Hidden gems: small woody landscape features on agricultural land are overlooked in current assessments of agroforestry in Germany.

Vol.: (0123456789) Agroforest Syst (2025) 99:267 https://doi.org/10.1007/s10457-025-01358-7 Hidden gems: small woody landscape features onagricultural land are overlooked incurrent assessments ofagroforestry inGermany NoraObladen · ZoeSchindler · JonathanP.Sheppard · KatjaKröner · ElenaLarysch · PawanDatta · ThomasSeifert · ChristopherMorhart Received: 11 September 2024 / Accepted: 30 September 2025 © The Author(s) 2025 Abstract In the light of escalating climate challenges, agroforestry is receiving renewed attention for its potential to mitigate greenhouse gas emissions while enhancing the resilience of agricultural systems. The main difference between agroforestry and conventional agricultural systems is the presence and management of woody landscape features (WLF). The few datasets assessing WLF on agricultural land are limited in their spatial resolution and apply minimum mapping thresholds, potentially biasing derived estimates of WLF characteristics. Our study aimed to assess the current extent of WLF in Germany with high spatial resolution, including size variability and WLF type composition. We investigated WLF on agricultural land across seven federal states. In each of the seven states, 100 grid cells totalling 25 km2 of agricultural land were selected, amounting to a total area of 175 km2. Within this area, WLF were manually identified, delineated and classified using digital orthophotos. The results were compared with stateof-the-art datasets, particularly the Digital Basic Landscape Model (ATKIS). We identified a total of 4.8 km2 land hosting WLF, covering 2.7% of the investigated area. The extent of WLF estimated in our study was twice as large as the estimate derived from the ATKIS dataset. Overall, we identified a much larger number of WLF, which were on average much smaller in size compared to the WLF in the ATKIS Communicated by Gerardo Moreno. Nora Obladen and Zoe Schindler contributed equally. Supplementary Information The online version contains supplementary material available at https:// doi. org/ 10. 1007/ s1045702501358-7. N.Obladen· Z.Schindler· J.P.Sheppard· K.Kröner· E.Larysch· T.Seifert· C.Morhart(*) Chair ofForest Growth andDendroecology, University ofFreiburg, Freiburg, Germany e-mail: christopher[email protected] N. Obladen e-mail: [email protected] Z. Schindler e-mail: [email protected]g.de J. P. Sheppard e-mail: [email protected] K. Kröner e-mail: [email protected]g.de E. Larysch e-mail: elena.larysc[email protected] T. Seifert e-mail: thomas.seifer[email protected] P.Datta Thüringen Forst AöR, Forstliches Forschungsund Kompetenzzentrum (FFK), Gotha, Germany e-mail: [email protected] T.Seifert Department ofForest andWood Science, Stellenbosch University, Stellenbosch, SouthAfrica Agroforest Syst (2025) 99:267 267 Page 2 of 21 Vol:. (1234567890) dataset. Extrapolating our results to the national level, we estimate WLF coverage at 4,899 km2, corresponding to 57.2 Tg above-ground biomass or 28.6 Tg carbon. The biomass and carbon content were estimated based on literature-derived biomass densities. Our study indicates that accurately assessing WLF requires higher spatial resolutions than previously used. Datasets based on low-resolution images and those using minimal mapping units are not suited for capturing small WLF. Keywords Trees outside forests· Woody perennials· Above-ground biomass· Remote sensing· Orthophotos· ATKIS Introduction In recent years, agroforestry as a dynamic and sustainable land management practice has gained increasing attention (Abbas et al. 2017; Terasaki Hart etal. 2023; Kumar etal. 2024; Jovanelly etal. 2025) for its potential to address economic and ecological challenges impacting modern agricultural practices posed by climate change (Hernández-Morcillo etal. 2018). Simultaneously, in the context of climate change mitigation, there has been growing interest towards the assessment of above-ground biomass (AGB) production (Thapa etal. 2023) and the increased opportunity for carbon storage (Nair 2012; Schnell etal. 2015; Golicz etal. 2022; Sheppard etal. 2024). Agroforestry systems (AFS), as a form of climate-smart agriculture, exist in many temporal and spatial arrangements but are primarily combinations of trees and shrubs with livestock or cropland (Nair etal. 2021). Such woody structures are often referred to as woody landscape features (WLF) and can be considered an integral component of AFS. Integrating WLF in agriculture not only enhances biodiversity and ecosystem health but also aligns with broader environmental goals. The European Green Deal, for example, seeks to restore high-diversity landscape features such as tree lines, tree groups, and hedges on a minimum of 10% of agricultural land by 2030 (DG RTD 2021). Although various ecological benefits of AFS and WLF are widely recognised (Jose 2009; Udawatta et al. 2019; Sollen-Norrlin et al. 2020; Pantera et al. 2021), there is currently a lack of detailed, quantitative knowledge about AFS and WLF (Castle etal. 2022; Sattler etal. 2024; Schaffer etal. 2024), for instance regarding their spatial extent and distribution, as well as their biomass (Schnell et al. 2015; Golicz etal. 2022). The provision of baseline estimates regarding the current status of WLF could yield valuable information that could be applied for multiple purposes. For example, estimates of the AGB and carbon content of WLF could be used for estimating their climate change mitigation potential. Such information is relevant for ecological analyses, but also as an aid for political decisions. To assess the carbon content or other characteristics of WLF, information on the extent of the WLF is required. And the more accurate the input data, the better the derived estimates of carbon content and WLF characteristics. Despite large technological advances in geospatial data collection and remote sensing over the last decades, reliable information about the extent of WLF on agricultural land on national scales remains scarce. Currently the most prevalent land cover and land use datasets (LULC) available at the European level which contain information concerning WLF are the Land Use/Cover Area frame Survey (LUCAS), the CORINE land cover (CLC), the CLC + Backbone (CLC +), and the High Resolution Layer Small Woody Features (HRL-SWF) datasets (Table 1). LUCAS is a data collection based on upscaled point samples and field observations. The CLC, CLC + and HRL-SWF, on the other hand, are based on satellite data. On a German national level, the Digital basic landscape model (Basis-DLM), provides various detailed geospatial information and includes information on land cover and land use (AdV 2024b). The Basis-DLM is a high-resolution dataset which is part of the Authoritative Topographic Cartographic Information System (ATKIS). It is based on topographic maps, digital orthophotos (DOP) and some additional data, and is updated regularly (Jäger 2003). The dataset is compiled from the data collected by individual federal states. Due to differing sampling approaches, the spatial resolution and the time of data collection are not uniform across all states. While there may be other land use datasets available for specific regions within Germany, the Basis-DLM is one of the most prominent and widely used datasets that includes both topographic and land use information at the German national scale. Agroforest Syst (2025) 99:267 Page 3 of 21 267 Vol.: (0123456789) Such LULC datasets present various limitations concerning either their topicality, accessibility and/ or spatial resolution. For instance, LUCAS does not spatially map WLF but instead provides area estimates of WLF presence and types across Europe based on field-sampled points (Buck et al. 2015; Rubio-Delgado etal. 2024). Most other datasets apply varied minimum and maximum thresholds for mapping units, and some for the widths and lengths of linear structures (see Table1). While LULC datasets are not designed to represent biomass directly, their spatial limitations can result in the systematic omission of small-scale landscape features such as WLF. As a result, the extent of WLF, including the biomass and ecological functions they contribute to, may be underestimated when such datasets are used for applications such as land-use planning or carbon accounting, also in the context of AFS. Earlier literature has estimated the extent of either AFS or WLF on agricultural land in Europe and Germany based on these datasets. However, these estimates differ considerably between studies, likely explained by the different characteristics and limitations of the underlying data as described above (see Table1). For example, den Herder etal. (2017) estimated that 8.8% of the agricultural area in the EU is under AFS management, with the estimate for Germany amounting to 1.6% of the total agricultural area. According to Golicz etal. (2021), the current extent of WLF in Germany totals 4.6% of the agricultural land. Using an updated version of the dataset, the EEA (2024c) determined a comparable extent of 4.8%. Even though the Basis-DLM is one of the most used land cover datasets within Germany, is publicly available, and has a higher spatial resolution than most satellite datasets, we are not aware of any studies to date using this data to estimate the extent of WLF in Germany. The objective of this study was to assess the current extent of WLF on agricultural land in Germany. Specifically, we analysed the total WLF area as well as the variability in size and shares of different WLF types across seven federal states. This was accomplished by utilising high-resolution orthophotos with a spatial resolution of 20cm for a manual delineation and classification of WLF features in sample grid cells. An estimation of WLF extent on these same sample grid cells based on the Basis-DLM allowed Table 1 Overview on land use and land cover datasets containing information on woody landscape features. The minimum and maximum mapping sizes of the relevant features for each dataset are given as minimum/maximum mapping area (MMA), minimum/maximum mapping width (MMW), and minimum/maximum mapping length (MML). For rasterbased datasets, the MMA corresponds to the ground area covered by one single pixel. The full names of the datasets are: Land Use/Cover Area frame Survey (LUCAS), CORINE land cover (CLC), CORINE land cover plus Backbone (CLC+), High Resolution Layer–Small Woody Features (HRL-SWF), Digital basic landscape model (Basis-DLM). The datasets were used for AFS or WLF estimation by 1den Herder etal. (2017), 2Golicz etal. (2021) and 3the EEA (2024c) Dataset (release) References Coverage Sampling MMA MMW MML LUCAS (2012)1Buck etal. (2015) Europe Sample points with field data ≥ 2,500 m2 or ≥ 10,000 m2 n.a n.a CLC (2018) Büttner etal. (2017) Europe Satellite data ≥ 250,000 m2 ≥ 100m n.a CLC + vector (2018) EEA (2022, 2024b) Europe Satellite data ≥ 5,000 m2 ≥ 20m n.a CLC + raster (2021) EEA (2022, 2024a) Europe Satellite data ≥ 100 m2n.a n.a HRL-SWF (2015)2EEA (2020) Europe Satellite data ≥ 200 m2 ≤ 5,000 m2 ≤ 30m ≥ 50m HRL-SWF (2018)3CLMS (2018) Europe Satellite data ≥ 200 m2 ≤ 5,000 m2 ≤ 30m ≥ 30m Basis-DLM (n.d.) AdV (2021, 2024b) Germany Topographic maps, digital orthophotos ≥ 1,000 m2 or ≥ 5,000 m2 or ≥ 10,000 m2n.a ≥ 200m Agroforest Syst (2025) 99:267 267 Page 4 of 21 Vol:. (1234567890) for comparisons between digital orthophotos and the Basis-DLM at an individual federal state level. Additionally, we aimed for an upscaling of WLF area from sampling areas to federal state and an extrapolation to national levels alongside an estimation of AGB and carbon storage within WLF on agricultural land in Germany. Material & methods Research area The study was carried out within seven federal states of Germany (Fig. 1): Lower Saxony (NI), North Rhine-Westphalia (NW), Brandenburg (BB), Hesse (HE), Saxony-Anhalt (ST), Saxony (SN), and Thuringia (TH). All other states were excluded, either due to unavailable LULC data or insignificance due to a low extent of agricultural area, i.e. city states. An overview over the seven selected states and their shares of land under agricultural use are shown in Fig.1. The west of Germany generally has a temperate oceanic climate (Cfb), while higher elevations in the south and eastern regions experience a warm-summer humid continental climate (Dfb), with localized subarctic conditions (Dfc) in alpine areas (Beck et al. 2018). Germany has a mean annual precipitation sum of 729mm and an average mean surface air temperature of 9.6°C (World Bank 2011). The seven sampled federal states experience a higher average mean surface air temperature of 10 °C with little variability between states, while the mean annual precipitation Fig. 1 a Map of Germany illustrating the selected (light green) and not selected states (dark green). b Detailed overview of the selected states included in this study. The doughnut plots depict the share of agricultural land (dark orange) and the share of land under other land uses (light orange) (Map data: Bundesamt für Kartographie und Geodäsie (2021), Geo data: Esri (2024), Graphics data: Statistisches Bundesamt (2024)) Agroforest Syst (2025) 99:267 Page 5 of 21 267 Vol.: (0123456789) sum of 678 mm is lower than the national average and shows decreasing precipitation on a west to east gradient. Data sources To define the seven selected states, standardised administrative borders (Bundesamt für Kartographie und Geodäsie 2021) were used. Additional data were sourced from the ATKIS database, depending on open-access availability, which varied between federal states. Specifically, we used the Basis-DLM (referred to as the ATKIS-DLM hereafter) as digital landscape model data, alongside digital orthophotos (DOPs). The ATKIS-DLM describes topographical objects regarding their spatial position and arrangement, type and descriptive attributes, i.e. additional thematic or qualitative information (AdV 2024b). DOPs are distortion-free, true-to-scale photographic images of the Earth’s surface which are derived from aerial photographs. In the ATKIS data, they are available at different ground resolutions (AdV 2024a). For our purpose, we used DOPs with a ground resolution of 20cm. The underlying aerial photos were acquired between 2019 and 2023, depending on the state and the tiles’ topicality. Additionally, Google Maps imagery (Google 2023) was used to occasionally validate the results or to obtain a different seasonal picture of the landscape. Study design & data collection A grid with a cell size of 500m × 500m (0.25 km2) was overlaid on each federal state (Fig.2a, Fig.3a). Within the AdV object catalogue, the different types of topographic objects are described in detail (AdV Fig. 2 Visualisation of the study workflow. a Sampling of the grid cells within the selected federal states. b Extraction of ATKIS WLF in the selected grid cells. c Manual delineation and classification (MDC) of the WLF in the selected grid cells. d Estimating the total WLF area per selected federal state from the ATKIS-WLF and the MDC-WLF. e Estimating the total WLF biomass per selected federal state from the MDC WLF Agroforest Syst (2025) 99:267 267 Page 6 of 21 Vol:. (1234567890) 2021). In terms of topicality, some objects are updated annually (peak topicality, e.g.traffic). To investigate only agricultural areas, all grid cells intersecting with topographic object types designated as urban area (code 41,000), forest (code 43,002) and water bodies (code 44,000) were excluded from the grid, and thus, also from the analysis (Fig.2a, Fig.3b). Additionally, grid cells intersecting with objects of the value type streets (code 42,002) under the broader object type traffic (code 42,000) were excluded to eliminate large, sealed areas from our survey. From the remaining grid, 100 grid cells per state were randomly selected for further analysis (Fig.2a, Fig.3c), resulting in 700 grid cells in total. This resulted in a total study area of 25 km2 per state, amounting to a total study area of 175 km2. Within the selected grid cells, WLF were identified and mapped manually on DOPs as vector polygons (Fig. 2c). The WLF were visually classified into seven different feature types, namely hedgerow, hedgerow with gaps, tree row, grove, shrub, tree and orchard (Table2, Fig.4). In the following, the data obtained within this study is referred to as “manual delineation and classification” (MDC). In order to compare the mapped WLF from the MDC to the ATKIS-DLM dataset, the object types Fig. 3 a Example of the full grid before exclusions, b after exclusion of urban area, forest, water bodies and federal highways and c one randomly selected grid cell as used for our study (Map data: Esri (2024)) Table 2 Definitions of the woody landscape feature (WLF) types used in the manual delineation and classification WLF type Definition Hedgerow Shrubs with or without trees arranged in rows Hedgerow with gaps Shrubs with or without trees arranged in rows with small gaps in-between Tree row Trees planted in rows with < 10m distance to each other Grove Tree groups with > 2 individuals, together with shrubs but dominated by trees Shrub Shrubs with or without single trees in-between which are not arranged in rows Tree Individual trees Orchard Collection of trees managed for e.g. fruit or nut production, with regular or irregular spacing and varied management intensity Agroforest Syst (2025) 99:267 Page 7 of 21 267 Vol.: (0123456789) woody features (code 43,003) and vegetation feature (code 54,001) were aggregated to describe the WLF within the ATKIS dataset (Fig. 2b). In the following, this aggregated data will be referred to as ATKIS-WLF. In the ATKIS-DLM, scattered orchards and orchards are listed under agriculture (code 43,001). Since these value types were not accessible for all seven selected federal states, they were not included in the ATKIS-WLF. Fig. 4 Examples of the seven woody landscape feature types obtained in the manual delineation and classification. The images are illustrating the types a hedgerow, b hedgerow with gaps, c tree row, d grove, e shrub, f tree and g orchard Agroforest Syst (2025) 99:267 267 Page 8 of 21 Vol:. (1234567890) Area and biomass estimation We upscaled both total and type-specific WLF areas to the federal state level by computing their shares in the sampled agricultural area and scaling them using the total agricultural area of each state (see Fig.2d). The dry woody AGB within the different federal states and WLF categories was estimated by applying biomass densities reported in previous studies to the area of the respective WLF categories (see Table3, Fig.2e). The AGB for the categories hedgerow, hedgerow with gaps, tree rows, groves and shrub were estimated using insights from Green etal. (2021). In order to calculate the AGB of individual trees, open grown walnut trees (Juglans regia L.; a common species employed within AFS, c.f. Reisner etal. (2007) and Pardon etal. (2020)) were used as a representative tree species (Table3). The tree biomass was derived by first estimating the diameter at breast height (DBH, in cm) from the crown projection area (CPA, in m2) using a rearranged equation given by Schindler etal. (2023b) (Eq.1). This step was applied since there was no function available which directly inferred AGB from CPA. Using the estimated DBH (in cm), the AGB (in kg) was calculated for each tree using the corresponding equation by Schindler etal. (2023b) (Eq.2). For example, a tree with a CPA of 71.5 m2 and a corresponding diameter of 30cm would represent 745kg AGB. (1) DBH =e( ln ( CPA 1.0638 ) +2.038 ) ∕ 1.819 (2) AGB =e−1.8324+2.4675∗ln(DBH)∗1.0542 For the category orchard, a density of 100 trees per hectare with a mean DBH of 20cm was assumed. In Germany, it is common to plant trees in orchards with a tree spacing of 10m (e.g. Kompetenzzentrum Ökolandbau Niedersachsen 2019). AGB was calculated drawing on the insights of Schindler etal (2023a) regarding cherry trees (Prunus avium L.). This methodology ensures a nuanced and context-specific approach to AGB estimation across diverse landscape features utilising a representative tree species that has wide applicability. To provide an outlook, we estimated dry woody AGB on agricultural land within WLF at the national level by extrapolating estimates from the federal state level up to the national level. This extrapolation relies on the assumption that the relative extent of WLF within agricultural land across Germany mirrors that of the sampled federal states, with the proportions of WLF types approximated based on the composition of WLF types in the seven analysed states. The carbon stored in AGB was estimated assuming a 50% content in dry woody biomass (Thomas and Martin 2012). Carbon content values were converted to CO2 equivalents (CO2e) by multiplying them by 3.67, which corresponds to the difference in atomic weight (Guest etal. 2013). The data preparation, collection as well as geospatial processing (Fig. 2a,b,c) were carried out with the open-source software QGIS version 3.28.1 (QGIS Development Team 2022). Data analyses (Fig.2d,e) were conducted using the statistics software R version 4.5.1(R Core Team 2025). Table 3 Overview of the used dry above-ground biomass (AGB) per area for each woody landscape feature (WLF) type. The reference column refers to the reference from which the estimate was obtained. For each category except for trees, a single value was applied to the WLF area. For trees, the biomass was derived based on the crown projection area of individual trees WLF type AGB (Mg km−2) References Hedgerow 9,900 Unmanaged hedgerows (Green etal. 2021) Hedgerows with gaps 5,560 Managed hedgerows (Green etal. 2021) Tree rows 24,800 Tree lines (Green etal. 2021) Groves 14,460 Woodland (Green etal. 2021) Shrub 100 Scrub (Green etal. 2021) Tree – Biomass function for Juglans regia (Schindler etal. 2023b) Orchard 1,860 Biomass function for Prunus avium based on 100 trees per hectare each with a DBH of 20cm (Schindler etal. 2023a) Agroforest Syst (2025) 99:267 Page 9 of 21 267 Vol.: (0123456789) Results WLF detection & area As a result of the MDC methodology, 88.6% of the investigated grid cells were found to contain one or more WLF (Fig.5a). Within these grid cells containing at least one WLF, 11,136 features were detected, covering a total area of 4.8 km2 constituting 2.7% of the investigated area (Fig.5b). Upscaling our results to the total agricultural area of the seven selected states yields a sum of 2,640 km2 WLF area (Appendix Table4). Both the total extent and the relative extent of WLF within agricultural land designation (WLF%) varies considerably between states (Fig.5, Appendix Table4, Appendix Fig.8). For instance, the WLF% in HE was more than twice as large as in BB, NW, ST, or SN. As the total agricultural area in Germany amounts to around 180,207 km2 (Statistisches Bundesamt 2024), we estimate WLF in Germany to cover an area of 4,899 km2 when assuming a similar WLF distribution for all federal states in Germany. Using the ATKIS-DLM, on the other hand, WLF were detected only in 26.9% of the grid cells. Within these, 312 features (2.8% of the MDC) were detected, covering a total area of 2.1 km2 (44.5% of the MDC), totalling 1.2% of the investigated area. The large differences in the number and total area of WLF between the MDC and the data derived from the ATKIS-DLM can be observed in all investigated states (Fig.5). The sizes of individual WLF in the MDC ranged from 0.44 m2 to 122,560 m2 (Fig. 6a). The overall interquartile range (IQR) spanned from 24 m2 to 139 m2. In comparison, the sizes of the WLF identified from the ATKIS-WLF ranged between 0.04 m2 and 177,629 m2 (Fig.6b). The IQR of the ATKISWLF dataset ranged between 1,297 m2 and 6,862 m2. The average size of the WLF identified in the MDC differed significantly between the states (Fig.6a). In the case of the ATKIS-WLF, we found no significant differences between states (Fig.6b). A comparison of the distribution of WLF sizes shows that the ATKISWLF were generally larger than the WLF identified in the MDC (Fig.6c). WLF type composition & carbon storage Across all investigated states, a total of 52,942 Mg of dry woody AGB was estimated for the surveyed area within the grid cells, whereby HE, NI and TH show the highest amounts (Fig.7a). The proportions of the different WLF types vary considerably between states, e.g., HE displays a high share of hedgerows and groves, whereas NW shows the largest tree row area among the seven investigated states. In BB, hedgerows accounted for more than half of the WLF area, while in NI only a quarter of the WLF could be attributed to hedgerow structures. This density of AGB, i.e. the AGB per agricultural area, varied between states and increased in the order of SN < BB < ST < NW < TH < NI < HE. Again, the differences were considerable, with HE storing approximately twice as much AGB per agricultural area compared to SN, BB and ST. When considering the shares of area and WLF type composition, hedgerows have the highest area share followed by groves, whereas groves have the highest AGB share followed by hedgerows (Fig.7b). Tree rows, on the other hand, occupy only a small area while accounting for a high proportion of biomass. An extrapolation to the total area of each state shows that the states differed considerably in their WLF biomass content. For example, NI boasts significantly more WLF biomass in comparison with the other states (Fig.7c). Conversely, the quantities of biomass in NW and HE are relatively similar, although they are composed of different WLF types. Summing up these extrapolations, the seven investigated states contain in total 30.8Tg of AGB, which is equivalent to 15.4 Tg carbon or 56.6 Tg CO2e. Under the assumption that the selected federal states are representative of Germany as a whole, in terms of their WLF extent and type composition (see Fig.7b), an extrapolated national estimation of AGB was calculated. In total, we estimate that Germany contains 57.2Tg of AGB within WLF. This is equivalent to 28.6Tg carbon or 105.0Tg CO2e. Moreover, we can suggest that at present WLF covers a land area of 4,899 km2 in Germany. Discussion This study aimed to evaluate whether applying MDC methods to WLF in Germany produces different estimates of WLF extent and biomass compared to existing data available in the ATKISDLM. WLF identified within the ATKIS-DLM Agroforest Syst (2025) 99:267 267 Page 16 of 21 Vol:. (1234567890) of dry woody AGB may be held within WLF in Germany. This equates to 28.6 Tg carbon or 105.0 Tg CO2e, which is a significant amount in the context of carbon budgeting. Since there are even fewer publications on the extent than on the biomass and carbon of WLF, there is little opportunity to compare our estimate with those of other studies. Of the previously mentioned studies, only Golicz etal. (2021) provided a corresponding estimate (36.3Tg carbon). However, as they included both aboveand belowground biomass in their analysis, their results are not comparable with ours. It should be emphasised that our estimate was derived under a set of assumptions. Specifically, the limited number of federal states analysed and the use of biomass densities from other geographic regions (e.g. Green etal. (2021) from Ireland and Schindler (2023a; b) from the Rhine Valley in Southern Germany) should be critically considered. Using specific biomass densities from the sampled regions or regions with similar growing conditions (e.g. climatic conditions and management), could improve the accuracy of AGB estimates. Moreover, applying a single biomass density per WLF type is a strong simplification. As for example, the AGB in hedgerows varies considerably depending on hedgerow structure, species composition, height and management (Axe etal. 2017; Black etal. 2023). Example scenario Evidently. a key limitation in existing estimates of WLF extent and carbon estimates based on satellite imagery is the previously mentioned underrepresentation of small WLF due to limited spatial resolution and minimum mapping units. To highlight the relevance of these often-overlooked features, we present an example scenario based on the allometries presented by Schindler etal. (2023b), which were used in this study to estimate tree AGB. Assuming that every hectare of agricultural land in Germany would have an additional five walnut trees with a trunk circumference of 20cm each, this would result in an additional 3,082 km2 of WLF, i.e. 1.7% of the agricultural area in Germany. In terms of biomass, this would correspond to 24.7Tg AGB (i.e. 12.3Tg carbon, 45.3Tg CO2e). As walnut trees of this diameter have a crown width of about 6m, they would likely go undetected in datasets with a low spatial resolution, unless they were grouped together. This example illustrates how the cumulative impact of many small WLF can significantly influence estimates at a larger scale. Thus, while our approach used higher resolution data, there is a continuing need to better resolve small features in carbon budgeting. Methodological improvements & outlook Using the presented methodology, we were able to estimate the extent, biomass and carbon content of WLF in Germany at both state and national level. We demonstrated the importance of using high-resolution imagery to detect small WLF that would otherwise be overlooked. Nevertheless, our results are subject to various uncertainties. In the following, we recognise the limitations of the applied methodology and make suggestions for further improvements. Regarding sampling method, the number of analysed grid cells is the main source of uncertainty. To ensure comparability between states at the grid cell level, the same number of grid cells was sampled in each state, regardless of the total agricultural area of the respective states. However, this means that our results are more representative for smaller states than for larger states, as a larger proportion of the total agricultural area is sampled. In larger states, we are more likely to have randomly underrepresented less common types of agriculture, e.g. agricultural land that is cultivated with low intensity. Another source of uncertainty is the extrapolation of the results from the selected states to the whole of Germany, because the relative extent and type composition of WLF varies among agroecological regions and this is likely to be reflected in differences among federal states. An inclusion of all states in the analysis would improve the reliability of our results. While data availability prevented a full national analysis, we assume that the surveyed states are somewhat representative for Germany. The seven analysed federal states, covering only about half of Germany’s land area, captured the west–east gradient between former West and East Germany, which we assume has a large influence on WLF distribution. The MDC methodology may also have introduced some uncertainty in the results. Although we used very high-resolution data compared to previous datasets, the manual delineation and classification was not unambiguous for all WLF. Depending Agroforest Syst (2025) 99:267 Page 17 of 21 267 Vol.: (0123456789) on the feature size and the DOPs, some WLFs were more difficult to delineate and classify than others. For example, the timing of the data collection of the DOPs had a major influence due to the seasonal influence on the foliage and the influence of the sun’s position on shade cast. Although humans are very good at recognising patterns (Parikh and Zitnick 2010), any MDC is likely subjective to some degree. In this study, both delineation and classification of WLF were likely biased to some extent by this subjectivity. Therefore, to minimise inconsistencies, all MDC was conducted by a single person. To further improve the reliability of our results, orthophotos with a higher spatial resolution could be used to facilitate the MDC of WLF. However, such data were neither available nor affordable in the context of this study. In addition, the inclusion of 3D data as opposed to 2D data could aid feature classification and improve the accuracy of AGB estimates (Lingner et al. 2018). Finally, AGB estimates could be improved by incorporating shrub and tree species identity into the calculation. Although automated approaches for classifying tree species from remote sensing data are already available, these approaches typically use data at a much higher spatial resolution than was available (Egli and Höpke 2020; Schiefer etal. 2020). In conclusion, once data with a higher resolution or dimension are available, AGB estimates can be improved. As an outlook, we propose that future research should focus on the development and optimisation of automatic classification of WLFs to improve estimates of WLF extent. Although we are positive that our approach is a viable option for assessing the extent of WLFs on small scales, it is not practical for detecting WLFs on larger scales. Once larger areas can be automatically classified, the agricultural area could be fully classified, and thus, regional differences could be better captured than with a sampling approach. For automatic classification, deep learning approaches could be used. As summarised by Christin etal. (2019), deep learning has recently revolutionised various research areas and can be a powerful tool for researchers. However, training deep learning algorithms requires high computing power and large amounts of training data. In the future, at least the challenge of computationally intensive training is likely to be reduced thanks to technological progress. Conclusion Although the numerous potential benefits of WLFs in the agricultural landscape are well known, there is little detailed, high-resolution data on the extent of these structures. Existing datasets are mostly based on low-resolution satellite data and apply minimum mapping units, which can lead to small-scale WLF being overlooked in these surveys. In this study, we mapped WLF in Germany using high-resolution orthophotos and compared them with an established dataset. Compared to the ATKIS dataset, we identified considerably more, and also, considerably smaller WLFs, in sum covering an area about twice as large as the extent estimated in the ATKIS dataset. These differences emphasise that the use of high-resolution data is necessary for more accurate estimations of WLF extent, because a large number of small elements can have substantial effects in aggregate. Our study also indicates that the status quo of the WLF distribution in Germany, and probably also in other countries, may be underestimated. The insights we have gained are particularly relevant for accurate carbon budgeting, which is currently of high relevance in the context of climate change. They can also be useful for policy makers involved in the promotion of AFS. For funding guidelines to be effective, WLF distribution and carbon budget estimates cannot be disconnected from reality. An accurate assessment of the current status could possibly help to prevent such a disconnection. Future studies on WLFs should, for this reason, aim to include small WLFs. Furthermore, future research could include additional WLF characteristics to allow for a more accurate assessment of carbon storage in these structures. Including small WLF will make accounting for the current area and volume of woody components on agricultural land more accurate. This would allow a more realistic prediction of the development of long-term carbon stocks in trees and hedges outside forests, leading to a more realistic timeframe for achieving the EU Green Deal target of ensuring that at least 10% of agricultural land is covered by woody vegetation by 2030. Moreover, in-situ or 3D data could be used to incorporate the vitality and internal structure of WLF. Such information would not only be relevant for a more accurate assessment of carbon storage but could also be used to assess other effects of WLFs, such as the enhancement of biodiversity or socio-economic benefits. Agroforest Syst (2025) 99:267 267 Page 18 of 21 Vol:. (1234567890) Acknowledgements We would like to thank Janusch Jehle for his support and his ideas regarding the methodology. Authors’ contributions NO and ZS contributed equally. Conceptualization: NO, CM; Methodology: NO, CM, PD; Investigation: NO; Formal analysis: NO, ZS; Visualization: NO; Writing—original draft preparation: NO, ZS, JS, KK, EL; Writing—review and editing: NO, ZS, KK, JS, EL, PD, CM, TS; Funding acquisition: CM, TS; Supervision: CM. Funding Open Access funding enabled and organized by Projekt DEAL. This study is funded by the following projects: The project INTEGRA, grant number 2819NA071, is supported by funds of the German Federal Ministry of Food and Agriculture (BMEL) based on a decision of the parliament of the Federal Republic of Germany via the Federal Office for Agriculture and Food (BLE) under the Federal Programme for Ecological Farming and Other Forms of Sustainable Agriculture. The MODEMA project, grant number 2222NR061J, is supported by funds of the Federal Ministry of Agriculture, Food and Regional Identity (BMLEH) based on a decision of the Parliament of the Federal Republic of Germany via Agency for Renewable Resources (FNR) under the funding programme “Sustainable Renewable Resources”. The project MARVIC is co-funded by the European Union, grant number 101112942. Views and opinions expressed are however those of the authors only and do not necessarily reflect those of the European Union. Neither the European Union nor the granting authority can be held responsible for them. Data availability All data supporting the findings of this study are available within the paper and its Supplementary Information. Declarations Conflict of interests The authors declare no competing interests. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. References Abbas F, Hammad HM, Fahad S, Cerdà A, Rizwan M, Farhad W, Ehsan S, Bakhat HF (2017) Agroforestry: a sustainable environmental practice for carbon sequestration under the climate change scenarios-a review. Environ Sci Pollut Res Int 24:11177–11191. https:// doi. org/ 10. 1007/ s113560178687-0 AdV (Arbeitsgemeinschaft der Vermessungsverwaltungen der Länder der Bundesrepublik Deutschland) (2021) Dokumentation zur Modellierung der Geoinformationen des amtlichen Vermessungswesens (GeoInfoDok) - Erläuterungen zum ATKIS® Basis-DLM - Version 7.1.1. https:// www. advonline. de/ icc/ extdeu/ nav/ 35b/ binar ywr it erser vlet? imgUid= 3a130 14714204e71083212914 a39df 1f& uBasV ar iant= 11111 11111111111111111111 11111 11. Accessed 12 February 2025 AdV (Arbeitsgemeinschaft der Vermessungsverwaltungen der Länder der Bundesrepublik Deutschland) (2024a) Aerial Photographs. https:// www. advonline. de/ Produ cts/ Geoto pogra phy/ AerialPhoto graphs/. Accessed 5 June 2024 AdV (Arbeitsgemeinschaft der Vermessungsverwaltungen der Länder der Bundesrepublik Deutschland) (2024b) ATKIS® Digital Landscape Models. https:// www. advonline. de/ Produ cts/ Geoto pogra phy/ Digit alLands capeModels/ BasisDLM/. Accessed 12 February 2025 Axe MS, Grange ID, Conway JS (2017) Carbon storage in hedge biomass - a case study of actively managed hedges in England. Agric Ecosyst Environ 250:81–88. https:// doi. org/ 10. 1016/j. agee. 2017. 08. 008 Batáry P, Gallé R, Riesch F, Fischer C, Dormann CF, Mußhoff O, Császár P, Fusaro S, Gayer C, Happe A-K, Kurucz K, Molnár D, Rösch V, Wietzke A, Tscharntke T (2017) The former Iron Curtain still drives biodiversity-profit tradeoffs in German agriculture. Nat Ecol Evol 1:1279–1284. https:// doi. org/ 10. 1038/ s415590170272-x Beck HE, Zimmermann NE, McVicar TR, Vergopolan N, Berg A, Wood EF (2018) Present and future Köppen-Geiger climate classification maps at 1-km resolution. Scientific Data. https:// doi. org/ 10. 1038/ sdata. 2018. 214 Black K, Lanigan G, Ward M, Kavanagh I, hUallacháin DÓ, Sullivan LO (2023) Biomass carbon stocks and stock changes in managed hedgerows. Sci Total Environ 871:162073. https:// doi. org/ 10. 1016/j. scito tenv. 2023. 162073 Buck O, Haub C, Woditsch S, Lindemann D, Kleinewillinghöfer L, Hazeu G, Kosztra B, Kleeschulte S, Arnold S, Hölzl M (2015) Task 1.9 - Analysis of the LUCAS nomenclature and proposal for adaptation of the nomenclature in view of its use by the Copernicus land monitoring services. Technical Report. https:// land. coper nicus. eu/ usercor ner/ techn icallibra ry/ LUCAS_ Coper nicus_ Report_ v22. pdf Bundesamt für Kartographie und Geodäsie (2021) Verwaltungsgebiete - NUTS-Gebiete 1:5000000, Stand 31.12. (NUTS5000 31.12.). https:// gdz. bkg. bund. de/ index. php/ defau lt/ digit alegeoda ten/ ver wa ltung sgebi ete/ nutsgebie te-15000000stand3112nuts5 0003112. html Bundesministerium für Ernährung und Landwirtschaft (2022) Daten und Fakten: Land-, Forstund Ernährungswirtschaft mit Fischerei und Weinund Gartenbau. https:// www. bmel. de/ Share dDocs/ Downl oads/ DE/ Brosc hueren/ datenfakten2022. pdf. Accessed 14 February 2025 Agroforest Syst (2025) 99:267 Page 19 of 21 267 Vol.: (0123456789) Büttner G, Kosztra B, Soukup T, Sousa A, Langanke T (2017) CLC 2018 Technical Guidelines. https:// land. coper nicus. eu/ en/ techn icallibra ry/ clc2018techn icalguide lines Castle SE, Miller DC, Merten N, Ordonez PJ, Baylis K (2022) Evidence for the impacts of agroforestry on ecosystem services and human well-being in high-income countries: a systematic map. Environ Evid 11:10. https:// doi. org/ 10. 1186/ s1375002200260-4 Christin S, Hervet É, Lecomte N (2019) Applications for deep learning in ecology. Methods Ecol Evol 10:1632–1644. https:// doi. org/ 10. 1111/ 2041210X. 13256 CLMS (Copernicus Land Monitoring Service) (2018) User manual - Small Woody Features 2018 and Small Woody Features Changes 2015–2018 den Herder M, Moreno G, Mosquera-Losada RM, Palma JH, Sidiropoulou A, Santiago Freijanes JJ, Crous-Duran J, Paulo JA, Tomé M, Pantera A, Papanastasis VP, Mantzanas K, Pachana P, Papadopoulos A, Plieninger T, Burgess PJ (2017) Current extent and stratification of agroforestry in the European Union. Agric Ecosyst Environ 241:121– 132. https:// doi. org/ 10. 1016/j. agee. 2017. 03. 005 DG Rtd (European Commission: Directorate-General for Research and Innovation) (2021) European Green Deal – Research & innovation call. Publications Office of the European Union. https:// doi. org/ 10. 2777/ 33415 EEA (European Environment Agency) (2020) Copernicus Land Monitoring Service – High Resolution Layer Small Woody Features - 2015 reference year. Product Specifications & User Guidelines EEA (European Environment Agency) (2022) CLC+ Backbone Product Specification And User Manual (Issue 3.0). https:// land. coper nicus. eu/ en/ techn icallibra ry/ clcbackb oneprodu ctusermanual. Accessed 12 February 2025 EEA (European Environment Agency) (2024a) CLC+Backbone 2021 (raster 10 m), Europe, 3-yearly, Jun. 2024 EEA (European Environment Agency) (2024b) CLCplus Backbone 2018 (vector), Europe, Nov. 2024 EEA (European Environment Agency) (2024c) Woody landscape features on agricultural land in Europe. https:// www. eea. europa. eu/ en/ analy sis/ indic ators/ woodylands capefeatu resonagric ultur alland Egli S, Höpke M (2020) CNN-based tree species classification using high resolution RGB image data from automated UAV observations. Remote Sens (Basel) 12:3892. https:// doi. org/ 10. 3390/ rs122 33892 Esri (2024) World Imagery. https:// www. arcgis. com/ home/ item. html? id= 10df2 279f9 684e4 a9f6a 7f08f ebac2 a9. Accessed 14 February 2025 FAO (Food and Agriculture Organization) (2001) Global Forest Resources Assessment 2000. https:// www. fao. org/4/ y1997e/ y1997 e00. htm Federal Agency for Nature Conservation (2025) Monitoring of high nature value farmland. https:// www. bfn. de/ en/ monit oringhighnaturevaluefarml and# anchor2331. Accessed 12 February 2025 Funk R, Völker L, Deumlich D (2023) Landscape structure model based estimation of the wind erosion risk in Brandenburg, Germany. Aeolian Res 62:100878. https:// doi. org/ 10. 1016/j. aeolia. 2023. 100878 Golicz K, Ghazaryan G, Niether W, Wartenberg AC, Breuer L, Gattinger A, Jacobs SR, Kleinebecker T, Weckenbrock P, Große-Stoltenberg A (2021) The role of small woody landscape features and agroforestry systems for national carbon budgeting in Germany. Land (Basel) 10:1028. https:// doi. org/ 10. 3390/ land1 01010 28 Golicz K, Bellingrath-Kimura S, Breuer L, Wartenberg AC (2022) Carbon accounting in European agroforestry systems – key research gaps and data needs. Curr Res Environ Sustain 4:100134. https:// doi. org/ 10. 1016/j. crsust. 2022. 100134 Google (2023) Map Data. https:// www. google. com/ maps Green S, Martin S, Gharechelou S, Cawkwell F, Black K (2021) Briar: biomass retrieval in Ireland using active remote sensing (2014-Ccrp-Ms.17). https:// www. epa. ie/ publi catio ns/ resea rch/ clima techange/ Resea rch_ Repor t_ 305. pdf. Accessed 14 February 2025 Guest G, Bright RM, Cherubini F, Strømman AH (2013) Consistent quantification of climate impacts due to biogenic carbon storage across a range of bio-product systems. Environ Impact Assess Rev 43:21–30. https:// doi. org/ 10. 1016/j. eiar. 2013. 05. 002 Hernández-Morcillo M, Burgess P, Mirck J, Pantera A, Plieninger T (2018) Scanning agroforestry-based solutions for climate change mitigation and adaptation in Europe. Environ Sci Policy 80:44–52. https:// doi. org/ 10. 1016/j. envsci. 2017. 11. 013 Herzog F (1998) Streuobst: a traditional agroforestry system as a model for agroforestry development in temperate Europe. Agrofor Syst 42:61–80. https:// doi. org/ 10. 1023/A: 10061 52127 824 Jäger E (2003) ATKIS® als Gemeinschaftsaufgabe der Länder und des Bundes. j. Cartogr Geogr Inf 53:113–119. https:// doi. org/ 10. 1007/ BF035 45526 Jose S (2009) Agroforestry for ecosystem services and environmental benefits: an overview. Agroforest Syst 76(1):1–10. https:// doi. org/ 10. 1007/ s104570099229-7 Jovanelly K, Su C, Ong TW (2025) Managing agroforestry transitions in a rapidly changing climate. Agroecol Sustainable Food Syst 49:228–268. https:// doi. org/ 10. 1080/ 21683 565. 2024. 24160 09 Kompetenzzentrum Ökolandbau Niedersachsen (2019) Streuobstwiesen pflanzen: Artenreiche Lebensräume und „Regenwälder Deutschlands“. https:// roten burg. bund. net/ filea dmin/ roten burg/ Landn atur/ koen_ streu obstw iese_ pflan zen. pdf. Accessed 19 February 2025 Kumar S, Alam B, Taria S, Singh P, Yadav A, Arunachalam A (2024) Agroforestry solutions for climate change and environmental restoration. Springer Nature Singapore, Singapore. https:// doi. org/ 10. 1007/ 978981975004-7 Lal Shrestha H (2020) Assessment of above ground biomass of Trees outside Forest (TOF) in the Context of Climate Change. JENR 4. https:// doi. org/ 10. 23880/ JENR16000 186 Lechner AM, Stein A, Jones SD, Ferwerda JG (2009) Remote sensing of small and linear features: quantifying the effects of patch size and length, grid position and detectability on land cover mapping. Remote Sens Environ 113:2194–2204. https:// doi. org/ 10. 1016/j. rse. 2009. 06. 002 Lingner S, Thiessen E, Hartung E (2018) Aboveground biomass estimation in linear forest objects: 2Dvs. 3D-data. Agroforest Syst (2025) 99:267 267 Page 20 of 21 Vol:. (1234567890) J Forest Sci 64:523–532. https:// doi. org/ 10. 17221/ 106/ 2018JFS Liu S, Brandt M, Nord-Larsen T, Chave J, Reiner F, Lang N, Tong X, Ciais P, Igel C, Pascual A, Guerra-Hernandez J, Li S, Mugabowindekwe M, Saatchi S, Yue Y, Chen Z, Fensholt R (2023) The overlooked contribution of trees outside forests to tree cover and woody biomass across Europe. Sci Adv 9:eadh4097. https:// doi. org/ 10. 1126/ sciadv. adh40 97 Nair PKR (2012) Carbon sequestration studies in agroforestry systems: a reality-check. Agrofor Syst 86:243–253. https:// doi. org/ 10. 1007/ s104570119434-z Nair PKR, Kumar BM, Nair VD (2021) Definition and Concepts of Agroforestry. In: Nair PKR, Kumar BM, Nair VD (eds) An Introduction to Agroforestry. Springer International Publishing, Cham, pp 21–28. https:// doi. org/ 10. 1007/ 978-303075358-0_2 Pantera Α, Mosquera-Losada MR, Herzog F, den Herder M (2021) Agroforestry and the environment. Agroforest Syst 95(5):767–774. https:// doi. org/ 10. 1007/ s1045702100640-8 Pardon P, Mertens J, Reubens B, Reheul D, Coussement T, Elsen A, Nelissen V, Verheyen K (2020) Juglans regia (walnut) in temperate arable agroforestry systems: effects on soil characteristics, arthropod diversity and crop yield. Renewable Agric Food Syst 35:533–549. https:// doi. org/ 10. 1017/ S1742 17051 90001 76 Parikh D, Zitnick CL (2010) The role of features, algorithms and data in visual recognition. 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition:2328–2335 Plieninger T (2011) Capitalizing on the carbon sequestration potential of agroforestry in Germany’s agricultural landscapes: realigning the climate change mitigation and landscape conservation agendas. Landsc Res 36:435–454. https:// doi. org/ 10. 1080/ 01426 397. 2011. 582943 QGIS Development Team (2022) QGIS Geographic Information System. https:// www. qgis. org/ R Core Team (2025) R: A Language and Environment for Statistical Computing. https:// www.Rproje ct. org/ Reisner Y, de Filippi R, Herzog F, Palma J (2007) Target regions for silvoarable agroforestry in Europe. Ecol Eng 29:401– 418. https:// doi. org/ 10. 1016/j. ecole ng. 2006. 09. 020 Rubio-Delgado J, Schnabel S, Lavado-Contador JF, Schmutz U (2024) Small woody features in agricultural areas: agroforestry systems of overlooked significance in Europe. Agric Syst 218:103973. https:// doi. org/ 10. 1016/j. agsy. 2024. 103973 Sattler C, Schrader J, Hüttner M-L, Henle K (2024) Effects of management, habitat and landscape characteristics on biodiversity of orchard meadows in Central Europe: a brief review. Nat Conserv 55:103–134. https:// doi. org/ 10. 3897/ natur econs ervat ion. 55. 108688 Schaffer C, Elbakidze M, Björklund J (2024) Motivation and perception of farmers on the benefits and challenges of agroforestry in Sweden (Northern Europe). Agrofor Syst 98:939– 958. https:// doi. org/ 10. 1007/ s1045702400964-1 Schiefer F, Kattenborn T, Frick A, Frey J, Schall P, Koch B, Schmidtlein S (2020) Mapping forest tree species in high resolution UAV-based RGB-imagery by means of convolutional neural networks. ISPRS J Photogramm Remote Sens 170:205–215. https:// doi. org/ 10. 1016/j. isprs jprs. 2020. 10. 015 Schindler Z, Seifert T, Sheppard JP, Morhart C (2023a) Allometric models for above-ground biomass, carbon and nutrient content of wild cherry (Prunus avium L.) trees in agroforestry systems. Ann for Sci 80:28. https:// doi. org/ 10. 1186/ s1359502301196-6 Schindler Z, Morhart C, Sheppard JP, Frey J, Seifert T (2023b) In a nutshell: exploring single tree parameters and aboveground carbon sequestration potential of common walnut (Juglans regia L.) in agroforestry systems. Agrofor Syst 97:1007–1024. https:// doi. org/ 10. 1007/ s1045702300844-0 Schnell S, Kleinn C, Ståhl G (2015) Monitoring trees outside forests: a review. Environ Monit Assess 187:600. https:// doi. org/ 10. 1007/ s106610154817-7 Schumacher U, Meinel G (2009) ATKIS, ALK(IS), Orthobild – Vergleich von Datengrundlagen eines Flächenmonitorings. In: Schumacher U, Meinel G (eds) Flächennutzungsmonitoring. 1: Konzepte, Indikatoren, Statistik. Shaker, pp 47–67 Seidel D, Busch G, Krause B, Bade C, Fessel C, Kleinn C (2015) Quantification of biomass production potentials from trees outside forests - a case study from Central Germany. Bioenergy Res 8:1344–1351. https:// doi. org/ 10. 1007/ s121550159596-z Sharma P, Bhardwaj DR, Singh MK, Nigam R, Pala NA, Kumar A, Verma K, Kumar D, Thakur P (2023) Geospatial technology in agroforestry: status, prospects, and constraints. Environ Sci Pollut Res Int 30:116459–116487. https:// doi. org/ 10. 1007/ s1135602220305-y Sheppard JP, Larysch E, Cuaranhua CJ, Schindler Z, Du Toit B, Malherbe GF, Kunneke A, Morhart C, Bohn Reckziegel R, Seifert T, Kahle H-P (2024) Assessment of biomass and carbon storage of a Populus simonii windbreak located in the Western Cape Province, South Africa. Agroforest Syst 98:697–714. https:// doi. org/ 10. 1007/ s1045702300940-1 Sollen-Norrlin M, Ghaley BB, Rintoul NLJ (2020) Agroforestry benefits and challenges for adoption in Europe and beyond. Sustainability 12:7001. https:// doi. org/ 10. 3390/ su121 77001 Statistisches Bundesamt G-O (2024) Bodenfläche (tatsächliche Nutzung). https:// wwwgenes is. desta tis. de/ daten bank/ online/ stati stic/ 33111/ table/ 331110005. Accessed 22 April 2024 Terasaki Hart DE, Yeo S, Almaraz M, Beillouin D, Cardinael R, Garcia E, Kay S, Lovell ST, Rosenstock TS, Sprenkle-Hyppolite S, Stolle F, Suber M, Thapa B, Wood S, Cook-Patton SC (2023) Priority science can accelerate agroforestry as a natural climate solution. Nat Clim Chang 13:1179–1190. https:// doi. org/ 10. 1038/ s4155802301810-5 Thapa B, Lovell S, Wilson J (2023) Remote sensing and machine learning applications for aboveground biomass estimation in agroforestry systems: a review. Agrofor Syst 97:1097–1111. https:// doi. org/ 10. 1007/ s1045702300850-2 Thomas SC, Martin AR (2012) Carbon content of tree tissues: a synthesis. Forests 3:332–352. https:// doi. org/ 10. 3390/ f3020 332 Udawatta RP, Rankoth L, Jose S (2019) Agroforestry and biodiversity. Sustainability 11(10):2879. https:// doi. org/ 10. 3390/ su111 02879 World Bank (2011) Climate Change Knowledge Portal: Observed Climate Data, CRU Ts4.07 0.5-Degree Agroforest Syst (2025) 99:267 Page 21 of 21 267 Vol.: (0123456789) Zomer RJ, Neufeldt H, Xu J, Ahrends A, Bossio D, Trabucco A, van Noordwijk M, Wang M (2016) Global tree cover and biomass carbon on agricultural land: the contribution of agroforestry to global and national carbon budgets. Sci Rep 6:29987. https:// doi. org/ 10. 1038/ srep2 9987 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.