schattler, U., Bitzer, H. W., Gassmann, A., Damrath, U., and Gregoric, G.: Meso-gamma scale forecasts using the nonhydrostatic model LM, Meteorol. Atmos. Phys., 82, 75–96, 2003. Stoy, P. C., Mauder, M., Foken, T., Marcolla, B., Boegh, E., Ibrom, A., Arain, M. A., Arneth, A., Aurela, M., Bernhofer, C., Cescatti, A., Dellwik, E., Duce, P., Gianelle, D., van Gorsel, E., Kiely, G., Knohl, A., Margolis, H., McCaughey, H., Merbold, L., Montagnani, L., Papale, D., Reichstein, M., Saunders, M., SerranoOrtiz, P., Sottocornola, M., Spano, D., Vaccari, F., and Varlagin, A.: A datadriven analysis of energy balance closure across FLUXNET research sites: The role of landscape scale heterogeneity, Agr. Forest Meteorol., 171–172, 137–152, doi: 10.1016/j.agrformet.2012.11.004, 2013. Su, Z., Wen, J., Dente, L., van der Velde, R., Wang, L., Ma, Y., Yang, K., and Hu, Z.: The Tibetan Plateau observatory of plateau scale soil moisture and soil temperature (Tibet-Obs) for quantifying uncertainties in coarse resolution satellite and model products, Hydrol. Earth Syst. Sci., 15, 2303–2316, doi:10.5194/hess-15-2303-2011, 2011. Takayabu, I., Takata, K., Yamazaki, T., Ueno, K., Yabuki, H., and Haginoya, S.: Comparison of the Four Land Surface Models Driven by a Common Forcing Data Pre51 References pared from GAME/Tibet POP’97 Products – Snow Accumulation and Soil Freezing Processes –, J. Meteorol. Soc. Jpn., 79, 535–554, doi:10.2151/jmsj.79.535, 2001. Thiem, E.: Modeling of the energy exchange above lake and land surfaces, Master thesis, University of Bayreuth, 73pp., 2011. Twine, T. E., Kustas, W. P., Norman, J. M., Cook, D. R., Houser, P. R., Meyers, T. P., Prueger, J. H., Starks, P. J., and Wesely, M. L.: Correcting eddy-covariance flux underestimates over a grassland, Agr. Forest Meteorol., 103, 279–300, doi:10. 1016/S0168-1923(00)00123-4, 2000. van der Velde, R., Su, Z., Ek, M., Rodell, M., and Ma, Y.: Influence of thermodynamic soil and vegetation parameterizations on the simulation of soil temperature states and surface fluxes by the Noah LSM over a Tibetan Plateau site, Hydrol. Earth Syst. Sci., 13, 759–777, doi:10.5194/hess-13-759-2009, 2009. Wallace, J., Wright, I., Stewart, J., and Holwill, C.: The Sahelian Energy Balance Experiment (SEBEX): Ground based measurements and their potential for spatial extrapolation using satellite data, Adv. Space Res., 11, 131–141, doi: 10.1016/0273-1177(91)90413-E, 1991. Warrach, K., Mengelkamp, H.-T., and Raschke, E.: Treatment of frozen soil and snow cover in the land surface model SEWAB, Theor. Appl. Climatol., 69, 23–37, doi: 10.1007/s007040170033, 2001. Wendisch, M. and Foken, T.: Sensitivity test of a multilayer energy exchange model, Z. Meteorol., 39, 36–39, 1989. Wilson, K., Goldstein, A., Falge, E., Aubinet, M., Baldocchi, D., Berbigier, P., Bernhofer, C., Ceulemans, R., Dolman, H., Field, C., Grelle, A., Ibrom, A., Law, B. E., Kowalski, A., Meyers, T., Moncrieff, J., Monson, R., Oechel, W., Tenhunen, J., Valentini, R., and Verma, S.: Energy balance closure at FLUXNET sites, Agr. Forest Meteorol., 113, 223–243, 2002. Xu, J. and Haginoya, S.: An Estimation of Heat and Water Balances in the Tibetan Plateau, J. Meteorol. Soc. Jpn., 79, 485–504, doi:10.2151/jmsj.79.485, 2001. Xu, J., Yu, S., Liu, J., Haginoya, S., Ishigooka, Y., Kuwagata, T., Hara, M., and Yasunari, T.: The Implication of Heat and Water Balance Changes in a Lake Basin on the Tibetan Plateau, Hydrol. Res. Lett., 3, 1–5, doi:10.3178/hrl.3.1, 2009. Yanai, M. H., Li, C. F., and Song, Z. S.: Seasonal Heating of the Tibetan Plateau and its Effects on the Evolution of the Asian Summer Monsoon, J. Meteorol. Soc. Jpn., 70, 319–351, 1992. 52 References Yang, K. and Wang, J.: A temperature prediction-correction method for estimating surface soil heat flux from soil temperature and moisture data, Sci. China Ser. D, 51, 721–729, doi:10.1007/s11430-008-0036-1, 2008. Yang, K., Koike, T., and Yang, D.: Surface flux parameterization in the Tibetan Plateau, Bound.-Lay. Meteorol., 116, 245–262, doi:10.1023/A:1021152407334, 2003. Yang, K., Koike, T., Ye, B., and Bastidas, L.: Inverse analysis of the role of soil vertical heterogeneity in controlling surface soil state and energy partition, J. Geophys. Res., 110, D08101, doi:10.1029/2004JD005500, 2005. Yang, K., Koike, T., Ishikawa, H., Kim, J., Li, X., Liu, H., Wang, J., Liu, S., and Ma, Y.: Turbulent Flux Transfer over Bare-Soil Surfaces: Characteristics and Parameterization, J. Appl. Meteorol. Clim., 47, 276–290, doi:10.1175/2007JAMC1547.1, 2008. Yang, K., Chen, Y.-Y., and Qin, J.: Some practical notes on the land surface modeling in the Tibetan Plateau, Hydrol. Earth Syst. Sci., 13, 687–701, doi: 10.5194/hess-13-687-2009, 2009. Yang, K., Guo, X., and Wu, B.: Recent trends in surface sensible heat flux on the Tibetan Plateau, SCIENCE CHINA Earth Sciences, pp. 1–10, doi:10.1007/ s11430-010-4036-6, 2010. Ye, D.-Z. and Wu, G.-X.: The role of the heat source of the Tibetan Plateau in the general circulation, Meteorol. Atmos. Phys., 67, 181–198, doi:10.1007/BF01277509, 1998. Yu, S., Liu, J., Xu, J., and Wang, H.: Evaporation and energy balance estimates over a large inland lake in the Tibet-Himalaya, Environ. Earth Sci., 64, 1169–1176, doi:10.1007/s12665-011-0933-z, 2011. Zhang, X.: Improvement of a soil-atmosphere-transfer model for the simulation of bare soil surface energy balances in semiarid areas, Asia Pac. J. Atmos. Sci., 48, 97–105, doi:10.1007/s13143-012-0009-3, 2012. Zhou, D., Eigenmann, R., Babel, W., Foken, T., and Ma, Y.: The study of near-ground free convection conditions at Nam Co station on the Tibetan Plateau, Theor. Appl. Climatol., 105, 217–228, doi:10.1007/s00704-010-0393-5, 2011. Zilitinkevich, S. S.: Non-local turbulent transport: Pollution dispersion aspects of coherent structure of convective flows, in: Air Pollution – Volume I. Air Pollution Theory and Simulation, edited by Power, H., Moussiopoulos, N., and Brebbia, C. A., pp. 53–60, Computational Mechanics Publications, 1995. 53 A. Individual contributions to the joint publications This cumulative thesis consists of publications and manuscripts listed hereafter. Other authors contributed to these papers as well. Therefore my own contribution to the individual manuscripts is specified in this section. Nam Co experiment 2009 and data preparation Some of the following publications is based on the Nam Co experiment in 2009, further described in section 3.1. Together with Tobias Biermann I was responsible for the realisation of the experiment. The setup and design was planned and prepared together with Tobias Biermann in equal shares. We both assembled the measurement complex at the site, while Tobias Biermann alone was responsible for data collection and maintenance during the whole campaign. The data from the Nam Co experiment (NamUBT) as well as data provided from the permanent station of the ITP (NamITP) have been post-processed as follows: Tobias Biermann did quality checks on low frequency data and turbulent flux processing with the TK2/TK3 software package, including sector-wise planar-fit for NamUBT. I myself elaborated the actual land use distribution and contributed with the footprint analysis as well as the calculation of the ground heat flux and energy balance, including the EBC correction methods. Appendix B Gerken, T., Babel, W., Hoffmann, A., Biermann, T., Herzog, M., Friend, A. D., Li, M., Ma, Y., Foken, T., and Graf, H.-F.: Turbulent flux modelling with a simple 2-layer soil model and extrapolated surface temperature applied at Nam Co Lake basin on the Tibetan Plateau, Hydrol. Earth Syst. Sci., 16, 1095–1110, doi:10.5194/hess-16-1095-2012, 2012. •Tobias Gerken developed the idea of the manuscript and coordinated individual contributions. He conducted the Hybrid modelling and the analysis. He wrote the whole publication and acted as corresponding author. 54 •I provided the NamUBT data from the Nam Co experiment together with Tobias Biermann (see p. 54). I myself conducted SEWAB simulations for ITP and UBT land surface and gathered respective model parameters. •Alex Hoffmann contributed to the Hybrid model development relevant to this paper. •Michael Herzog contributed with technical advice on the manuscript. •Andrew Friend is the original author of Hybrid. The provided assistance with the handling of Hybrid. •Li Maoshan and Ma Yaoming provided data from the ITP station and supported the field trip. •Thomas Foken and Hans-F. Graf contributed to the manuscript at various stages with fruitful discussions. Appendix C Babel, W., Chen, Y., Biermann, T., Yang, K., Ma, Y., and Foken, T.: Adaptation of a land surface scheme for modeling turbulent fluxes on the Tibetan Plateau under different soil moisture conditions, submitted to J. Geophys. Res. •I myself developed the idea of this manuscript. It was me who adapted the SEWAB code, gathered the model parameters and calculated all simulations. Together with Tobias Biermann I provided the NamUBT data from the Nam Co experiment (p. 54). I alone conducted the whole statistical analysis and wrote the manuscript. Finally I act as corresponding author for the submitted manuscript. •Chen Yingying provided source code for the roughness length parametrisation of the TP version and for the ground heat flux calculation at NamITP station. He conducted laboratory measurements of soil samples he took over various places on the TP, which I used to derive the measured parameters. •Yang Kun was the developer of the new roughness length parametrisation and the new soil thermal conductivity calculation. Together with him I had fruitful discussions about model simulation results. •Ma Yaoming provided data from the ITP station and gave support during the field trip. •As my supervisor Thomas Foken contributed with scientific discussions throughout all stages of analysis and manuscript preparation 55 A. Individual contributions to the joint publications Appendix D Biermann, T., Babel, W., Ma, W., Chen, X., Thiem, E., Ma, Y., and Foken, T.: Turbulent flux observations and modelling over a shallow lake and a wet grassland in the Nam Co basin, Tibetan Plateau, Theor. Appl. Climatol., accepted •The idea of this manuscript was developed in equal parts by Tobias Biermann and me. We both realised the Nam Co experiment and data preparation (p. 54) and wrote the text in equal shares. While Tobias Biermann focused more on literature and scope of the manuscript as well as the experimental part, I put my emphasis on data analysis and the modelling parts. Tobias Biermann acts as corresponding author for the submitted manuscript. •Ma Yaoming, Chen Xuelong and Ma Weiqiang supported us during the field trip in 2009 and gave access to data from the ITP station •Elisabeth Thiem contributed to the manuscript’s workload within a master thesis under the supervision of me and Thomas Foken. Thereby she filled gaps in the model forcing data, implemented the shallow water term into the lake model HM, and conducted a preliminary sensitivity analysis for the lake model. •Thomas Foken acted as supervisor and liberally shared his experience in encouraging and fruitful discussions. Appendix E Charuchittipan, D., Babel, W., Mauder, M., Leps, J.-P., and Foken, T.: Extension of the averaging time of the eddy-covariance measurement and its effect on the energy balance closure, submitted to Bound.-Lay. Meteorol. •Doojdao Charuchittipan conducted the whole data analysis and wrote the text of the manuscript. She acts as corresponding author for the submitted manuscript. •I developed the proposed new energy balance closure correction algorithm together with DC and realised the respective part in the conclusions. •Matthias Mauder initiated the use of wavelets to analyse the long-term fluctuations. •Jens-Peter Leps provided data from several LITFASS-2003 stations and generated land-use classified data. •Thomas Foken encouraged the structure of the manuscript and contributed with many scientific discussions. Furthermore, he initiated the project related to the manuscript. 56 Appendix F Li, M., Babel, W., Tanaka, K., and Foken, T.: Note on the application of planar-fit rotation for non-omnidirectional sonic anemometers, Atmos. Meas. Tech., 6, 221–229, doi:10.5194/amt-6-221-2013, 2013. •Li Maoshan was in charge of the data used in this publication. He did the data preparation and analysis according to my instructions. Furthermore he wrote the description of BJ station. •I myself coordinated the analysis and wrote the main text passages. Finally I acted as the corresponding author. •Kenji Tanaka conducted first investigations of irregular friction occurring for the Kaijo Denki DAT 600 at BJ site. He did this work in Bayreuth during a sabbatical, his results are not published yet. •Due to his experience with the Kaijo Denki DAT 600 over decades, Thomas Foken initiated the investigation on this topic and contributes to its progress with several scientific discussions. 57 B. Gerken et al. (2012) Gerken, T., Babel, W., Hoffmann, A., Biermann, T., Herzog, M., Friend, A. D., Li, M., Ma, Y., Foken, T., and Graf, H.-F.: Turbulent flux modelling with a simple 2-layer soil model and extrapolated surface temperature applied at Nam Co Lake basin on the Tibetan Plateau, Hydrol. Earth Syst. Sci., 16, 1095–1110, doi:10.5194/hess-16-10952012, 2012. 58 Hydrol. Earth Syst. Sci., 16, 1095–1110, 2012 www.hydrol-earth-syst-sci.net/16/1095/2012/ doi:10.5194/hess-16-1095-2012 © Author(s) 2012. CC Attribution 3.0 License. Hydrology and Earth System Sciences Turbulent flux modelling with a simple 2-layer soil model and extrapolated surface temperature applied at Nam Co Lake basin on the Tibetan Plateau T. Gerken1,2, W. Babel2, A. Hoffmann1, T. Biermann2, M. Herzog1, A. D. Friend3, M. Li4, Y. Ma5, T. Foken2,6, and H.-F. Graf1 1Centre for Atmospheric Science, Department of Geography, University of Cambridge, UK 2Department of Micrometeorology, University of Bayreuth, Germany 3Department of Geography, University of Cambridge, UK 4Cold and Arid Region Environmental and Engineering Research Institute, Chinese Academy of Sciences, Lanzhou, China 5Institute of Tibetan Plateau Research, Chinese Academy of Sciences, Beijing, China 6Member of Bayreuth Center of Ecology and Environmental Research (BayCEER), University of Bayreuth, Germany Correspondence to: T. Gerken ([email protected]) Received: 19 October 2011 – Published in Hydrol. Earth Syst. Sci. Discuss.: 21 November 2011 Revised: 8 March 2012 – Accepted: 27 March 2012 – Published: 3 April 2012 Abstract. This paper introduces a surface model with two soil-layers for use in a high-resolution circulation model that has been modified with an extrapolated surface temperature, to be used for the calculation of turbulent fluxes. A quadratic temperature profile based on the layer mean and base temperature is assumed in each layer and extended to the surface. The model is tested at two sites on the Tibetan Plateau near Nam Co Lake during four days during the 2009 Monsoon season. In comparison to a two-layer model without explicit surface temperature estimate, there is a greatly reduced delay in diurnal flux cycles and the modelled surface temperature is much closer to observations. Comparison with a SVAT model and eddy covariance measurements shows an overall reasonable model performance based on RMSD and cross correlation comparisons between the modified and original model. A potential limitation of the model is the need for careful initialisation of the initial soil temperature profile, that requires field measurements. We show that the modified model is capable of reproducing fluxes of similar magnitudes and dynamics when compared to more complex methods chosen as a reference. 1 Introduction Turbulent fluxes of momentum, latent heat (QE) and sensible heat (QH) are some of the most important interactions between land surface and atmosphere. These fluxes are responsible for the development or modification of mesoscale circulations and the generation of clouds feed back on surface fluxes through the modification of solar radiation. The effects of vegetation influencing boundary layer structure and moisture are widely acknowledged (i.e. Freedman et al., 2001; van Heerwaarden et al., 2009), while the feedback from short-lived clouds is less understood, but important. Shallow cumulus-surface interactions were shown in an LES (large eddy simulation) study to impact surface temperature and fluxes on very short time scales (Lohou and Patton, 2011). For improved process understanding, it is necessary to use: (1) atmospheric models with sufficiently high resolution (O(100m)) to resolve boundary layer processes as well as clouds and (2) surface models capable of reproducing the system’s surface flux dynamics. Our research focuses on surface-atmosphere interactions on the Tibetan Plateau (TP) in the Nam Co Lake region. With more than 4700ma.s.l., a semi-arid climate and with a highly adapted Kobresia pygmea alpine steppe (Miehe et al., 2011), the TP proves to be a difficult environment for surface models (Yang et al., 2003, 2009). Specific problems Published by Copernicus Publications on behalf of the European Geosciences Union. 59 1102 T. Gerken et al.: A simple two-layer soil model with extrapolated surface temperature 0 4 8 12 16 20 0 10 20 30 40 50 60 70 80 T2 [°C] Tbase1 [°C] a) Tbase1 = Tbase2 + a2 d22 −30 −20 −10 0 10 20 30 40 50 b) T0 [°C] Tbase1 [°C] T1 [°C] 0 4 8 12 16 20 0 4 8 12 16 20 0 4 8 12 16 20 0 4 8 12 16 20 −100 −70 −70 −70 −40 −40 −40 −10 −10 −10 0 00 0 20 20 30 30 40 c) T0 [°C] T1 [°C] T2 [°C] Fig. 5. Dependency of soil temperature parameters: (a) relationship between mean temperature of layer 2 ( ¯ T2) and bottom temperature of layer 1 [Tbase1, Eq. (7)] – a2is calculated according to Eq. (6); (b) surface temperature (T0) contour plot as function of Tbase1and layer 1 mean temperature ( ¯ T1) and (c) contours of (T0) as function of ¯ T2and ¯ T1. The black rectangle at the intersection of the layer temperature ranges (yellow) indicates the theoretical parameter space given by the temperature values used in this study and the black crosses mark the actual configurations. 4.2 Statistical evaluation measures Model quality was assessed by Root Mean Square Deviation (RMSD) RMSD =v u u t1 N N X i=1Pp−Pr2 i(12) and Cross Correlation according to the coefficient of determination (R2): R2(j) = covPp(1+j:N),Pr(1:N−j) σPp(1+j:N)σPr(1:N−j) !2 (13) with R2(j) being the coefficients of determination for the predicted (Pp) and reference (Pr) flux time series shifted by jelements, the total number of elements in each time series (N) and σas their respective standard deviations. Both SEWAB and EC measurements produce 30-min flux averages, whereas Hybrid was set to 10-min averaged fluxes. Therefore the reference fluxes were linearly interpolated to Hybrid’s output times before statistical evaluation. Periods when no energy balance corrected EC measurements were available (see Figs. 6and7 for details) were excluded from the calculation of the statistical measures. 5 Results and discussion The following section presents and discusses the improvements that are achieved for a simple two-layer model when a new algorithm for the surface temperature was implemented. The original two-layer model Hybrid fails to reproduce the diurnal dynamics observed at UBT (Figs. 6and7) due to the thermal inertia of the top-layer. The delayed response in surface temperature leads to a shift in the resulting turbulent surface fluxes. This causes an underestimation of QEand QH until ∼18:00BST and later to an overestimation due to delayed surface cooling. The improvement of the modified Hybrid over the original formulation is discussed in more detail in Sects. 5.1 and 5.4. The latent (Fig. 6 – left column) and sensible heat fluxes (right column) estimated with the modified Hybrid model are generally in good agreement with the reference fluxes derived by EC and SEWAB. The diagnostic surface temperature (right column) also shows a close agreement. In some instances there remains a small shift in fluxes compared to the reference values, but this has been greatly improved compared to the original Hybrid. The surface temperatures are also in good agreement after sunrise, despite the fact that during the clear sky days in August excessive night-time surface cooling is simulated. This is less of an issue during the overcast nights. Hydrol. Earth Syst. Sci., 16, 1095–1110, 2012 www.hydrol-earth-syst-sci.net/16/1095/2012/ B. Gerken et al. (2012) 66 T. Gerken et al.: A simple two-layer soil model with extrapolated surface temperature 1103 0 100 200 300 400 500 600 a) QE UBT 10−Jul−2009 Turbulent Flux [W m−2] 03:00 06:00 09:00 12:00 15:00 18:00 21:00 −100 0 100 200 300 400 500 600 b) QH UBT −10 0 10 20 30 40 50 60 T0 [°C] 03:00 06:00 09:00 12:00 15:00 18:00 21:00 0 100 200 300 400 500 600 c) QE UBT 27−Jul−2009 −100 0 100 200 300 400 500 600 d) QH UBT −10 0 10 20 30 40 50 60 0 100 200 300 400 500 600 e) QE UBT 05−Aug−2009 LHyb,new LHyb,org WCOARE LEC LEC,EBC LSEWAB −100 0 100 200 300 400 500 600 f) QH UBT −10 0 10 20 30 40 50 60 LHyb,new LHyb,org WCOARE LEC LEC,EBC LSEWAB 03:00 06:00 09:00 12:00 15:00 18:00 21:00 0 100 200 300 400 500 600 g) QE UBT 06−Aug−2009 Time BST L+WEC WEC WHM −100 0 100 200 300 400 500 600 h) QH UBT −10 0 10 20 30 40 50 60 L+WEC WEC WHM T0,Hyb T0,Ref 03:00 06:00 09:00 12:00 15:00 18:00 21:00 Time BST Fig. 6. Model results for the modified Hybrid at UBT for 10 July 2009 (a–b), 27 July 2009 (c–d), 5 August 2009 (e–f) and 6 August 2009 (g–h). Left column: latent heat flux (QE); right column: sensible heat flux (QH) and surface temperature T0[◦C]. Land Wrefer to “land” and “water” as origin of the fluxes. L+W is the complete available time series. The subscripts Hyb,mod and Hyb,org refer to fluxes from the modified and original Hybrid and COARE are fluxes from the lake derived by TOGA-COARE whereas SEWAB is a SVAT model and HM refers to a hydrodynamic multi-layer lake model after Foken (1984) and Panin et al. (2006). EC and EC,EBC refer to measurements by eddy covariance method where in the latter the energy balance has been closed by distributing the residual according to Bowen-ratio (this requires good data quality and fluxes and can only be done for fluxes that are attributed to land). The circles indicate poor data quality of the EC system according to Foken et al. (2004). Gray shading indicates times where the flux footprint of UBT was over the lake. The situation at ITP is quite similar to UBT. The modified model agrees well with the EC and SEWAB reference data. On 5 August the turbulent flux dynamics, but not the magnitude of the fluxes, match the EC measurements closely (Fig. 7), while the original Hybrid showed a strong delay in the flux response as the soil remained frozen during the morning. While the magnitude of the latent heat flux is close to EC measurements, QHproduced by Hybrid are of a similar magnitude as QHfrom SEWAB. These are considerably larger than the fluxes measured by EC and corrected for energy balance closure. For 6 August the modelled maximum of QEis larger than the maximum QEEC,EBC and much greater throughout most of the day compared to SEWAB. QHin contrast shows similar diurnal dynamics as QHEC,EBC, but with its magnitude between the sensible heat flux derived by SEWAB and QHEC,EBC. Around 18:00h the QH-fluxes from the different methods become more similar. A large negative QH-flux in the morning hours is apparent but greatly improved compared to the unmodified Hybrid version. Figure 6a and b also highlights some limitations of ecosystem research as a large portion of the data had to be rejected due to limitations described in Sect. 4.1. During lake breeze events the surface fluxes over water derived from TOGA-COARE are displayed. Sensible heat fluxes are in close agreement with EC data and fluxes derived by a hydrodynamic multi-layer lake model (Foken, 1984; Panin et al., 2006). Latent heat fluxes show a similar behaviour and are of similar magnitude on 10 July www.hydrol-earth-syst-sci.net/16/1095/2012/ Hydrol. Earth Syst. Sci., 16, 1095–1110, 2012 67 1104 T. Gerken et al.: A simple two-layer soil model with extrapolated surface temperature 0 100 200 300 400 500 600 a) QE ITP 10−Jul−2009 Turbulent Flux [W m−2] 03:00 06:00 09:00 12:00 15:00 18:00 21:00 −100 0 100 200 300 400 500 600 b) QH ITP −10 0 10 20 30 40 50 60 T0 [°C] 03:00 06:00 09:00 12:00 15:00 18:00 21:00 0 100 200 300 400 500 600 c) QE ITP 27−Jul−2009 −100 0 100 200 300 400 500 600 d) QH ITP −10 0 10 20 30 40 50 60 0 100 200 300 400 500 600 e) QE ITP 05−Aug−2009 −100 0 100 200 300 400 500 600 f) QH ITP −10 0 10 20 30 40 50 60 03:00 06:00 09:00 12:00 15:00 18:00 21:00 0 100 200 300 400 500 600 g) QE ITP 06−Aug−2009 Time BST LHyb,new LHyb,org LEC LEC,EBC LSEWAB −100 0 100 200 300 400 500 600 h) QH ITP −10 0 10 20 30 40 50 60 LHyb,new LHyb,org LEC LEC,EBC LSEWAB T0,Hyb T0,Ref 03:00 06:00 09:00 12:00 15:00 18:00 21:00 Time BST Fig. 7. Same as Fig. 6, but for ITP. There are no contributions from the lake. and 6 August. On 5 August there is at least a qualitative agreement between COARE and EC measurements. 5.1 Discussion of turbulent fluxes The original two layer model reacts only slowly to the atmospheric forcing, delaying the fluxes’ response. Such a time lag leads to a shift in the diurnal cycle and is problematic for the coupling to atmospheric models since surface fluxes are one of the main drivers of regional and local circulation as well as cloud development. These will certainly be affected by erroneous surface flux dynamics. In our specific case, the dampening of the diurnal temperature cycle and the delay in surface fluxes may reduce the intensity of the land-lake breeze or may delay its development through a reduction of differential heating between land and lake surface. However, there is still a minor delay visible in the modified Hybrid as the surface temperature is purely diagnostic and dependent on ¯ T1. This is discussed in more detail in Sect. 5.4. Table 3 shows the results of the RMSD between the modelled results and the reference quantities. With the modified Hybridmodelthere isa 40–60%improvementin theRMSDs compared to the original Hybrid, when both are compared against SEWAB. The only notable exception for this is 6 August at ITP, where a strong deviation of turbulent fluxes derived by SEWAB and measured fluxes was encountered. This is due to an underestimation of soil water content by SEWAB as 6 August falls into a dry interval between rainy periods, where SEWAB underestimates the soil water content. The picture is more diverse for the comparison between the energy balance corrected EC fluxes and Hybrid. There is a reduction in the error for all cases, except QHon 6 August at ITP, but the reductions cover a much larger range from less than 1 to 80%. Due to data quality concerns the number of comparable elements is much lower (Ngiven in Table 3) and probably too small for meaningful statistics in case of UBT. This is especially true as the daytime lake breeze influence coincides with the times with periods of usually higher quality of EC fluxes. As flux qualities are usually lower during conditions with limited vertical exchange (stable stratification), EC fluxes at ITP mainly reflect the daytime model performance whereas the comparison with SEWAB also takes Hydrol. Earth Syst. Sci., 16, 1095–1110, 2012 www.hydrol-earth-syst-sci.net/16/1095/2012/ B. Gerken et al. (2012) 68 T. Gerken et al.: A simple two-layer soil model with extrapolated surface temperature 1105 01234 0 0.25 0.5 0.75 1 a) Hyb−EC Qh(UBT) R2(t) 10−Jul 27−Jul 05−Aug 06−Aug 01234 0 0.25 0.5 0.75 1 c) Hyb−EC QE(UBT) orig mod 01234 0 0.25 0.5 0.75 1 e) Hyb−SEWAB QH(UBT) 01234 0 0.25 0.5 0.75 1 g) Hyb−SEWAB QE(UBT) tlag [h] 01234 0 0.25 0.5 0.75 1 b) Hyb−EC QH(ITP) 01234 0 0.25 0.5 0.75 1 d) Hyb−EC QE(ITP) 01234 0 0.25 0.5 0.75 1 f) Hyb−SEWAB QH(ITP) 01234 0 0.25 0.5 0.75 1 h) Hyb−SEWAB QE(ITP) tlag [h] Fig. 8. Cross correlation R2(t) of simulated fluxes against flux reference shifted by tlag as multiples of 10 minutes for each of the four days simulated with the original and modyfied Hybrid. The maximum number of elements used in the calculation of R2for each curve can be taken from Table 3. into account the night-time, where fluxes and therefore absolute differences are smaller. The small improvement of RMSD of QHand QHEC,EBC at ITP can be explained by the fact that the modified Hybrid follows the dynamics of EC, but flux estimates are larger and of the same magnitude as fluxes calculated by SEWAB. Mauder et al. (2006) have estimated the error or EC measurements to be 5% or <10Wm2for QHand 15% or <30Wm−2for QE. Additional uncertainty is added to the measured fluxes by the lack in energy balance closure. When this is taken into account there is a significant difference between the QHHybrid and QHEC,EBC for ITP on 6 August. On 5 August (ITP) and 6 August (UBT) the deviation of fluxes may still be explained by measurement errors and by shortcomings in the energy balance closure scheme. Indeed, there is no indication to assume scalar similarity between temperature and moisture transport (Ruppert et al., www.hydrol-earth-syst-sci.net/16/1095/2012/ Hydrol. Earth Syst. Sci., 16, 1095–1110, 2012 69 1106 T. Gerken et al.: A simple two-layer soil model with extrapolated surface temperature Table 3. Root mean square deviation (RMSD) between the modelled quantities of the original and modified Hybrid and reference values. The reference quantities used are either measured by EC and corrected for energy balance closure (EC,EBC) or modelled with SEWAB for fluxes or taken from longwave outgoing radiation for T0. The values in parenthesis (N) correspond to the number of elements used for calculation of RMSD and R2(l =0)in Fig. 8. RMSD Site Date Run QEQHQEQHT0 EC,EBC [Wm−2] SEWAB [Wm−2] [◦C] UBT 10 July orig 318 117 (8) 94 74 (94) 4.3 (139) 27 July 97 58 (19) 60 59 (139) 4.5 (143) 5 August 168 139 (11) 90 64 (110) 4.3 (143) 6 August 159 84 (52) 87 71 (128) 3.7 (143) ITP 10 July orig 182 93 (25) 97 69 (143) 3.7 (143) 27 July 43 64 (72) 58 75 (143) 3.8 (143) 5 August 224 103 (64) 179 68 (143) 8.3 (143) 6 August 118 80 (52) 130 119 (143) 5.1 (143) UBT 10 July mod 214 43 (8) 51 36 (94) 2.3 (139) 27 July 79 44 (19) 32 28 (139) 2.9 (143) 5 August 93 62 (11) 36 26 (110) 3.4 (143) 6 August 78 57 (52) 39 32 (128) 3.2 (143) ITP 10 July mod 74 73 (25) 42 32 (143) 1.6 (143) 27 July 42 58 (72) 55 36 (143) 2.6 (143) 5 August 44 80 (64) 64 30 (143) 2.6 (143) 6 August 68 82 (52) 113 77 (143) 3.5 (143) UBT all orig 170 92 (90) 83 67 (471) 4.2 (568) mod 100 54 (90) 39 31 (471) 3.0 (568) ITP all orig 152 84 (213) 125 86 (572) 5.6 (572) mod 54 73 (213) 74 48 (572) 2.7 (572) 2006; Mauder et al., 2007). Therefore, additional research, such as high-resolution atmospheric modelling studies, need to be carried out in order to determine the contributions of QHand QEto the “missing” energy. It should be noted that all modelled fluxes and measurements have errors, so that there is no absolute way of knowing which method produces the best flux estimates. The incorporation of surface fluxes into a regional circulation model may give some insight into whether modelled surface atmosphere interactions lead to realistic atmospheric flow patterns. The large negative and potentially unreasonable night-time QH-fluxes that are modelled for ITP on 6 August are owed to a frozen soil and strong surface winds that lead to an overestimation of the temperature gradient, delayed reaction of the surface model and resulted in a potential underestimation of modelled surface temperatures and thus surface fluxes. 5.2 Discussion of surface temperature For surface temperature there is a notable decrease in RMSD for all cases. Additionally, the source of the error changes. In the original model the error in T0was mainly due to the time-lag and a general underestimation of daytime maximum surface temperatures. In the new model daytime T0matches a lot better with observations except for ITP 6 August, where evaporative cooling due to excessive evapotranspiration contributes to too small warming rates. In return, the cooling during the nighttime is overestimated. This may either be due to errors in soil moisture, surface emissivity () or due to the surface temperature extrapolation function used in this work. 5.3 Soil moisture variation and evapotranspiration After a 24h run the moisture content of the first model layer (last two columns of Table 2) is smaller than measurements suggest. For UBT, measured soil moisture content does hardly vary on a day to day scale and is kept well above FC due to groundwater influence. This is not reflected by the model as it lacks the capability to include groundwater tables. The true soil moisture at ITP has a much larger variation due to its low FC and comparatively large pore volume. During the dry day of 27 July the upper soil layer loses 1.5mm of water whereas during the moist days of August there is a total loss from layer one of 6.7 and 5.3mmd−1, respectively. Comparing θ1end of 5 August with θ1obs of the next day shows that the model would perform considerably worse Hydrol. Earth Syst. Sci., 16, 1095–1110, 2012 www.hydrol-earth-syst-sci.net/16/1095/2012/ B. Gerken et al. (2012) 70 T. Gerken et al.: A simple two-layer soil model with extrapolated surface temperature 1107 if it were not restarted every day. This is caused by a very limited soil hydrology included in Hybrid. Hu et al. (2008) have estimated the summer evapotranspiration on a central Tibetan grassland site to be in the order of 4–6mmd−1. An experiment conducted within the framework of TiP has estimated bare soil evaporation and evapotranspiration of a very dry soil at Kema in 2010 (∼150km northeast of Nam Co Lake) at 2mmd−1rising to at least 6mmd−1and possibly more for a vegetated Kobresia pasture during an irrigation experiment (H. Coners – University of G¨ ottingen, personal communication, 27 June 2011). Even though the soils are not directly comparable this suggests similar dynamics in QE to the ITP site. One factor likely to play a role in the local water cycle that is not included is dew fall in the early morning hours. Direct absorption of atmospheric moisture on bare soil (Agam (Ninari) and Berliner, 2004) and dew fall are often considered a significant moisture input for semi-arid environments (Agam and Berliner, 2006). Heavy dewfall in the vicinity of Nam Co Lake is frequently observed, but has, at least to our knowledge, never been quantified. This additional source of water and the associated local recycling of water may account for a significant fraction of the missing water. In addition to this, the too simplistic representation of soil hydrology is very likely responsible for the remaining water deficit in the upper layer of the soil model. 5.4 Cross correlation of turbulent fluxes A different way of looking at the model performance is cross correlation of the modelled surface fluxes against EC measurements and SEWAB (Fig. 8). These measures give an insight into the reasons for the delayed response of the surface model and the amount of flux-variance explained, but does not yield information whether the model and the reference fluxes show a true one-to-one correlation. As with RMSD the quality of the analysis is limited by the number of data points that can be correlated, which is comparatively small for the energy balance corrected EC measurements at ITP and even smaller at UBT due to lake breeze influences (Fig. 8a–e). Hence, it is very difficult to interpret the cross correlations for EC. It is probably fair to say that there is a tendency for smaller time lags during the time series with higher number of elements, notably UBT 6 August and all days of ITP and that the total explained variances are at the same level of determination, when comparing the maximum R2(j). A notable exception is ITP 5 August. For the comparison with SEWAB (Fig. 8f–h), it becomes notable that for many cases the maximum R2(j) of the modified Hybrid approach R2→1 and that their maxima are usually found at lags of 10–30min (j=1−3). Solar radiation rapidly modifies the skin temperature that is governing turbulent fluxes. As SEWAB has an instantaneous surface temperature solver for each model time step, one would expect a direct response of SEWAB to changes in solar radiation. This may even be faster than in reality, especially for QEflux that is not only dependent on the actual skin temperature, but also on the vegetation’s response. Including negative values of j into Fig. 8 would show a gradual decrease of correlations with decreasing j, showing that the flux dynamics of Hybrid never precede EC measurements or SEWAB. 5.5 Natural variability of fluxes Atmospheric quantities and turbulent surface fluxes have a large natural variability that is difficult to measure or to model. The EC approach is dependent on averaging procedures and most standard measurements will yield mean values. In order to use high-frequency measurements for flux estimation, less common techniques such as conditional sampling or wavelet-spectra have to be used. Even if models are capable of reproducing variability on realistic scales it is difficult to supply forcing data with similar resolution. The forcing data used in this study, sampled and averaged 10 or 30min means, are used for SEWAB. Running Hybrid at time steps comparable to a high-resolution mesoscale model requires interpolation of the forcing data and therefore potentially causes a smoothing of the model’s response compared to the actual weather forcing as it would be provided by a coupled model. As surface models share a similar approach to the parameterisation of surface fluxes and close the surface energy balance locally, SEWAB and Hybrid fluxes are more similar to each other than they are to field measurements. 6 Conclusions The accurate generation of surface fluxes is a necessary prerequisite for studies of surface-atmosphere interactions and local to mesoscale circulations. In order to gain a better process understanding of the interaction between atmospheric circulation, clouds, radiation and surface fluxes, the generated diurnal flux cycles have to be of realistic magnitude and without temporal shift. The original two-layer surface model without a specific formulation for T0produced both a considerable time lag and failed to capture the full diurnal dynamics due to its unresponsiveness. We have demonstrated that the introduction of an extrapolated surface temperature enables even a quite simplistic soil model to realistically simulate skin temperatures and thus to generate more realistic surface fluxes. The delay of fluxes during the daily cycle was greatly reduced, making the model usable for diurnal process studies. The total magnitude of fluxes is also much improved, when few and computationally cheap additional physically based processes are introduced. Comparing SEWAB with Hybrid, the RMSD for both fluxes and surface temperature is decreased by generally 40–60%. The improvement in quality was somewhat more varied in comparison to EC measurements, as comparison of models and measurements is not straight forward. The improved R2(j) for smaller values of jshows that temporal shifts of www.hydrol-earth-syst-sci.net/16/1095/2012/ Hydrol. Earth Syst. Sci., 16, 1095–1110, 2012 71 1108 T. Gerken et al.: A simple two-layer soil model with extrapolated surface temperature the flux time series have been greatly reduced and the overall correlations are high. As with any natural system it is impossible to obtain complete data sets that capture the full amount of natural variability. However, the modified model has been tested for a larger spectrum of environmental conditions on the TP and produced reasonable results for both dry and moist conditions. We have shown that a rather simple soil surface model can efficiently calculate turbulent fluxes at a high temporal resolution when driven by realistic atmospheric conditions. Nevertheless, it is quite clear that such an approach with extrapolated surface temperature needs careful model initialisation. The initial soil heat contents and therefore knowledge of soil temperature profiles is necessary. Due to the fact that the surface temperature in this study is a purely diagnostic quantity, there may still be some limitations such as a delayed or smoothed response to atmospheric forcing on very short timescales, such as the feedback between passing boundarylayer clouds and the surface fluxes. The influence of surface fluxes and their dynamics to regional circulation will be investigated in a future study. Acknowledgements. This research was funded by the German Research Foundation (DFG) Priority Programme 1372 “Tibetan Plateau: Formation, Climate, Ecosystems” as part of the Atmosphere – Ecology – Glaciology Cluster (TiP-AEG). ITP data was provided through CEOP-AEGIS, which is a EU-FP7 Collaborative Project/ Small or medium-scale focused research project – Specific International Co-operation Action coordinated by the University of Strasbourg, France under the call ENV.2007.4.1.4.2: “Improving observing systems for water resource management.” The authors would like to thank everyone who has contributed to the collection of field data in one of the world’s most remote regions. ADF acknowledges support from the European Community’s Seventh Framework Programme (FP7/2007-2013) under grant agreement no. 238366. AH acknowledges funding by the Fonds National de la Recherche (FNR-Luxembourg), under the grant BFR07-089 and support by the Cambridge European Trust (CET-UK). Edited by: C. de Michele References Agam (Ninari), N. and Berliner, P. R.: Diurnal Water Content Changes in the Bare soil of a coastal desert, J. Hydrometeorol., 5, 922–933, doi:10.1175/15257541(2004)005<0922:DWCCIT>2.0.CO;2, 2004. Agam, N. and Berliner, P. R.: Dew formation and water vapor adsorption in semi-arid environments – A review, J. Arid Environ., 65, 572–590, doi:10.1016/j.jaridenv.2005.09.004, 2006. Aubinet, M., Grelle, A., Ibrom, A., Rannik, U., Moncrieff, J., Foken, T., Kowalski, A., Martin, P., Berbigier, P., Bernhofer, C., Clement, R., Elbers, J., Granier, A., Gr¨ unwald, T., Morgenstern, K., Pilegaard, K., Rebmann, C., Snijders, W., Valentini, R., Vesala, T., Fitter, A., and Raffaelli, D.: Estimates of the annual net carbon and water exchange of forests: The EUROFLUX Methodology, Adv. Ecol.Res., 30, 113–175, doi:10.1016/S00652504(08)60018-5, 1999. Biermann, T., Babel, W., Olesch, J., and Foken, T.: Mesoscale circulations and energy and gas exchange over the Tibetan Plateau – Documentation of the micrometeorological experiment, Nam Tso, Tibet – 25 June–8 August 2009, Arbeitsergebnisse 41, University of Bayreuth, ISSN 1614-8616, Bayreuth, 2009. Blackadar, A.: High resolution models of the planetary boundary layer, in: Advances in Environmental Science and Engineering, edited by: Pfafflin, J. and Ziegler, E., Vol. 1, 50–85, Gordon and Breach, New York, 1979. Camillo, P. J. and Gurney, R. J.: A resistance parameter for bare-soil evaporation models, Soil Sci., 141, 95–105, 1986. Cong, Z., Kang, S., Smirnov, A., and Holben, B.: Aerosol optical properties at Nam Co, a remote site in central Tibetan Plateau, Atmos. Res., 92, 42–48, doi:10.1016/j.atmosres.2008.08.005, 2009. Cui, X., Langmann, B., and Graf, H.: Summer monsoonal rainfall simulation on the Tibetan Plateau with a regional climate model using a one-way double-nesting system, SOLA, 3, 49–52, 2007. Deardorff, J. W.: Dependence of air-sea transfer coefficients on bulk stability, J. Geoph. Res., 73, 2549–2557, 1968. Fairall, C. W., Bradley, E. F., Godfrey, J. S., Wick, G. A., Edson, J. B., and Young, G. S.: Cool-skin and warm-layer effects on sea surface temperature, J. Geophys. Res., 101, 1295–1308, 1996a. Fairall, C. W., Bradley, E. F., Rogers, D. P., Edson, J. B., and Young, G. S.: Bulk parameterization of air-sea fluxes for Tropical Ocean-Global Atmosphere Coupled-Ocean Atmosphere Response Experiment, J. Geophy. Res., 101, 3747–3764, 1996b. Foken, T.: The parameterisation of the energy exchange across the air-sea interface, Dynam. Atmos. Ocean, 8, 297–305, doi:10.1016/0377-0265(84)90014-9, 1984. Foken, T.: The energy balance closure problem: An overview, Ecol. Appl., 18, 1351–1367, 2008. Foken, T., G¨ ockede, M., Mauder, M., Mahrt, L., Amiro, B., and Munger, J.: Post field data quality control, in: Handbook of Micrometeorology: A guide for surface flux measurements and analysis, edited by: Lee, X., Massman, W., and Law, B., 181– 208, Kluwer, Dordrecht, 2004. Foken, T., Aubinet, M., Finnigan, J. J., Leclerc, M. Y., Mauder, M., and Paw U, K. T.: Results Of A Panel Discussion About The Energy Balance Closure Correction For Trace Gases, B. Am. Meteorol. Soc., 92, ES13–ES18, doi:10.1175/2011BAMS3130.1, 2011. Freedman, J. M., Fitzjarrald, D. R., Moore, K. E., and Sakai, R. K.: Boundary layer clouds and vegetation-atmosphere feedbacks, J. Climate, 14, 180–197, doi:10.1175/15200442(2001)013<0180:BLCAVA>2.0.CO;2, 2001. Friend, A. D.: Terrestrial plant production and climate change, J. Exp. Bot., 61, 1293–1309, doi:10.1093/jxb/erq019, 2010. Friend, A. D. and Kiang, N. Y.: Land surface model development for the GISS GCM: effects of improved canopy physiology on simulated climate, J. Climate, 18, 2883–2902, 2005. Friend, A. D., Stevens, A. K., Knox, R. G., and Cannell, M. G. R.: A process-based, terrestrial biosphere model of ecosystem dynamics (Hybrid v3.0), Ecol. Model., 95, 249–287, 1997. Hansen, J., Russell, G., Rind, D., Stone, P., Lacis, A., Lebedeff, S., Ruedy, R., and Travis, L.: Efficient three-dimensional global models for climate studies: Models I and II, Mon. Weather Rev., Hydrol. Earth Syst. Sci., 16, 1095–1110, 2012 www.hydrol-earth-syst-sci.net/16/1095/2012/ B. Gerken et al. (2012) 72 T. Gerken et al.: A simple two-layer soil model with extrapolated surface temperature 1109 111, 609–662, 1983. Herzog, M., Graf, H., Textor, C., and Oberhuber, J. M.: The effect of phase changes of water on the development of volcanic plumes, J. Volcanol. Geoth. Res., 87, 55–74, doi:10.1016/S03770273(98)00100-0, 1998. Hu, Z., Yu, G., Fu, Y., Sun, X., Li, Y., Shi, P., Wang, Y., and Zheng, Z.: Effects of vegetation control on ecosystem water use efficiency within and among four grassland ecosystems in China, Glob. Change Biol., 14, 1609–1619, doi:10.1111/j.13652486.2008.01582.x, 2008. Hu, Z., Yu, G., Zhou, Y., Sun, X., Li, Y., Shi, P., Wang, Y., Song, X., Zheng, Z., Zhang, L., and Li, S.: Partitioning of evapotranspiration and its controls in four grassland ecosystems: Application of a two-source model, Agric. For. Meteorol., 149, 1410–1420, doi:10.1016/j.agrformet.2009.03.014, 2009. Kanda, M., Inagaki, A., Letzel, M. O., Raasch, S., and Watanabe, T.: LES Study of the energy imbalance problem with eddy covariance fluxes, Bound.-Lay. Meteorol., 110, 381–404, doi:10.1023/B:BOUN.0000007225.45548.7a, 2004. Keil, A., Berking, J., M¨ ugler, I., Sch¨ utt, B., Schwalb, A., and Steeb, P.: Hydrological and geomorphological basin and catchment characteristics of Lake Nam Co, South-Central Tibet, Quart. Int., 218, 118–130, 2010. Kracher, D., Mengelkamp, H., and Foken, T.: The residual of the energy balance closure and its influence on the results of three SVAT models, Meteorologische Z., 18, 647–661, doi:10.1127/0941-2948/2009/0412, 2009. Kuwagata, T., Kondo, J., and Sumioka, M.: Thermal effect of the sea breeze on the structure of the boundary layer and the heat budget over land, Bound.-Lay. Meteorol., 67, 119–144, doi:10.1007/BF00705510, 1994. Li, M., Ma, Y., Hu, Z., Ishikawa, H., and Oku, Y.: Snow distribution over the Namco lake area of the Tibetan Plateau, Hydrol. EarthSyst.Sci., 13, 2023–2030, doi:10.5194/hess-13-20232009, 2009. Lohou, F. and Patton, E. G.: Land-surface response to shallow cumulus, EGU General Assembly, Vienna, Austria, 3– 8 April 2011, EGU2011-10280-1, 2011. Ma, Y., Wang, Y., Wu, R., Hu, Z., Yang, K., Li, M., Ma, W., Zhong, L., Sun, F., Chen, X., Zhu, Z., Wang, S., and Ishikawa, H.: Recent advances on the study of atmosphere-land interaction observations on the Tibetan Plateau, Hydrol. Earth Syst. Sci., 13, 1103–1111, doi:10.5194/hess-13-1103-2009, 2009. Mauder, M. and Foken, T.: Documentation and instruction manual of the eddy covariance software package TK3, Arbeitsergebnisse 46, University of Bayreuth, ISSN 1614-8916, Bayreuth, 2011. Mauder, M., Oncley, S. P., Vogt, R., Weidinger, T., Ribeiro, L., Bernhofer, C., Foken, T., Kohsiek, W., Bruin, H. A. R., and Liu, H.: The energy balance experiment EBEX-2000. Part II: Intercomparison of eddy-covariance sensors and post-field data processing methods, Bound.-Lay. Meteorol., 123, 29–54, doi:10.1007/s10546-006-9139-4, 2006. Mauder, M., Desjardins, R. L., and MacPherson, I.: Scale analysis of airborne flux measurements over heterogeneous terrain in a boreal ecosystem, J. Geophys. Res., 112, D13112, doi:10.1029/2006JD008133, 2007. Mauder, M., Foken, T., Clement, R., Elbers, J. A., Eugster, W., Gr¨ unwald, T., Heusinkveld, B., and Kolle, O.: Quality control of CarboEurope flux data – Part 2: Inter-comparison of eddy-covariance software, Biogeosciences, 5, 451–462, doi:10.5194/bg-5-451-2008, 2008. Mengelkamp, H., Warrach, K., and Raschke, E.: SEWAB – a parameterization of the surface energy and water balance for atmospheric and hydrologic models, Adv. Wat. Resour., 23, 165–175, doi:10.1016/S0309-1708(99)00020-2, 1999. Metzger, S., Ma, Y., Markkanen, T., G¨ ockede, M., Li, M., and Foken, T.: Quality assessment of Tibetan Plateau Eddy covariance measurements utilizing footprint modeling, Adv. Earth Sci., 21, 1260–1267, 2006. Miehe, G., Miehe, S., Bach, K., N¨ olling, J., Hanspach, J., Reudenbach, C., Kaiser, K., Wesche, K., Mosbrugger, V., Yang, Y., and Ma, Y.: Plant communities of central Tibetan pastures in the Alpine Steppe/Kobresia pygmaea ecotone, J. Arid Environ., 75, 711–723, doi:10.1016/j.jaridenv.2011.03.001, 2011. Oberhuber, J. M., Herzog, M., Graf, H., and Schwanke, K.: Volcanic plume simulation on large scales, J. Volcanol. Geoth. Res., 87, 29–53, doi:10.1016/S0377-0273(98)00099-7, 1998. Panin, G. N., Tetzlaff, G., and Raabe, A.: Inhomogeneity of the land surface and problems in the parameterization of surface fluxes in natural conditions, Theor. Appl. Climatol., 60, 163–178, doi:10.1007/s007040050041, 1998. Panin, G. N., Nasonov, A. E., and Foken, T.: Evaporation and heat exchange of a body of water with the atmosphere in a shallow zone, Izvestiya, Atmos. Ocean. Phys., 42, 337–352, doi:10.1134/S0001433806030078, 2006. Ruppert, J., Thomas, C., and Foken, T.: Scalar similarity for relaxed eddy accumulation methods, Bound.-Lay. Meteorol., 120, 39– 63, doi:10.1007/s10546-005-9043-3, 2006. Sellers, P. J., Mintz, Y., Sud, Y. C., and Dalcher, A.: A Simple Biosphere Model (SIB) for use within general circulation models, J. Atmos. Sci., 43, 505–531, 1986. Su, Z., Wen, J., Dente, L., van der Velde, R., Wang, L., Ma, Y., Yang, K., and Hu, Z.: The Tibetan Plateau observatory of plateau scale soil moisture and soil temperature (Tibet-Obs) for quantifying uncertainties in coarse resolution satellite and model products, Hydrol. Earth Syst. Sci., 15, 2303–2316, doi:10.5194/hess15-2303-2011, 2011. Twine, T. E., Kustas, W. P., Norman, J. M., Cook, D. R., Houser, P. R., Meyers, T. P., Prueger, J. H., Starks, P. J., and Wesely, M. L.: Correcting eddy-covariance flux underestimates over a grassland, Agric. For. Meteorol., 103, 279–300, doi:10.1016/S0168-1923(00)00123-4, 2000. van Heerwaarden, C. C., de Arellano, J. V., Moene, A. F., and Holtslag, A. A. M.: Interactions between dry-air entrainment, surface evaporation and convective boundary-layer development, Q. J. Roy. Meteorol. Soc., 135, 1277–1291, doi:10.1002/qj.431, 2009. Xue, Y., Zeng, F. J., and Schlosser, C. A.: SSiB and its sensitivity to soil properties – a case study using HAPEX-Mobilhy data, Glob. Planet. Change, 13, 183–194, doi:10.1016/09218181(95)00045-3, 1996. Yang, K., Koike, T., and Yang, D.: Surface flux parameterization in the Tibetan Plateau, Bound.-Lay. Meteorol., 106, 245–262, doi:10.1023/A:1021152407334, 2003. Yang, K., Chen, Y.-Y., and Qin, J.: Some practical notes on the land surface modeling in the Tibetan Plateau, Hydrol. Earth Syst. Sci., 13, 687–701, doi:10.5194/hess-13-687-2009, 2009. www.hydrol-earth-syst-sci.net/16/1095/2012/ Hydrol. Earth Syst. Sci., 16, 1095–1110, 2012 73 1110 T. Gerken et al.: A simple two-layer soil model with extrapolated surface temperature Yee, S. Y. K.: The force-restore method revisited, Bound.-Lay. Meteorol., 43, 85–90, doi:10.1007/BF00153970, 1988. You, Q. L., Kang, S. C., Li, C. L., Li, M. S., and Liu, J. S.: Features of meteorological parameters at Nam Co station, Tibetan Plateau, in: Annual Report of Nam Co Monitoring and Research Station for Multisphere Interactions, edited by: Nam Co Monitoring and Research Station for Multisphere Interaction, 1–8, Chinese Academy of Sciences, Beijing, 2006 (in Chinese with English abstract). Zhou, D., Eigenmann, R., Babel, W., Foken, T., and Ma, Y.: The study of near-ground free convection conditions at Nam Co station on the Tibetan Plateau, Theor. Appl. Climatol., 105, 217– 228, doi:10.1007/s00704-010-0393-5, 2011. Hydrol. Earth Syst. Sci., 16, 1095–1110, 2012 www.hydrol-earth-syst-sci.net/16/1095/2012/ B. Gerken et al. (2012) 74 C. Babel et al. (2013) Babel, W., Chen, Y., Biermann, T., Yang, K., Ma, Y., and Foken, T.: Adaptation of a land surface scheme for modeling turbulent fluxes on the Tibetan Plateau under different soil moisture conditions, submitted to J. Geophys. Res. 75 BABEL ET AL.: ADAPTATION OF A LAND SURFACE SCHEME X - 7 NamITP NamUBT (a) (b) TP version 0 10 20 30 40 50 0 10 20 30 40 50 x n = 1872 Tsfc in °C, NamITP modeled observed −200 0 100 200 300 400 −200 −100 0 100 200 300 400 x n = 1872 QG in Wm−2, NamITP observed 0 10 20 30 40 50 0 10 20 30 40 50 x n = 1585 Tsfc in °C, NamUBT modeled observed −200 0 100 200 300 400 −200 −100 0 100 200 300 400 x n = 1528 QG in Wm−2, NamUBT observed (c) (d) original version 0 10 20 30 40 50 0 10 20 30 40 50 x n = 1872 Tsfc in °C, NamITP modeled observed −200 0 100 200 300 400 −200 −100 0 100 200 300 400 x n = 1872 QG in Wm−2, NamITP observed 0 10 20 30 40 50 0 10 20 30 40 50 x n = 1585 Tsfc in °C, NamUBT modeled observed −200 0 100 200 300 400 −200 −100 0 100 200 300 400 x n = 1528 QG in Wm−2, NamUBT observed Figure 6. Observations of surface temperature and ground heat flux versus model simulations, measured parameters at NamITP (a,c) and NamUBT (b,d); simulations are displayed for TP version (a,b) and original version (c,d). 1 observed simulated −10cm −20cm −40cm −7cm −20cm −40cm NamITP 1 observed simulated −10cm −30cm −50cm −7cm −20cm −40cm NamUBT 0 5 10 15 20 dates Θ in % (a) measured parameter 0 10 20 30 40 50 dates Θ in % (b) measured parameter Jul 05 Jul 15 Jul 25 Aug 04 0 5 10 15 20 Θ in % (c) default parameter Jul 05 Jul 15 Jul 25 Aug 04 0 10 20 30 40 50 Θ in % (d) default parameter Figure 7. Timeseries of observed and simulated volumetric soil moisture content at various depths, TP version, at NamITP (a,c) and NamUBT (b,d); simulations are displayed for measured parameters (a,b) and default parameters (c,d). Note, that simulated soil depths do not match the observed soil depths in case of NamUBT. The Taylor diagram handles measures of performance which largely ignore model bias. Therefore an overview of the model bias B=ξsim −ξobs is given in Figure 4, displaying bias forsensible heat, latent heat and the sum of turbulent fluxes, for the same combinations as in Figure 3. It should be noted that the bias is calculated with EBCBo corrected observations, therefore “observed” turbulent fluxes also equal the observed available energy. The turbulent flux bias is in general large and positive at NamITP, while it is nearly vanishing at NamUBT. This can be attributed to the ground heat flux, which is discussed in the next section. In general the TP version reduces the sensible heat flux and therefore its bias as well as, to a smaller extent, the bias of latent heat. Simulations with default parameters show less bias than those with measured parameters, which is more pronounced at NamUBT than NamITP. This is caused by parameter differences of field capacity and wilting point: The high measured wilting point suppresses evapotranspiration compared to the lower default values at NamUBT. At NamITP, this effect of different field capacity and wilting point can be partly compensated by bare soil evaporation, which is even higher due to the changed parameterization in the TP version. In Figure 5 the model performance is summed up with theNash-Sutcliffe coefficient [Nash and Sutcliffe, 1970], furC. Babel et al. (2013) 82 X - 8 BABEL ET AL.: ADAPTATION OF A LAND SURFACE SCHEME ther abbreviated as NS. The NS coefficient can be similarly interpreted as the common coefficient of determination R2, but in contrast, the NS is sensitive to both correlation and bias. Again, observations are EBC-Bo corrected. In general, model simulations perform better at NamUBT than at NamITP, as a result of smaller bias (Figure 4). This can be attributed to SEWAB formulations which have not been validated for such dry conditions before, such as the scheme to calculate stomatal resistance by Noilhan and Planton [1989]. It is further shown for NamITP that the TP version performs considerably better with the sensible heat flux, and negative effects on latent heat are very small and can be neglected. This result agrees with previous studies over dry surfaces [Yang et al., 2008, 2009; Chen et al., 2010]. Furthermore, turbulent fluxes at NamUBT are not compromised by the TP version. Predictions of sensible heat flux even improve, partly at the expense of latent heat flux performance. This suggests that the implemented z0h-scheme is not limited to dry surfaces. The simulations with the default parameters perform significantly better than the ones with the measured parameters, but this difference is larger for the original version than for the TP version. Thus the TP version seems to be less sensitive to soil parameters. 3.2.2. Soil Moisture and Ground Heat Flux In order to judge if “meaningful” parameters have been set, it is important to also look into accompanying variables. Most prominent are ground heat flux and surface temperature, rendering the budget for turbulent fluxes, and soil moisture, which determines the water availability for evapotranspiration. Figure 6 compares simulations of soil heat flux and surface temperature (original and TP version, measured parameters) with observations. As expected, the drier NamITP site exhibits higher surface temperatures than NamUBT with maximum temperatures around 50 ◦C and 35 ◦C, respectively. In general, simulations reflect these observations reasonably well, except that the maximum surface temperature at NamITP in the TP version (Figure 6a) performs slightly better than the original version (Figure 6c). However, the soil heat flux is moderately or even poorly represented. It clearly explains the observed bias of turbulent fluxes for measured parameters in Figure 4: Underestimation of surface temperature (and therefore long-wave upwelling radiation) as well as underestimation of the soil heat flux lead to an overestimation of available energy at NamITP, while the overestimation of soil heat flux at NamUBT leads to an underestimation of turbulent fluxes. At NamITP, the new parameterization of thermal conductivity (TP version) shows less bias, but more scatter (Figure 6a). This can be attributed to the different model formulations for the thermal conductivity (Figure 8). Within the observed soil moisture rangeat NamITP the TP version shows 0.0 0.2 0.4 0.6 0.2 0.4 0.6 0.8 1.0 1.2 Θ in m3m−3 1 NamITP orig TP λs in Wm−1K−1 0.0 0.2 0.4 0.6 Θ in m3m−3 1 NamUBT Figure 8. Simulated thermal conductivity λsin dependence on soil moisture Θ; grey rectangles indicate the observed range of upper soil moisture for the respective station. a much larger sensitivity than the original version, therefore errors in soil moisture produce a larger scatter. Another handicap for ground heat flux simulation at NamITP might be the thermal decoupling of the upper soil layer from deeper layers under very dry conditions. At NamUBT, the TP version predicts the surface temperature accurately, but overestimates the soil heat flux, as a consequence of higher thermal conductivities within the relevant soil moisture range (Figure 8). However, the measured thermal conductivity fits well to calculated values in the TP version, therefore this overestimation is somewhat surprising. Looking at soil moisture, both measured and default parameters slightly overestimate the observed soil moisture at NamITP (Figure 7). This illustrates the large range of realistic estimates for soil properties: When using a field capacity of 0.05 and wilting point of 0.02, taken from a general soil water retention curve for sandy soil [Blume et al., 2010, p. 228], the model resembles the measured soil moisture much better (not shown). However we stick here to the parameters from the laboratory measurements for consistency. Here soil moisture and temperature was initialized by the measured profile. Additional simulations, using a 3year spin-up forcing data set, yield similar results for soil moisture at NamITP (not shown). At NamUBT, measured parameters roughly approximate the special moisture profile near the lake. Default parameters, however, show a different profile, despite the better performance for turbulent fluxes. This characteristic profile could not be realized with longer spin-up periods because of the near, but unknown, position of the groundwater table and its seasonal variation. This problem has to be solved when using the model for a longer period at such a location. Furthermore, modeling with detailed parameters for different soil layers may improve the moisture profile, as measured soil parameters differ substantially within the first 50 cm depth. 3.3. Influence of the Method for Energy Balance Closure Correction Model simulations were compared to observations corrected for energy balance closure according to the Bowen ratio (EBC-Bo). This correction assumes scalar similarity of sensible and latent heat fluxes with respect to the observed residual. Some studies suggest, however, that a large part of the residual should be attributed to the sensible heat flux [Mauder and Foken, 2006; Ingwersen et al., 2011]. Recent discussion hypothesizes an influence of secondary circulations on the energy balance closure [Foken et al., 2011, 2010; Foken, 2008a]. If secondary circulations in the longwave part of the turbulence spectra – not measured with the eddy-covariance NSQH Nash−Sutcliffe coefficient NS NSQE 0.0 0.2 0.4 0.6 0.8 1.0 NamITP NamITP NamUBT od om Td Tm od om Td Tm Figure 9. Similar to Figure 5, but for EBC-HB corrected observations. 83 BABEL ET AL.: ADAPTATION OF A LAND SURFACE SCHEME X - 9 QH σsim σobs 0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5 0.2 0.4 0.6 0.8 1 1.2 1.4 ● 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.95 0.99 Correlation ● ● ● ● QE σsim σobs Standard deviation 0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5 0.2 0.4 0.6 0.8 1 1.2 1.4 ● 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.95 0.99 Correlation ● ● ● ● Station NamITP NamUBT ●● Version, par od om Td Tm Figure 10. Normalized Taylor diagram similar to Figure 3, but for EBC-HB corrected observations. The results with EBC-Bo corrected observations are displayed in light colors for comparison. NamITP NamUBT (a) (b) EBC-Bo correction 0 100 200 300 400 500 600 0 100 200 300 400 500 600 Qh_EB[index.Qh] sim_Qh[index.Qh] ● ● ● ●● ● ● ●● ● ●● ●● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ●● ● ●● ●● ● ●● ● NS coef = 0.269 n = 623 QH in Wm−2, NamITP modelled EB−corrected obs 0 100 200 300 400 500 600 Qe_EB[index.Qe] sim_Qe[index.Qe] ● ●● ● ● ● ● ● ●● ●● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ●● ●● ● ● ● ● ● ● ● ●● ● ● ● ●● ● NS coef = 0.703 n = 623 QE in Wm−2, NamITP EB−corrected obs 0 100 200 300 400 500 600 0 100 200 300 400 500 600 Qh_EB[index.Qh] sim_Qh[index.Qh] ●● ● ● ● ● ●● ●●● ● ● ●●● ●● NS coef = 0.605 n = 81 QH in Wm−2, NamUBT modelled EB−corrected obs 0 100 200 300 400 500 600 Qe_EB[index.Qe] sim_Qe[index.Qe] ● ●● ● ●●● ● ● ●●●● ● ● ● ●● NS coef = 0.598 n = 81 QE in Wm−2, NamUBT EB−corrected obs (c) (d) EBC-HB correction 0 100 200 300 400 500 600 0 100 200 300 400 500 600 Qh_EB[index.Qh] sim_Qh[index.Qh] ● ●● ●● ● ● ● ●● ● ● ●● ●● ● ●● ● ● ●●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ●● ●● ●● ●● ● ●●● ●● ●● ●● ● ●● ● NS coef = 0.359 n = 572 QH in Wm−2, NamITP modelled EB−corrected obs 0 100 200 300 400 500 600 Qe_EB[index.Qe] sim_Qe[index.Qe] ● ●● ● ● ● ●● ●●● ●● ●● ● ●● ● ●● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●●● ● ●● ● ● ● ● ●● ● ● ● ●● ● ●● ● ● ● ● ● ●● ● ●● ●● ● NS coef = 0.677 n = 572 QE in Wm−2, NamITP EB−corrected obs 0 100 200 300 400 500 600 0 100 200 300 400 500 600 Qh_EB[index.Qh] sim_Qh[index.Qh] ● ●● ● ●● ●● ●●● ●● ●●● ●● NS coef = 0.376 n = 71 QH in Wm−2, NamUBT modelled EB−corrected obs 0 100 200 300 400 500 600 Qe_EB[index.Qe] sim_Qe[index.Qe] ●●● ● ●●● ● ● ● ●●● ● ● ● ● ● NS coef = 0.817 n = 71 QE in Wm−2, NamUBT EB−corrected obs Figure 11. Turbulent flux observations versus model simulations, measured parameters, TP version, at NamITP (a,c) and NamUBT (b,d); Observations are displayed using the EBC-Bo correction (a,b) and the EBC-HB correction (c,d). The red points indicate data with −Res > 150 W m−2(not included in the analysis). method – due to convection are the reason for the unclosed energy balance, the residual should not simply be added to the sensible heat flux. Because density is also affected by moisture Charuchittipan et al. [2013] propose a correction with the buoyancy flux QHB, further named EBC-HB QEBC−HB H=QH+fHB ·Res (5) QEBC−HB E=QE+ (1 −fHB)·Res, with (6) fHB =QH QHB =1 + 0.61Tcp λ·Bo−1(7) where cpis the air heat capacity and λis the heat of evaporation. There is a weak dependency of EBC-HB to air temperature T. For Bo = 1, more than 90 % of the residual is added to the sensible heat flux, while for Bo = 0.1 it is approximately 60 %. A general effect on model performance in terms of pattern statistics is summarized in a Taylor diagram in FigC. Babel et al. (2013) 84 X - 10 BABEL ET AL.: ADAPTATION OF A LAND SURFACE SCHEME ure 10. The EBC-HB correction correlates less with the simulations than EBC-Bo in the case of sensible heat, but the correlation is similar for latent heat fluxes. The increase of observed sensible heat fluxes due to the EBC-HB correction leads to larger “observed” standard deviations and therefore smaller normalized standard deviations (and, vice versa, the decrease of latent heat fluxes leads to larger normalized standard deviations). The NS coefficients still attribute better performance to the TP version with respect to the original version (Figure 9). But in contrast, themeasured parameters now show slightly higher NS values for NamITP than the default parameters and significantly better performance for NamUBT, which is the opposite result than is obtained from Figure 5. The EBC-HB correction in general shifts sensible heat flux bias from EBC-Bo corrected observations (Figure 4) in the negative direction and latent heat flux bias in the positive direction. This leads to the pattern change for the NS. Considering an example in detail, Figure 11 shows observed vs. simulated fluxes (TP version, measured parameters) at both sites and for both correction methods. Looking at the black dots, the simulations of sensible heat flux show more scatter with the EBC-HB corrected observations than for EBC-Bo. This problem is not so pronounced for the latent heat flux; at NamUBT simulations fit even better. The red points indicate values with residuals −Res > 150 W m−2, which were not used in the analysis. Obviously, these points were most strongly affected by the choice of the correction method, particularly because these large residuals occur nearly exclusively for low Bowen ratios (not shown). This is no surprise for NamUBT, where only wet conditions occur, but somehow unexpected for NamITP. The reason might be uncertainty in the ground heat flux calculation affecting the storage term [Leuning et al., 2012], as already mentioned in Sect. 2.4. Also Kracher et al. [2009] show with another data set, that the energy balance implementation in SEWAB roughly preserves the Bowen ratio measured by eddy-covariance. From their calculations it follows that models which determine the fluxes independently, such as TERRA [part of the “Lokalmodell” LM, Steppeler et al., 2003] and REMO [Jacob and Podzun, 1997], would follow the proposed algorithm better, but at the expense of the ground heat flux, which is not determined by these models and in fact equals the residual of their energy balance. Other familiar land surface models such as, for example, the common land model CLM [Dai et al., 2003], or the simple biosphere model SiB2 [Sellers et al., 1996] solve this step in a way similar to SEWAB. 4. Discussion and Conclusions Two data sets of energy fluxes over a wet and a dry grassland site on the Tibetan Plateau were derived for a six-week period in summer 2009. The carefully quality-checked eddycovariance measurements have been corrected for energy balance closure. Furthermore, the turbulent fluxes have been modeled with the land surface model SEWAB in its original version and a version adapted to Tibetan Plateau conditions (TP version). Both versions have been run with two different parameter sets, a default set and one with soil physical parameters from laboratory investigations of field samples. Both parameter sets – default and measured –, perform reasonably well, taking into account that no calibration algorithm has been applied. However, measured parameters do not prove to give better results; rather the opposite is true. This might be attributed to the high spatial variability of soil properties in the field, which are not adequately reflected by the taken sample. Another aspect is that a given model structure might require effective parameter values to yield optimal performance, which deviate from measured ones. Nevertheless, some field investigations will be inevitable to derive a high quality parameter set (measured or default parameters), namely soil moisture measurements and field based knowledge of the soil type, derived at least by a conventional pedological description. SEWAB has been adapted to Tibetan Plateau conditions, modifying formulations for the soil thermal conductivity, bare soil evaporation and z0h. In agreement with previous studies, this version improves predictions of sensible heat flux on the dry surface, although the original model already performs quite well. Interestingly, predictions of fluxes over the wet surface are not (or not substantially) compromised, thus the observed spatial differences in surface fluxes due to small scale heterogeneity in soil moisture could be reproduced with a consistent model version. Therefore this study suggests that the z0h-scheme by Yang et al. [2008] is not limited to dry or bare soil surfaces. Moreover, the sensitivity due to soil parameters seems to decrease with the TP version, implying more robust results. Investigations concerning the energy balance closure correction clearly show that SEWAB is more compatible with the observations corrected by preserving the Bowen ratio. However, there is no proof that this closure method reflects reality. In contrast, the recent discussion of the closure problem tends to attribute the residual more – or in total – to the sensible heat flux as the driver for advective fluxes not resolved by EC systems [Mauder and Foken, 2006; Foken, 2008a; Foken et al., 2011; Ingwersen et al., 2011; Charuchittipan et al., 2013]. While the TP version outperforms the original version for both correction methods, the ranking between the parameter sets change. In summary, this case study yields huge differences in model performance dependent on the closure method. Thus it follows that the closure issue should not be neglected when models were validated with EC measurements. Typically for case studies, this work is limited by the amount of data sets and sites enclosed, and in the first instance the results are representative for the Nam Co region. They could be strengthened by extending the investigations over a larger variety of measurement periods and sites to test the robustness of the model performance. Furthermore, the errors of observations are not explicitly taken into account. Beside the usual errors to be expected, there are some difficulties at the NamITP site due to a large gravel content in the soil and the surface as well as missing top soil temperature measurements. Firstly, this compromises quality and representativeness of the measured available energy; secondly, laboratory investigations of soil physical properties become very uncertain and difficult to carry out. Even “realistic” parameters have obviously to be considered within a range due to model and measurement uncertainty as well as soil heterogeneity. Especially for the wet site, soil properties change a lot within the vertical profile, which should be taken into account by land surface models. This study shows that a successful modeling of turbulent fluxes requires a careful model selection and parameter investigation. However, when comparing simulations with eddy-covariance derived turbulent fluxes, the performance is strongly affected by the measured energy balance closure gap. This is especially important in a sensitive region like the Tibetan Plateau, in which Kobresia pastures and alpine vegetation are supposed to be highly susceptible to climate change and livestock grazing [Miehe et al., 2008, 2011]. The selected data sets also imply that surface flux heterogeneity, generated by landscape heterogeneity, could be remarkable in a short distance. This should be taken into account when regarding flux measurements as representative for a larger area. This study prepares the SEWAB model to be usable for such issues. Acknowledgments. The authors acknowledge the staff members of the ITP Nam Co Monitoring and Research Station for 85 BABEL ET AL.: ADAPTATION OF A LAND SURFACE SCHEME X - 11 their work on the Tibetan Plateau, and Heinz-Theo Mengelkamp for making SEWAB available. The work described in this publication has been supported by the European Commission (Call FP7-ENV-2007-1 Grant nr. 212921) as part of the CEOP-AEGIS project (http://www.ceop-aegis.org/) coordinated by the University of Strasbourg. Furthermore, this work has been funded by the DFG Priority Programme 1372 “Tibetan Plateau: Formation, Climate, Ecosystems”. Dr. Yingying Chen gratefully acknowledges the support of the National Natural Science Foundation of China (Grant No. 41105003). References Agam, N., P. R. Berliner, A. Zangvil, and E. Ben-Dor (2004), Soil water evaporation during the dry season in an arid zone, J. Geophys. Res.,109(D16), D16,103–, doi:10.1029/ 2004JD004802. Alapaty, K., J. E. Pleim, S. Raman, D. S. Niyogi, and D. W. Byun (1997), Simulation of atmospheric boundary layer processes using localand nonlocal-closure schemes, J. Appl. Meteorol.,36(3), 214–233, doi:10.1175/1520-0450(1997)036h0214: SOABLPi2.0.CO;2. Balsamo, G., S. Boussetta, E. Dutra, A. Beljaars, P. Viterbo, and B. van den Hurk (2011), Evolution of land surface processes in the IFS, ECMWF Newslett.,127, 17–22. Blume, H.-P., G. W. Br¨ummer, R. Horn, E. Kandeler, I. K¨ogelKnabner, R. Kretzschmar, K. Stahr, and B.-M. Wilke (2010), Scheffer/Schachtschabel Lehrbuch der Bodenkunde, 16 ed., 569 pp., Spektrum Akademischer Verlag Heidelberg. Brilli, F., L. H¨ortnagl, A. Hammerle, A. Haslwanter, A. Hansel, F. Loreto, and G. Wohlfahrt (2011), Leaf and ecosystem response to soil water availability in mountain grasslands, Agr. Forest Meteorol.,151(12), 1731–1740, doi:10.1016/j. agrformet.2011.07.007. Charuchittipan, D., W. Babel, M. Mauder, J.-P. Leps, and T. Foken (2013), Extension of the averaging time of the eddycovariance measurement and its effect on the energy balance closure, submitted to Bound.-Lay. Meteorol. Chen, Y., K. Yang, D. Zhou, J. Qin, and X. Guo (2010), Improving the Noah land surface model in arid regions with an appropriate parameterization of the thermal roughness length, J. Hydrometeorol.,11(4), 995–1006, doi:10.1175/2010JHM1185. 1. Chen, Y., K. Yang, J. He, J. Qin, J. Shi, J. Du, and Q. He (2011), Improving land surface temperature modeling for dry land of china, J. Geophys. Res.,116(D20), D20,104–, doi: 10.1029/2011JD015921. Chen, Y., K. Yang, W. Tang, J. Qin, and L. Zhao (2012), Parameterizing soil organic carbon’s impacts on soil porosity and thermal parameters for eastern tibet grasslands, Sci. China Earth Sci.,55, 1001–1011, doi:10.1007/s11430-012-4433-0. Clapp, R. B., and G. M. Hornberger (1978), Empirical equations for some soil hydraulic properties, Water Resour. Res.,14 (4), 601–604. Dai, Y., et al. (2003), The common land model, B. Am. Meteorol. Soc.,84(8), 1013–1023, doi:10.1175/BAMS-84-8-1013. Duan, A., and G. Wu (2005), Role of the tibetan plateau thermal forcing in the summer climate patterns over subtropical asia, Clim. Dynam.,24, 793–807, doi:10.1007/s00382-004-0488-8. Foken, T. (2008a), The energy balance closure problem: An overview, Ecol. Appl.,18(6), 1351–1367. Foken, T. (2008b), Micrometeorology, 308 pp., Springer, Berlin, Heidelberg. Foken, T., et al. (2010), Energy balance closure for the LITFASS2003 experiment, Theor. Appl. Climatol.,101, 149–160, doi: 10.1007/s00704-009-0216-8. Foken, T., M. Aubinet, J. J. Finnigan, M. Y. Leclerc, M. Mauder, and K. T. Paw U (2011), Results of a panel discussion about the energy balance closure correction for trace gases, B. Am. Meteorol. Soc.,92(4), ES13–ES18, doi:10.1175/ 2011BAMS3130.1. Foken, T., R. Leuning, S. R. Oncley, M. Mauder, and M. Aubinet (2012), Corrections and data quality control, in Eddy Covariance: A Practical Guide to Measurement and Data Analysis, edited by M. Aubinet, T. Vesala, and D. Papale, Springer Atmospheric Sciences, pp. 85–131, Springer Netherlands, doi: 10.1007/978-94-007-2351-1 4. Gao, Y. X. (1985), Tibetan soil, Science Press, Beijing, pp 2–15. Gerken, T., et al. (2012), Turbulent flux modelling with a simple 2-layer soil model and extrapolated surface temperature applied at nam co lake basin on the tibetan plateau, Hydrol. Earth Syst. Sci.,16(4), 1095–1110, doi:10.5194/ hess-16-1095-2012. G¨ockede, M., C. Rebmann, and T. Foken (2004), A combination of quality assessment tools for eddy covariance measurements with footprint modelling for the characterisation of complex sites, Agr. Forest Meteorol.,127(3-4), 175–188, doi: 10.1016/j.agrformet.2004.07.012. G¨ockede, M., et al. (2008), Quality control of carboeurope flux data - part 1: Coupling footprint analyses with flux data quality assessment to evaluate sites in forest ecosystems, Biogeosciences,5(2), 433–450, doi:10.5194/bg-5-433-2008. Hong, J., and J. Kim (2010), Numerical study of surface energy partitioning on the Tibetan Plateau: comparative analysis of two biosphere models, Biogeosciences,7(2), 557–568, doi:10.5194/bg-7-557-2010. Hsu, H.-H., and X. Liu (2003), Relationship between the tibetan plateau heating and east asian summer monsoon rainfall, Geophys. Res. Lett.,30(20), 2066, doi:10.1029/2003GL017909. Hu, Z., et al. (2009), Partitioning of evapotranspiration and its controls in four grassland ecosystems: Application of a twosource model, Agr. Forest Meteorol.,149(9), 1410–1420, doi: 10.1016/j.agrformet.2009.03.014. Ingwersen, J., et al. (2011), Comparison of noah simulations with eddy covariance and soil water measurements at a winter wheat stand, Agr. Forest Meteorol.,151 (3), 345–355, doi: 10.1016/j.agrformet.2010.11.010. Jacob, D., and R. Podzun (1997), Sensitivity studies with the regional climate model REMO, Meteorol. Atmos. Phys.,63, 119–129. Kang, S., Y. Xu, Q. You, W.-A. Fl¨ugel, N. Pepin, and T. Yao (2010), Review of climate and cryospheric change in the tibetan plateau, Environ. Res. Lett.,5(1), 015,101, doi:10.1088/ 1748-9326/5/1/015101. Kracher, D., H. T. Mengelkamp, and T. Foken (2009), The residual of the energy balance closure and its influence on the results of three svat models, Meteorol. Z.,18 (6), 1–15, doi: 10.1127/0941-2948/2009/0412. Larcher, W. (2001), ¨ Okophysiologie der Pflanzen: Leben, Leistung und Streßbew¨altigung der Pflanzen in ihrer Umwelt, 6th ed., 408 pp., Verlag Eugen Ulmer Stuttgart. Leuning, R., E. van Gorsel, W. J. Massman, and P. R. Isaac (2012), Reflections on the surface energy imbalance problem, Agr. Forest Meteorol.,156(0), 65–74, doi:10.1016/j.agrformet. 2011.12.002. Liebethal, C., B. Huwe, and T. Foken (2005), Sensitivity analysis for two ground heat flux calculation approaches, Agr. Forest Meteorol.,132, 253–262, doi:10.1016/j.agrformet.2005.08.001. Liu, J., S. Kang, T. Gong, and A. Lu (2010), Growth of a highelevation large inland lake, associated with climate change and permafrost degradation in tibet, Hydrol. Earth Syst. Sci., 14(3), 481–489, doi:10.5194/hess-14-481-2010. Liu, Y., Q. He, H. Zhang, and A. Mamtimin (2012), Improving the colm in taklimakan desert hinterland with accurate key parameters and an appropriate parameterization scheme, Adv. Atmos. Sci.,29 (2), 381–390, doi:10.1007/s00376-011-1068-6. Louis, J. (1979), A parametric model of vertical eddy fluxes in the atmosphere, Bound.-Lay. Meteorol.,17, 187–202, doi: 10.1007/BF00117978. Ma, Y., O. Tsukamoto, J. Wang, H. Ishikawa, and I. Tamagawa (2002), Analysis of aerodynamic and thermodynamic parameters on the grassy marshland surface of tibetan plateau, Prog. Nat. Sci.,12, 36–40. Mauder, M., and T. Foken (2004), Documentation and instruction manual of the eddy covariance software package TK2, Work Report University of Bayreuth, Dept. of Micrometeorology, ISSN 1614-8916, 26, 42 pp. Mauder, M., and T. Foken (2006), Impact of post-field data processing on eddy covariance flux estimates and energy balance closure, Meteorol. Z.,15 (13), 597–609, doi:10.1127/ 0941-2948/2006/0167. Mauder, M., and T. Foken (2011), Documentation and instruction manual of the eddy-covariance software package TK3, Work Report University of Bayreuth, Dept. of Micrometeorology, ISSN 1614-8916, 46, 58 pp. C. Babel et al. (2013) 86 X - 12 BABEL ET AL.: ADAPTATION OF A LAND SURFACE SCHEME Mengelkamp, H. T., K. Warrach, and E. Raschke (1997), A land surface scheme for atmospheric and hydrologic models: SEWAB (Surface Energy and Water Balance), GKSS Report 97/E/69, 40pp., GKSS Geesthacht. Mengelkamp, H. T., K. Warrach, and E. Raschke (1999), SEWAB - a parameterization of the surface energy and water balance for atmospheric and hydrologic models, Adv. Water Resour., 23(2), 165–175, doi:10.1016/S0309-1708(99)00020-2. Mengelkamp, H. T., G. Kiely, and K. Warrach (2001), Evaluation of the hydrological components added to an atmospheric landsurface scheme, Theor. Appl. Climatol.,69(3-4), 199–212, doi: 10.1007/s007040170025. Miehe, G., S. Miehe, K. Kaiser, L. Jianquan, and X. Zhao (2008), Status and dynamics of the kobresia pygmaea ecosystem on the tibetan plateau, Ambio,37(4), 272–279, doi: 10.1579/0044-7447(2008)37[272:SADOTK]2.0.CO;2. Miehe, G., K. Bach, S. Miehe, J. Kluge, Y. Yongping, L. Duo, S. Co, and K. Wesche (2011), Alpine steppe plant communities of the tibetan highlands, Applied Vegetation Science, 14(4), 547–560, doi:10.1111/j.1654-109X.2011.01147.x. Mihailovi´c, D. T., R. A. Pielke, B. Rajkovi´c, T. J. Lee, and M. Jefti´c (1993), A resistance representation of schemes for evaporation from bare and partly plant-covered surfaces for use in atmospheric models, J. Appl. Meteorol.,32(6), 1038–1054, doi:10.1175/1520-0450(1993)032h1038:ARROSFi2.0.CO;2. Mihailovi´c, D. T., B. Rajkovi´c, B. Lali´c, and L. Deki´c (1995), Schemes for parameterizing evaporation from a non-plantcovered surface and their impact on partitioning the surface energy in land-air exchange parameterization, J. Appl. Meteorol.,34(11), 2462–2475, doi:10.1175/1520-0450(1995) 034h2462:SFPEFAi2.0.CO;2. M¨ugler, I., et al. (2010), A multi-proxy approach to reconstruct hydrological changes and Holocene climate development of Nam Co, Central Tibet, J. Paleolimnol.,43, 625–648, doi: 10.1007/s10933-009-9357-0. Nash, J. E., and J. V. Sutcliffe (1970), River flow forecasting through conceptual models part I – A discussion of principles, J. Hydrol.,10(3), 282–290, doi:10.1016/0022-1694(70) 90255-6. Noilhan, J., and S. Planton (1989), A simple parameterization of land surface processes for meteorological models, Mon. Weather Rev.,117(3), 536–549, doi:10.1175/1520-0493(1989) 117h0536:ASPOLSi2.0.CO;2. Rannik, U., M. Aubinet, O. Kurbanmuradov, K. K. Sabelfeld, T. Markkanen, and T. Vesala (2000), Footprint analysis for measurements over a heterogeneous forest, Bound.-Lay. Meteorol.,97(1), 137–166, doi:10.1023/A:1002702810929. Sellers, P., D. Randall, G. Collatz, J. Berry, C. Field, D. Dazlich, C. Zhang, G. Collelo, and L. Bounoua (1996), A revised land surface parameterization (SiB2) for atmospheric GCMs. part I: Model formulation, J. Climate,9(4), 676–705, doi: 10.1175/1520-0442(1996)009h0676:ARLSPFi2.0.CO;2. Steppeler, J., G. Doms, U. Schattler, H. W. Bitzer, A. Gassmann, U. Damrath, and G. Gregoric (2003), Meso-gamma scale forecasts using the nonhydrostatic model LM, Meteorol. Atmos. Phys.,82, 75–96. Su, Z., J. Wen, L. Dente, R. van der Velde, L. Wang, Y. Ma, K. Yang, and Z. Hu (2011), The Tibetan Plateau observatory of plateau scale soil moisture and soil temperature (tibet-obs) for quantifying uncertainties in coarse resolution satellite and model products, Hydrol. Earth Syst. Sci.,15(7), 2303–2316, doi:10.5194/hess-15-2303-2011. Takayabu, I., K. Takata, T. Yamazaki, K. Ueno, H. Yabuki, and S. Haginoya (2001), Comparison of the four land surface models driven by a common forcing data prepared from GAME/Tibet POP’97 products – snow accumulation and soil freezing processes –, J. Meteorol. Soc. Jpn.,79 (1B), 535–554, doi:10.2151/jmsj.79.535. Taylor, K. E. (2001), Summarizing multiple aspects of model performance in a single diagram, J. Geophys. Res.,106(D7), 7183–7192, doi:10.1029/2000JD900719. Twine, T. E., W. P. Kustas, J. M. Norman, D. R. Cook, P. R. Houser, T. P. Meyers, J. H. Prueger, P. J. Starks, and M. L. Wesely (2000), Correcting eddy-covariance flux underestimates over a grassland, Agr. Forest Meteorol.,103(3), 279–300, doi: 10.1016/S0168-1923(00)00123-4. van der Velde, R., Z. Su, M. Ek, M. Rodell, and Y. Ma (2009), Influence of thermodynamic soil and vegetation parameterizations on the simulation of soil temperature states and surface fluxes by the Noah LSM over a Tibetan Plateau site, Hydrol. Earth Syst. Sci.,13(6), 759–777, doi:10.5194/ hess-13-759-2009. Wallace, J., I. Wright, J. Stewart, and C. Holwill (1991), The sahelian energy balance experiment (SEBEX): Ground based measurements and their potential for spatial extrapolation using satellite data, Adv. Space Res.,11(3), 131–141, doi: 10.1016/0273-1177(91)90413-E. Wilczak, J., S. Oncley, and S. Stage (2001), Sonic anemometer tilt correction algorithms, Bound.-Lay. Meteorol.,99, 127–150, doi:10.1023/A:1018966204465. Xu, J., and S. Haginoya (2001), An estimation of heat and water balances in the Tibetan Plateau, J. Meteorol. Soc. Jpn., 79(1B), 485–504, doi:10.2151/jmsj.79.485. Yanai, M. H., C. F. Li, and Z. S. Song (1992), Seasonal heating of the Tibetan Plateau and its effects on the evolution of the Asian Summer Monsoon, J. Meteorol. Soc. Jpn.,70(1B), 319–351. Yang, K., and J. Wang (2008), A temperature predictioncorrection method for estimating surface soil heat flux from soil temperature and moisture data, Sci. China Ser. D,51, 721–729, doi:10.1007/s11430-008-0036-1. Yang, K., T. Koike, and D. Yang (2003), Surface flux parameterization in the Tibetan Plateau, Bound.-Lay. Meteorol.,116, 245–262, doi:10.1023/A:1021152407334. Yang, K., T. Koike, B. Ye, and L. Bastidas (2005), Inverse analysis of the role of soil vertical heterogeneity in controlling surface soil state and energy partition, J. Geophys. Res.,110 (D8), D08101, doi:10.1029/2004JD005500. Yang, K., T. Koike, H. Ishikawa, J. Kim, X. Li, H. Liu, J. Wang, S. Liu, and Y. Ma (2008), Turbulent flux transfer over baresoil surfaces: Characteristics and parameterization, J. Appl. Meteorol. Clim.,47, 276–290, doi:10.1175/2007JAMC1547.1. Yang, K., Y.-Y. Chen, and J. Qin (2009), Some practical notes on the land surface modeling in the Tibetan Plateau, Hydrol. Earth Syst. Sci.,13(5), 687–701, doi:10. 5194/hess-13-687-2009. Yang, K., X. Guo, and B. Wu (2010), Recent trends in surface sensible heat flux on the tibetan plateau, Sci. China Earth Sci., pp. 1–10, doi:10.1007/s11430-010-4036-6. Ye, D.-Z., and G.-X. Wu (1998), The role of the heat source of the Tibetan Plateau in the general circulation, Meteorol. Atmos. Phys.,67(1), 181–198, doi:10.1007/BF01277509. Zhang, X. (2012), Improvement of a soil-atmosphere-transfer model for the simulation of bare soil surface energy balances in semiarid areas, Asia Pac. J. Atmos. Sci.,48(1), 97–105, doi:10.1007/s13143-012-0009-3. Zhou, D., R. Eigenmann, W. Babel, T. Foken, and Y. Ma (2011), The study of near-ground free convection conditions at nam co station on the tibetan plateau, Theor. Appl. Climatol.,105, 217–228, doi:10.1007/s00704-010-0393-5. W. Babel, Abt. Mikrometeorologie, Universit¨at Bayreuth, 95440 Bayreuth, Germany. (wolfgang.bab[email protected]) T. Biermann, Abt. Mikrometeorologie, Universit¨at Bayreuth, 95440 Bayreuth, Germany. Y. Chen, Institute of Tibetan Plateau Research, Chinese Academy of Sciences, 4A Datun Road, Chaoyang District, Beijing 100101, China. T. Foken, Abt. Mikrometeorologie, Universit¨at Bayreuth, 95440 Bayreuth, Germany. Y. Ma, Institute of Tibetan Plateau Research, Chinese Academy of Sciences, 4A Datun Road, Chaoyang District, Beijing 100101, China. K. Yang, Institute of Tibetan Plateau Research, Chinese Academy of Sciences, 4A Datun Road, Chaoyang District, Beijing 100101, China. 87 D. Biermann et al. (2013) Biermann, T., Babel, W., Ma, W., Chen, X., Thiem, E., Ma, Y., and Foken, T.: Turbulent flux observations and modelling over a shallow lake and a wet grassland in the Nam Co basin, Tibetan Plateau, Theor. Appl. Climatol., accepted 88 1 Turbulent flux observations and modelling over a shallow lake and a wet grassland in the Nam Co basin, Tibetan Plateau Tobias Biermann 1 , Wolfgang Babel 1 , Weiqiang Ma 2 , Xuelong Chen 3 , Elisabeth Thiem 4 , Yaoming Ma 5 , Thomas Foken 1,6 5 1 Department of Micrometeorology, University of Bayreuth, Bayreuth, Germany 2 Cold and Arid Regions Environment and Engineering Research Institute, Chinese Academy of Sciences, Lanzhou, China 3 Faculty of Geo-Information Science and Earth Observation, University of Twente, Enschede, The Netherlands 10 4 Department of Geography, Ludwig-Maximilian University, Munich, Germany 5 Laboratory of Tibetan Environment Changes and Land Surface Processes, Institute of Tibetan Plateau Research, Chinese Academy of Sciences, Beijing, China 6 Member of Bayreuth Center of Ecology and Ecosystem Research (BayCEER) 15 Corresponding author address: Tobias Biermann, Abt. Mikrometeorologie, Universität Bayreuth, Universitätsstr. 30 , 95440 Bayreuth, Germany Phone: +49 921 552180 20 Fax: +49 921 552366 E-mail: [email protected] 89 2 Abstract The Tibetan Plateau plays an important role in the global water cycle and is strongly influenced by climate change. While energy and matter fluxes have been more intensely studied over land 25 surfaces, a large proportion of lakes have been either neglected or parameterised with simple bulk approaches. Therefore turbulent fluxes were measured over wet grassland and a shallow lake with a single eddy-covariance complex at the shoreline in the Nam Co basin in summer 2009. Footprint analysis was used to split observations according to the underlying surface, and two sophisticated surface models were utilised to derive gap-free time series. Results were then compared with 30 observations and simulations from a nearby eddy-covariance station over dry grassland, yielding pronounced differences. Observations and footprint integrated simulations compared well, even for situations with flux contributions including grassland and lake. It is shown that the accessibility problem for EC measurements on lakes can be overcome by combining standard meteorological measurements at the shoreline with model simulations, only requiring representative estimates of 35 lake surface temperature. Keywords: Eddy-covariance, Lake, Nam Co, Surface modelling, Tibetan Plateau, Turbulent fluxes D. Biermann et al. (2013) 90 3 1. Introduction 40 The role of the Tibetan Plateau in the global water cycle and its reaction to climate change has become a topic of strong scientific interest (e.g. Immerzeel et al. 2011, Ni 2011). Representing a unique geological formation, the Tibetan Plateau is considered the largest and highest plateau on earth, with an average elevation greater than 4000 m a.s.l.. Furthermore, the Tibetan Plateau is the source of a 45 large number of major rivers in Asia. Its role in the modulation of the Asian Monsoon and the climate for large parts of Asia, due to its heat budget caused by its elevation in conjunction with the bordering Himalayan mountain range, has been of major research interest (Molnar et al. 2010, Boos and Kuang 2010). To understand the role of the Tibetan Plateau for the global heat and water budget, 50 much effort has been put into the estimation of energy balance and turbulent flux measurements within international campaigns like GAME/Tibet (GEWEXGlobal Energy and Water cycle Experiment Asian Monsoon Experiment) and (CAMP -Coordinated Enhanced Observing Period Asia-Australia Monsoon Project) (Xu and Haginoya 2001; Ma et al. 2003; Ma et al. 2005) and in the 55 framework of the Tibetan Observation and Research Platform, TORP (Ma et al. 2009). Despite these efforts, observations on the Tibetan Plateau are sparse due to its remote location (Frauenfeld et al. 2005, Kang et al. 2010, Maussion et al. 2011). The importance of evaporation for the hydrological cycle under the influence of 60 climate change has been highlighted by Yang et al. (2011). Most long-term observation stations focus on the major land cover types such as alpine steppe, Kobresia pastures and wetlands (Zhao et al. 2010), however approximately 45.000 km² of the plateau are covered by lakes (Xu et al. 2009). This lake area has been subject to changes in the last decades, the reasons are not 65 well understood due to lack of observational data (Xu et al. 2009). Although Huang et al. (2008) report a general decrease of lake volume in Qinghai-Tibet Plateau, the Nam Co lake area has been increasing (Liu et al. 2010, Wu and Zhu 2008, Zhu et al. 2010). They attribute this change to increasing precipitation as well as thawing permafrost and glacial melt due to rising mean annual 70 temperatures , nevertheless the relative contribution of the balance components, especially the role of evaporation, is discussed controversially. Consequently fluxes over lake surfaces on the Tibetan Plateau should not be neglected, since various studies have shown the contribution of lakes to the regional energy balance and water cycle in different catchments around the world (Rouse et al. 75 2005, Nordbo et al. 2011). Until now, estimations of evaporation over lake surfaces on the Tibetan Plateau have been modelled using remote sensing or land surface observations as forcing (Xu et al. 2009, Haginoya et al. 2009), whereas no direct measurements of turbulent fluxes over a lake surface have been conducted so far. The installation of a flux station in a lake on the Tibetan Plateau is nearly 80 impossible, due to problems of accessibility, strong winds and waves during the summer, as well as ice cover during winter. Nevertheless, it is known from model estimations that evaporation over lake surfaces differs from evapotranspiration over land throughout the year due to the heat storage capacity of the lakes, and has a strong effect on convection and thus 85 on local climates (Haginoya et al. 2009). The landscape on the Tibetan Plateau is fairly heterogeneous, including alpine steppe, Kobresia pygmea mats, wetlands and open water surfaces in various sizes. Therefore high quality evaporation measurements over water surfaces on the Tibetan Plateau need to be considered 91 10 Table 2. Governing equations for the hydrodynamic multilayer (HM) model (Foken 1979, 1984) with shallow water extension (Panin and Foken 2005) Variable/component Equation Sensible heat flux     Γ       with Γ  ! " # $ " % & ! " Pr   ) * " + , - . 5 . ln 1 $ "  23 4 5   Latent heat flux Analogue to    , assuming + , - 6 + 7 - , Δ  - 6 Δ 9 - , and replacing Pr with Sc Stability dependence Monin-Obukhov Similarity Theory, universal function after Foken and Skeib (1983) Shallow water term   , ; <=    , ;   "  1 . >  , ; <= " ? @   with mean square wave height ? 6 0 . 07 #  B  C@ " D #  B E 3 . ) " C  (Davidan et al., 1985) and >  , ; <= 6 2 (Panin et al. 2006b) Symbols C gravity acceleration [ms - 2 ] @ lake depth [m] ? mean square wave height [m] > ,; <= empirical correction factor [-] FG Prandtl number [-]  ,;  sensible (H) and latent (L) heat flux without shallow water correction [Wm -2 ]  ,; <= sensible (H) and latent (L) heat flux with shallow water correction [Wm -2 ] 9 specific humidity [-] HI Schmid number [-]  temperature [K]  - dimensionless temperature [-] # $ friction velocity [ms -1 ] #  wind velocity in height z [ms -1 ]  measurement height [m] Γ profile coefficient [ms -1 ] + , - dimensionless thickness of the molecular temperature boundary layer [-] ! von Kármán constant [-] J kinematic viscosity [m 2 s -1 ] 2.3.1. Description of the models used For the lake surface a hydrodynamic multilayer model (HM) by Foken (1979, 300 1984) was utilised. In order to account for multiple layers within the surface layer, turbulent fluxes are parameterised in HM using an integrated profile coefficient. As opposed to a single bulk coefficient the integrated profile coefficient resolves the molecular boundary layer, the viscous buffer layer, and the turbulent layer. Therefore near-surface exchange conditions are reflected according to 305 hydrodynamic theory. Originally designed for exchange over the ocean, a correction term for shallow water (Panin and Foken 2005) was added, resulting in increased turbulent fluxes due to an enhanced mixing by higher waves in shallow water. The model has been successfully applied to simulate fluxes above ocean surfaces and lakes with a large fetch as well as over arctic snow fields (Panin et al. 310 2006b, Foken 1986, Lüers and Bareiss 2010). Details of the governing equations can be found in Table 2. D. Biermann et al. (2013) 98 11 Turbulent fluxes over the land surface were simulated with the one-dimensional Surface Energy and WAter Balance scheme (SEWAB, Mengelkamp et al. 1999, 315 2001), a soil-vegetation-atmosphere-transfer model. All energy balance components are given separately. Turbulent fluxes are formulated with bulk approaches, atmospheric stability is considered. The main features are summarized in Table 3. The energy balance is then closed by iteration of the surface temperature. Evapotranspiration from vegetation is calculated with a 320 single leaf concept in a Jarvis-type scheme after Noilhan and Planton (1989). Emphasis is placed on the description of soil processes. Soil temperature distribution and vertical soil water movement are described by the diffusion equation and the Richards equation, respectively. Soil moisture characteristics are inter-related following Clapp and Hornberger (1978). 325 Both models were forced with standard meteorological in-situ measurements. In order to provide gap-free input data, the small gaps within the forcing data from NamUBT were filled by linear interpolation while larger gaps were filled by linear regression using the data from NAMORS. 330 2.3.2. Application of the HM model to a shallow lake The forcing data set for the HM model includes the standard meteorological parameters wind velocity, air temperature, humidity and air pressure. Radiation measurements are not required for the HM model, instead water surface temperature has to be supplied instead. In this study we used the measured water 335 temperature (Table 1) as an estimate for the water surface temperature. Wendisch and Foken (1989) investigated the relative error contribution of model parameter, amongst others water temperature, air temperature, air humidity and wind velocity, to the model output with a sensitivity analysis (Fourier Amplitude 340 Sensitivity Test by Cukier et al., 1978). Assuming typical measurement errors for the initial parameter distribution they estimated that water temperature contributed up to 50% of the overall error while the influence of wind velocity, air temperature and humidity are comparatively small, each contributing 10-20% to the error. Since the temperature probe was only shielded against direct 345 (downward) radiation, the effect of diffuse radiation on the accuracy of the water temperature measurements was evaluated. The radiation error has been estimated with a graphical analysis of short term temperature perturbations as related to rapid changes in downward shortwave radiation. Caused by the small fraction of diffuse radiation in the low air density of the Tibetan Plateau and sudden cloud 350 cover changes, the shortwave radiation observations occasionally drop from 1000 Wm -2 to 150 Wm -2 (or increase in reverse) within a few minutes. The corresponding shifts in water temperature suggest a possible radiation error of approximately 0.2 K. Therefore water temperature measurements have been accepted for model forcing without correction. 355 The shallow water parameterisation included in the current version of the HM model accounts for an enhanced turbulent exchange due to increased wave heights in shallow water. Consequently, the turbulent fluxes increase with the mean square wave height (Table 2). Together with the wave height parameterisation 360 after Davidan et al. (1985), additional parameters influence the model results. These are the wind velocity, lake depth and an empirical coefficient, which was 99 12 set to 2 in this study following Panin et al. (2006b). In this study the water depth has been estimated as 1.5 m within the average footprint area of the measurement period. 365 Table 3. Governing equations for SEWAB (Mengelkamp et al. 1999, 2001) and adaptations to the Tibetan Plateau as used in Babel et al. (2013). Variable/component Equation Net radiation K L   K MN D 1  O E  K PMN . QR   S K MN and K PMN in forcing data set Ground heat flux  T  U       <  " Δ  <    Sensible heat flux    V  W I X # D  E      D  E Latent heat flux Composed of bare soil Y  , wet foliage Y  and plant transpiration Y LZ after Noilhan and Planton (1989) Y  V ; W#DE[9    9DE Y   V ; W# D  E  9      9 D  E Y LZ  D K   K  E   W  9      9 D  E Stability dependence V  after Louis (1979), V ;  V  Adaptations to TP: I) Soil thermal conductivity U  D \ E  U NZ] .  U L  U NZ] exp a > , D 1  Θ L / Θ E d , > ,  0 . 36 (Yang et al. 2005) II) thermal roughness length  3 f  70 J " # $   " exp D  g # $ 3 . h |  $ | 3 . Bh E g  7 . 2 s 0.5 m -0.5 K -0.25 (Yang et al. 2008) III) bare soil evaporation [  j 1  & 1  k k lm * B , Θ  Θ no 1 , Θ  Θ no p (Mihailović et al. 1993) Symbols O albedo [-] V  Stanton number [-] V ; Dalton number [-] I X air heat capacity [J kg -1 K -1 ] 9 specific humidity [-] 9  saturation specific humidity [-] K  turbulent atmospheric resistance [s m -1 ] K  stomata resistance [s m -1 ] K PMN long wave downward radiation [Wm -2 ] K MN short wave downward radiation [Wm -2 ]  temperature [K]  $ dynamic temperature scale [K]   surface temperature [K]  < temperature in first soil layer [K] # $ friction velocity [ms -1 ]  measurement height [m] Δ < thickness of first soil layer [m] [ dependence factor of soil air humidity to soil water content [-] Q emissivity [-] Θ volumetric soil water content [-] Θ L volumetric soil water content at saturation, porosity [-] Θ no volumetric soil water content at field capacity [-] U  soil thermal conductivity [Wm -1 K -1 ] U NZ] soil thermal conductivity for dry soil [Wm -1 K -1 ] U L soil thermal conductivity, soil moisture at saturation [Wm -1 K -1 ] W air density [kg m -3 ] R Stefan Boltzmann constant [Wm -2 K -4 ] J kinematic viscosity [m 2 s -1 ] D. Biermann et al. (2013) 100 13 The influence of the shallow water term becomes dominant with increasing wind velocity and decreasing water depth. Calculation of the shallow water equations described in Table 2 with the estimated water depth of 1.5 m and the average wind 370 velocity of 4 ms -1 yields an increase in turbulent fluxes of 14.5 % compared to deep water conditions. Consideration of small changes in wind velocity and water depth yields local sensitivities of roughly 2.9% of the deep water fluxes per ms -1 and -3.9% per m water depth, respectively. For high wind velocities (10 ms -1 ) the shallow water extension causes an increase of 30.1% with sensitivities of 2.4% 375 per ms -1 and -8.0% per m water depth. Assuming a typical error of 0.3 ms -1 for wind velocity and variability of the lake depth up to 1 m within the footprint leads to flux uncertainties of 1% and 4%, respectively. These errors, although not negligible, are within the uncertainty range of the EC flux measurements. 2.3.3. Adaptation of SEWAB 380 On the Tibetan Plateau a strong diurnal cycle of the surface temperature during dry periods over bare soil and short grassland have been observed, which typically leads to an overestimation of surface sensible heat flux (Yang et al. 2009, Hong and Kim 2010). To account for these conditions, SEWAB has been adapted for the Tibetan Plateau by (i) a revised calculation of the soil thermal conductivity as 385 used in Yang et al. (2005), (ii) a different formulation of the thermal roughness length after Yang et al. (2008) and (iii), by changing the parameterisation of bare soil evaporation according to Mihailović et al. (1993). The formulations can be seen in Table 3. These changes have been implemented using flux data from NamITP station and 390 flux data from NamUBT corresponding to land surface (Babel et al. 2013), who evaluated this adaptation as an improvement compared to the original version. SEWAB has been run offline, forced by measurements of precipitation, air temperature, wind velocity, air pressure, relative humidity and downwelling 395 shortwave and longwave radiation. The respective parameters for both land surface types were estimated by a combination from the in-situ measurements and laboratory investigation of soil characteristics (Chen et al., 2012). Surface emissivity, leaf area index and minimum stomatal resistence have been derived from various sources (Yang et al. 2009, Hu et al. 2009, Alapaty et al. 1997) 400 2.4. Statistics For evaluation of model performance, simple comparisons were carried out using the bias q  ∑DF s  t s E u vw and the mean absolute error xyY q  ∑|F s  t s | u vw , with O as the observations and P the model predictions. In equivalence to the MAE the differences between two time series of predictions 405 can be quantified and we define the desired measure as + sz { q   | | F  , s  F B , s | u v w  (1) with P 1 and P 2 as predictions from the respective land use types 1 and 2. The Nash-Sutcliffe coefficient NS serves as a goodness of fit measure qH  1  ∑ D F s  t s E B } v w  ∑ D t s  t ~ E B } v w  (2) with t ~ as the mean of the observations. 101 14 410 Fig. 5 Mean diurnal energy fluxes for the whole measurement period, separated for (a) grass - , (b) grass + and (c) lake; for land surfaces, all components are measured; lake fluxes: the net radiation is calculated from measured downwelling radiation and using an albedo of 0.06 and the lake surface temperature with an emissivity of 0.96; the lower panel shows diurnal surface and air temperature. 415 The time axis is displayed in Beijing standard time (CST), mean local solar noon during the observation period is at 1400 CST 3. Results 3.1. Flux measurements 420 The measured energy fluxes over the lake surface and land (grass + ) show pronounced differences in their magnitude and dynamics. The daytime net radiation is substantially higher over the lake surface, caused by a lower albedo and decreased upwelling longwave radiation due to damped surface temperatures over the lake (Fig. 5c). However, upward radiation components were only 425 measured over the land surface; for the lake surface they were parameterised using an albedo of 0.06 and the lake surface temperature with an emissivity of 0.96. As expected for the monsoon season on the Tibetan Plateau, the latent heat flux 430 over the land surface was larger than the sensible heat flux (Fig. 5a,b). This observation is in agreement with e.g. Gu et al. (2005) and Ma and Ma (2006). The mean diurnal cycles of surface and air temperature also show the typical dynamics above land surface, with unstable stratification during daytime but higher surface temperatures are observed over grass - . Ground heat flux and sensible heat flux are 435 in the same order of magnitude for each land surface again with higher values for grass - . In consequence the latent heat flux is lower over this land surface. The turbulent fluxes over the lake, however, do not show a diurnal cycle, but remain constant over the day. The energy input from radiation is stored in the lake 440 body and is available at any time as indicated by the lake surface temperature in D. Biermann et al. (2013) 102 15 Table 4: Model performance of turbulent fluxes for the three land use types and two energy balance correction methods for the land observations: n (number of observations), bias, MAE (mean absolute error), offset and slope from linear regression (mean geometric regression) as well as NS (Nash-Sutcliffe coefficient) and R 2 . 445 Flux Land use EBC n Bias MAE Offset Slope NS R² [Wm - 2 ] [Wm - 2 ] [Wm - 2 ] [-] [-] [-] Q H grass - Bo 627 52.2 55.5 36.8 1.13 0.26 0.80 HB 572 38.3 55.5 36.9 1.01 0.36 0.62 grass + Bo 81 18.5 23.8 17.2 1.02 0.61 0.78 HB 71 -24.4 40.5 12.9 0.7 0.38 0.52 lake - 327 -2.7 7.6 5.3 0.72 0.75 0.79 Q E grass - Bo 627 -10.8 45.3 -8.6 0.99 0.70 0.73 HB 572 -0.4 42.0 -14.3 1.1 0.68 0.74 grass + Bo 81 -28.6 50.8 13.5 0.77 0.60 0.69 HB 71 1.7 23.4 5.4 0.98 0.82 0.82 lake - 392 -23.3 30.3 -8.5 0.9 0.50 0.64 Fig. 5c. No complete energy balance could be estimated over the lake surface, as no measurements exist for the heat storage in the water body and heat fluxes into the sediment. 450 Evaporation is comparably high for lake surfaces due to high wind velocities of 4 ms -1 on average. In addition, high lake surface temperatures, caused by the shallow water table and the small extent of this lake, lead to unstable stratification even during daytime (Fig. 5c). Therefore turbulent exchange is enhanced compared to stable stratification typically found over lake surfaces during daytime 455 (e.g. Beyrich et al. 2006, Nordbo et al. 2011). 3.2. Model performance Different measures for model performance are summarised in Table 4. The results from grass + simulations show reasonable performance, although there are only few observations left after filtering, separation and energy balance closure 460 correction (Fig. 6). In case of EBC-Bo corrected observations, the latent heat flux is slightly underestimated while the simulation of the sensible heat flux resembles the measurements quite well. The opposite is true when using the buoyancy flux method for energy balance closure correction (EBC-HB). The correlation is affected in a similar way. Good R 2 values are achieved for the sensible heat flux 465 corrected using EBC-Bo and latent heat flux corrected using EBC-HB, and they decrease for the other two cases. Lake surface modelling yields reasonable coherence to the EC observations within the footprint of the measurements, with a bias of -23.3 W m -2 and -2.7 W m -2 for the latent heat flux Q E and the sensible heat flux Q H , respectively. For grass - at NamITP, mean absolute errors for 470 sensible heat flux are larger, mainly caused by the bias, although a good correlation is obtained in the case of EBC-Bo corrected observations. Aside from model deficiencies, the reason for the remaining bias can be attributed to uncertainties in estimation of the observed ground heat flux due to high gravel content in the soil and a lack of temperature measurements in the topmost soil 475 layer. 103 16 Fig. 6 Scatterplots of modelled and observed turbulent fluxes. For land surface flux observations (grass + ) vs. SEWAB model simulations, observations are energy balance corrected with the 480 Bowen ratio method (a,b) and with the buoyancy method (c,d). Turbulent fluxes without EBC correction over lake vs. HM model runs (e,f). Model performance is indicated with the NashSutcliffe coefficient (NS coef), bias and the squared Pearson correlation coefficient. Red crosses indicated data excluded due to residuals K•€150 •‚ B 485 D. Biermann et al. (2013) 104 17 3.3. Footprint and spatial integration In the previous section we have shown that eddy-covariance measurements, selected according to their footprint as pure fluxes from each surface type, can be represented by SEWAB in case of grassland and by the HM model in case of the lake surface. However, a part of the measurements show contributions from more 490 than one land use type as well. The footprint concept enables us to link the simulations even with such observations. For each time step, the footprint approach provides the relative contribution of all involved surfaces to the measured fluxes. The simulations are then related to the observations by calculating a weighted mean from the output of both models according to the 495 actual land use contribution. This is shown with the footprint integrated simulations for lake and grass + together with the EC observations at NamUBT in Fig. 7 for three different situations: 17 July – changing conditions under moderate wind velocities, 5 August – typical day with land - lake circulation and moderate winds of about 2-6 ms -1 , and 6 August – situation with larger than average wind 500 speeds of about 6 ms -1 . In all selected situations the eddy-covariance measurements can be closely modelled by the footprint integrated simulation. This also holds for measurements with contributions from both surfaces, seen in some events on 17 July and 5 August. Due to the different exchange of each surface with its atmosphere over the course of the day, instantaneous turbulent 505 fluxes can show differences of up to 200 Wm -2 . The performance of the footprint integrated simulation is displayed in Fig. 8 for the whole period. Situations with contributions from both surface types larger than 20% (misc) are highlighted. The simulations for such situations follow the same pattern as simulations for the pure surface types (contribution of a single surface greater than 80%). Miscellaneous 510 footprints, however, did not occur for situations with very high fluxes. 3.4. Flux heterogeneity at Nam Co It is well known that heterogeneous surfaces affect the landscape scale fluxes. The presented measurements have shown that the fluxes over land and lake surfaces behave differently. To consider the most abundant surfaces near Nam Co station, 515 grass - at NamITP has also been included in addition to the observations and simulations of grass + and lake at NamUBT. Fig. 9 shows the mean diurnal cycles of measured fluxes, corrected with EBC-Bo for land surfaces, and modelled fluxes. The model simulations resemble the observed characteristics of the different surfaces in a reasonable sense. Since simulations overestimated the 520 sensible heat flux for both land surfaces, the differences between land use types were maintained. The obvious differences in characteristics of the investigated surfaces, especially between land and lake, are also reflected in mean fluxes for the whole period (Fig. 525 10). The two land surface types already differ in the longwave radiation balance. As expected, the mean latent heat flux became more dominant with increasing soil moisture for the land surfaces. The evaporation over the small lake is even higher, due to its shallow water table resulting in comparatively high surface temperatures. Mean differences of sensible and latent heat flux between grass + 530 and grass - are 24.0 Wm -2 and -33.5 Wm -2 , respectively, and between grass + and lake are -27.3 Wm -2 and 22.3 Wm -2 , respectively. 105 18 Fig. 7 Source weight integrated modelled fluxes at NamUBT, 17 July (a,b), 5 August (c,d), 6 August (e,f). Displayed are simulated fluxes with SEWAB (dashed line) and HM (solid grey line) 535 and integrated simulations (solid black line) according to contributions of lake or land within the footprint. Observations (not energy balance corrected) are shown as black circles. The land use contribution in % is indicated as bar plot, with upwind situations from the land in green and upwind situations from lake in blue. The time axis is displayed in Beijing standard time (CST), mean local solar noon during the observation period is at 1400 CST 540 D. Biermann et al. (2013) 106 19 Fig. 8 Observations vs. footprint integrated model simulations at NamUBT. Three classes of data 545 are presented, cases with a greater contribution than 80% of one land use type are considered as representative. The misc cases contain all fluxes with a contribution less than 80% for both land use types. Since no EBC correction could be performed for the lake data, all data is shown without correction for better inter comparison 550 Fig. 9 Mean diurnal cycles for the whole measurement period. Observed fluxes (corrected with EBC-Bo) are denoted by black solid lines, the horizontal bars indicate the respective standard deviation; grey lines show the modelled fluxes with standard deviations given by the grey shaded area. The time axis is displayed in Beijing standard time (CST), mean local solar noon during the 555 observation period is at 1400 CST 107 26 Mauder M, Foken T (2006) Impact of post-field data processing on eddy covariance flux estimates and energy balance closure. Meteorol Z 15(13):597–609, DOI 10.1127/0941-2948/2006/0167 Mauder M, Foken T (2011) Documentation and instruction manual of the eddy-covariance software package TK3. Work Report University of Bayreuth, Dept. of Micrometeorology, ISSN 1614-8916, 46, 58 pp., URL http://opus.ub.uni-bayreuth.de/opus4800 ubbayreuth/frontdoor/index/index/docId/681 Mauder M, Foken T, Clement R, Elbers JA, Eugster W, Grünwald T, Heusinkveld B, Kolle O (2008) Quality control of CarboEurope flux data – Part 2: Inter-comparison of eddy-covariance software. Biogeosciences 5(2):451–462. doi:10.5194/bg-5-451-2008 Maussion F, Scherer D, Finkelnburg R, Richters J, Yang W, Yao T (2011) WRF simulation of a 805 precipitation event over the Tibetan Plateau, China – an assessment using remote sensing and ground observations. Hydrol Earth Syst Sci 15(6):1795–1817, DOI 10.5194/hess-15-1795-2011 Mengelkamp HT, Warrach K, Raschke E (1999) SEWAB - a parameterization of the surface energy and water balance for atmospheric and hydrologic models. Adv Water Resour 23(2):165– 175, DOI 10.1016/S0309-1708(99)00020-2 810 Mengelkamp HT, Kiely G, Warrach K (2001) Evaluation of the hydrological components added to an atmospheric land-surface scheme. Theor Appl Climatol 69(3-4):199–212, DOI 10.1007/s007040170025 Metzger S, Ma Y, Markkanen T, Göckede M, Li M, Foken T (2006) Quality Assessment of Tibetan Plateau Eddy Covariance Measurements Utilizing Footprint Modeling. Advances in Earth 815 Sciences 21(12):1260–1267 Mihailović DT, Pielke RA, Rajković B, Lee TJ, Jeftić M (1993) A resistance representation of schemes for evaporation from bare and partly plant-covered surfaces for use in atmospheric models. J Appl Meteorol 32(6):1038–1054, DOI 10.1175/15200450(1993)032<1038:ARROSF>2.0.CO;2 820 Molnar P, Boos WR, Battisti DS (2010) Orographic Controls on Climate and Paleoclimate of Asia: Thermal and Mechanical Roles for the Tibetan Plateau. Annu. Rev. Earth Planet. Sci. 38(1):77–102, DOI 10.1146/annurev-earth-040809-152456 Mügler I, Gleixner G, Günther F, Mäusbacher R, Daut G, Schütt B, Berking J, Schwalb A, Schwark L, Xu B, Yao T, Zhu L, Yi C (2010) A multi-proxy approach to reconstruct hydrological 825 changes and Holocene climate development of Nam Co, Central Tibet. J Paleolimnol 43(4):625– 648, DOI 10.1007/s10933-009-9357-0 Ni J (2011) Impacts of climate change on Chinese ecosystems: key vulnerable regions and potential thresholds. Regional Environmental Change 11(1):49-64, DOI 10.1007/s10113-0100170-0 830 Noilhan J, Planton S (1989) A simple parameterization of land surface processes for meteorological models. Mon Weather Rev 117(3):536–549, DOI 10.1175/15200493(1989)117<0536:ASPOLS>2.0.CO;2 Nordbo A, Launiainen S, Mammarella I, Lepparanta M, Huotari J, Ojala A, Vesala T (2011) Longterm energy flux measurements and energy balance over a small boreal lake using eddy covariance 835 technique. J. Geophys. Res 116(D2):D02119 D. Biermann et al. (2013) 114 27 Panin GN, Foken T (2005) Air–sea interaction including a shallow and coastal zone. Journal of Atmospheric & Ocean Science 10(3):289–305, DOI 10.1080/17417530600787227 Panin GN, Nasonov AE, Foken T (2006a) Evaporation and heat exchange of a body of water with the atmosphere in a shallow zone. Izv. Atmos. Ocean. Phys. 42(3):337–352. 840 doi:10.1134/S0001433806030078 Panin GN, Nasonov AE, Foken T, Lohse H (2006b) On the parameterization of evaporation and sensible heat exchange for shallow lakes. Theor Appl Climatol 85:123–129, DOI 10.1007/s00704005-0185-5 Paw U K, Baldocchi D, Meyers T, Wilson K (2000) Correction of eddy-covariance measurements 845 incorporating both advective effects and density fluxes. Bound-Lay Meteorol 97:487–511, DOI 10.1023/A:1002786702909 Rannik U, Aubinet M, Kurbanmuradov O, Sabelfeld KK, Markkanen T, Vesala T (2000) Footprint analysis for measurements over a heterogeneous forest. Bound-Lay Meteorol 97(1):137–166, DOI 10.1023/A:1002702810929 850 Rebmann C, Kolle O, Heinesch B, Queck R, Ibrom A, Aubinet M (2012) Data Acquisition and Flux Calculations. In: Aubinet M, Vesala T, Papale D (eds) Eddy Covariance. Springer Netherlands, pp 59-83 Rouse WR, Oswald CJ, Binyamin J, Spence C, Schertzer WM, Blanken PD, Bussières N, Duguay CR (2005) The Role of Northern Lakes in a Regional Energy Balance. Journal of 855 Hydrometeorology. J. Hydrometeor 6(3):291–305, DOI 10.1175/JHM421.1 Twine TE, Kustas WP, Norman JM, Cook DR, Houser PR, Meyers TP, Prueger JH, Starks PJ, Wesely ML (2000) Correcting eddy-covariance flux underestimates over a grassland. Agr Forest Meteorol 103(3):279–300. doi:10.1016/S0168-1923(00)00123-4 Wendisch M, Foken T (1989) Sensitivity test of a multilayer energy exchange model. Z Meteorol 860 39:36–39 Wilczak J, Oncley S, Stage S (2001) Sonic anemometer tilt correction algorithms. Bound-Lay Meteorol 99:127–150, DOI 10.1023/A:1018966204465 Wu Y, Zhu L (2008) The response of lake-glacier variations to climate change in Nam Co Catchment, central Tibetan Plateau, during 1970–2000. J. Geogr. Sci. 18(2):177–189, DOI 865 10.1007/s11442-008-0177-3 Xu J, Haginoya S (2001) An estimation of heat and water balances in the tibetan plateau. J Meteorol Soc Jpn 79(1B):485–504, DOI 10.2151/jmsj.79.485 Xu J, Yu S, Liu J, Haginoya S, Ishigooka Y, Kuwagata T, Hara M, Yasunari T (2009) The Implication of Heat and Water Balance Changes in a Lake Basin on the Tibetan Plateau. 870 Hydrological Research Letters 3:1–5, DOI 10.3178/hrl.3.1 Yang K, Koike T, Ye B, Bastidas L (2005) Inverse analysis of the role of soil vertical heterogeneity in controlling surface soil state and energy partition. J Geophys Res 110(D8):D08101, DOI 10.1029/2004JD005500 Yang K, Chen YY, Qin J (2009) Some practical notes on the land surface modeling in the Tibetan 875 Plateau. Hydrol Earth Syst Sci 13(5):687–701, DOI 10.5194/hess-13-687-2009 115 28 Yang K, Koike T, Ishikawa H, Kim J, Li X, Liu H, Wang J, Liu S, Ma Y (2008) Turbulent flux transfer over bare-soil surfaces: Characteristics and parameterization. J Appl Meteorol Clim 47:276–290, DOI 10.1175/2007JAMC1547.1 Yang K, Ye B, Zhou D, Wu B, Foken T, Qin J, Zhou Z (2011) Response of hydrological cycle to 880 recent climate changes in the Tibetan Plateau. Climatic Change 109(3-4):517–534, DOI 10.1007/s10584-011-0099-4 Yu S, Liu J, Xu J, Wang H (2011) Evaporation and energy balance estimates over a large inland lake in the Tibet-Himalaya. Environ Earth Sci 64(4):1169–1176. doi:10.1007/s12665-011-0933-z Zhao L, Li J, Xu S, Zhou H, Li Y, Gu S, Zhao X (2010) Seasonal variations in carbon dioxide 885 exchange in an alpine wetland meadow on the Qinghai-Tibetan Plateau. Biogeosciences 7(4):1207–1221, DOI 10.5194/bg-7-1207-2010 Zhou D, Eigenmann R, Babel W, Foken T, Ma Y (2011) The study of near-ground free convection conditions at Nam Co station on the Tibetan Plateau. Theor Appl Climatol 105(1-2):217-228, DOI 10.1007/s00704-010-0393-5 890 Zhu L, Xie M, Wu Y (2010) Quantitative analysis of lake area variations and the influence factors from 1971 to 2004 in the Nam Co basin of the Tibetan Plateau. Chin. Sci. Bull. 55(13):1294–1303. DOI 10.1007/s11434-010-0015-8 D. Biermann et al. (2013) 116 E. Charuchittipan et al. (2013) Charuchittipan, D., Babel, W., Mauder, M., Leps, J.-P., and Foken, T.: Extension of the averaging time of the eddy-covariance measurement and its effect on the energy balance closure, submitted to Bound.-Lay. Meteorol. 117 Noname manuscript No. (will be inserted by the editor) Extension of the averaging time of the eddy-covariance1 measurement and its effect on the energy balance closure2 Doojdao Charuchittipan ·Wolfgang Babel ·3 Matthias Mauder ·Jens-Peter Leps ·4 Thomas Foken5 6 Received: date / Accepted: date7 Abstract In this study, the modified ogive analysis and block ensemble average were8 employed to investigate the impact of the averaging time extension on the energy bal-9 ance closure over six different land use types. The modified ogive analysis suggests10 that the standard averaging time of 30 minutes is still generally enough for the eddy-11 covariance measurement. The block ensemble average reveals that the averaging time12 extension over several days can improve energy balance closure for some sites and13 over some specific time, when secondary circulations exist in the vicinity of the sen-14 sor. These near-surface secondary circulations mainly transport sensible heat, and15 when near-ground warm air is transported upward, the sensible heat flux observed by16 the block ensemble average will increase at longer averaging times. The close rela-17 tionship between near-surface secondary circulations and sensible heat flux suggests18 an alternative energy balance correction for a near-surface eddy-covariance measure-19 ment by using the buoyancy flux ratio, which is the larger fraction of the residual20 attribute to the sensible heat flux.21 Keywords Energy balance closure ·Ensemble average ·LITFASS ·Ogive analysis22 D. Charuchittipan (B)·W. Babel ·T. Foken Department of Micrometeorology, University of Bayreuth, D-95440 Bayreuth, Germany Tel.: +49-921-552176 Fax: +49-921-552366 E-mail: [email protected] T. Foken Member of Bayreuth Center of Ecology and Environmental Research (BayCEER), University of Bayreuth, Germany M. Mauder Institute of Meteorology and Climate Research, Karlsruhe Institute of Technology, D-82467 GarmischPartenkirchen, Germany J.-P. Leps German Meteorological Service, Richard-Aßmann-Observatory, D-15848 Lindenberg, Germany E. Charuchittipan et al. (2013) 118 2 D. Charuchittipan et al. 1 Introduction23 The imbalance of the measured fluxes at the earth’s surface is known as the energy24 balance closure problem and micrometeorologists have been aware of it since the late25 1980s (Foken, 2008a; Leuning et al., 2012). Many micrometeorological experiments26 over low vegetation reveal that the available energy, which is the sum of the net27 radiation and the ground heat flux, is larger than the sum of the sensible and latent28 heat fluxes. These experiments include the EBEX experiment (Oncley et al., 2007),29 which was especially designed to study the energy balance at the earth’s surface, and30 the LITFASS-2003 experiment (Beyrich and Mengelkamp, 2006), which aimed to31 study the effect of surface heterogeneity. To conserve energy, the residual was added32 to the energy budget equation over low vegetation at the earth’s surface,33 Res =−Q∗−(QG+QH+QE),(1) where Res is the residual or missing energy, Q∗is the net radiation, QGis the ground34 heat flux, QHis the sensible heat flux, and QEis the latent heat flux. Each term in Eq.35 1 is positive, as the energy is transported away from the ground.36 In the past few years, despite improvements in measuring and data processing37 techniques, this closure problem still remains. It is believed that the residual is caused38 by large scale eddies or secondary circulations. These secondary circulations are gen-39 erated by surface heterogeneity and normally move away from the ground (Kanda40 et al., 2004; Inagaki et al., 2006). Due to their large size and slow motion, their con-41 tributions to the low frequency part of the turbulent spectrum cannot be detected by42 the eddy-covariance (EC) measurement, which is typically averaged over a period of43 30 minutes. This results in the underestimation of QHand QE, which are normally44 measured by the EC technique.45 An extension of the averaging time was suggested and expected to result in a46 greater contribution from the low frequency parts. Two different approaches were47 introduced for this task: the ogive analysis (Desjardins et al., 1989; Oncley et al.,48 1990) and the block ensemble average (Finnigan et al., 2003). The ogive analysis uses49 the turbulent spectrum to estimate the turbulent fluxes at different frequency ranges,50 allowing assessment of the contribution of the low frequency parts to the turbulent51 fluxes measured by the EC method. In Foken et al. (2006), the ogive analysis was52 applied to the data measured over a maize field of the LITFASS-2003 experiment. It53 was focused mainly on data from three selected days, where the averaging time was54 extended up to 4 hours. It was found that the time extension would not significantly55 increase the turbulent fluxes overall.56 For the block ensemble average, low frequency contributions were added to the57 turbulent fluxes from long term fluctuations over several hours to days. In Mauder58 and Foken (2006), this treatment was also applied to the dataset from the same maize59 field of the LITFASS-2003 experiment. The length of the selected dataset was 1560 days, while the block ensemble averaging period varied from 5 minutes to 5 days.61 This study showed that the block ensemble average can close the energy balance62 at longer averaging time. Extensive discussion of the energy balance closure of the63 LITFASS-2003 experiment can be found in Foken et al. (2010).64 119 Extension of the averaging time of eddy-covariance measurements 3 In our study, to investigate whether the averaging time extension would have the65 same impact over different types of surface, we extended both ogive analysis and66 block ensemble average to cover more land use types of the LITFASS-2003 experi-67 ment. Since the LITFASS area is composed of many distinct agricultural fields, this68 surface heterogeneity could induce secondary circulations, some of which may still69 exist in the vicinity of the measuring stations. However, since the data obtained from70 different measuring stations do not always have the same sampling rate, some minor71 modifications in both ogive analysis and block ensemble average were made. These72 modifications were validated by repeating the ogive analysis in Foken et al. (2006)73 and the block ensemble average in Mauder and Foken (2006). Then we continued74 our investigations, which was more oriented toward the low frequency contributions.75 Finally, we came up with the appropriate method for correcting the energy balance of76 a near-surface eddy-covariance measurement.77 2 Material and Method78 2.1 LITFASS-2003 experiment and data processing79 The LITFASS-2003 experiment was performed between 19 May 2003 and 18 June80 2003 near the meteorological observatory of the German Meteorological Service in81 Lindenberg, Germany. The local time zone in this area is UTC+1. During the ex-82 periment, there were 14 ground-based micrometeorological measuring stations over83 13 sites, and 2 elevated measuring stations on the tower at 50 and 90 metre heights.84 This experiment covered an area of 20x20 km2and made up of 5 major land use85 types: grass, maize, rape, cereals (include rye, barley and triticale), lake and forest.86 More information about the LITFASS-2003 experiment can be found in Beyrich and87 Mengelkamp (2006).88 To cover the most important land use types of the LITFASS-2003 experiment,89 we selected the following measuring stations for our study: grass (NV2 and NV4),90 maize (A6), rape (A7), rye (A5), lake (FS) and forest (HV). Note that NV2 and NV491 were actually installed on the same field. They were oriented to different wind sectors92 to monitor turbulence at this field from all wind directions. We combined these two93 stations according to the wind direction and they were reported as the single station94 NV. Information of these selected stations can be found in Table 1.95 All selected stations were equipped with EC systems as listed in Table 1. Four-96 component net radiometers and soil heat flux plates were also installed in each station.97 Each measuring station can measure all the energy balance components in Eq. 1. De-98 tails of these measurements were well described in Mauder et al. (2006) and Liebethal99 et al. (2005). These measurements allow estimation of the residual, which—on aver-100 age—reached its maximum during 1000 -1200 UTC. For low vegetation, its average101 value during this time ranged from 75 to 145 W m−2(or 20% - 30% of the available102 energy as shown in Table 1).103 Over the forest, an additional term, for canopy heat storage, needs to be added to104 Eq. 1. For all plants, the canopy heat storage has two main contributions, the plant105 material (or biomass) and the air between plants. Over low vegetation, e.g. cotton106 E. Charuchittipan et al. (2013) 120 4 D. Charuchittipan et al. Table 1 Summary of selected measuring stations from the LITFASS-2003 experiment during 20 May 2003, 1200 UTC - 18 June 2003, 0000 UTC. Notations: hc=canopy height; zm=measurement height; θ =accepted wind direction; ∆ t=timestep of short term time series; Res = Mean residual between 1000-1200 UTC, this is when the residual is normally reach its maximum; %Res = 100·Res/(−Q∗−QG) between 1000-1200 UTC. Full details can be found in Beyrich and Mengelkamp (2006) and Mauder et al. (2006) Station Canopy hczm θ Turbulence ∆ t Res %Res (m) (m) (degree) Sensors (minutes) (Wm−2) (%) HV Pine forest 14 30.5 30 - 330 USA-1/LI-7500 10 133 24% A5 Rye 0.75-1.50 2.80 60 - 30 USA-1/KH20 5 144 30% A6 Maize 0.05-0.70 2.70 90 - 270 CSAT3/LI-7500 5 122 31% A7 Rape 0.70-0.90 3.40 30 - 240 CSAT3/KH20 5 87 22% NV2 Grass 0.05-0.25 2.40 60 - 180 USA-1/LI-7500 5 75 22% NV4 Grass 0.05-0.25 2.40 150 - 330 USA-1/LI-7500 5 82 24% FS Lake 0 3.85 180 - 30 USA-1/LI-7500 10 247 63% M50 Grass 0.05-0.25 50.7 90 -300 USA-1/LI-7500 5 - - M90 Grass 0.05-0.25 90.7 90 - 300 USA-1/LI-7500 5 - - fields, both contributions of canopy heat storage are relatively small and negligible107 (Oncley et al., 2007). The canopy heat storage becomes significant over forest and108 must be included in the energy budget equation (Lindroth et al., 2010). Unfortunately,109 we did not collect all required biomass properties of the forest during the LITFASS-110 2003 experiment, so the forest’s canopy heat storage could not be precisely estimated.111 Hence, all analyses of this site were done without a canopy heat storage term. Over112 the lake, due to its large heat capacity, QGwas calculated according to Nordbo et al.113 (2011).114 During the LITFASS-2003 campaign, the raw data were processed and averaged115 over 30 minutes. For this task, all the participating groups agreed to use the software116 package TK2 (Mauder and Foken, 2004), which has been tested and compared inter-117 nationally (Mauder et al., 2008). During flux calculation processes, several flux cor-118 rections were applied. Cross-correlation analysis was used for fixing the time delay119 between the sonic anemometer and hygrometer. Moore correction was used to correct120 the spectral loss in the high frequency range (Moore, 1986). The planar-fit rotation121 was used to align the sonic anemometer with a long term mean streamline (Wilczak122 et al., 2001). SND correction was used to convert the sonic temperature, as recorded123 by the sonic anemometer, to the actual temperature (Schotanus et al., 1983). WPL124 correction was used to correct the density fluctuation (Webb et al., 1980). Crosswind125 correction was used to account for a different type of sonic anemometer (Liu et al.,126 2001). Tanner correction was used to correct the cross sensitivity between H2O and127 O2molecules (Tanner et al., 1993), which was only applied for the Krypton Hygrom-128 eter KH20 (deployed in A5 and A7). More details of these corrections can be found129 in Mauder et al. (2006) and Foken et al. (2012).130 After all these flux corrections, quality flags were assigned to each 30 minute131 period. These quality flags are the steady state flag, the integral turbulence charac-132 teristic (ITC) flag (Foken and Wichura, 1996) and combined flag. The steady state133 121 Extension of the averaging time of eddy-covariance measurements 5 flag is a result of the steady state test and represents the stationarity of the data. The134 ITC flag represents the development of turbulent conditions, which is the result of the135 flux variance similarity test. The combined flag is the combination of the steady state136 and ITC flags. All these flags range from 1-9 (from best to worst). High quality data,137 considered suitable for fundamental scientific research, have flag values of 1-3. More138 details of the data quality analysis can be found in Foken et al. (2012, 2004).139 Besides flux calculations, flux corrections and assignment of data quality flags,140 TK2 can also generate short term averages and covariances at 5 or 10 minute inter-141 vals. Due to the limited storage capacity, these short term average data points were142 stored instead of the raw data in some measuring stations. However, the statistics143 for longer periods can be reconstructed from these short term information with the144 following relations (Foken, 2008b),145 a′b′=1 M−1"(U−1) N ∑ j=1a′b′j+UN ∑ j=1 ajbj−U2 M−1 N ∑ j=1 aj N ∑ j=1 bj#,(2) where a′b′is the long term covariance and Mis the number of measurement points146 of the long term time series. This long term time series consists of Nshort term147 time series, whose number of measurement points is U.a′b′jis the short term148 covariance, and ajand bjare the short term averages. These short term averages are149 derived from raw data, to which no flux corrections have been applied. Therefore, any150 necessary flux corrections must be included when using these short term averages for151 flux calculations. These short term average data points from selected stations were152 used for both ogive analysis and block ensemble average calculations. Short term153 averaging intervals of selected stations are shown in Table 1.154 2.2 Data selection155 In most selected measuring stations, ground heat flux and radiation data are only156 available since 20 May 2003, 1200 UTC, so the period during 20 May 2003, 1200157 UTC - 18 June 2003, 0000 UTC was used in this study. To ensure high data quality158 as well as to minimize the irrelevant factors which might influence turbulent fluxes,159 we imposed sets of data selection criteria to the ogive analysis and block ensemble160 average separately. For the ogive analysis, we increased the averaging time to up to161 4 hours. This 4 hour period consists of 8 consecutive subperiods (or blocks) of 30162 minutes. We performed the ogive analysis over any 4 hour period only if all blocks163 satisfied the selection criteria.164 The first selection criterion is identical to Mauder et al. (2006), which is that the165 sonic anemometers must not be disturbed by either the internal boundary layer result-166 ing from the heterogeneity of the surface, or the flow distortion caused by obstacles.167 The internal boundary layer height was estimated from168 zm≤ δ =0.3√x,(3) (Raabe, 1983) where zmis the measurement height, δ is the internal boundary layer169 height and xis the distance from the sensor to boundary of the next land use class.170 E. Charuchittipan et al. (2013) 122 6 D. Charuchittipan et al. To keep the measurement undisturbed, zmmust not exceed δ . Hence, we rejected any171 wind direction whose corresponding xdid not satisfy Eq. 3. The undisturbed wind172 sectors ( θ ) of each measuring station, from both internal boundary layer and flow173 distortion, are listed in Table 1. Additionally, footprint analysis was used to confirm174 that the target land use type has a significant contribution to our measurement. This175 contribution varied over the stability range. We further rejected any wind sectors176 whose contribution from target land use type is less than 80%.177 The next data selection criterion is a steady state condition of the time series,178 which is indicated by the steady state flag (section 2.1). We only accepted data with179 high quality flags (flag 1-3). In this study, we did the ogive analysis of the energy180 balance components (QHand QE) and CO2flux (Fc=w′c′CO2) separately. For the181 energy balance components, we only considered the steady state flags of friction ve-182 locity (u∗), QHand QE. We only performed the ogive analysis on any periods during183 which these three steady state flags qualified simultaneously. For Fc, we considered184 only steady state flags of u∗and CO2flux, and performed the ogive analysis on any185 periods where these two steady state flags were accepted simultaneously.186 We avoided the transition period by excluding from our analysis the time period187 covering one hour before to one hour after both sunrise and sunset. We also specified188 the threshold value of each turbulent flux as a minimum requirement in our analy-189 sis. For u∗, which indicates the level of turbulence (Massman and Lee, 2002), the190 threshold value is 0.1 m s−1. This was set to rule out very small turbulent fluxes,191 which can result from instrumentation noise. This limit normally excludes periods192 with very weak wind as well. For QH,QEand Fc, threshold values were formulated193 to avoid complication with their measurement errors. According to Mauder et al.194 (2006), based on 30 minutes averaging time, the measurement errors of QHand QE 195 are 10% - 20% of the turbulent flux at 30 minutes, or 10 - 20 W m−2, whichever is196 larger. For u∗and Fc, the measurement errors are 0.02 - 0.04 m s−1and 0.5 - 1 µ mol197 m−2s−1, respectively (Meek et al., 2005). Therefore, we set the threshold values of198 QHand QEto be 20 W m−2, and the threshold value of Fcto be 1 µ mol m−2s−1.199 Unusually large uncertainty of Fcduring the night time was taken into account by200 using only data periods with u∗greater than 0.25 m s−1(Hollinger and Richardson,201 2005).202 Similar selection criteria cannot apply to the block ensemble average, as it in-203 volves averaging times of several hours to days. Therefore, the quality control of this204 part was done by discarding any periods with more than 10% of missing raw data.205 This missing data could result from various factors, e.g. electrical black out.206 2.3 Modified ogive analysis207 The ogive analysis was introduced by Desjardins et al. (1989) and Oncley et al. (1990)208 to investigate the flux contribution from each frequency range as well as to determine209 suitable averaging periods to capture most of the turbulent fluxes. The ogive function210 of the turbulent flux (ogw,c) is defined as the cumulative integral of the cospectrum of211 123 Extension of the averaging time of eddy-covariance measurements 13 Table 3 Results from the modified ogive analysis of the energy balance components (QHand QE) from selected stations of the LITFASS-2003 experiment between 20 May 2003, 1200 UTC - 18 June 2003, 0000 UTC. Notations: η is the width of error band, which is set to be 10% and 20% of the turbulent flux at 30 minute period and has a minimum value equals to the measurement error of each turbulent flux; F30 is the average size of turbulent flux at 30 minute period; ∆ max is the average of maximum flux difference; Runs(%) is number of runs in each ogive case, which is also available in percentage in the parenthesis. Station Flux η Case 1 Case 2 Case 3 F30 Runs(%) F30 ∆ max Runs(%) F30 ∆ max Runs(%) Forest QH10% 261 92(74.8%) 205 -33 4(3.3%) 224 33 27(22.0%) (Wm−2) 20% 252 119(96.7%) 237 -56 1(0.8%) 217 70 3(2.4%) (HV) QE10% 107 53(43.1%) 128 -33 12(9.8%) 119 27 58(47.2%) (Wm−2) 20% 112 93(75.6%) 126 -45 6(4.9%) 125 40 24(19.5%) Rye QH10% 148 192(88.1%) 99 -15 6(2.8%) 85 19 20(9.2%) (Wm−2) 20% 143 213(97.7%) - - 0(0.0%) 61 36 5(2.3%) (A5) QE10% 145 196(89.9%) 118 -20 10(4.6%) 131 23 12(5.5%) (Wm−2) 20% 143 212(97.2%) 116 -26 2(0.9%) 132 30 4(1.8%) Maize QH10% 106 99(84.6%) 98 -12 3(2.6%) 116 28 15(12.8%) (Wm−2) 20% 108 111(94.9%) - - 0(0.0%) 92 39 6(5.1%) (A6) QE10% 134 97(82.9%) 77 -20 14(12.0%) 80 18 6(5.1%) (Wm−2) 20% 127 112(95.7%) 91 -37 3(2.6%) 57 22 2(1.7%) Rape QH10% 127 85(90.4%) 83 -13 8(8.5%) 94 12 1(1.1%) (Wm−2) 20% 123 94(100.0%) - - 0(0.0%) - - 0(0.0%) (A7) QE10% 181 93(98.9%) - - 0(0.0%) 141 16 1(1.1%) (Wm−2) 20% 181 94(100.0%) - - 0(0.0%) - - 0(0.0%) Grass QH10% 117 187(93.0%) 101 -15 12(6.0%) 132 23 2(1.0%) (Wm−2) 20% 116 200(99.5%) 99 -27 1(0.5%) - - 0(0.0%) (NV) QE10% 131 173(86.1%) 95 -19 4(2.0%) 118 19 24(11.9%) (Wm−2) 20% 140 196(97.5%) 94 -31 1(0.5%) 114 27 4(2.0%) Lake QH10% 40 69(95.8%) - - 0(0.0%) 31 14 3(4.2%) (Wm−2) 20% 40 72(100.0%) - - 0(0.0%) - - 0(0.0%) (FS) QE10% 197 69(95.8%) 93 -15 1(1.4%) 121 14 2(2.8%) (Wm−2) 20% 193 72(100.0%) - - 0(0.0%) - - 0(0.0%) and 3 of both QHand QEfrom rye, grass, maize and—remarkably—forest stations.413 These periods of Case 2 and 3 of rye, grass and maize sites were closely related to the414 stationarity of QHand QEover a 4 hour period. For these three sites, periods of Case415 1 usually had a four hour steady state flag of 1 for QHand QE, while Case 2 and 3416 usually had steady state flags of 2 or more. This relation was not readily apparent in417 the forest site, implying that the averaging time extension has imposed unsteadiness418 on the turbulence over low vegetation. If we restrict our consideration to rye, grass,419 maize and forest sites, we found that the number of Case 3s was normally greater than420 the number of Case 2s in both QHand QE. This would tell us that the averaging time421 extension most likely increases QHand QE. The average maximum flux difference422 ( ∆ max) for QHwas mostly higher than for QE. ∆ max is increased with larger size of423 E. Charuchittipan et al. (2013) 130 14 D. Charuchittipan et al. Table 4 Results from the modified ogive analysis of friction velocity (u∗) and CO2flux (Fc). The description is similar to Table 3 Station Flux η Case 1 Case 2 Case 3 F30 Runs(%) F30 ∆ max Runs(%) F30 ∆ max Runs(%) Forest u∗10% 0.64 191(99.5%) - - 0(0.0%) 0.38 0.06 1(0.5%) (ms−1) 20% 0.64 192(100.0%) - -2.48 0(0.0%) - - 0(0.0%) (HV) Fc10% 8.68 112(58.3%) 8.25 -1.57 24(12.5%) 7.43 1.54 56(29.2%) µ mol m−2s−120% 8.29 171(89.1%) 7.73 -2.48 8(4.2%) 8.21 3.23 13(6.8%) Maize u∗10% 0.31 111(97.4%) 0.26 -0.03 1(0.9%) 0.15 0.03 2(1.8%) (ms−1) 20% 0.31 114(100.0%) - -1.70 0(0.0%) - - 0(0.0%) (A6) Fc10% 9.09 71(62.3%) 7.13 -1.56 16(14.0%) 7.34 2.40 27(23.7%) µ mol m−2s−120% 8.69 90(78.9%) 7.10 -1.70 12(10.5%) 7.52 4.09 12(10.5%) Grass u∗10% 0.33 183(88.8%) - - 0(0.0%) 0.27 0.04 23(11.2%) (ms−1) 20% 0.33 199(96.6%) - - 0(0.0%) 0.22 0.05 7(3.4%) (NV) Fc10% 9.95 153(74.3%) 8.80 -1.67 29(14.1%) 7.65 1.27 24(11.7%) µ mol m−2s−120% 9.57 195(94.7%) 8.47 -2.74 8(3.9%) 9.41 2.82 3(1.5%) an error band ( η ), while lower numbers of Case 2 and 3 were observed. This would424 indicate that the fewer periods left had larger ∆ max. However, even with the greatest425 ∆ max added on top of flux corrections, the energy increase is still not enough to close426 the energy balance. Furthermore, from scalar similarity of QHand QE, we expected427 these fluxes to increase or decrease together. This means we should see Case 2 or428 Case 3 in both QHand QEsimultaneously, which was rarely observed.429 F30 of u∗had the highest value over the forest and the smallest value over the lake,430 and they were closely grouped together over low vegetation. Our MOG classified431 most periods from all sites as Case 1. This suggests that the time extension has almost432 no impact on u∗regardless of canopy types.433 For Fc, all sites gave compatible values of F30. Case 1 was still in the majority,434 with a larger fraction of Case 2 and 3 than the energy balance components. Forest435 also had larger fraction of Case 2 and 3 than did low vegetation. Overall, the number436 of Case 3s was greater than number of Case 2s, and ∆ max was also increased with η .437 The four hour steady state flags were normally 1 for Case 1 and higher for Case 2 and438 Case 3. However, Case 2s generally had higher steady state flags than Case 3. This439 suggests that if the averaging time extension does not impose much unsteadiness, it440 tends to increase Fc, and decrease it when more unsteadiness has been imposed.441 3.2 Block ensemble average442 The block ensemble average (Eq. 17) of all selected sites during 2 June 2003, 1800443 UTC - 18 June 2003, 0000 UTC, are shown in Fig. 2. We chose this period as our444 observation period NP to repeat Mauder and Foken (2006) with some minor mod-445 ifications (section 2.4). We found that our result from the maize station (Fig. 2 d)446 differed from the original by less than the measurement errors of QHand QE. There-447 131 Extension of the averaging time of eddy-covariance measurements 15 fore, these modifications still give the same results and we can confidently apply them448 with other selected measuring stations.449 log(averaging time in minute) Energy flux density (W m−2) 101102103104 −100 −50 0 50 100 150 200 (a)Lake 101102103104 0 20 40 60 80 100 (b)Forest 101102103104 −20 0 20 40 60 80 100 (c)Rye 101102103104 0 20 40 60 80 (d)Maize 101102103104 0 50 100 (e)Grass Qh Qe Res 101102103104 −20 0 20 40 60 80 100 120 (f)Rape Fig. 2 Block ensemble averages of sensible heat flux and latent heat flux (Eq. 17 with cis temperature and absolute humidity), and their corresponding residuals, during 2 June 2003, 1800 UTC - 18 June 2003, 0000 UTC of selected sites in the LITFASS-2003 experiment: (a) lake, (b) forest, (c)rye, (d) maize (reproduction of Mauder and Foken (2006)), (e) grass and (f) barley. The outcome of the block ensemble average was quite unexpected to us. It could450 close the energy balance only over maize, rye and rape sites. For maize and rye sites,451 the closures were at around 15 - 30 hours, which is close to the results obtained in452 Mauder and Foken (2006). These closures were mainly caused by the increasing of453 E. Charuchittipan et al. (2013) 132 16 D. Charuchittipan et al. hQHiwith longer block ensemble averaging period P. For the rape site, both hQHi 454 and hQEiwere approximately constant at all P. During the observation period, this455 site was also influenced by rain events in the southern part of the LITFASS area.456 Therefore, its closure at very long Pwas not enhanced by the block ensemble average.457 For grassland and lake, hQHiwas decreased with longer P, which was compensated458 by the increase in hQEi, and caused the residual to be approximately constant at all459 P. For lake and forest sites, we must interpret the results carefully, because the lake460 has different characteristics from other terrain sites and we cannot precisely estimate461 the canopy heat storage (section 2.1) of the forest from our data.462 At all sites, both hQHiand hQEiwere approximately constant within the first463 few hours. Over longer P,hQEiwas more steady than hQHi. The inflection at the464 diurnal scale was found at all sites of both hQHiand hQEi. As all these selected sites465 are practically in the same 20x20 km2area, the diurnal effects should not be much466 different and the degree of inflection should be compatible. Therefore, the stronger467 inflection over some sites and fluxes may not be entirely caused by the diurnal effects.468 As the block ensemble average could not close the energy balance for all selected469 sites from 2 June 2003, 1800 UTC to 18 June 2003, 0000 UTC, we need to determine470 the reason behind this and whether it would be the same in a different observation471 period NP. We know that the ˜w˜cterm of the block ensemble average is related to472 the low frequency flux contribution. In principle, ˜w˜crepresents the flux contributions473 beyond the averaging period P. If we set Pto be 30 minutes, ˜w˜cwould represent474 additional flux after the 30 minute averaging time. Hence, longterm observation of ˜w˜c475 would show variation of additional fluxes from low frequency contributions, which476 may be related to observed block ensemble average fluxes. These variations can be477 observed more clearly when the observation period NP is long enough to suppress478 any transient effects in the block ensemble average fluxes.479 Our observation period NP, which covered an entire period of the LITFASS-2003480 experiment, was 20 May 2003, 1200 UTC - 18 June 2003, 0000 UTC. We used ˜w˜c481 from all 30 minute non-overlapping blocks (P=30 minutes) within this period NP482 to construct the Hovmøller diagrams of ˜ QH( ˜w˜ Tin energetic units, Tis temperature)483 and ˜ QE( ˜w˜ain energetic units, ais absolute humidity ). These diagrams would show484 the variation of additional fluxes beyond 30 minute averaging time. According to485 section 2.4, ˜w˜ccan be very large in any arbitrary blocks. Therefore, we expected to486 observe some random large ˜ QHand ˜ QEin these diagrams.487 TheHovmøllerdiagrams of ˜ QHforrye and grasslandare shown in Fig.3. Through-488 out the entire experiment, we found large ˜ QHmore often than large ˜ QE. We firstly489 started with the period during 2 June 2003, 1800 UTC - 18 June 2003, 0000 UTC.490 Within this period, large ˜ QHwere mainly positive for the rye (Fig. 3 a) and maize (not491 shown) sites, and mainly negative over the grassland (Fig. 3 b). These observations492 are consistent with the observed block ensemble average fluxes, in which hQHiwas493 increasing at longer Pfor the rye and maize sites, and vice versa for the grassland494 (Fig. 2). The lake site is more dominated by large negative ˜ QH(not shown), which495 is consistent with the decreasing of hQHiat longer P. There were only few large496 ˜ QEat all sites, which are consistent with approximately constant hQEiat all Pover497 rye, maize, rape and forest sites. However, over lake and grassland, these few large498 133 Extension of the averaging time of eddy-covariance measurements 17 ˜ QEwere extremely large when compared to their own block ensemble averages, and499 caused their hQEito increase at longer P.500 Date Time of day 24/05 29/05 03/06 08/06 13/06 0 3 6 9 12 15 18 21 0 −800 −600 −400 −200 0 200 400 600 800 (a) Rye Date Time of day 24/05 29/05 03/06 08/06 13/06 0 3 6 9 12 15 18 21 0 −800 −600 −400 −200 0 200 400 600 800 (b) Grass Date Time of day 24/05 29/05 03/06 08/06 13/06 0 3 6 9 12 15 18 21 0 −800 −600 −400 −200 0 200 400 600 800 (c) 90 m tower Fig. 3 The Hovmøller diagrams of ˜ QHfrom (a) Rye, (b) Grass and (c) 90 m tower, they represent ˜w˜ Tof each 30 minute block in the energetic units. Series of large ˜ QHwere observed in rye (positive) and grass (mainly negative) during 1 June 2003 - 5 June 2003. These patterns are related to secondary circulations, which is consistent with frequent observations of large ˜ QHat 90 m height. Each colour depicts the flux in W m−2. More interestingly, large ˜ QHwere observed consecutively for a few days during501 1 June 2003 - 5 June 2003, over rye, maize, grass and lake. This period was the dry502 period between the rain events and was not influenced by any significant synoptic503 E. Charuchittipan et al. (2013) 134 18 D. Charuchittipan et al. events. These large ˜ QHwere positive for rye and maize, and mainly negative for504 grass and lake. Large ˜ QEwas not found in this same period. As decribed in section505 2.4, large ˜ QH(or large ˜w˜ T) could compensate a strong horizontal divergence in an506 individual block. However, consecutive occurrences indicate that they were certainly507 not transient effects. A strong horizontal divergence would imply to a strong hori-508 zontal advection, which is related to secondary circulations. Hence, we believe that509 these patterns of large ˜ QHwere caused by near-surface secondary circulations. To510 support this statement, we inspected the Hovmøller diagram of ˜ QHobtained from the511 measurement at 90 metre height (M90). At this height, there always exist secondary512 circulations, which means we should observe series of large ˜ QHmore often than in513 ground measurements. We did actually observe series of large positive and negative514 ˜ QHthroughout the entire period of the LITFASS-2003 experiment (Fig. 3 c).515 To observe the effect of near-surface secondary circulations more clearly, we516 chose 1 June 2003, 1500 UTC - 5 June 2003, 1500 UTC as the new observation517 period NP. We used the beginning and ending time of 1500 UTC to avoid gaps in the518 data from the maize field on the morning of 1 June 2003 and the rain event on the519 evening of 5 June 2003. Additionally, we wanted to complete a daily cycle as well.520 Since this long period only lasted for 4 days, the block ensemble averaging period521 Pwas varied from 10 minutes to 3 days. The block ensemble averages of this new522 observation period are shown in Fig. 4. As expected, we found that during this period,523 the energy balance is closed using averaging times of half a day for rye and maize524 sites. Over grassland and lake, hQHiwas decreased at longer P, which corresponds to525 the large negative ˜ QH.hQEiwere approximately constant at all Pat all sites, which526 were consistent with the absence of large ˜ QE.527 Accepted models state that secondary circulations can only reach down to levels528 near the earth’s surface under the free convection condition, which occurs when the529 buoyancy term dominates the shear production term as z/L≤ −1. This situation is530 also accompanied by low friction velocity (Eigenmann et al., 2009). As we did not531 observe any free convection during 1 June 2003, 1500 UTC - 5 June 2003, 1500 UTC,532 we believe that these near-surface secondary circulations were caused by the thermal533 heterogeneity between different land use types (Stoy et al., 2013).534 3.3 Scale analysis535 We used the wavelet analysis to resolve the scales of motion during 1 June 2003, 1500536 UTC - 5 June, 2003 1500 UTC with data from rye, maize and grassland stations.537 The wavelet analysis of rye and grassland stations are shown in Fig. 5 and Fig. 6538 respectively. Results for the maize field and the rye field are very similar. From these539 wavelet cross-scalograms, we found small and large scales of motion. The size of the540 small one is around a few minutes, which should be captured by the eddy-covariance541 measurement over 30 minute averaging time. It appears during the daytime at all542 sites and transports both QHand QE. The size of the larger scale is approximately543 a day, and mainly transports QH. It tends to increase QHin the maize and rye sites,544 while decreasing QHover grass. This conforms to the patterns of ˜ QHand the block545 135 Extension of the averaging time of eddy-covariance measurements 19 log(averaging time in minute) Energy flux density (W m−2) 101102103104 −200 −100 0 100 200 (a)Lake 101102103104 −20 0 20 40 60 80 100 120 140 (b)Forest 101102103104 −50 0 50 100 (c)Rye 101102103104 −50 0 50 100 150 (d)Maize 101102103104 −20 0 20 40 60 80 (e)Grass Qh Qe Res 101102103104 0 20 40 60 80 (f)Rape Fig. 4 Block ensemble averages of sensible heat flux and latent heat flux (Eq. 17 with cis temperature and absolute humidity), and their corresponding residuals, during 1 June 2003, 1500 UTC - 5 June 2003, 1500 UTC of selected sites in the LITFASS-2003 experiment: (a) lake, (b) forest, (c) rye, (d) maize, (e) grass and (f) barley. ensemble average fluxes. This scale of motion would not be captured by the eddy-546 covariance measurement averaging over the 30 minute period.547 E. Charuchittipan et al. (2013) 136 20 D. Charuchittipan et al. Fig. 5 Wavelet cross-scalograms over the rye field during 1 June 2003, 1500 UTC - 5 June 2003, 1500 UTC of (a) sensible heat flux and (b) latent heat flux. The colour represents the value in W m−2. The black solid line represents the cone of influence. 137 Extension of the averaging time of eddy-covariance measurements 21 Fig. 6 Wavelet cross-scalograms over the grassland during 1 June 2003, 1500 UTC - 5 June 2003, 1500 UTC of (a) sensible heat flux and (b) latent heat flux. The colour represents the value in W m−2. The black solid line represents the cone of influence. E. Charuchittipan et al. (2013) 138 22 D. Charuchittipan et al. Both patterns from the Hovmøller diagram and wavelet analysis show the in-548 crease or decrease of ˜ QH. However, they do not actually show what contributes to549 these changes. For the turbulent fluxes (w′c′), which are caused by instantaneous550 fluctuations, we can carry out a quadrant analysis by dividing instantaneous contri-551 butions of w′c′into four quadrants of w′and c′(Shaw, 1985). Our findings suggest552 that the main contribution for closing the energy balance is ˜ QH, which is caused by553 block to block fluctuations ( ˜wand ˜c). We therefore divided block to block contribu-554 tions of ˜w˜cinto four quadrants of ˜wand ˜c. We used ˜ T(temperature) and ˜a(absolute555 humidity) as horizontal axes and ˜w(vertical velocity) as a vertical axis, which gave556 our four quadrants (Qi,i=1,...,4) as557 Q1: ˜w>0 and ˜ T>0 or ˜a>0 warm air rising or moist air rising,558 Q2: ˜w>0 and ˜ T<0 or ˜a<0 cold air rising or dry air rising,559 Q3: ˜w<0 and ˜ T<0 or ˜a<0 cold air sinking or dry air sinking,560 Q4: ˜w<0 and ˜ T>0 or ˜a>0 warm air sinking or moist air rising.561 Q1and Q3contribute to the positive flux, while Q2and Q4contribute to the negative562 flux. We then normalized each axis by its standard deviation and set the hyperbolic563 hole size to 0.5 (H=0.5). We can neglect the weak contribution by excluding the564 contribution inside the hole and only considering any points which satisfy565  ˜w˜ T σ ˜w σ ˜ T or  ˜w˜a σ ˜w σ ˜a >H.(18) With the quadrant analysis, we expect to see which types of turbulence actually566 contribute to the increasing or decreasing of ˜w˜c. To make it consistent with our567 Hovmøller diagrams, we used the same observation period NP, which is 20 May568 2003,1200 UTC -18 June 2003,0000 UTC, and set Pto 30 minutes (non-overlapped).569 Therefore, any points on the quadrant analysis diagram represent the normalized ˜w˜c570 from each non-overlapped 30 minute period.571 The results of the quadrant analysis of rye and grassland stations are shown in572 Fig. 7. In this figure, we distinguished all points during 1 June 2003, 1500 UTC -573 5 June 2003, 1500 UTC from the rest by the use of red colour dots. By considering574 only strong contribution outside a hyperbolic hole (blue line), we found that during575 this period, ˜ QH(via ˜w˜ T) has more contribution from Q1(warm air rising) for the rye576 sites (Fig. 7 a), while there are more contributions from Q4(warm air sinking) for577 the grassland (Fig. 7 b). There was no significant contribution outside the hyperbolic578 hole for ˜ QE(via ˜w˜a) in both rye and grass stations. Over the maize field, the quadrant579 analysis is similar to that of the rye field, while the lake is similar to the grassland.580 For both rape and forest (not shown), Q1and Q4equally contributed to ˜ QH, with no581 significant contribution outside the hole for ˜ QE. These results tell us that the increase582 of hQHiat longer Pof rye and maize fields were caused by warm air near the surface583 rising, while the decreasing of hQHiof grassland and lake were caused by warm584 air aloft sinking. For forest and rape stations, both contributions from Q1and Q4 585 canceled each other and keep hQHiapproximately constant at all P. The absence586 of significant contributions outside the hyperbolic hole keeps hQEiapproximately587 constant at all sites.588 139