Novel approach to assessing residual biomass from pruning: A case study in Atlantic Pinus pinaster Ait. timber forests

Accepted Manuscript Novel approach to assessing residual biomass from pruning: A case study in Atlantic Pinus pinaster Ait. timber forests Andrea Hevia, Alejandra Crabiffosse, Juan Gabriel Álvarez-González, Ana Daria Ruiz-González, Juan Majada PII: S0960-1481(17)30115-5 DOI: 10.1016/j.renene.2017.02.029 Reference: RENE 8536 To appear in: Renewable Energy Received Date: 5 October 2016 Accepted Date: 12 February 2017 Please cite this article as: Hevia A, Crabiffosse A, Álvarez-González JG, Ruiz-González AD, Majada J, Novel approach to assessing residual biomass from pruning: A case study in Atlantic Pinus pinaster Ait. timber forests, Renewable Energy (2017), doi: 10.1016/j.renene.2017.02.029. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Copyright © 2017 Elsevier Ltd. This manuscript version is made available under the CCBY-NC-ND 4.0 license (http://creativecommons.org/licenses/by-nc-nd/4.0/) MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 1 Article 1 Renewable energy 2 3 Novel approach to assessing residual biomass from pruning: A case 4 study in Atlantic Pinus pinaster Ait. timber forests 5 6 7 Andrea Hevia 1* , Alejandra Crabiffosse 2 , Juan Gabriel Álvarez-González 2 , Ana 8 Daria Ruiz-González 2 and Juan Majada 1 9 1 Forest and Wood Technology Research Centre (CETEMAS). Pumarabule, 10 Carbayín, s/n, 33936 Siero - Asturias, Spain. E-mail: [email protected] , 11 [email protected] 12 2 Sustainable Forest Management Unit (UXFS). Agroforestry Engineering 13 Department. High Polytechnic School. University of Santiago de Compostela. 14 Campus Universitario s/n, 27002 Lugo, Spain. E-mail: [email protected], 15 [email protected] , [email protected] 16 17 *Corresponding author: 18 Andrea Hevia 19 Sustainable Forest Management Area, Forest and Wood Technology Research 20 Centre (CETEMAS). 21 Pumarabule, Carbayín, s/n, 33936 Siero - Asturias. 22 Tel: +34 984 50000; fax: +34 984 50001. 23 E-mail address: [email protected] 24 25 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 2 Abstract 26 Forestry residual biomass from pruning operations is an important, though little 27 studied, potential resource. Residues normally remain in the stand, since tools 28 for their accurate quantification do not exist and it has no particular end use. 29 Traditional tree biomass estimation models consider the whole-tree, but 30 estimating pruned biomass requires the development of more specific 31 equations. This work provides a modelling approach for assessing biomass 32 along the stem and the corresponding residual biomass from forest pruning, 33 and quantitative results from different pruning intensities in Pinus pinaster Ait. 34 are presented. Two types of models were considered: allometric biomass 35 equations (whole-tree) and biomass ratio equations (tree by height along the 36 stem), and the 2-parameter Weibull distribution function resulted in the best 37 characterization. Diameter at breast height was the best explanatory variable in 38 all equations, and model accuracy increased when models were combined with 39 total tree height for the tree stem and thicker branches, or with crown ratio for 40 the remaining tree crown components. This study provides a powerful tool to 41 estimate residual pruned biomass, enabling its better management as a 42 valuable source of bioenergy, as well as the importance in nutrient balance and 43 fire risk which it plays in a sustainable forestry production. 44 45 46 Keywords: maritime pine, forest biomass waste, Weibull distribution function, 47 forestry modelling, silviculture. 48 49 50 51 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 3 Highlights 52 53 • Final models combine allometric equations and Weibull probability density 54 function. 55 • Biomass equations by height along the stem allow accurate estimation of 56 pruned biomass. 57 • Model can be adapted to ascertain residual biomass from different pruning 58 intensities. 59 • Work provides a valuable tool to manage processing of residual pruned 60 biomass. 61 62 63 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 4 1. Introduction 64 Maritime pine (Pinus pinaster Aiton) is one of the most important conifer 65 timber species in the Atlantic area and Southern Europe, as well as having 66 great potential as a source of residual forest biomass. However, there is a 67 distinct lack of information about the forest management of this species in the 68 Atlantic region, which affects the optimal utilization of both timber and residual 69 forest biomass. It is therefore of enormous value to expand the knowledge base 70 for the silvicultural management of this species and its associated residues in 71 order to improve and optimise the sustainability and profitability of P. pinaster 72 stands. 73 In timber forests, thinning and pruning are two of the key interventions to 74 consider due to their role in enhancing the quality of wood. Residual tree 75 biomass from thinning operations has been estimated for some of the main 76 forest species [1-3] but biomass extracted by pruning has not, however, been 77 quantified very frequently in forest species. Indeed, only a few recent studies in 78 urban forests [4-6] and agricultural systems [7] focusing on pruned biomass can 79 be found. Consequently, there is little information on the availability and 80 characteristics of the residual biomass resulting from forest pruning, which is 81 usually left in the forest stand without any further processing or proper 82 management. 83 The spatial distribution of biomass within crowns is generally ignored, 84 although more intensive silvicultural methods are now occasioning the need for 85 information on the vertical distribution of branch wood biomass in some regions 86 (e.g. [8]). Particularly, distinct crown components have been studied at 87 disparate levels of resolution in different fields, but recent work has been both 88 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 5 more attentive to developments in other disciplines and increasingly concerned 89 with the vertical structure of the crown. There remains, however, comparatively 90 little information concerning the magnitude of intrinsic variation in crown 91 architecture or on the effects of stand manipulations [9]. 92 Biomass estimation of forests has been subject to research for over a century 93 (c.f. Ref. [10]) and the importance of monitoring and assessing forest biomass 94 for governments worldwide has gained increased attention in the last few 95 decades against the background of global climate change [11,12]. To quantify 96 tree and forest biomass, mathematical equations that transform tree or stand 97 variables into biomass estimates have been shown to be important and 98 powerful tools for forest management. Most tree biomass estimates rely on 99 allometric relationships developed using traditional forest measurements and 100 include the whole-tree aboveground biomass for each tree component. 101 However, forest pruning only removes part of the tree crown biomass, meaning 102 that new studies and models are needed to better understand and quantify the 103 biomass extracted by pruning to different heights along the stem. This data is 104 essential to both establish and quantify the role of these residues in forest 105 systems with respect to aspects such as their use as biofuel [13] and the 106 important role they play as regards nutrient stability, silviculture and ecology 107 [13,14] as well as fire hazard [15-18]. What is more, the use of forestry biomass 108 for bioenergy is one of the flagship initiatives within the Europe 2020 framework 109 to develop a low carbon economy by 2050 (COM(2011)112 final) [19]. 110 Several previous studies of Pinus pinaster in Atlantic forests have developed 111 biomass equations [1,2,18,20]. However, most have been carried out in adult 112 stands and employ traditional allometric equations for the whole-tree. In 113 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 6 addition, these studies are limited by the lack of research on forest residual 114 biomass from silvicultural interventions [1,2], and those studies which do exist 115 mainly rely upon data from thinning treatments or final harvesting, rather than 116 pruning. Thus, it can be seen that research directed at filling these gaps in 117 information would be of great value when taking forestry management decisions 118 in timber forests. 119 In the present study, the amount of biomass from young pruned trees was 120 estimated using dendrometric parameters obtained from field measurements 121 and data from destructive sampling in P. pinaster stands in Northwest Spain. 122 The objectives were: (i) to develop biomass prediction tools for tree fractions for 123 the whole-tree and at different heights along the stem and (ii) to assess the 124 residual biomass extracted under different intensities of pruning. This 125 constitutes a novel methodological approach which can be applied to a variety 126 of silvicultural interventions in timber forests where pruning is an essential 127 treatment. 128 2. Material and methods 129 2.1. Study sites and tree data collection 130 The present study took place in the north-western region of Asturias, Spain 131 (Fig. 1a), which has an Atlantic climate, with mild temperatures (annual average 132 12-14 ºC) and abundant rainfall (930-1475 mm) spread throughout the year. 133 Mean elevation ranged between 101 and 296 m above sea level. Soils were 134 acid (pH from 3.75 to 4.33), and the average slope of the plots was between 135 15% and 31 %. 136 The biomass study was carried out in three temporary plots (trees being 7 to 137 11 years at the time of the study) across the area of distribution of P. pinaster. 138 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 7 They were previously unmanaged (i.e. were without silviculture interventions) 139 and each was located close to one of a permanent network of silvicultural 140 research sites (Figure 1c) established in the region [21]. 141 a) b) c) Fig. 1. (a) Location of the study area in Southern Europe (Northwestern Spain), (b) example of a trial from the permanent network of silvicultural research sites (light pruning intensity), and (c) example of location of one temporary (yellow) and one permanent (red) plot in Pinus pinaster stands. 142 A total of 28 non-pruned trees (from the three temporary plots) were 143 exhaustively characterized and destructively sampled. First, the following 144 dendrometric variables were collected from the trees while they were still 145 standing: diameter at breast height (d), total tree height (h), crown base height 146 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 8 (h base ), defined as the height from ground to the point on the stem of the lowest 147 branch (live or dead), height to live crown base (h base_v ), defined as the height 148 from ground to the point on the stem of the lowest live branch and crown 149 diameter (d crown ). Next, certain of these measurements were used to calculate 150 the following tree variables: crown length, cl (total and live crown) and crown 151 ratio, CR, defined as the ratio between crown length and total tree height (cl/h). 152 In addition, age of the stand (t), number of trees per hectare (N), stand basal 153 area (G), quadratic mean diameter (D g ), mean diameter (D m ), mean height (H m ), 154 dominant diameter (D o ) and dominant height (H o ), defined as the mean 155 diameter and mean height of the 100 thickest trees per hectare, respectively, 156 were estimated in each plot. Tables 1 and 2 show the main tree and stand 157 variables. Additional information about the study sites can be found in Hevia 158 [21]. 159 Table 1 Details of the temporary plots at the time of the destructive sampling 160 (winter 2008-2009). 161 162 t: age of the stand (years); N: number of trees per hectare (stems ha -1 ); G: stand basal 163 area (m 2 ha -1 ); D g : quadratic mean diameter (cm); D m : mean diameter (cm); D o : 164 dominant diameter (cm); H m : mean height (m); H 0 : dominant height (m). 165 166 2.2. Aboveground tree biomass sampling 167 After felling the 28 selected trees, destructive sampling was undertaken, and 168 the tree variables d, h, h base and h base_v were verified. Conventional 169 Experiment site t N G D g D m D o H m H o Valsera 11 1352 15.31 12.02 11.53 18.44 7.89 9.68 Monteagudo 11 1670 17.88 11.68 11.38 16.51 6.49 7.67 Loris 14 1450 25.77 15.05 14.55 19.42 12.75 14.95 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 15 3. Results and discussion 279 3.1. Aboveground tree biomass modelling 280 The final equations of each biomass component (stem wood, stem bark, 281 needles, twigs, thin branches and thick branches), and the value of RMSE and 282 R 2adj for each equation, are given in Table 3. All the parameters were found to 283 be significant at the 5% level. 284 Diameter at breast height (d) was found to be the best tree-level predictor 285 variable. The strong relationship between d and various of the different tree 286 fractions studied, together with the practicability and low cost of its 287 measurement in the field ensure that d is one of the most widely used variables 288 in biomass functions [e.g. 2,18, 26-30]. Some authors, however, assume that d 289 is not sufficient for accurate predictions of tree biomass [31] and as a result d 290 has often been combined with other tree variables which are considered more 291 accurate in biomass equations. In the present study, total tree height (h) was 292 the variable selected together with d for the stem fractions (W sw , W sb ), and thick 293 branches (W b7 ) models. This concurs with other studies [e.g. 32-33] which also 294 found a significant improvement of biomass equations when h was used. In 295 contrast, crown ratio (CR) was the variable selected for the rest of the crown 296 biomass models (crown components of W n , W b06 , W b2 ), results in line with other 297 different works [e.g. 15,18,29-31,34-36] which obtained the most accurate 298 estimations of crown biomass components when crown variables were 299 included. Moreover, the use of CR as an independent variable in the model 300 ensures it is sensitive to changes in the crown parameters when silviculture 301 techniques such as pruning are applied. The CR is also important in explaining 302 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 16 the variability of crown biomass fractions, which has also been observed for 303 needles of P. pinaster [37] and other conifers such as Pinus taeda L. [38]. 304 Most of the models obtained in the present work showed accurate predictions 305 of tree biomass, particularly the equations by height along the stem, which 306 comparatively fitted at least slightly better than the whole-tree models for the 307 same biomass component, except for thick branches (Table 3). Specifically, 308 models for the whole-tree explained between 96% and 97% of the observed 309 variability for stem wood, stem bark and thick branches, and 74% to 76% for 310 needles and thin branches, respectively, with lower accurate predictions for 311 twigs (40%). On the another hand, equations by height along the stem 312 explained 96% to 99% of variability for stem wood, stem bark and thick 313 branches, and much higher variability for needles and thin branches (both 93%) 314 as well as twigs (95%). The goodness-of-fit data obtained in our study were 315 generally similar to those given in most biomass studies for the whole-tree 316 which use allometric equations for the same species [e.g. 18,37,39,40]. 317 Moreover, similar goodness-of-fit statistics have been obtained in other studies 318 in which the Weibull distribution equation was used, e.g., for crown fuel biomass 319 of P. pinaster trees [18] and P. ponderosa trees [41] and branches and needles 320 of P. taeda [42]. In line with this, the 2-parameter Weibull distribution function 321 has also been used to adequately describe the foliage distribution of different 322 pine species [43-45]. 323 In general, the models obtained for stem components fitted better than those 324 for crown fractions (Table 3). Specifically, and with respect to all components, 325 stem wood was the most accurate and twigs was the least, and this effect was 326 greater for the whole-tree equation. This result is to be expected due to the 327 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 17 difficulties of estimating the thinner fractions in relation to the whole-tree. Our 328 results confirm previous studies with P. pinaster [46], Pinus radiata D. Don [47] 329 and comparing both conifers [2], where crown biomass models were less 330 accurate than those pertaining to stem components. These differences may be 331 related to factors such as the random method of selecting crown components 332 during the sampling, or differences in tree age, tree species or tree distribution 333 in the stand [2]. It has been stated that the crowns of conifers may be highly 334 variable for the same tree as well as for different trees, particularly with respect 335 to foliar biomass and surface area [48]. Moreover, the prediction of crown 336 biomass is more difficult compared to that of stem biomass, mainly because of 337 the lack of homogeneity of the components and the complexity of crown 338 composition [6,38]. 339 In accordance with the literature, the inclusion of other independent variables 340 (i.e. with respect to age, silviculture, provenance, site quality, etc.) in future 341 works could substantially improve the precision of the models, particularly age, 342 which has been shown to be the variable most strongly associated with biomass 343 in P. pinaster (e.g. for site-specific studies [2,18,49], and more generalized 344 studies across contrasting environments [40]). The inclusion of age, as well as 345 employing long-term studies, will also provide more dynamic equations, to 346 better describe the effect of silviculture on biomass allocation. 347 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 18 Table 3 Biomass equations and goodness-of-fit statistics for the whole-tree and the tree by height along the stem models for P. 348 pinaster obtained by simultaneous fitting. 349 Tree Level Tree fraction Tree component Biomass Equation RMSE (kg) R 2adj Whole-tree Stem Wood 381.0360.2 024.0 hdW sw ⋅⋅= 1.6873 0.9681 Bark 729.0557.1 019.0 hdW sb ⋅⋅= 0.3812 0.9615 Crown Needles 617.2574.2 015.0 CRdW n ⋅⋅= 1.3864 0.7439 Twigs (up to 0.6 cm) 482.1226.1 06 031.0 − ⋅⋅= CRdW b 0.3759 0.3953 Thin branches (0.6 to 2.0 cm) 017.1601.1 2 062.0 CRdW b ⋅⋅= 0.5039 0.7555 Thick branches (2.0 cm to 7 cm) 4416 572106 7 10473 . . . − − ⋅⋅⋅= hdW b 0.2042 0.9746 Tree by height along the stem Stem Wood                      −−⋅= 181.1 _ 275.0 ´ exp1 i swswh h WW 1.0822 0.9823 Bark                      −−⋅= 977.0 _ 210.0 ´ exp1 i sbsbh h WW 0.2070 0.9865 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 19 Crown Needles                      −−⋅= 398.3 _ 638.0 ´ exp1 i nnh h WW 0.5627 0.9295 Twigs (up to 0.6 cm)                      −−⋅= 573.2 0606_ 552.0 ´ exp1 i bbh h WW 0.0963 0.9457 Thin branches (0.6 to 2.0 cm)                      −−⋅= 343.2 22_ 521.0 ´ exp1 i bbh h WW 0.2476 0.9318 Thick branches (2.0 cm to 7 cm)                      −−⋅= 753.3 77_ 530.0 ´ exp1 i bbh h WW 0.1819 0.9639 W i : biomass (kg) of the tree component i ( W sw = stem wood, W sb =stem bark, W b7 = thick branches, W b2 = thin branches, W b06 = twigs, W n = 350 needles); d : diameter at breast height (cm); h : total tree height (m); CR : crown ratio; RMSE : root mean square error; R 2adj : the adjusted 351 coefficient of determination of the model. All coefficients were significant at p<0.05. 352 353 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 20 3.2. Assessment of performance of biomass equations for forest pruning 354 The aboveground biomass removed by simulated pruning for each tree 355 crown component in a representative tree and stand of Pinus pinaster in the 356 Atlantic region are presented in Table 4. Quantification of this residual biomass 357 gave a mean quantity of dry total crown biomass per tree of around 0 - 5 kg, 358 which represents approximately 0 – 4 t per hectare of dry tree residues in the 359 average stand (Table 4). This low value at stand level can be explained by the 360 low weight of the (principally fine) tree crown components extracted in a forest 361 pruning intervention. 362 Table 4 Biomass extracted at tree- (kg) and stand- (t ha -1 ) levels, in each 363 simulated pruning intensity (from 10% to 60% of crown length), for an average 364 Pinus pinaster tree and stand, based on data from Hevia [21]. 365 Simulated pruning W n W b06 W b2 W b7 W fc W cb W c_tot At tree level 10% 0.008 0.008 0.048 0.001 0.016 0.049 0.065 20% 0.085 0.048 0.233 0.008 0.133 0.241 0.374 30% 0.326 0.129 0.555 0.037 0.454 0.592 1.047 40% 0.814 0.242 0.964 0.097 1.055 1.061 2.116 50% 1.556 0.369 1.381 0.183 1.925 1.564 3.489 60% 2.444 0.486 1.740 0.264 2.929 2.003 4.933 At stand level 10% 0.0002 0.0003 0.0029 0.0000 0.0005 0.0029 0.0034 20% 0.0195 0.0137 0.0801 0.0046 0.0332 0.0847 0.1179 30% 0.1318 0.0596 0.2953 0.0368 0.1914 0.3320 0.5234 40% 0.4292 0.1412 0.6263 0.1302 0.5703 0.7565 1.3268 50% 0.9611 0.2454 1.0042 0.2935 1.2065 1.2976 2.5042 60% 1.6722 0.3498 1.3499 0.4749 2.0220 1.8247 3.8467 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 21 W n : needle biomass (kg); W b06 : twig biomass (kg); W b2 : thin branches biomass (kg); 366 W b7 : thick branches biomass (kg); W fc : fine crown biomass (kg); W cb : crown branches 367 biomass (kg); W c_tot : total crown biomass (kg). 368 369 Our simulations showed that pruning in a representative tree removes more 370 than 50% of twigs and crown branches (thin and thick) when pruning involves 371 removal of 50% of crown length. Needles, on the other hand, need heavier 372 pruning intensities to be implemented in order for half of this tree fraction to be 373 extracted (as illustrated in Fig. 4). These values are consistent with those 374 obtained in previous studies on pine and other conifer species which found 375 maximum foliage density to be located in the upper half of the tree [39,42,50-376 52]. 377 Fig. 4 . Relative removal of crown biomass components (W n = needles; W b06 = twigs; W b2 = thin branches, W b7 = thick branches) in relation to the total quantity of each crown fraction in a representative tree of 10.5 cm of diameter at breast height for different pruning intensity scenarios (0% (control) to 60% of pruned crown length). 378 Regarding the total crown biomass ( W c_tot ), the heaviest pruning simulated 379 removed around 20% of the total crown biomass (Fig. 5). Our results show that 380 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 22 fine crown biomass ( W fc ) is more affected by pruning than crown branches (both 381 thin and thick branches) when removing over 40% of crown length. 382 The profitability of exploiting these resources is dependent upon the amount 383 of biomass that exists [5] and the intensity of the silvicultural treatment (see 384 Figs. 4 and 5; Table 4). For example, higher intensity pruning and larger trees 385 would provide a higher quantity of crown biomass [21]. 386 Fig. 5. Proportion of crown biomass ––fine crown biomass (W fc , pale grey bars), crown branches (W cb , dark grey) and total crown biomass (W c_tot , black)–– removed in each simulated pruning intensity ––0% (control) to 60% of crown length–– in relation to the total tree biomass in an average Pinus pinaster tree with a 10.5 cm diameter at breast height. 387 The results obtained cannot be compared directly with other studies, since, 388 as far as we are aware, no data on residual biomass obtained from forest 389 pruning of Atlantic timber conifers has been published. Thus, the orientative 390 values illustrated here, together with the equations developed in the present 391 study, can be considered as a valuable precedent for future research in timber 392 forest species. Furthermore, and in accordance with previous studies in urban 393 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 23 forests [e.g. 4,6] the quantification of pruning residual biomass is of 394 considerable importance for the best planning for and management of these 395 biomass residues to achieve energy (biofuel) and environmental targets. 396 4. Conclusions 397 The development of biomass equations which allow the estimation of the 398 residual biomass resulting from forest pruning is crucial for timber species. 399 Moreover, the results obtained in this work extend the basic information 400 available to date for assessing forestry biomass in Atlantic managed P. pinaster 401 forests. This research has demonstrated that the residual pruned biomass of P. 402 pinaster can be determined with high precision from dendrometric variables and 403 biomass equations by height along the stem. In fact, the 2-parameter Weibull 404 distribution function selected showed comparatively better predictive reliability 405 than the classic allometric functions developed for the whole-tree. As such, the 406 equations developed here allow both total biomass and pruned biomass 407 residues to be determined from data contained in classic forest inventories. 408 Thus the equations and methodology presented here constitute a novel tool to 409 improve and optimise the management of timber forests which takes into 410 consideration the residues extracted by pruning operations, as well as the best 411 option for their further processing or proper management. 412 Acknowledgments 413 We thank the Forest Services of the Government of the Principality of Asturias 414 for access to the forests used in this study and for financial support. Thank you 415 to people from CETEMAS (L. González, M. García, P. Vallejo) and SERIDA 416 (J.C. Hernández) for their participation in the fieldwork. Thank you to Ronnie 417 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 24 Lendrum for reviewing the English and FORRISK project (Interreg IV B SUDOE 418 2007-2013) for its support during data analysis. Andrea Hevia was financially 419 supported during fieldwork and data analysis by the Spanish Ministry of 420 Education and Science through the FPU scholarship program (Reference 421 AP2006-03890). 422 References 423 [1] Balboa-Murias MA. Biomasa arbórea y estabilidad nutricional de los sistemas 424 forestales de Pinus pinaster Ait., Eucalyptus globulus Labill. y Quercus robur L. 425 en Galicia. PhD thesis. Escuela Politécnica Superior de Lugo, Universidad de 426 Santiago de Compostela (in Spanish); 2005. 427 [2] Balboa-Murias MA, Rodríguez-Soalleiro R, Merino A, Álvarez-González JG. 428 Temporal variations and distribution of carbon stocks in aboveground biomass 429 of radiata pine and maritime pine pure stands under different silvicultural 430 regimes. For Ecol Manage 2006; 237: 29-38. DOI:10.1016/j.foreco.2006.09.024 431 [3] Gómez-García E, Diéguez-Aranda U, Cunha M, Rodríguez-Soalleiro R. Comparison 432 of harvest-related removal of aboveground biomass, carbon and nutrients in 433 pedunculate oak stands and in fast-growing tree stands in NW Spain. For Ecol 434 Manage 2016.;365:119–127. DOI: 10.1016/j.foreco.2016.01.021 435 [4] Sajdak M, Velázquez-Martí B. Estimation of pruned biomass through the adaptation 436 of classic dendrometry on urban forests: case study of Sophora japonica. 437 Renew Energy 2012;47:188-193. DOI:10.1016/j.renene.2012.04.002 438 [5] Velázquez-Martí B, Sajdak M, López-Cortés I. Available residual biomass obtained 439 from pruning of Morus alba L. trees cultivated in urban forest. Renew Energy 440 2013;60:121-128. DOI:10.1016/j.renene.2013.04.001 441 [6] Sajdak M, Velázquez-Martí B, López-Cortés I, Fernández-Sarriá A, Estornell J. 442 Prediction models for estimating pruned biomass obtained from Platanus 443