Pathways to productivity: how functional ecosystem service flow determines crop yield in a prairie landscape

Pathways to productivity: how functional ecosystem service flow determines crop yield in a prairie landscape | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Pathways to productivity: how functional ecosystem service flow determines crop yield in a prairie landscape Ehsan Pashanejad, Brian Robinson, Lael Parrott This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7467778/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 21 Mar, 2026 Read the published version in Landscape Ecology → Version 1 posted 9 You are reading this latest preprint version Abstract Context. Sustaining agricultural productivity while maintaining ecological integrity requires understanding the spatial dynamics of ecosystem services (ES). In the Canadian prairies—an intensively modified agricultural landscape—the degradation of natural habitats has impacted ES flows crucial for food security. Objectives. We investigated how internal ES flows, mediated by landscape structure, influence crop yield at the Soil Landscape of Canada (SLC) scale, an ecologically meaningful delineation based on natural features. Our primary objective was to determine the relative importance of landscape composition versus configuration in predicting agricultural productivity. Method. We conducted a biophysical assessment of key ES (pollination, carbon storage, habitat quality, soil erosion control) for the year 2020. We quantified landscape composition and configuration metrics at the SLC scale to represent ES flow pathways. Generalized additive models (GAMs) were used to analyze the non-linear effects of these variables on a composite crop yield index. Results. Our findings reveal that landscape configuration—notably connectivity (positive linear effect) and crop diversity (complex non-linear effect)—significantly predicts crop yield, often exerting greater influence than the mere amount of natural habitat. A secondary analysis showed that yield in specific crops like canola, which depends on pollination, responded positively to natural habitat extent. The models explained a substantial portion of yield variance (Adjusted R² ≈ 0.66–0.67). Conclusion . Our analysis highlights that agricultural output is not solely a function of field-level inputs but is deeply embedded within, and responsive to the landscape matrix at SLC scale and the ecological processes it mediates. Strategically enhancing landscape cohesion and crop diversity may therefore offer greater yield benefits than focusing on increasing isolated natural habitat, guiding a shift towards spatially explicit, multifunctional landscape planning. agroecosystem landscape structure multifunctionality connectivity diversity crop yield resilience Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 1. Introduction Reconciling the competing pressures of agriculture with the need to conserve biodiversity and ecosystem services is a central challenge, especially in intensive farming systems (Tamburini et al. 2020 ). The concept of sustainable multifunctional landscapes has emerged as a key solution to addressing global challenges in food security while also safeguarding biodiversity and ecosystem services (Richter et al. 2024 ). Maintaining farmland biodiversity and the flow of ecosystem services in farmland is essential to the long-term sustainability of agroecosystems and food production. For instance, studies show that practices supporting on farm biodiversity such as reduced tillage, crop rotation, and organic farming, can maintain or even improve crop yield over the long term (Cozim-Melges et al. 2024 ). However, the homogenization of landscape structure within modern production systems has caused significant biodiversity loss, undermining the stability and resilience of these systems (Galpern et al. 2020 ). The link between landscape structure and ecological benefits has been well-established in previous studies. A recent global meta-analysis found that increasing landscape complexity, particularly through changes in composition, configuration, or heterogeneity, has significant positive effects on biodiversity and the associated ecosystem services, thereby enhancing the potential benefits to agricultural production (Estrada-Carmona et al. 2022 ). Empirical evidence indicates that various aspects of landscape configuration can influence yield outcomes. For instance, greater landscape diversity has been linked to improved natural pest control (Madin and Nelson 2023), while higher connectivity can either facilitate beneficial species movement (Wood et al. 2022 ) or, conversely, spread pests and diseases (Margosian et al. 2009 ). Additionally, factors such as natural habitat patch size and distance to edges have been associated with crop productivity within agricultural landscapes (Nguyen, Robinson, and Galpern 2022). Beyond yield benefits, spatial arrangement also acts as a mediator and facilitator of ecosystem service interactions, shaping how services co-occur, reinforce, or compete across space (Pashanejad et al. 2024 ; Rieb and Bennett 2020). As such, landscape structure is central to promoting ecosystem multifunctionality and resilience in agricultural systems. Despite growing recognition of the role of ES in sustaining resilient agricultural landscapes, our understanding remains limited regarding how spatial and ecological patterns influence the optimal delivery of these services—particularly in relation to crop productivity and long-term sustainability outcomes (Rehman et al. 2022 ). Recent advances in landscape ecology have emphasized a “pattern–process–service–sustainability” paradigm (Fu et al. 2025 ), which calls for deeper integration of spatial configuration and ecological processes to better understand how landscape structure supports service flows and functional outcomes. Yet, a critical and underexplored frontier lies in understanding how internal functional connectivity—that is, the spatial transfer of ecosystem service benefits within ecologically meaningful units—mediates agricultural performance. While previous research has shown that landscape composition and structure influence service supply and movement, few empirical studies have examined how these internal flows accumulate and interact to affect integrated outcomes such as crop yield (Fu et al. 2025 ; Qiu and Mitchell 2024). This is particularly relevant in large-scale, intensively managed systems like the Canadian prairies, where service delivery and agricultural performance are shaped by multiple, interacting, and often nonlinear spatial dynamics. In this study, we address this gap by using the Soil Landscapes of Canada (SLCs), an ecologically meaningful spatial delineation based on natural features, to investigate how internal ES flows within these units—mediated by landscape structure and composition—influence yield as a cumulative, emergent property of multifunctional landscapes. We apply generalized additive models (GAMs) to examine how key ecosystem service flows and structural landscape attributes—such as habitat connectivity and crop diversity—jointly influence agricultural productivity. To this, we integrate administrative crop yield data with modeled ES estimates for the year 2020, analyzing their relationship within SLC units. Our main objective is to disentangle whether spatial and functional dimensions of service flows within multifunctional landscapes contribute to crop yield as an integrated performance metric. In so doing, this research aims to identify spatial trade-offs and synergies that emerge across agricultural mosaics. Ultimately, our findings are intended to inform data-driven, landscape-level strategies for enhancing ecosystem multifunctionality and supporting sustainable agricultural management in intensively farmed regions. 2. Method The methodology of this study was structured in three main steps. First, we collated both ecosystem service and crop yield data for each SLC in the study area. To do this, we mapped and quantified key ecosystem services including pollination, habitat quality, carbon storage, nutrient retention, and soil erosion control across the Canadian prairies using the InVEST modeling tool. A complete and detailed description of each ecosystem service model, its data requirements, and specific parameterization is provided in the Supplementary Information (S1). This methodological description for ES mapping is provided in SI file in detail. Following the ES mapping, we aggregated pixel-level ES values to the SLC aligning with the scale at which crop yield data were available. We then developed a composite index of crop productivity as our primary response variable by standardizing and integrating yield data for four major crops: canola, barley, wheat (both spring and winter wheat). To do this we first standardized the yield values (kg/acre) for the four aforementioned major crops in the study area by converting them to z-score. This process rescales each crop's yield distribution to have a mean of zero and a standard deviation of one, ensuring that each crop contributes equally to the index regardless of differences in absolute yield. The final composite crop yield index for each SLC unit was then calculated as the mean of these four standardized yield values. In the second step, we quantified landscape structure for each SLC unit using a set of composition and configuration metrics. These landscape metrics were used as proxies for ecosystem service flows, conceptualized as the internal spatial transfer of ecological benefits within SLC units —from service-providing areas (e.g., natural habitats) to service-demanding areas (e.g., agricultural lands). Specifically, we linked individual metrics to distinct ecological processes: landscape connectivity, for example, was used to represent the potential for pollinator movement from natural habitats to croplands, while edge density proxied the spillover of beneficial insects like pest predators from field margins. The SLC unit was chosen as the appropriate scale to capture these flows for two key reasons. First, with an average size of approximately 46,500 hectares, SLCs are large enough to encompass the heterogenous mosaic of service-providing and service-demanding areas. This spatial extent is ecologically relevant as it aligns with the typical dispersal and foraging distances of mobile service providers like pollinators. Second, SLCs are delineated by stable biophysical features like soil and topography, creating coherent landscape units where the cumulative effects of these internal ecological flows on agricultural productivity can be meaningfully assessed. This approach helped us to examine how internal functional connectivity, local spatial heterogeneity, and spatial spillover effects within landscape units may influence agricultural productivity. In the final step, we employed generalized additive models (GAMs) to investigate the relationship between ES provisioning, ES flows, and crop productivity. To disentangle the influence of biophysical functions from that of spatial structure, we developed two distinct analytical models. The first, our Ecosystem Service (ES) Model, examined the direct predictive power of key regulating services such as pollination and erosion control on crop yield. The second, our Landscape Structure (LSM) Model, focused instead on how landscape composition and configuration metrics explain yield variation. Both models used the composite crop index as the dependent variable and included a consistent control variables to account for management, climate, and topography (Table 1 ). The following subsections describe each step in greater detail, including the study area and data sources (Sections 2.1 and 2.2), and the methodological rationale for the selection and application of landscape structure metrics and GAMs (Sections 2.3 and 2.4). Table 1 Variables included in the Generalized Additive Models (GAMs). Role in Model Category Variable Data source Response Productivity Composite Crop Yield Estimated based on Agriculture and Agri-Food Canada (AAFC, 2024) crop yield data Predictor Ecosystem Service (ES Model) Pollination InVEST model output Habitat Quality Soil Erosion Control Carbon Storage Predictor Landscape Structure (LSM Model) % Natural Habitat Calculated based on AAFC land cover and land use (annual crop inventory for the year 2020) Habitat Diversity (Shannon Diversity) Crop Diversity Connectivity (patch cohesion) Edge Density Aggregation (Clumpiness) Control (Both Models) Environment & Management Soil Organic Matter ISRIC–World Soil Information Elevation Digital Elevation Model FABDEM V1-2 Temperature Climate NA platform (Wang et al. 2016 ) Precipitation Tillage Census of Agriculture: Agri-Environmental Spatial Data (AESD) Spatial Term Longitude and latitude centroid of each SLC polygon 2.1. Study area The Canadian prairies, a vast agricultural region spanning the provinces of Alberta, Saskatchewan, and Manitoba in central Canada, were selected as the study region to investigate whether internal functional connectivity of ecosystem services within landscapes influences crop productivity. Specifically, we test the hypothesis that landscapes with greater habitat protection, higher diversity, and increased connectivity can enhance agricultural outcomes by facilitating the flow of ecosystem services. The prairies are a highly significant region for multiple ecosystem functions and agricultural production, with over 80% of Canada’s farmland located in this area. This landscape supports several strategic crops, including oilseeds such as canola, cereals like barley and wheat (both spring and winter), oats, and many others that are essential to the resilience and sustainability of Canada’s agri-food economy. However, the prairie landscape is increasingly shaped by pressures such as land-use intensification, climate variability, and habitat fragmentation. These changes are largely driven by the expansion of cultivated land and the simplification of landscape structure. While such trends have boosted production in the short term, they have also contributed to biodiversity loss and a decline in ecosystem service capacity. Despite its productivity, the region has experienced significant ecological degradation. Intensification practices including the removal of natural buffers, expansion of field sizes, wetland drainage, and conversion of marginal lands have disrupted hydrological functions, lowered water quality, and undermined biodiversity (Baulch et al. 2021 ). Nevertheless, the prairies hold considerable potential for climate adaptation and ecological restoration. With intentional land management, natural and semi-natural areas can serve as green infrastructure that supports carbon sequestration, species movement, and the sustained delivery of ecosystem services crucial to resilient agriculture (Pashanejad 2025 ). 2.2. Data 2.2.1. Crop yield data We obtained crop yield data from Agriculture and Agri-Food Canada (AAFC, 2024) at the SLC unit scale for the major crops cultivated across the prairie region. While the dataset provides a temporal record of annual crop yields, we focused on the year 2020 to align with the ecosystem service (ES) mapping outputs. Our analysis concentrated on four strategically important crops in the region: canola, barley, and wheat (including both spring and winter varieties). The spatial distribution of these crops across SLC units is illustrated in Fig. 1 . Based on these data, we developed a single composite crop yield index to serve as the response variable in both GAM models (ES and LSM). To create this, we first standardized the raw yield values (kg/acre) of the four major crops by converting them z-scores. This index reflects the productivity per acre of cultivated land within an SLC, not the total production of the entire area. This distinction allows our analysis to focus on the factors driving on-farm productivity, independent of how much land within the SLC unit is allocated to agriculture versus natural habitat. To calculate landscape structure metrics, we used the AAFC Annual Crop Inventory for the year 2020. This land cover/land use dataset was also used as a primary input in the ES mapping process, ensuring consistency across analyses. The crop inventory provides detailed information on land use and crop type, enabling the quantification of both compositional and configurational aspects of landscape heterogeneity. 2.2.2. ES provisioning data The ecosystem service variables used as predictors in our GAMs were derived from a suite of biophysical models. These models were initially run at a 300 m resolution, a scale chosen to balance the need for fine-grained landscape detail with computational feasibility across the large prairies. We then aggregated pixel-level outputs to the SLC unit level for analysis. Full methodological details of the ES mapping are provided in SI file but we summarize the key outputs here. Pollination service was estimated using the InVEST Crop Pollination model, which generates an index of pollinator abundance based on nesting sites and floral resources. Habitat Quality was also modeled with InVEST, providing an index of the landscape's capacity to support terrestrial biodiversity based on habitat suitability and threats. Soil Erosion Control was quantified as the amount of avoided erosion (t/ha/yr) using the InVEST Sediment Delivery Ratio (SDR) model. Finally, Carbon Storage represents the total carbon (tonnes/ha) stored in aboveground biomass, belowground biomass, dead organic matter, and soil, as estimated by the InVEST Carbon model. The spatial distribution of these key ecosystem services across the study area is shown in Fig. 2 . 2.3. Landscape structure metrics To assess the spatial structure of landscapes and their influence on ES flow and crop productivity, we calculated a series of landscape metrics at the SLC unit level. Metrics were derived from the 2020 Annual Crop Inventory (AAFC 2020), resampled to 300 m resolution to match the scale of analysis and for computational efficiency at the large area of the Canadian prairies. Calculations were performed using the landscapemetrics package in R (Hesselbarth et al. 2019 ). We organized the metrics into two main categories: composition and configuration, representing the quantity and spatial arrangement of land cover types, respectively. 2.3.1. Composition Metrics A composition metric reflects the abundance and diversity of natural habitats within each SLC unit. To assess the influence of landscape composition on crop productivity, we calculated several key composition-based indicators as proxies based on the evidence that the amount of habitat is a critical determinant of biodiversity and functional support for ES provisioning (Fahrig 2013 ; Tscharntke et al. 2005 ). First, we used the percentage of natural habitat, which quantifies the proportion of land within each SLC unit classified as natural or semi-natural (e.g., wetlands, grasslands, forests, shrublands). This metric provides an estimate of habitat availability for essential regulating services such as pollination, natural pest control, and water regulation (Chaplin-Kramer et al. 2011; Tscharntke et al. 2005 ). To capture overall landscape heterogeneity, we calculated the Shannon diversity index (McGarigal and Marks 1995 ) for habitat. This index incorporates both the richness and evenness of all natural land cover types within each SLC unit, serving as an indicator of spatial multifunctionality. More heterogeneous landscapes are generally associated with greater biodiversity and more stable ecological processes. In addition, we calculated the Shannon diversity index for croplands to evaluate diversity specifically within agricultural land use classes. This metric reflects the variety and distribution of crop types across each unit. Higher crop diversity has been linked to enhanced ecological resilience, improved pest suppression, and more diverse ecosystem service provisioning (Kremen and Miles 2012). 2.3.2. Configuration Metrics A configuration metric evaluates the spatial arrangement and structural connectivity of landscapes. We selected configuration metrics that capture the fragmentation, continuity, and potential spillover dynamics of land cover patches within each SLC unit. These metrics help account for the internal flow of ecosystem services by assessing how natural and managed elements are organized in space. The “clumpiness” index was used to measure the degree of spatial aggregation of natural habitat patches (He, DeZonia, and Mladenoff 2001 ). Higher clumpiness values indicate more contiguous blocks of habitat, which can enhance within-field ecosystem service flows, particularly those related to pollination and natural pest regulation. This metric quantifies the compactness of habitat areas and their capacity to support consistent service provision. We also included the patch cohesion index, which assesses the physical connectedness of land cover patches belonging to the same class. Greater cohesion suggests stronger ecological connectivity across the landscape, supporting species movement, seed dispersal, and the delivery of services (Mitchell, Bennett, and Gonzalez 2013). It serves as an indicator of functional habitat networks, which are essential for maintaining ecological resilience in agricultural systems. Finally, edge density was calculated to quantify the length of habitat edges per unit area (McGarigal and Marks 1995 ). This metric reflects the extent of interface between natural and agricultural areas, where ecosystem service spillover is likely to occur. High edge density can signal increased interaction zones between habitats and crop fields, potentially influencing service delivery dynamics such as pollination, pest suppression, and nutrient flow (Blitzer et al. 2012 ; Ries et al. 2004 ). 2.4. Generalized Additive Model Generalized Additive Models (GAMs) are flexible, nonparametric extensions of traditional Generalized Linear Models (GLMs) designed to model complex, nonlinear relationships between predictor variables and a response variable (Hastie & Tibshirani, 1990). Unlike GLMs, which assume linearity between covariates and the response, GAMs employ smooth functions (e.g., splines) to model the effects of each predictor, making them particularly suitable for ecological data where relationships are often nonlinear, context-dependent, and shaped by interacting factors. Mathematically, a GAM can be written as: $$\:{y}_{i}={\beta\:}_{0}+{f}_{1}{(x}_{1i})+{f}_{2}{(x}_{2i})+\dots\:+{f}_{k}{(x}_{ki})+\:{\epsilon\:}_{i}$$ where \(\:{y}_{i}\) is the response variable in this case crop productivity index, \(\:{f}_{1}(.)\) represent smooth functions for the predictor \(\:{(x}_{i}\) ), and \(\:{\epsilon\:}_{i}\:\sim\:N(0,\:{\sigma\:}^{2})\) is the normally distributed error term. We fitted two GAMs to explain spatial variation in crop productivity across Soil Landscapes of Canada (SLC) units in the Prairie region. The Ecosystem Services (ES) model examined the influence of pollination, habitat quality, erosion control, and carbon storage. The Landscape Structure (LSM) model included structural configuration variables such as natural habitat cover, crop diversity, connectivity, and patch aggregation. Both models included a two-dimensional smooth function of geographic coordinates (longitude and latitude) to account for residual spatial autocorrelation not explained by the predictors. This approach allows GAMs to flexibly capture spatial structure without relying on the user pre-define a spatial weight matrix (Wood, 2017). The model's error distribution was specified based on the nature of the response variable. Since the crop productivity index is a continuous measure, we used the Gaussian family with an identity link. This structure, which assumes normally distributed residuals, was confirmed to be appropriate through a thorough examination of residual diagnostics. It models the mean of the response as a linear (or smoothed) function of predictors and is appropriate when residuals are approximately symmetrically distributed. To isolate the contribution of ecosystem service and landscape structure variables, we incorporated a consistent set of control covariates into both models. We specifically considered soil organic matter accounting for soil fertility and microbial activity, elevation capturing topographic and microclimate variation, mean temperature and precipitation representing climate suitability and conventional tillage as representative of management practice influencing soil health and crop performance. We conducted an analysis of Variance Inflation Factors (VIF) to explore the multicollinearity among predictors before fitting the final model (See Supplementary Information file). Nutrient retention for nitrogen and phosphorus and conservation tillage has shown a high multicollinearity, therefore we retained only conventional tillage, which showed a moderate VIF value and consistent significant importance (confidence intervals for point estimates did not overlap zero) across models. Our primary analysis focused on the composite crop productivity index to identify overarching patterns across the prairie system. However, to investigate potential crop-specific responses and ensure the robustness of our findings, we also conducted a secondary analysis. This involved fitting the ES model and the LSM model separately for each of the major individual crops: canola, barley, and wheat (both spring and winter). The predictor variables and model specifications remained identical to those described below, with only the response variable changing for each run. Ecosystem Service model: $$\:{Y}_{i}^{ES}={\beta\:}_{0}+{\beta\:}_{1}.{Carbon}_{i}+{\beta\:}_{2}.SoilOM+{\beta\:}_{3}.{Elevation}_{i}+{\beta\:}_{4}.{Temp}_{i}+{\beta\:}_{5}.{Precp}_{i}+\:{\beta\:}_{6}.{Tillage}_{i}+{f}_{1}{(Pollination}_{i})+{f}_{2}{(Habitat}_{i})+{f}_{3}{(Erosion}_{i})+{f}_{4}{(lon}_{i},\:{lat}_{i})+\:{\epsilon\:}_{i}$$ Landscape Structure Model: $$\:{Y}_{i}^{LSM}={\beta\:}_{0}+{\beta\:}_{1}.SoilOM+{\beta\:}_{2}.{Elevaion}_{i}+{\beta\:}_{3}.{Temp}_{i}+{\beta\:}_{4}.{Precp}_{i}+\:{\beta\:}_{5}.{Tillage}_{i}+{\beta\:}_{6}.{EdgeDensity}_{i}+{\beta\:}_{6}.{PatchAggregation}_{i}+{f}_{1}{(\%NaturalHabitat}_{i})+{f}_{2}{(CropDiversity}_{i})+{f}_{3}{(Connectivity}_{i})+{f}_{4}{(lon}_{i},\:{lat}_{i})+\:{\epsilon\:}_{i}$$ where \(\:Y\) is the response variable, representing the composite crop productivity index and \(\:{\beta\:}_{0}\) is the model intercept, representing the baseline crop productivity. To indicate whether a variable has a linear and non-linear response in the crop productivity, we followed a model-building approach based on preliminary diagnostics and ecological theory. Variables for which we hypothesized a complex, non-linear relationship with crop yield (e.g., Pollination, % Natural Habitat) were modeled using smooth functions (f). Conversely, for predictors where initial diagnostic plots suggested a primarily linear response, or where there was no strong a priori theoretical reason to expect non-linear, we modeled them as linear terms ( \(\:\beta\:\) ). These include the control variables for soil organic matter (SoilOM), elevation (Elevation), temperature (Temp), precipitation (Precp), and conventional tillage (Tillage). The model also estimates linear effects for specific ES or LSM predictors like carbon, edge density, and patch aggregation as initial diagnostic plots for these variables suggested a primarily linear response. The \(\:\beta\:\) coefficients represent the estimated slope for each of these variables. The variables expected to have a complex, non-linear relationship with crop productivity are modeled using smooth functions ( \(\:f\) ). These smoothers (by default thin plate splines), allow the model to flexibly capture the shape of the relationship from the data itself. To account for the spatial autocorrelation, a two-dimensional smooth term over geographic coordinates (longitude and latitude) is included in both models. \(\:{\epsilon\:}_{i}\) is the residual error term, assumed to be normally distributed. 3. Results 3.1. Landscape structure analysis We observed clear spatial gradients and strong correlations among landscape structure metrics across the study area. Areas with a high percentage of natural habitat were more spatially cohesive (r = 0.76) and exhibited higher edge density (r = 0.67), but these landscapes often contained fewer habitat types. Notably, habitat cohesion was also moderately associated with edge density (r = 0.34), suggesting some alignment between connectedness and edge exposure. Interestingly, landscapes with greater habitat diversity tended to support more diverse cropping systems (r = 0.60), highlighting potential synergies between natural and managed diversity. However, our findings reveal a strong negative correlation between natural habitat extent and habitat diversity (r = − 0.76), indicating that larger areas of contiguous habitat often result in reduced habitat heterogeneity likely due to the dominance of single habitat types, such as continuous native grasslands or forest patches. Similarly, higher cohesion was strongly associated with lower habitat diversity (r = − 0.90), reinforcing the idea that well-connected landscapes can be structurally simplified, which may constrain ecosystem service multifunctionality. The spatial distribution of each landscape metric across the prairie region is presented in Fig. 3 , illustrating clear regional contrasts in natural cover, habitat and crop diversity, clumpiness, cohesion, and edge density. 3.2. GAM model output: ES and landscape structure effects on crop yield Both the Ecosystem Services (ES) and Landscape Structure (LSM) models explained a comparable proportion of variance in crop productivity (Adjusted R² = 0.667 and 0.662, respectively; see Table 2 ) underscoring that both biophysical functions and spatial landscape attributes are important in shaping agricultural outcomes. In the ES model, key regulating services such as pollination, habitat quality, and soil erosion control showed strong, significant non-linear effects (p < 0.001), while tillage emerged as the most influential linear management variables. Carbon storage, on the other hand, showed a marginally negative association (p = 0.07). Most of the control variables, including soil organic matter, elevation, temperature, and precipitation, were not statistically significant (p > 0.1) in this model (see Table 2 ). In contrast, the LSM model, designed to capture how landscape configuration and composition potentially mediate ecosystem service flows to agricultural areas, highlighted the importance of structural attributes. Specifically, crop diversity and landscape connectivity were both significant predictors of crop yields. However, while these structural configuration metrics demonstrated significance, the overall proportion of natural habitat (%) was non-significant when considered alongside configuration metrics. Among the linear terms, habitat aggregation and tillage were positively associated with yield, whereas soil organic matter had a significant negative effect and precipitation was marginally negative. Habitat diversity and edge density—two landscape structure metrics—were not significant in this model. These two models reveal distinct but complementary mechanisms: the ES model reflects the biophysical functioning of the landscape, while the LSM model highlights how landscape structure and spatial organization condition the flow and impact of those services (Table 2 ). As a robustness check we assessed whether these patterns held true for individual crops separately. These crop-specific models, detailed in Supplementary Information, largely reinforce our main findings. Across all crops, both the ES and LSM models explained a substantial portion of yield variance, and the spatial term remained a dominant, highly significant factor. However, this secondary analysis also revealed important nuances. For instance, while the amount of natural habitat was not significant in the composite models, canola yield showed a unique and significant positive response to the percentage of natural habitat, likely reflecting its high dependence on pollinators whose population and foraging activities are supported by those areas. This highlights that while general principles apply, specific crop needs can alter the relative importance of certain landscape attributes in facilitating beneficial ES flows. Table 2 Summary of generalized additive models (GAMs) explaining variation in crop productivity across the Canadian Prairies. The table compares the performance of two models: one based on ecosystem services (ES) and another on landscape structure metrics (LSM). Both models include spatial smooth terms and control for climatic, topographic, and management-related variables. Metric Ecosystem Services (ES) Model Landscape Structure (LSM) Model Adjusted R² 0.667 0.662 Deviance Explained (%) 68.0 67.5 Significant Smooth Terms Pollination (+, ***) Habitat Quality (+, ***) Soil Erosion Control (+, ***) Spatial Coordinates (Smooth, ***) Crop Diversity (+, *) Landscape Connectivity (+, *) Spatial Coordinates (Smooth, ***) Non-Significant Smooth terms — Percentage of Natural Habitat (NS) Significant Linear Terms Tillage (+, ***) Carbon Storage (−, .) Aggregation of Natural Habitat Patches (+, *) Soil Organic Matter Content (−, *) Tillage (+, ***) Mean Annual Precipitation (−, .) Non-Significant Linear terms Soil Organic Matter Content (NS)Elevation (NS)Mean Annual Temperature (NS)Mean Annual Precipitation (NS) Habitat Diversity (NS)Edge Density (NS)Elevation (NS)Mean Annual Temperature (NS) Note: Significance of non-linear effects (smooth terms) and linear terms is indicated as follows: (***) for p < 0.001, (**) for p < 0.01, (*) for p < 0.05, and (.) for p < 0.1 (marginally significant). Terms with p ≥ 0.1 are considered not significant (NS). Linear terms include a plus (+) or minus (−) symbol to denote the direction of the relationship (positive or negative). “Spatial (Smooth)” refers to the two-dimensional smooth over spatial coordinates s(X, Y), with its significance also indicated using the aforementioned symbols. 3.3. Partial effect plots for the Ecosystem Services (ES) and Landscape Structure model Figure 4, (panel A) shows that pollination had a non-linear relationship with crop yield, characterized by an initial decline followed by a steady increase in partial effect values beyond moderate pollination levels. Habitat quality exhibited a consistently negative effect across its range, while erosion control showed a strongly non-linear association, with partial effects decreasing at low to moderate levels and increasing substantially at higher values. The significance of the spatial smooth term (X, Y coordinates) reveals strong spatial autocorrelation in crop yield unexplained by the other model covariates. The model identified distinct geographic clusters of positive and negative partial effects, corresponding to regions where underlying spatial processes either enhance or reduce crop productivity. The Landscape Structure model similarly revealed complex predictor effects. For instance, crop diversity showed a non-linear effect, with an initial decrease in its partial effect on yield as diversity increased to moderate levels (approx. SHDI 1.5-2.0), followed by a slight positive trend at higher diversity values. Landscape connectivity, a key facilitator of potential ES flows like pollinator movement, showed a near-linear positive relationship with crop productivity. In contrast, percentage of natural habitat had a weak and slightly negative effect that remained largely flat across the range. The spatial smoother captured spatial heterogeneity not explained by the included variables, indicating consistent spatial structure in the residuals across the study region. The model diagnostics plots for both models are shown in Fig. 5. These plots indicated that the assumptions of our GAM models were fitted at an acceptable level. The Q-Q plots showed an approximately normal distribution of residuals with slight deviations at the tails (Fig. 5, panel A, top-left), and residual histograms were unimodal and centered around zero (Fig. 5, panel A, top-right), indicating minimal skewness. The residuals vs. fitted plots revealed no pronounced heteroscedasticity or structure, suggesting a good fit across the range of predicted values (Fig. 5, bottom-left). The response vs. linear predictor plots further confirmed the non-linear relationship captured by the model (Fig. 5, bottom-right). In addition, spatial diagnostic maps of residuals (Fig. 5, panels C and D) revealed geographically patterned residuals in some parts of the study area, particularly in underpredicted and overpredicted areas. This highlights localized model uncertainty and suggests the presence of spatial heterogeneity not fully captured by the selected predictors. 4. Discussion The complex relationship between spatial ecological patterns, the ecosystem services they generate, and agricultural productivity is central to designing sustainable and resilient agricultural landscapes. In this research, we examined how landscape patterns and the mechanism they support including both ecological functions and land management practices interact to influence crop yields across large, intensively managed systems like the Canadian prairies. While previous research has increasingly recognized that landscape structure underpins service flows and functional outcomes, the specific mechanism by which internal functional connectivity—the spatial transfer of ecosystem service benefits within defined ecological units—mediates agricultural performance remains underexplored. We addressed this critical gap by investigating how internal ES flows, as mediated by landscape composition and configuration within Soil Landscape of Canada units, collectively shape crop productivity as an emergent property of these multifunctional units. Our findings demonstrate that both key regulating ecosystem services such as pollination and soil erosion control and the structural attributes of the landscape play complementary and significant roles in determining crop yield performance. This dual influence highlights a fundamental principle: agricultural output is not merely a function of inputs and field-level conditions but is deeply embedded within, and responsive to, the broader landscape matrix and the ecological processes it supports (Kremen and Miles 2012). Notably, our results revealed that landscape configuration metrics, particularly connectivity and crop diversity were stronger predictors of crop yield than the simple composition measures such as the amount of natural habitat. However, crop-specific analysis reveals a critical exception to this pattern. For canola, which is a high insect pollination-dependent crop, the amount of natural habitat was a significant positive predictor (See SI file). This suggests that while functional connectivity may drive the overall system, for specific services tightly linked to habitat-dependent organisms, the total habitat area is foundational to support species richness and persistence (Fahrig 2013 ). Overall, our results suggest that at the broad SLC scale, spatial arrangement and functional connectivity to croplands are more influential in shaping aggregated agricultural performance. However, this does not diminish the importance of the habitat amount; rather, it refines its role, indicating that its effect can be enhanced or mediated by landscape configuration (Tscharntke et al. 2005 ). Previous studies (Blitzer et al. 2012 ; Kremen, Williams, and Thorp 2002 ; Mitchell et al. 2013 ) have shown that metrics like landscape connectivity serve as more effective proxies for the potential ES flows — such as pollinator spillover or pest predator movement — between natural areas and farmlands. Our modeling reinforces the perspective that once ES flow-mediating attributes are incorporated, the simple percentage of habitat may provide limited additional explanatory power. This aligns with previous findings that emphasize habitat interfaces as critical interaction zones for service delivery (Blitzer et al. 2012 ) Heterogeneous landscapes tend to support greater habitat availability and higher species diversity, which can enhance ecosystem resilience, stability, and recovery following disturbances (Feit et al. 2021 ; Tscharntke et al. 2012). In such complex agricultural systems, increased availability of habitats for pollinators can improve ecosystem functioning and service delivery, ultimately leading to higher crop yields and quality through enhanced pollination and natural pest regulation (Estrada-Carmona et al. 2022 ). However, our findings reveal a critical distinction between local service benefits and landscape-scale trade-offs. The observed negative relationship between aggregated habitat quality and crop yield at the SLC scale likely reflects a spatial trade-off rather than a direct detrimental effect of habitat on adjacent crop performance. SLC units with extensive, high-quality natural remnants, such as native grasslands, will inherently have a smaller proportion of their area dedicated to, or lower overall output from, intensive crop production when aggregated. The agricultural land that does exist within these high-conservation value landscapes may therefore be inherently less productive. Thus, a high average habitat quality score for an SLC unit can be spatially correlated with lower overall agricultural yield for that unit, reflecting broad patterns of land suitability and agricultural marginality across the prairies. This highlights that relationship between ES indicators and agricultural productivity at this scale are heavily influenced by underlaying environmental gradients. While this study provides practical insights across a globally significant agricultural region, we acknowledge its limitations. Specifically, the use of landscape configuration to infer ecosystem service interactions implies relationship that are correlational, not necessarily causal. Further research is needed to disentangle the underlying mechanisms driving these observed associations. The primary objective of this study was to explore the spatial dynamics of landscape structure in a heavily intensified agroecosystem, such as the Canadian prairies. Although we included key controlling factors such as land management, climate, and topography, we did not fully account for the complex, multi-layered nature of the agri-food system. Our analysis provides a robust snapshot using 2020 data. While this cross-sectional approach reveals critical spatial relationships, future work using a multi-year panel data analysis could further strengthen these findings by controlling for time-invariant factors. Such extensions could significantly enhance the explanatory and predictive capacity of the current spatial modeling framework. However, while our study highlights the critical spatial relationship in a given year, future temporal analyses are essential for disentangling these complex drivers and confirming the causal pathways of landscape influence on agricultural outcome. It is crucial to recognize that the landscape patterns we measured are not solely the result of ecological forces, but are profoundly shaped by social processes (Berkes, Colding, and Folke 2008) given high intensity of agriculture in the study area. The strong predictive power of conventional tillage, for instance, directly reflects farmer management decisions, which are influenced by economic pressures, tradition, and policy incentives. Similarly, while our results show a clear benefit to crop diversity, the prevalence of monocultures across the prairies is driven by social and economic systems that favor specialization (Qualman 2019 ; Qualman et al. 2020). The very configuration of natural habitats—which our study shows is critical for mediating ES flows—is also constrained by human-defined property lines that dictate land management decisions at the parcel level. This delineation of ownership often leads to habitat fragmentation which influences the success and spatial extent of conservation programs. Therefore, understanding and improving these multifunctional landscapes requires an approach that extends beyond ecological knowledge to actively engage with the complex social and economic and political drivers that shapes their current state and future trajectories (Berkes et al. 2008; Ostrom 2009 ). Landscape patterns, quantified through land cover composition and structure, influence underlying ecological processes such as pollination or soil stability. These, in turn, affect the provisioning of crop yield—a key service in agroecosystems. Our modeling approach demonstrates how these interactions manifest across space and can inform decisions about where and how to intervene. For instance, enhancing landscape cohesion and crop diversity may offer greater returns than simply increasing the amount of semi-natural habitat. By operationalizing this pattern-process-service framework, our research supports more informed, spatially explicit planning for multifunctional landscapes and a paradigm shift. Through this research, we tried to shift our perspective from a static collection of patches to dynamic functional landscapes defined by ES flows (Fu et al. 2025 ). In addition, the use of generalized additive models allowed us to capture the non-linear relationship between key ecosystem services. Thus, reinforcing agricultural systems respond to ecological support in complex, non-linear ways, an essential insight for effective management. 5. Conclusion The future of agricultural sustainability depends on maintaining landscape complexity rather than reducing it. Homogenized landscapes may maximize short-term yield but risk long-term ecological degradation and declining resilience. Our results show that multifunctionality can support both productivity and conservation. This calls for a paradigm shift in agricultural policy and planning: from maximizing yield per hectare to optimizing landscape configuration for long-term resilience and service provision. Moving forward, integrating detailed management data, socio-economic drivers, and farmer decision-making into spatial models will be crucial to developing actionable, context-sensitive strategies for sustainable agriculture. Our results show that connectivity and diversity—two core principles of resilient and sustainable agricultural systems—play a critical role in enabling ecosystem service flows and improving agricultural outcomes such as crop yield. The observed synergy between biodiversity, ecosystem service provisioning, and crop productivity, often strengthened by improved spatial configuration and landscape heterogeneity, underscores the potential for ecological intensification. By quantitatively linking landscape-scale patterns and internal service flows to crop productivity, our study advances this understanding and offers practical insights for sustainable landscape planning. Declarations The authors have no relevant financial or non-financial interests to disclose. Author Contribution All authors contributed to the study conception and design. Material preparation, data collection and analysis and the discussion were performed by Ehsan Pashanejad. The first draft of the manuscript was written by Ehsan Pashanejad. Lael Parrott and Brian E. Robinson supervised the project and provided feedback and edits on previous versions of the manuscript. All authors read and approved the final manuscript. Acknowledgement We acknowledge the support of the Natural Sciences and Engineering Research Council of Canada (NSERC), [funding reference number NSERC NETGP 523374–18]. Cette recherche a été financée par le Conseil de recherches en sciences naturelles et en génie du Canada (CRSNG), [numéro de référence NSERC NETGP 523374–18]. We are grateful to the Agroclimate, Geomatics and Earth Observation Division of Agriculture and Agri-Food Canada for providing crop yield data, and we extend our thanks to Bahram Daneshfar for his generous support in supplying the major crop yield data at the Soil Landscape of Canada scale. Data Availability The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request. The primary crop yield data used in this study are available from Agriculture and Agri-Food Canada (AAFC) as cited in the manuscript. References Agriculture and Agri-Food Canada, 2024, “Interpolated Crop Yield by Soil Landscape of Canada (SLC), 2010–2023”, Agroclimate, Geomatics and Earth Observation Division, Science and Technology Baulch, Helen, Colin Whitfield, Jared Wolfe, Nandita Basu, Angela Bedard-Haughn, Kenneth Belcher, Robert Clark, Grant Ferguson, Masaki Hayashi, Andrew Ireson, Patrick Lloyd-Smith, Phil Loring, John W. Pomeroy, Kevin Shook, and Christopher Spence. 2021. “Synthesis of Science: Findings on Canadian Prairie Wetland Drainage.” Canadian Water Resources Journal / Revue Canadienne Des Ressources Hydriques 46(4):229–41. doi: 10.1080/07011784.2021.1973911 . Berkes, Fikret, Johan Colding, and Carl Folke. 2008. Navigating Social-Ecological Systems: Building Resilience for Complexity and Change. Cambridge university press. Blitzer, Eleanor J., Carsten F. Dormann, Andrea Holzschuh, Alexandra-Maria Klein, Tatyana A. Rand, and Teja Tscharntke. 2012. “Spillover of Functionally Important Organisms between Managed and Natural Habitats.” Agriculture, Ecosystems & Environment 146(1):34–43. doi: 10.1016/j.agee.2011.09.005 . Chaplin-Kramer, Rebecca, Megan E. O’Rourke, Eleanor J. Blitzer, and Claire Kremen. 2011. “A Meta-Analysis of Crop Pest and Natural Enemy Response to Landscape Complexity.” Ecology Letters 14(9):922–32. doi: 10.1111/j.1461-0248.2011.01642.x . Cozim-Melges, Felipe, Raimon Ripoll-Bosch, G. F. Veen, Philipp Oggiano, Felix J. J. A. Bianchi, Wim H. van der Putten, and Hannah H. E. van Zanten. 2024. “Farming Practices to Enhance Biodiversity across Biomes: A Systematic Review.” Npj Biodiversity 3(1):1. doi: 10.1038/s44185-023-00034-2 . Estrada-Carmona, Natalia, Andrea C. Sánchez, Roseline Remans, and Sarah K. Jones. 2022. “Complex Agricultural Landscapes Host More Biodiversity than Simple Ones: A Global Meta-Analysis.” Proceedings of the National Academy of Sciences 119(38). doi: 10.1073/pnas.2203385119 . Fahrig, Lenore. 2013. “Rethinking Patch Size and Isolation Effects: The Habitat Amount Hypothesis” edited by K. Triantis. Journal of Biogeography 40(9):1649–63. doi: 10.1111/jbi.12130 . Feit, Benjamin, Nico Blüthgen, Eirini Daouti, Cory Straub, Michael Traugott, and Mattias Jonsson. 2021. “Landscape Complexity Promotes Resilience of Biological Pest Control to Climate Change.” Proceedings of the Royal Society B: Biological Sciences 288(1951):20210547. doi: 10.1098/rspb.2021.0547 . Fu, Bojie, Yanxu Liu, Wenwu Zhao, and Jianguo Wu. 2025. “The Emerging ‘Pattern-Process-Service-Sustainability’ Paradigm in Landscape Ecology.” Landscape Ecology 40(3):54. doi: 10.1007/s10980-025-02063-7 . Galpern, Paul, Jess Vickruck, James H. Devries, and Michael P. Gavin. 2020. “Landscape Complexity Is Associated with Crop Yields across a Large Temperate Grassland Region.” Agriculture, Ecosystems & Environment 290:106724. doi: 10.1016/j.agee.2019.106724 . He, H. S., B. E. DeZonia, and D. J. Mladenoff. 2001. “Erratum: An Aggregation Index (AI) to Quantify Spatial Patterns of Landscapes (Landscape Ecology (200) 15 (591–601)).” Landscape Ecology 16(1):87. doi: 10.1023/A:1017308405507 . Hesselbarth, Maximilian H. K., Marco Sciaini, Kimberly A. With, Kerstin Wiegand, and Jakub Nowosad. 2019. “Landscapemetrics: An Open-source R Tool to Calculate Landscape Metrics.” Ecography 42(10):1648–57. doi: 10.1111/ecog.04617 . Kremen, Claire, and Albie Miles. 2012. “Ecosystem Services in Biologically Diversified versus Conventional Farming Systems: Benefits, Externalities, and Trade-Offs.” Ecology and Society 17(4):art40. doi: 10.5751/ES-05035-170440 . Kremen, Claire, Neal M. Williams, and Robbin W. Thorp. 2002. “Crop Pollination from Native Bees at Risk from Agricultural Intensification.” Proceedings of the National Academy of Sciences 99(26):16812–16. doi: 10.1073/pnas.262413599 . Madin, Michael Biwalib, and Katherine S. Nelson. 2023. “Effects of Landscape Simplicity on Crop Yield: A Reanalysis of a Global Database” edited by M. Gaglio. PLOS ONE 18(12):e0289799. doi: 10.1371/journal.pone.0289799 . Margosian, Margaret L., Karen A. Garrett, J. M. Shawn Hutchinson, and Kimberly A. With. 2009. “Connectivity of the American Agricultural Landscape: Assessing the National Risk of Crop Pest and Disease Spread.” BioScience 59(2):141–51. doi: 10.1525/bio.2009.59.2.7 . McGarigal, Kevin, and Barbara J. Marks. 1995. FRAGSTATS: Spatial Pattern Analysis Program for Quantifying Landscape Structure. Mitchell, Matthew G. E., Elena M. Bennett, and Andrew Gonzalez. 2013. “Linking Landscape Connectivity and Ecosystem Service Provision: Current Knowledge and Research Gaps.” Ecosystems 16(5):894–908. doi: 10.1007/s10021-013-9647-2 . Nguyen, Lan H., Samuel V. J. Robinson, and Paul Galpern. 2022. “Effects of Landscape Complexity on Crop Productivity: An Assessment from Space.” Agriculture, Ecosystems & Environment 328:107849. doi: 10.1016/j.agee.2021.107849 . Ostrom, Elinor. 2009. “A General Framework for Analyzing Sustainability of Social-Ecological Systems.” Science 325(5939):419–22. Pashanejad, Ehsan. 2025. Balancing Agriculture and Conservation in the Canadian Prairies. Ottawa, Ontario. Pashanejad, Ehsan, Ali Kharrazi, Zuelclady M. F. Araujo-Gutierrez, Brian E. Robinson, Brian D. Fath, and Lael Parrott. 2024. “A Functional Connectivity Approach for Exploring Interactions of Multiple Ecosystem Services in the Context of Agricultural Landscapes in the Canadian Prairies.” Ecosystem Services 68:101639. doi: 10.1016/j.ecoser.2024.101639 . Qiu, Jiangxiao, and Matthew Mitchell. 2024. “Understanding Biodiversity – Ecosystem Service Linkages in Real Landscapes.” Landscape Ecology 39(11):188. doi: 10.1007/s10980-024-01980-3 . Qualman, Darrin. 2019. “Tackling the Farm Crisis and the Climate Crisis: A Transformative Strategy for Canadian Farms and Food Systems.” 102. Qualman, Darrin, Annette Aurélie Desmarais, André Magnan, and Mengistu Wendimu. 2020. “Concentration Matters: Farmland Inequality on the Prairies.” Rehman, Abdul, Muhammad Farooq, Dong-Jin Lee, and Kadambot H. M. Siddique. 2022. “Sustainable Agricultural Practices for Food Security and Ecosystem Services.” Environmental Science and Pollution Research 29(56):84076–95. doi: 10.1007/s11356-022-23635-z . Richter, Franziska J., Matthias Suter, Andreas Lüscher, Nina Buchmann, Nadja El Benni, Rafaela Feola Conz, Martin Hartmann, Pierrick Jan, and Valentin H. Klaus. 2024. “Effects of Management Practices on the Ecosystem-Service Multifunctionality of Temperate Grasslands.” Nature Communications 15(1):3829. doi: 10.1038/s41467-024-48049-y . Rieb, Jesse T., and Elena M. Bennett. 2020. “Landscape Structure as a Mediator of Ecosystem Service Interactions.” Landscape Ecology 35(12):2863–80. doi: 10.1007/s10980-020-01117-2 . Ries, Leslie, Robert J. Fletcher, James Battin, and Thomas D. Sisk. 2004. “Ecological Responses to Habitat Edges: Mechanisms, Models, and Variability Explained.” Annual Review of Ecology, Evolution, and Systematics 35(1):491–522. doi: 10.1146/annurev.ecolsys.35.112202.130148 . Tamburini, Giovanni, Riccardo Bommarco, Thomas Cherico Wanger, Claire Kremen, Marcel G. A. van der Heijden, Matt Liebman, and Sara Hallin. 2020. “Agricultural Diversification Promotes Multiple Ecosystem Services without Compromising Yield.” Science Advances 6(45). doi: 10.1126/sciadv.aba1715 . Tscharntke, Teja, Alexandra M. Klein, Andreas Kruess, Ingolf Steffan-Dewenter, and Carsten Thies. 2005. “Landscape Perspectives on Agricultural Intensification and Biodiversity – Ecosystem Service Management.” Ecology Letters 8(8):857–74. doi: 10.1111/j.1461-0248.2005.00782.x . Tscharntke, Teja, Jason M. Tylianakis, Tatyana A. Rand, Raphael K. Didham, Lenore Fahrig, Péter Batáry, Janne Bengtsson, Yann Clough, Thomas O. Crist, Carsten F. Dormann, Robert M. Ewers, Jochen Fründ, Robert D. Holt, Andrea Holzschuh, Alexandra M. Klein, David Kleijn, Claire Kremen, Doug A. Landis, William Laurance, David Lindenmayer, Christoph Scherber, Navjot Sodhi, Ingolf Steffan-Dewenter, Carsten Thies, Wim H. van der Putten, and Catrin Westphal. 2012. “Landscape Moderation of Biodiversity Patterns and Processes ‐ Eight Hypotheses.” Biological Reviews 87(3):661–85. doi: 10.1111/j.1469-185X.2011.00216.x . Wang, Tongli, Andreas Hamann, Dave Spittlehouse, and Carlos Carroll. 2016. “Locally Downscaled and Spatially Customizable Climate Data for Historical and Future Periods for North America” edited by I. Álvarez. PLOS ONE 11(6):e0156720. doi: 10.1371/journal.pone.0156720 . Wood, Sylvia L. R., Kyle T. Martins, Véronique Dumais-Lalonde, Olivier Tanguy, Fanny Maure, Annick St-Denis, Bronwyn Rayfield, Amanda E. Martin, and Andrew Gonzalez. 2022. “Missing Interactions: The Current State of Multispecies Connectivity Analysis.” Frontiers in Ecology and Evolution 10. doi: 10.3389/fevo.2022.830822 . Additional Declarations No competing interests reported. Supplementary Files Supplementaryinformation.docx Cite Share Download PDF Status: Published Journal Publication published 21 Mar, 2026 Read the published version in Landscape Ecology → Version 1 posted Editorial decision: Revision requested 27 Oct, 2025 Reviews received at journal 10 Oct, 2025 Reviews received at journal 02 Oct, 2025 Reviewers agreed at journal 10 Sep, 2025 Reviewers agreed at journal 08 Sep, 2025 Reviewers invited by journal 07 Sep, 2025 Editor assigned by journal 29 Aug, 2025 Submission checks completed at journal 29 Aug, 2025 First submitted to journal 27 Aug, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7467778","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":513184647,"identity":"49d74af6-2873-4388-9d16-5f2bd62ce2d0","order_by":0,"name":"Ehsan Pashanejad","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA90lEQVRIiWNgGAWjYDACCShtwMB84MAHIIONnXgtbIkPZ4C0MBOvhcfYmAfEIqSFX7r94efCHXb25uxnzKRtfm2T52NmYPzwMQe3Fsk5Z4ylZ55JTtzZk1Ymndt327CNmYFZcuY23FoMbuQwSPO2MScYHEjeJp3bc5sRqIWNmRePFvsb6Y9/87bV2xucf2Ambdlz256gFgOJBDOgLYcZN9xIMTZm+HE7kaAWiRs5Zta8Z44nbrjxLPFhb8Pt5DZmxma8fuGfkf74Nu+OaqDDkg8c+PHntu389uaDHz7i0QIGjA0wRhsKlxgtDH8IKx4Fo2AUjIKRBwAmO1JvntMgJgAAAABJRU5ErkJggg==","orcid":"","institution":"University of British Columbia","correspondingAuthor":true,"prefix":"","firstName":"Ehsan","middleName":"","lastName":"Pashanejad","suffix":""},{"id":513184648,"identity":"a227fcf4-921b-466d-8810-f599458da217","order_by":1,"name":"Brian Robinson","email":"","orcid":"","institution":"McGill University","correspondingAuthor":false,"prefix":"","firstName":"Brian","middleName":"","lastName":"Robinson","suffix":""},{"id":513184649,"identity":"e516ae04-d5b5-4544-8391-66ce8cf4e6fa","order_by":2,"name":"Lael Parrott","email":"","orcid":"","institution":"University of British Columbia","correspondingAuthor":false,"prefix":"","firstName":"Lael","middleName":"","lastName":"Parrott","suffix":""}],"badges":[],"createdAt":"2025-08-27 04:53:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7467778/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7467778/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s10980-026-02333-y","type":"published","date":"2026-03-21T15:57:41+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":91182430,"identity":"ff38ff8c-ead4-4d2a-8991-e425670df8c5","added_by":"auto","created_at":"2025-09-12 13:18:47","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":838797,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial Distribution of Major Crop Yields across Soil Landscape of Canada (SLC) Units in the Canadian Prairies based on 2020 year data. Data source: Agriculture and Agri-Food Canada, 2024, “Interpolated Crop Yield by Soil Landscape of Canada (SLC), 2010-2023”, Agroclimate, Geomatics and Earth Observation Division, Science and Technology\u003c/p\u003e","description":"","filename":"Figure1.CropYieldMaps4Crops.png","url":"https://assets-eu.researchsquare.com/files/rs-7467778/v1/0c16541fdb5849267393fd60.png"},{"id":91182434,"identity":"6d96feba-cbd4-433c-bc17-97c62fae37b2","added_by":"auto","created_at":"2025-09-12 13:18:47","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":7391925,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial Distribution of key ES predictors used in this analysis.\u003c/p\u003e","description":"","filename":"Figure2.ES.png","url":"https://assets-eu.researchsquare.com/files/rs-7467778/v1/4ec4c11965d97fc36a5768f2.png"},{"id":91183421,"identity":"2eb10160-4b06-4ced-b1ec-077e33930f3d","added_by":"auto","created_at":"2025-09-12 13:26:47","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":385913,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial patterns and interrelationships among landscape structure metrics across the Canadian Prairies. A) Spatial distribution of six landscape metrics calculated at the Soil Landscape of Canada (SLC) unit level. B) Pearson correlation matrix depicting pairwise relationships among the six metrics. Circle size and color indicate the strength and direction of correlation (r), with blue for negative and red for positive associations.\u003c/p\u003e","description":"","filename":"Figure3.landscapeStrucuturemetric.png","url":"https://assets-eu.researchsquare.com/files/rs-7467778/v1/cbe018e3fde99e5db4bede4e.png"},{"id":91182431,"identity":"e10dc9fb-77e5-48d4-8eae-0bfee4c6fc59","added_by":"auto","created_at":"2025-09-12 13:18:47","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":872524,"visible":true,"origin":"","legend":"\u003cp\u003ePartial effects from the Generalized Additive Models (GAMs) relating ecosystem services and landscape structure to crop productivity across the study area. Panel A shows smooth partial effects from the Ecosystem Services (ES) model for (a) Pollination, (b) Habitat Quality, and (c) Erosion Control. Panel (d) displays the spatial smooth term over geographic coordinates (longitude, latitude), accounting for spatial autocorrelation in crop yield unexplained by other predictors. Panel B shows smooth partial effects from the Landscape Structure (LSM) model for (a) Percentage of Natural Habitat (PLAND), (b) Crop Diversity (SHDI), and (c) Landscape Connectivity (Cohesion). Panel (d) again presents the spatial smooth. The solid black line represents the estimated smooth function of each predictor’s effect on crop yield, with shaded bands indicating 95% confidence intervals. These plots illustrate the shape and strength of non-linear relationships between predictors and crop productivity, highlighting context-dependent responses across gradients of ecosystem service provisioning and landscape configuration. For the Spatial Smooth plots, the color scale represents the value of the partial effect at different geographic coordinates. Red areas indicate locations where yield is predicted to be higher than the model average (due to spatial factors alone), while blue areas indicate locations where it is predicted to be lower\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7467778/v1/6601acf03b61f36e0c737225.png"},{"id":91183423,"identity":"d18f5da5-948c-49cc-a011-327bd015aa5d","added_by":"auto","created_at":"2025-09-12 13:26:47","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":856726,"visible":true,"origin":"","legend":"\u003cp\u003eModel diagnostics and residual distributions for the GAM models of crop productivity.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-7467778/v1/9ee5db8edd0360901956ef22.png"},{"id":105223267,"identity":"e884f095-18a9-4183-b934-4401988ee3e9","added_by":"auto","created_at":"2026-03-23 16:01:29","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":7878835,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7467778/v1/0df4a294-eb36-4a2e-b59d-1d8821212ee4.pdf"},{"id":91183641,"identity":"6896af2b-7ef2-492d-8e40-6d7c4dd2bdc0","added_by":"auto","created_at":"2025-09-12 13:34:47","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":796175,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementaryinformation.docx","url":"https://assets-eu.researchsquare.com/files/rs-7467778/v1/bebe148d220c310feccee7f2.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003ePathways to productivity: how functional ecosystem service flow determines crop yield in a prairie landscape\u003c/p\u003e","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eReconciling the competing pressures of agriculture with the need to conserve biodiversity and ecosystem services is a central challenge, especially in intensive farming systems (Tamburini et al. \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). The concept of sustainable multifunctional landscapes has emerged as a key solution to addressing global challenges in food security while also safeguarding biodiversity and ecosystem services (Richter et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Maintaining farmland biodiversity and the flow of ecosystem services in farmland is essential to the long-term sustainability of agroecosystems and food production. For instance, studies show that practices supporting on farm biodiversity such as reduced tillage, crop rotation, and organic farming, can maintain or even improve crop yield over the long term (Cozim-Melges et al. \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). However, the homogenization of landscape structure within modern production systems has caused significant biodiversity loss, undermining the stability and resilience of these systems (Galpern et al. \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe link between landscape structure and ecological benefits has been well-established in previous studies. A recent global meta-analysis found that increasing landscape complexity, particularly through changes in composition, configuration, or heterogeneity, has significant positive effects on biodiversity and the associated ecosystem services, thereby enhancing the potential benefits to agricultural production (Estrada-Carmona et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Empirical evidence indicates that various aspects of landscape configuration can influence yield outcomes. For instance, greater landscape diversity has been linked to improved natural pest control (Madin and Nelson 2023), while higher connectivity can either facilitate beneficial species movement (Wood et al. \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) or, conversely, spread pests and diseases (Margosian et al. \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). Additionally, factors such as natural habitat patch size and distance to edges have been associated with crop productivity within agricultural landscapes (Nguyen, Robinson, and Galpern 2022). Beyond yield benefits, spatial arrangement also acts as a mediator and facilitator of ecosystem service interactions, shaping how services co-occur, reinforce, or compete across space (Pashanejad et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Rieb and Bennett 2020). As such, landscape structure is central to promoting ecosystem multifunctionality and resilience in agricultural systems.\u003c/p\u003e\u003cp\u003eDespite growing recognition of the role of ES in sustaining resilient agricultural landscapes, our understanding remains limited regarding how spatial and ecological patterns influence the optimal delivery of these services\u0026mdash;particularly in relation to crop productivity and long-term sustainability outcomes (Rehman et al. \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Recent advances in landscape ecology have emphasized a \u0026ldquo;pattern\u0026ndash;process\u0026ndash;service\u0026ndash;sustainability\u0026rdquo; paradigm (Fu et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), which calls for deeper integration of spatial configuration and ecological processes to better understand how landscape structure supports service flows and functional outcomes. Yet, a critical and underexplored frontier lies in understanding how internal functional connectivity\u0026mdash;that is, the spatial transfer of ecosystem service benefits within ecologically meaningful units\u0026mdash;mediates agricultural performance. While previous research has shown that landscape composition and structure influence service supply and movement, few empirical studies have examined how these internal flows accumulate and interact to affect integrated outcomes such as crop yield (Fu et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Qiu and Mitchell 2024). This is particularly relevant in large-scale, intensively managed systems like the Canadian prairies, where service delivery and agricultural performance are shaped by multiple, interacting, and often nonlinear spatial dynamics. In this study, we address this gap by using the Soil Landscapes of Canada (SLCs), an ecologically meaningful spatial delineation based on natural features, to investigate how internal ES flows within these units\u0026mdash;mediated by landscape structure and composition\u0026mdash;influence yield as a cumulative, emergent property of multifunctional landscapes.\u003c/p\u003e\u003cp\u003eWe apply generalized additive models (GAMs) to examine how key ecosystem service flows and structural landscape attributes\u0026mdash;such as habitat connectivity and crop diversity\u0026mdash;jointly influence agricultural productivity. To this, we integrate administrative crop yield data with modeled ES estimates for the year 2020, analyzing their relationship within SLC units. Our main objective is to disentangle whether spatial and functional dimensions of service flows within multifunctional landscapes contribute to crop yield as an integrated performance metric. In so doing, this research aims to identify spatial trade-offs and synergies that emerge across agricultural mosaics. Ultimately, our findings are intended to inform data-driven, landscape-level strategies for enhancing ecosystem multifunctionality and supporting sustainable agricultural management in intensively farmed regions.\u003c/p\u003e"},{"header":"2. Method","content":"\u003cp\u003eThe methodology of this study was structured in three main steps. First, we collated both ecosystem service and crop yield data for each SLC in the study area. To do this, we mapped and quantified key ecosystem services including pollination, habitat quality, carbon storage, nutrient retention, and soil erosion control across the Canadian prairies using the InVEST modeling tool. A complete and detailed description of each ecosystem service model, its data requirements, and specific parameterization is provided in the Supplementary Information (S1). This methodological description for ES mapping is provided in SI file in detail. Following the ES mapping, we aggregated pixel-level ES values to the SLC aligning with the scale at which crop yield data were available. We then developed a composite index of crop productivity as our primary response variable by standardizing and integrating yield data for four major crops: canola, barley, wheat (both spring and winter wheat). To do this we first standardized the yield values (kg/acre) for the four aforementioned major crops in the study area by converting them to z-score. This process rescales each crop's yield distribution to have a mean of zero and a standard deviation of one, ensuring that each crop contributes equally to the index regardless of differences in absolute yield. The final composite crop yield index for each SLC unit was then calculated as the mean of these four standardized yield values. In the second step, we quantified landscape structure for each SLC unit using a set of composition and configuration metrics. These landscape metrics were used as proxies for ecosystem service flows, conceptualized as the internal spatial transfer of ecological benefits within SLC units \u0026mdash;from service-providing areas (e.g., natural habitats) to service-demanding areas (e.g., agricultural lands). Specifically, we linked individual metrics to distinct ecological processes: landscape connectivity, for example, was used to represent the potential for pollinator movement from natural habitats to croplands, while edge density proxied the spillover of beneficial insects like pest predators from field margins. The SLC unit was chosen as the appropriate scale to capture these flows for two key reasons. First, with an average size of approximately 46,500 hectares, SLCs are large enough to encompass the heterogenous mosaic of service-providing and service-demanding areas. This spatial extent is ecologically relevant as it aligns with the typical dispersal and foraging distances of mobile service providers like pollinators. Second, SLCs are delineated by stable biophysical features like soil and topography, creating coherent landscape units where the cumulative effects of these internal ecological flows on agricultural productivity can be meaningfully assessed. This approach helped us to examine how internal functional connectivity, local spatial heterogeneity, and spatial spillover effects within landscape units may influence agricultural productivity. In the final step, we employed generalized additive models (GAMs) to investigate the relationship between ES provisioning, ES flows, and crop productivity. To disentangle the influence of biophysical functions from that of spatial structure, we developed two distinct analytical models. The first, our Ecosystem Service (ES) Model, examined the direct predictive power of key regulating services such as pollination and erosion control on crop yield. The second, our Landscape Structure (LSM) Model, focused instead on how landscape composition and configuration metrics explain yield variation. Both models used the composite crop index as the dependent variable and included a consistent control variables to account for management, climate, and topography (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The following subsections describe each step in greater detail, including the study area and data sources (Sections 2.1 and 2.2), and the methodological rationale for the selection and application of landscape structure metrics and GAMs (Sections 2.3 and 2.4).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eVariables included in the Generalized Additive Models (GAMs).\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cem\u003eRole in Model\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u003cem\u003eCategory\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eVariable\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cem\u003eData source\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eResponse\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eProductivity\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eComposite Crop Yield\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eEstimated based on Agriculture and Agri-Food Canada (AAFC, 2024) crop yield data\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e\u003cp\u003e\u003cb\u003ePredictor\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"3\" rowspan=\"4\"\u003e\u003cp\u003eEcosystem Service\u003c/p\u003e\u003cp\u003e(ES Model)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePollination\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"3\" rowspan=\"4\"\u003e\u003cp\u003eInVEST model output\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eHabitat Quality\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eSoil Erosion Control\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eCarbon Storage\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"5\" rowspan=\"6\"\u003e\u003cp\u003e\u003cb\u003ePredictor\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"5\" rowspan=\"6\"\u003e\u003cp\u003eLandscape Structure\u003c/p\u003e\u003cp\u003e(LSM Model)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e% Natural Habitat\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"5\" rowspan=\"6\"\u003e\u003cp\u003eCalculated based on AAFC land cover and land use (annual crop inventory for the year 2020)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eHabitat Diversity (Shannon Diversity)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eCrop Diversity\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eConnectivity (patch cohesion)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eEdge Density\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAggregation (Clumpiness)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"5\" rowspan=\"6\"\u003e\u003cp\u003e\u003cb\u003eControl\u003c/b\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003e(Both Models)\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"5\" rowspan=\"6\"\u003e\u003cp\u003eEnvironment \u0026amp;\u003c/p\u003e\u003cp\u003eManagement\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eSoil Organic Matter\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eISRIC\u0026ndash;World Soil Information\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eElevation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eDigital Elevation Model FABDEM V1-2\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eTemperature\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eClimate NA platform\u003c/p\u003e\u003cp\u003e(Wang et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2016\u003c/span\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePrecipitation\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eTillage\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eCensus of Agriculture: Agri-Environmental Spatial Data (AESD)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eSpatial Term\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eLongitude and latitude centroid of each SLC polygon\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1. Study area\u003c/h2\u003e\u003cp\u003eThe Canadian prairies, a vast agricultural region spanning the provinces of Alberta, Saskatchewan, and Manitoba in central Canada, were selected as the study region to investigate whether internal functional connectivity of ecosystem services within landscapes influences crop productivity. Specifically, we test the hypothesis that landscapes with greater habitat protection, higher diversity, and increased connectivity can enhance agricultural outcomes by facilitating the flow of ecosystem services. The prairies are a highly significant region for multiple ecosystem functions and agricultural production, with over 80% of Canada\u0026rsquo;s farmland located in this area. This landscape supports several strategic crops, including oilseeds such as canola, cereals like barley and wheat (both spring and winter), oats, and many others that are essential to the resilience and sustainability of Canada\u0026rsquo;s agri-food economy. However, the prairie landscape is increasingly shaped by pressures such as land-use intensification, climate variability, and habitat fragmentation. These changes are largely driven by the expansion of cultivated land and the simplification of landscape structure. While such trends have boosted production in the short term, they have also contributed to biodiversity loss and a decline in ecosystem service capacity. Despite its productivity, the region has experienced significant ecological degradation. Intensification practices including the removal of natural buffers, expansion of field sizes, wetland drainage, and conversion of marginal lands have disrupted hydrological functions, lowered water quality, and undermined biodiversity (Baulch et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Nevertheless, the prairies hold considerable potential for climate adaptation and ecological restoration. With intentional land management, natural and semi-natural areas can serve as green infrastructure that supports carbon sequestration, species movement, and the sustained delivery of ecosystem services crucial to resilient agriculture (Pashanejad \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2. Data\u003c/h2\u003e\u003cdiv id=\"Sec5\" class=\"Section3\"\u003e\u003ch2\u003e2.2.1. Crop yield data\u003c/h2\u003e\u003cp\u003eWe obtained crop yield data from Agriculture and Agri-Food Canada (AAFC, 2024) at the SLC unit scale for the major crops cultivated across the prairie region. While the dataset provides a temporal record of annual crop yields, we focused on the year 2020 to align with the ecosystem service (ES) mapping outputs. Our analysis concentrated on four strategically important crops in the region: canola, barley, and wheat (including both spring and winter varieties). The spatial distribution of these crops across SLC units is illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. Based on these data, we developed a single composite crop yield index to serve as the response variable in both GAM models (ES and LSM). To create this, we first standardized the raw yield values (kg/acre) of the four major crops by converting them z-scores. This index reflects the productivity per acre of cultivated land within an SLC, not the total production of the entire area. This distinction allows our analysis to focus on the factors driving on-farm productivity, independent of how much land within the SLC unit is allocated to agriculture versus natural habitat.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eTo calculate landscape structure metrics, we used the AAFC Annual Crop Inventory for the year 2020. This land cover/land use dataset was also used as a primary input in the ES mapping process, ensuring consistency across analyses. The crop inventory provides detailed information on land use and crop type, enabling the quantification of both compositional and configurational aspects of landscape heterogeneity.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section3\"\u003e\u003ch2\u003e2.2.2. ES provisioning data\u003c/h2\u003e\u003cp\u003eThe ecosystem service variables used as predictors in our GAMs were derived from a suite of biophysical models. These models were initially run at a 300 m resolution, a scale chosen to balance the need for fine-grained landscape detail with computational feasibility across the large prairies. We then aggregated pixel-level outputs to the SLC unit level for analysis. Full methodological details of the ES mapping are provided in SI file but we summarize the key outputs here. Pollination service was estimated using the InVEST Crop Pollination model, which generates an index of pollinator abundance based on nesting sites and floral resources. Habitat Quality was also modeled with InVEST, providing an index of the landscape's capacity to support terrestrial biodiversity based on habitat suitability and threats. Soil Erosion Control was quantified as the amount of avoided erosion (t/ha/yr) using the InVEST Sediment Delivery Ratio (SDR) model. Finally, Carbon Storage represents the total carbon (tonnes/ha) stored in aboveground biomass, belowground biomass, dead organic matter, and soil, as estimated by the InVEST Carbon model. The spatial distribution of these key ecosystem services across the study area is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e2.3. Landscape structure metrics\u003c/h2\u003e\u003cp\u003eTo assess the spatial structure of landscapes and their influence on ES flow and crop productivity, we calculated a series of landscape metrics at the SLC unit level. Metrics were derived from the 2020 Annual Crop Inventory (AAFC 2020), resampled to 300 m resolution to match the scale of analysis and for computational efficiency at the large area of the Canadian prairies. Calculations were performed using the \u003cem\u003elandscapemetrics package in R\u003c/em\u003e (Hesselbarth et al. \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). We organized the metrics into two main categories: composition and configuration, representing the quantity and spatial arrangement of land cover types, respectively.\u003c/p\u003e\u003cdiv id=\"Sec8\" class=\"Section3\"\u003e\u003ch2\u003e2.3.1. Composition Metrics\u003c/h2\u003e\u003cp\u003eA composition metric reflects the abundance and diversity of natural habitats within each SLC unit. To assess the influence of landscape composition on crop productivity, we calculated several key composition-based indicators as proxies based on the evidence that the amount of habitat is a critical determinant of biodiversity and functional support for ES provisioning (Fahrig \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Tscharntke et al. \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). First, we used the percentage of natural habitat, which quantifies the proportion of land within each SLC unit classified as natural or semi-natural (e.g., wetlands, grasslands, forests, shrublands). This metric provides an estimate of habitat availability for essential regulating services such as pollination, natural pest control, and water regulation (Chaplin-Kramer et al. 2011; Tscharntke et al. \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). To capture overall landscape heterogeneity, we calculated the Shannon diversity index (McGarigal and Marks \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e1995\u003c/span\u003e) for habitat. This index incorporates both the richness and evenness of all natural land cover types within each SLC unit, serving as an indicator of spatial multifunctionality. More heterogeneous landscapes are generally associated with greater biodiversity and more stable ecological processes. In addition, we calculated the Shannon diversity index for croplands to evaluate diversity specifically within agricultural land use classes. This metric reflects the variety and distribution of crop types across each unit. Higher crop diversity has been linked to enhanced ecological resilience, improved pest suppression, and more diverse ecosystem service provisioning (Kremen and Miles 2012).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec9\" class=\"Section3\"\u003e\u003ch2\u003e2.3.2. Configuration Metrics\u003c/h2\u003e\u003cp\u003eA configuration metric evaluates the spatial arrangement and structural connectivity of landscapes. We selected configuration metrics that capture the fragmentation, continuity, and potential spillover dynamics of land cover patches within each SLC unit. These metrics help account for the internal flow of ecosystem services by assessing how natural and managed elements are organized in space. The \u0026ldquo;clumpiness\u0026rdquo; index was used to measure the degree of spatial aggregation of natural habitat patches (He, DeZonia, and Mladenoff \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2001\u003c/span\u003e). Higher clumpiness values indicate more contiguous blocks of habitat, which can enhance within-field ecosystem service flows, particularly those related to pollination and natural pest regulation. This metric quantifies the compactness of habitat areas and their capacity to support consistent service provision. We also included the patch cohesion index, which assesses the physical connectedness of land cover patches belonging to the same class. Greater cohesion suggests stronger ecological connectivity across the landscape, supporting species movement, seed dispersal, and the delivery of services (Mitchell, Bennett, and Gonzalez 2013). It serves as an indicator of functional habitat networks, which are essential for maintaining ecological resilience in agricultural systems. Finally, edge density was calculated to quantify the length of habitat edges per unit area (McGarigal and Marks \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e1995\u003c/span\u003e). This metric reflects the extent of interface between natural and agricultural areas, where ecosystem service spillover is likely to occur. High edge density can signal increased interaction zones between habitats and crop fields, potentially influencing service delivery dynamics such as pollination, pest suppression, and nutrient flow (Blitzer et al. \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Ries et al. \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2004\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\u003ch2\u003e2.4. Generalized Additive Model\u003c/h2\u003e\u003cp\u003eGeneralized Additive Models (GAMs) are flexible, nonparametric extensions of traditional Generalized Linear Models (GLMs) designed to model complex, nonlinear relationships between predictor variables and a response variable (Hastie \u0026amp; Tibshirani, 1990). Unlike GLMs, which assume linearity between covariates and the response, GAMs employ smooth functions (e.g., splines) to model the effects of each predictor, making them particularly suitable for ecological data where relationships are often nonlinear, context-dependent, and shaped by interacting factors.\u003c/p\u003e\u003cp\u003eMathematically, a GAM can be written as:\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:{y}_{i}={\\beta\\:}_{0}+{f}_{1}{(x}_{1i})+{f}_{2}{(x}_{2i})+\\dots\\:+{f}_{k}{(x}_{ki})+\\:{\\epsilon\\:}_{i}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003ewhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{y}_{i}\\)\u003c/span\u003e\u003c/span\u003e is the response variable in this case crop productivity index, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{f}_{1}(.)\\)\u003c/span\u003e\u003c/span\u003e represent smooth functions for the predictor \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{(x}_{i}\\)\u003c/span\u003e\u003c/span\u003e), and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\epsilon\\:}_{i}\\:\\sim\\:N(0,\\:{\\sigma\\:}^{2})\\)\u003c/span\u003e\u003c/span\u003e is the normally distributed error term.\u003c/p\u003e\u003cp\u003eWe fitted two GAMs to explain spatial variation in crop productivity across Soil Landscapes of Canada (SLC) units in the Prairie region. The Ecosystem Services (ES) model examined the influence of pollination, habitat quality, erosion control, and carbon storage. The Landscape Structure (LSM) model included structural configuration variables such as natural habitat cover, crop diversity, connectivity, and patch aggregation. Both models included a two-dimensional smooth function of geographic coordinates (longitude and latitude) to account for residual spatial autocorrelation not explained by the predictors. This approach allows GAMs to flexibly capture spatial structure without relying on the user pre-define a spatial weight matrix (Wood, 2017). The model's error distribution was specified based on the nature of the response variable. Since the crop productivity index is a continuous measure, we used the Gaussian family with an identity link. This structure, which assumes normally distributed residuals, was confirmed to be appropriate through a thorough examination of residual diagnostics. It models the mean of the response as a linear (or smoothed) function of predictors and is appropriate when residuals are approximately symmetrically distributed.\u003c/p\u003e\u003cp\u003eTo isolate the contribution of ecosystem service and landscape structure variables, we incorporated a consistent set of control covariates into both models. We specifically considered soil organic matter accounting for soil fertility and microbial activity, elevation capturing topographic and microclimate variation, mean temperature and precipitation representing climate suitability and conventional tillage as representative of management practice influencing soil health and crop performance. We conducted an analysis of Variance Inflation Factors (VIF) to explore the multicollinearity among predictors before fitting the final model (See Supplementary Information file). Nutrient retention for nitrogen and phosphorus and conservation tillage has shown a high multicollinearity, therefore we retained only conventional tillage, which showed a moderate VIF value and consistent significant importance (confidence intervals for point estimates did not overlap zero) across models. Our primary analysis focused on the composite crop productivity index to identify overarching patterns across the prairie system. However, to investigate potential crop-specific responses and ensure the robustness of our findings, we also conducted a secondary analysis. This involved fitting the ES model and the LSM model separately for each of the major individual crops: canola, barley, and wheat (both spring and winter). The predictor variables and model specifications remained identical to those described below, with only the response variable changing for each run.\u003c/p\u003e\u003cp\u003eEcosystem Service model:\u003cdiv id=\"Equb\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e\n$$\\:{Y}_{i}^{ES}={\\beta\\:}_{0}+{\\beta\\:}_{1}.{Carbon}_{i}+{\\beta\\:}_{2}.SoilOM+{\\beta\\:}_{3}.{Elevation}_{i}+{\\beta\\:}_{4}.{Temp}_{i}+{\\beta\\:}_{5}.{Precp}_{i}+\\:{\\beta\\:}_{6}.{Tillage}_{i}+{f}_{1}{(Pollination}_{i})+{f}_{2}{(Habitat}_{i})+{f}_{3}{(Erosion}_{i})+{f}_{4}{(lon}_{i},\\:{lat}_{i})+\\:{\\epsilon\\:}_{i}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eLandscape Structure Model:\u003cdiv id=\"Equc\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equc\" name=\"EquationSource\"\u003e\n$$\\:{Y}_{i}^{LSM}={\\beta\\:}_{0}+{\\beta\\:}_{1}.SoilOM+{\\beta\\:}_{2}.{Elevaion}_{i}+{\\beta\\:}_{3}.{Temp}_{i}+{\\beta\\:}_{4}.{Precp}_{i}+\\:{\\beta\\:}_{5}.{Tillage}_{i}+{\\beta\\:}_{6}.{EdgeDensity}_{i}+{\\beta\\:}_{6}.{PatchAggregation}_{i}+{f}_{1}{(\\%NaturalHabitat}_{i})+{f}_{2}{(CropDiversity}_{i})+{f}_{3}{(Connectivity}_{i})+{f}_{4}{(lon}_{i},\\:{lat}_{i})+\\:{\\epsilon\\:}_{i}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003ewhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:Y\\)\u003c/span\u003e\u003c/span\u003e is the response variable, representing the composite crop productivity index and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\beta\\:}_{0}\\)\u003c/span\u003e\u003c/span\u003e is the model intercept, representing the baseline crop productivity. To indicate whether a variable has a linear and non-linear response in the crop productivity, we followed a model-building approach based on preliminary diagnostics and ecological theory. Variables for which we hypothesized a complex, non-linear relationship with crop yield (e.g., Pollination, % Natural Habitat) were modeled using smooth functions (f). Conversely, for predictors where initial diagnostic plots suggested a primarily linear response, or where there was no strong a priori theoretical reason to expect non-linear, we modeled them as linear terms (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\beta\\:\\)\u003c/span\u003e\u003c/span\u003e). These include the control variables for soil organic matter (SoilOM), elevation (Elevation), temperature (Temp), precipitation (Precp), and conventional tillage (Tillage). The model also estimates linear effects for specific ES or LSM predictors like carbon, edge density, and patch aggregation as initial diagnostic plots for these variables suggested a primarily linear response. The \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\beta\\:\\)\u003c/span\u003e\u003c/span\u003e coefficients represent the estimated slope for each of these variables. The variables expected to have a complex, non-linear relationship with crop productivity are modeled using smooth functions (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:f\\)\u003c/span\u003e\u003c/span\u003e). These smoothers (by default thin plate splines), allow the model to flexibly capture the shape of the relationship from the data itself. To account for the spatial autocorrelation, a two-dimensional smooth term over geographic coordinates (longitude and latitude) is included in both models. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\epsilon\\:}_{i}\\)\u003c/span\u003e\u003c/span\u003e is the residual error term, assumed to be normally distributed.\u003c/p\u003e\u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\n\u003ch2\u003e3.1. Landscape structure analysis\u003c/h2\u003e\n\u003cp\u003eWe observed clear spatial gradients and strong correlations among landscape structure metrics across the study area. Areas with a high percentage of natural habitat were more spatially cohesive (r\u0026thinsp;=\u0026thinsp;0.76) and exhibited higher edge density (r\u0026thinsp;=\u0026thinsp;0.67), but these landscapes often contained fewer habitat types. Notably, habitat cohesion was also moderately associated with edge density (r\u0026thinsp;=\u0026thinsp;0.34), suggesting some alignment between connectedness and edge exposure. Interestingly, landscapes with greater habitat diversity tended to support more diverse cropping systems (r\u0026thinsp;=\u0026thinsp;0.60), highlighting potential synergies between natural and managed diversity. However, our findings reveal a strong negative correlation between natural habitat extent and habitat diversity (r = \u0026minus;\u0026thinsp;0.76), indicating that larger areas of contiguous habitat often result in reduced habitat heterogeneity likely due to the dominance of single habitat types, such as continuous native grasslands or forest patches. Similarly, higher cohesion was strongly associated with lower habitat diversity (r = \u0026minus;\u0026thinsp;0.90), reinforcing the idea that well-connected landscapes can be structurally simplified, which may constrain ecosystem service multifunctionality. The spatial distribution of each landscape metric across the prairie region is presented in Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e, illustrating clear regional contrasts in natural cover, habitat and crop diversity, clumpiness, cohesion, and edge density.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\n\u003ch2\u003e3.2. GAM model output: ES and landscape structure effects on crop yield\u003c/h2\u003e\n\u003cp\u003eBoth the Ecosystem Services (ES) and Landscape Structure (LSM) models explained a comparable proportion of variance in crop productivity (Adjusted R\u0026sup2; = 0.667 and 0.662, respectively; see Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e) underscoring that both biophysical functions and spatial landscape attributes are important in shaping agricultural outcomes. In the ES model, key regulating services such as pollination, habitat quality, and soil erosion control showed strong, significant non-linear effects (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), while tillage emerged as the most influential linear management variables. Carbon storage, on the other hand, showed a marginally negative association (p\u0026thinsp;=\u0026thinsp;0.07). Most of the control variables, including soil organic matter, elevation, temperature, and precipitation, were not statistically significant (p\u0026thinsp;\u0026gt;\u0026thinsp;0.1) in this model (see Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e\n\u003cp\u003eIn contrast, the LSM model, designed to capture how landscape configuration and composition potentially mediate ecosystem service flows to agricultural areas, highlighted the importance of structural attributes. Specifically, crop diversity and landscape connectivity were both significant predictors of crop yields. However, while these structural configuration metrics demonstrated significance, the overall proportion of natural habitat (%) was non-significant when considered alongside configuration metrics. Among the linear terms, habitat aggregation and tillage were positively associated with yield, whereas soil organic matter had a significant negative effect and precipitation was marginally negative. Habitat diversity and edge density\u0026mdash;two landscape structure metrics\u0026mdash;were not significant in this model. These two models reveal distinct but complementary mechanisms: the ES model reflects the biophysical functioning of the landscape, while the LSM model highlights how landscape structure and spatial organization condition the flow and impact of those services (Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e\n\u003cp\u003eAs a robustness check we assessed whether these patterns held true for individual crops separately. These crop-specific models, detailed in Supplementary Information, largely reinforce our main findings. Across all crops, both the ES and LSM models explained a substantial portion of yield variance, and the spatial term remained a dominant, highly significant factor. However, this secondary analysis also revealed important nuances. For instance, while the amount of natural habitat was not significant in the composite models, canola yield showed a unique and significant positive response to the percentage of natural habitat, likely reflecting its high dependence on pollinators whose population and foraging activities are supported by those areas. This highlights that while general principles apply, specific crop needs can alter the relative importance of certain landscape attributes in facilitating beneficial ES flows.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\n\u003ctable id=\"Tab2\" border=\"1\"\u003e\u003ccaption\u003e\n\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n\u003cdiv class=\"CaptionContent\"\u003e\n\u003cp\u003eSummary of generalized additive models (GAMs) explaining variation in crop productivity across the Canadian Prairies. The table compares the performance of two models: one based on ecosystem services (ES) and another on landscape structure metrics (LSM). Both models include spatial smooth terms and control for climatic, topographic, and management-related variables.\u003c/p\u003e\n\u003c/div\u003e\n\u003c/caption\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003e\u003cem\u003eMetric\u003c/em\u003e\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003e\u003cem\u003eEcosystem Services (ES) Model\u003c/em\u003e\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003e\u003cem\u003eLandscape Structure (LSM) Model\u003c/em\u003e\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eAdjusted R\u0026sup2;\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0.667\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0.662\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eDeviance Explained (%)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e68.0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e67.5\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eSignificant Smooth Terms\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ePollination (+, ***) Habitat Quality (+, ***) Soil Erosion Control (+, ***) Spatial Coordinates (Smooth, ***)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eCrop Diversity (+, *) Landscape Connectivity (+, *) Spatial Coordinates (Smooth, ***)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eNon-Significant Smooth terms\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u0026mdash;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ePercentage of Natural Habitat (NS)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eSignificant Linear Terms\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eTillage (+, ***) Carbon Storage (\u0026minus;, .)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eAggregation of Natural Habitat Patches (+, *) Soil Organic Matter Content (\u0026minus;, *) Tillage (+, ***) Mean Annual Precipitation (\u0026minus;, .)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eNon-Significant Linear terms\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eSoil Organic Matter Content (NS)Elevation (NS)Mean Annual Temperature (NS)Mean Annual Precipitation (NS)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eHabitat Diversity (NS)Edge Density (NS)Elevation (NS)Mean Annual Temperature (NS)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003ctfoot\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"3\"\u003e\u003cem\u003eNote: Significance of non-linear effects (smooth terms) and linear terms is indicated as follows: (***) for p\u0026thinsp;\u0026lt;\u0026thinsp;0.001, (**) for p\u0026thinsp;\u0026lt;\u0026thinsp;0.01, (*) for p\u0026thinsp;\u0026lt;\u0026thinsp;0.05, and (.) for p\u0026thinsp;\u0026lt;\u0026thinsp;0.1 (marginally significant). Terms with p\u0026thinsp;\u0026ge;\u0026thinsp;0.1 are considered not significant (NS). Linear terms include a plus (+) or minus (\u0026minus;) symbol to denote the direction of the relationship (positive or negative). \u0026ldquo;Spatial (Smooth)\u0026rdquo; refers to the two-dimensional smooth over spatial coordinates s(X, Y), with its significance also indicated using the aforementioned symbols.\u003c/em\u003e\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tfoot\u003e\n\u003c/table\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\n\u003ch2\u003e3.3. Partial effect plots for the Ecosystem Services (ES) and Landscape Structure model\u003c/h2\u003e\n\u003cp\u003eFigure 4, (panel A) shows that pollination had a non-linear relationship with crop yield, characterized by an initial decline followed by a steady increase in partial effect values beyond moderate pollination levels. Habitat quality exhibited a consistently negative effect across its range, while erosion control showed a strongly non-linear association, with partial effects decreasing at low to moderate levels and increasing substantially at higher values. The significance of the spatial smooth term (X, Y coordinates) reveals strong spatial autocorrelation in crop yield unexplained by the other model covariates. The model identified distinct geographic clusters of positive and negative partial effects, corresponding to regions where underlying spatial processes either enhance or reduce crop productivity. The Landscape Structure model similarly revealed complex predictor effects. For instance, crop diversity showed a non-linear effect, with an initial decrease in its partial effect on yield as diversity increased to moderate levels (approx. SHDI 1.5-2.0), followed by a slight positive trend at higher diversity values. Landscape connectivity, a key facilitator of potential ES flows like pollinator movement, showed a near-linear positive relationship with crop productivity. In contrast, percentage of natural habitat had a weak and slightly negative effect that remained largely flat across the range. The spatial smoother captured spatial heterogeneity not explained by the included variables, indicating consistent spatial structure in the residuals across the study region.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003c/div\u003e\n\u003cp\u003eThe model diagnostics plots for both models are shown in Fig.\u0026nbsp;5. These plots indicated that the assumptions of our GAM models were fitted at an acceptable level. The Q-Q plots showed an approximately normal distribution of residuals with slight deviations at the tails (Fig.\u0026nbsp;5, panel A, top-left), and residual histograms were unimodal and centered around zero (Fig.\u0026nbsp;5, panel A, top-right), indicating minimal skewness. The residuals vs. fitted plots revealed no pronounced heteroscedasticity or structure, suggesting a good fit across the range of predicted values (Fig.\u0026nbsp;5, bottom-left). The response vs. linear predictor plots further confirmed the non-linear relationship captured by the model (Fig.\u0026nbsp;5, bottom-right). In addition, spatial diagnostic maps of residuals (Fig.\u0026nbsp;5, panels C and D) revealed geographically patterned residuals in some parts of the study area, particularly in underpredicted and overpredicted areas. This highlights localized model uncertainty and suggests the presence of spatial heterogeneity not fully captured by the selected predictors.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\n\u003cdiv class=\"colspec\" align=\"left\"\u003e\u0026nbsp;\u003c/div\u003e\n\u003c/div\u003e\n\u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThe complex relationship between spatial ecological patterns, the ecosystem services they generate, and agricultural productivity is central to designing sustainable and resilient agricultural landscapes. In this research, we examined how landscape patterns and the mechanism they support including both ecological functions and land management practices interact to influence crop yields across large, intensively managed systems like the Canadian prairies. While previous research has increasingly recognized that landscape structure underpins service flows and functional outcomes, the specific mechanism by which internal functional connectivity\u0026mdash;the spatial transfer of ecosystem service benefits within defined ecological units\u0026mdash;mediates agricultural performance remains underexplored. We addressed this critical gap by investigating how internal ES flows, as mediated by landscape composition and configuration within Soil Landscape of Canada units, collectively shape crop productivity as an emergent property of these multifunctional units.\u003c/p\u003e\u003cp\u003eOur findings demonstrate that both key regulating ecosystem services such as pollination and soil erosion control and the structural attributes of the landscape play complementary and significant roles in determining crop yield performance. This dual influence highlights a fundamental principle: agricultural output is not merely a function of inputs and field-level conditions but is deeply embedded within, and responsive to, the broader landscape matrix and the ecological processes it supports (Kremen and Miles 2012). Notably, our results revealed that landscape configuration metrics, particularly connectivity and crop diversity were stronger predictors of crop yield than the simple composition measures such as the amount of natural habitat. However, crop-specific analysis reveals a critical exception to this pattern. For canola, which is a high insect pollination-dependent crop, the amount of natural habitat was a significant positive predictor (See SI file). This suggests that while functional connectivity may drive the overall system, for specific services tightly linked to habitat-dependent organisms, the total habitat area is foundational to support species richness and persistence (Fahrig \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Overall, our results suggest that at the broad SLC scale, spatial arrangement and functional connectivity to croplands are more influential in shaping aggregated agricultural performance. However, this does not diminish the importance of the habitat amount; rather, it refines its role, indicating that its effect can be enhanced or mediated by landscape configuration (Tscharntke et al. \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). Previous studies (Blitzer et al. \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Kremen, Williams, and Thorp \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2002\u003c/span\u003e; Mitchell et al. \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2013\u003c/span\u003e) have shown that metrics like landscape connectivity serve as more effective proxies for the potential ES flows \u0026mdash; such as pollinator spillover or pest predator movement \u0026mdash; between natural areas and farmlands. Our modeling reinforces the perspective that once ES flow-mediating attributes are incorporated, the simple percentage of habitat may provide limited additional explanatory power. This aligns with previous findings that emphasize habitat interfaces as critical interaction zones for service delivery (Blitzer et al. \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2012\u003c/span\u003e)\u003c/p\u003e\u003cp\u003eHeterogeneous landscapes tend to support greater habitat availability and higher species diversity, which can enhance ecosystem resilience, stability, and recovery following disturbances (Feit et al. \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Tscharntke et al. 2012). In such complex agricultural systems, increased availability of habitats for pollinators can improve ecosystem functioning and service delivery, ultimately leading to higher crop yields and quality through enhanced pollination and natural pest regulation (Estrada-Carmona et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). However, our findings reveal a critical distinction between local service benefits and landscape-scale trade-offs. The observed negative relationship between aggregated habitat quality and crop yield at the SLC scale likely reflects a spatial trade-off rather than a direct detrimental effect of habitat on adjacent crop performance. SLC units with extensive, high-quality natural remnants, such as native grasslands, will inherently have a smaller proportion of their area dedicated to, or lower overall output from, intensive crop production when aggregated. The agricultural land that does exist within these high-conservation value landscapes may therefore be inherently less productive. Thus, a high average habitat quality score for an SLC unit can be spatially correlated with lower overall agricultural yield for that unit, reflecting broad patterns of land suitability and agricultural marginality across the prairies. This highlights that relationship between ES indicators and agricultural productivity at this scale are heavily influenced by underlaying environmental gradients.\u003c/p\u003e\u003cp\u003eWhile this study provides practical insights across a globally significant agricultural region, we acknowledge its limitations. Specifically, the use of landscape configuration to infer ecosystem service interactions implies relationship that are correlational, not necessarily causal. Further research is needed to disentangle the underlying mechanisms driving these observed associations. The primary objective of this study was to explore the spatial dynamics of landscape structure in a heavily intensified agroecosystem, such as the Canadian prairies. Although we included key controlling factors such as land management, climate, and topography, we did not fully account for the complex, multi-layered nature of the agri-food system. Our analysis provides a robust snapshot using 2020 data. While this cross-sectional approach reveals critical spatial relationships, future work using a multi-year panel data analysis could further strengthen these findings by controlling for time-invariant factors. Such extensions could significantly enhance the explanatory and predictive capacity of the current spatial modeling framework. However, while our study highlights the critical spatial relationship in a given year, future temporal analyses are essential for disentangling these complex drivers and confirming the causal pathways of landscape influence on agricultural outcome.\u003c/p\u003e\u003cp\u003eIt is crucial to recognize that the landscape patterns we measured are not solely the result of ecological forces, but are profoundly shaped by social processes (Berkes, Colding, and Folke 2008) given high intensity of agriculture in the study area. The strong predictive power of conventional tillage, for instance, directly reflects farmer management decisions, which are influenced by economic pressures, tradition, and policy incentives. Similarly, while our results show a clear benefit to crop diversity, the prevalence of monocultures across the prairies is driven by social and economic systems that favor specialization (Qualman \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Qualman et al. 2020). The very configuration of natural habitats\u0026mdash;which our study shows is critical for mediating ES flows\u0026mdash;is also constrained by human-defined property lines that dictate land management decisions at the parcel level. This delineation of ownership often leads to habitat fragmentation which influences the success and spatial extent of conservation programs. Therefore, understanding and improving these multifunctional landscapes requires an approach that extends beyond ecological knowledge to actively engage with the complex social and economic and political drivers that shapes their current state and future trajectories (Berkes et al. 2008; Ostrom \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2009\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eLandscape patterns, quantified through land cover composition and structure, influence underlying ecological processes such as pollination or soil stability. These, in turn, affect the provisioning of crop yield\u0026mdash;a key service in agroecosystems. Our modeling approach demonstrates how these interactions manifest across space and can inform decisions about where and how to intervene. For instance, enhancing landscape cohesion and crop diversity may offer greater returns than simply increasing the amount of semi-natural habitat. By operationalizing this pattern-process-service framework, our research supports more informed, spatially explicit planning for multifunctional landscapes and a paradigm shift. Through this research, we tried to shift our perspective from a static collection of patches to dynamic functional landscapes defined by ES flows (Fu et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). In addition, the use of generalized additive models allowed us to capture the non-linear relationship between key ecosystem services. Thus, reinforcing agricultural systems respond to ecological support in complex, non-linear ways, an essential insight for effective management.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eThe future of agricultural sustainability depends on maintaining landscape complexity rather than reducing it. Homogenized landscapes may maximize short-term yield but risk long-term ecological degradation and declining resilience. Our results show that multifunctionality can support both productivity and conservation. This calls for a paradigm shift in agricultural policy and planning: from maximizing yield per hectare to optimizing landscape configuration for long-term resilience and service provision. Moving forward, integrating detailed management data, socio-economic drivers, and farmer decision-making into spatial models will be crucial to developing actionable, context-sensitive strategies for sustainable agriculture. Our results show that connectivity and diversity\u0026mdash;two core principles of resilient and sustainable agricultural systems\u0026mdash;play a critical role in enabling ecosystem service flows and improving agricultural outcomes such as crop yield. The observed synergy between biodiversity, ecosystem service provisioning, and crop productivity, often strengthened by improved spatial configuration and landscape heterogeneity, underscores the potential for ecological intensification. By quantitatively linking landscape-scale patterns and internal service flows to crop productivity, our study advances this understanding and offers practical insights for sustainable landscape planning.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eThe authors have no relevant financial or non-financial interests to disclose.\u003c/p\u003e\n\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eAll authors contributed to the study conception and design. Material preparation, data collection and analysis and the discussion were performed by Ehsan Pashanejad. The first draft of the manuscript was written by Ehsan Pashanejad. Lael Parrott and Brian E. Robinson supervised the project and provided feedback and edits on previous versions of the manuscript. All authors read and approved the final manuscript.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eWe acknowledge the support of the Natural Sciences and Engineering Research Council of Canada (NSERC), [funding reference number NSERC NETGP 523374\u0026ndash;18]. Cette recherche a \u0026eacute;t\u0026eacute; financ\u0026eacute;e par le Conseil de recherches en sciences naturelles et en g\u0026eacute;nie du Canada (CRSNG), [num\u0026eacute;ro de r\u0026eacute;f\u0026eacute;rence NSERC NETGP 523374\u0026ndash;18]. We are grateful to the Agroclimate, Geomatics and Earth Observation Division of Agriculture and Agri-Food Canada for providing crop yield data, and we extend our thanks to Bahram Daneshfar for his generous support in supplying the major crop yield data at the Soil Landscape of Canada scale.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request. The primary crop yield data used in this study are available from Agriculture and Agri-Food Canada (AAFC) as cited in the manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAgriculture and Agri-Food Canada, 2024, \u0026ldquo;Interpolated Crop Yield by Soil Landscape of Canada (SLC), 2010\u0026ndash;2023\u0026rdquo;, Agroclimate, Geomatics and Earth Observation Division, Science and Technology\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBaulch, Helen, Colin Whitfield, Jared Wolfe, Nandita Basu, Angela Bedard-Haughn, Kenneth Belcher, Robert Clark, Grant Ferguson, Masaki Hayashi, Andrew Ireson, Patrick Lloyd-Smith, Phil Loring, John W. Pomeroy, Kevin Shook, and Christopher Spence. 2021. \u0026ldquo;Synthesis of Science: Findings on Canadian Prairie Wetland Drainage.\u0026rdquo; Canadian Water Resources Journal / Revue Canadienne Des Ressources Hydriques 46(4):229\u0026ndash;41. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1080/07011784.2021.1973911\u003c/span\u003e\u003cspan address=\"10.1080/07011784.2021.1973911\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBerkes, Fikret, Johan Colding, and Carl Folke. 2008. Navigating Social-Ecological Systems: Building Resilience for Complexity and Change. Cambridge university press.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBlitzer, Eleanor J., Carsten F. Dormann, Andrea Holzschuh, Alexandra-Maria Klein, Tatyana A. Rand, and Teja Tscharntke. 2012. \u0026ldquo;Spillover of Functionally Important Organisms between Managed and Natural Habitats.\u0026rdquo; Agriculture, Ecosystems \u0026amp; Environment 146(1):34\u0026ndash;43. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.agee.2011.09.005\u003c/span\u003e\u003cspan address=\"10.1016/j.agee.2011.09.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChaplin-Kramer, Rebecca, Megan E. O\u0026rsquo;Rourke, Eleanor J. Blitzer, and Claire Kremen. 2011. \u0026ldquo;A Meta-Analysis of Crop Pest and Natural Enemy Response to Landscape Complexity.\u0026rdquo; Ecology Letters 14(9):922\u0026ndash;32. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/j.1461-0248.2011.01642.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1461-0248.2011.01642.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCozim-Melges, Felipe, Raimon Ripoll-Bosch, G. F. Veen, Philipp Oggiano, Felix J. J. A. Bianchi, Wim H. van der Putten, and Hannah H. E. van Zanten. 2024. \u0026ldquo;Farming Practices to Enhance Biodiversity across Biomes: A Systematic Review.\u0026rdquo; Npj Biodiversity 3(1):1. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s44185-023-00034-2\u003c/span\u003e\u003cspan address=\"10.1038/s44185-023-00034-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEstrada-Carmona, Natalia, Andrea C. S\u0026aacute;nchez, Roseline Remans, and Sarah K. Jones. 2022. \u0026ldquo;Complex Agricultural Landscapes Host More Biodiversity than Simple Ones: A Global Meta-Analysis.\u0026rdquo; Proceedings of the National Academy of Sciences 119(38). doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1073/pnas.2203385119\u003c/span\u003e\u003cspan address=\"10.1073/pnas.2203385119\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFahrig, Lenore. 2013. \u0026ldquo;Rethinking Patch Size and Isolation Effects: The Habitat Amount Hypothesis\u0026rdquo; edited by K. Triantis. Journal of Biogeography 40(9):1649\u0026ndash;63. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/jbi.12130\u003c/span\u003e\u003cspan address=\"10.1111/jbi.12130\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFeit, Benjamin, Nico Bl\u0026uuml;thgen, Eirini Daouti, Cory Straub, Michael Traugott, and Mattias Jonsson. 2021. \u0026ldquo;Landscape Complexity Promotes Resilience of Biological Pest Control to Climate Change.\u0026rdquo; Proceedings of the Royal Society B: Biological Sciences 288(1951):20210547. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1098/rspb.2021.0547\u003c/span\u003e\u003cspan address=\"10.1098/rspb.2021.0547\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFu, Bojie, Yanxu Liu, Wenwu Zhao, and Jianguo Wu. 2025. \u0026ldquo;The Emerging \u0026lsquo;Pattern-Process-Service-Sustainability\u0026rsquo; Paradigm in Landscape Ecology.\u0026rdquo; Landscape Ecology 40(3):54. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10980-025-02063-7\u003c/span\u003e\u003cspan address=\"10.1007/s10980-025-02063-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGalpern, Paul, Jess Vickruck, James H. Devries, and Michael P. Gavin. 2020. \u0026ldquo;Landscape Complexity Is Associated with Crop Yields across a Large Temperate Grassland Region.\u0026rdquo; Agriculture, Ecosystems \u0026amp; Environment 290:106724. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.agee.2019.106724\u003c/span\u003e\u003cspan address=\"10.1016/j.agee.2019.106724\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHe, H. S., B. E. DeZonia, and D. J. Mladenoff. 2001. \u0026ldquo;Erratum: An Aggregation Index (AI) to Quantify Spatial Patterns of Landscapes (Landscape Ecology (200) 15 (591\u0026ndash;601)).\u0026rdquo; Landscape Ecology 16(1):87. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1023/A:1017308405507\u003c/span\u003e\u003cspan address=\"10.1023/A:1017308405507\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHesselbarth, Maximilian H. K., Marco Sciaini, Kimberly A. With, Kerstin Wiegand, and Jakub Nowosad. 2019. \u0026ldquo;Landscapemetrics: An Open-source R Tool to Calculate Landscape Metrics.\u0026rdquo; Ecography 42(10):1648\u0026ndash;57. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/ecog.04617\u003c/span\u003e\u003cspan address=\"10.1111/ecog.04617\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKremen, Claire, and Albie Miles. 2012. \u0026ldquo;Ecosystem Services in Biologically Diversified versus Conventional Farming Systems: Benefits, Externalities, and Trade-Offs.\u0026rdquo; Ecology and Society 17(4):art40. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.5751/ES-05035-170440\u003c/span\u003e\u003cspan address=\"10.5751/ES-05035-170440\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKremen, Claire, Neal M. Williams, and Robbin W. Thorp. 2002. \u0026ldquo;Crop Pollination from Native Bees at Risk from Agricultural Intensification.\u0026rdquo; Proceedings of the National Academy of Sciences 99(26):16812\u0026ndash;16. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1073/pnas.262413599\u003c/span\u003e\u003cspan address=\"10.1073/pnas.262413599\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMadin, Michael Biwalib, and Katherine S. Nelson. 2023. \u0026ldquo;Effects of Landscape Simplicity on Crop Yield: A Reanalysis of a Global Database\u0026rdquo; edited by M. Gaglio. PLOS ONE 18(12):e0289799. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1371/journal.pone.0289799\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0289799\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMargosian, Margaret L., Karen A. Garrett, J. M. Shawn Hutchinson, and Kimberly A. With. 2009. \u0026ldquo;Connectivity of the American Agricultural Landscape: Assessing the National Risk of Crop Pest and Disease Spread.\u0026rdquo; BioScience 59(2):141\u0026ndash;51. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1525/bio.2009.59.2.7\u003c/span\u003e\u003cspan address=\"10.1525/bio.2009.59.2.7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcGarigal, Kevin, and Barbara J. Marks. 1995. FRAGSTATS: Spatial Pattern Analysis Program for Quantifying Landscape Structure.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMitchell, Matthew G. E., Elena M. Bennett, and Andrew Gonzalez. 2013. \u0026ldquo;Linking Landscape Connectivity and Ecosystem Service Provision: Current Knowledge and Research Gaps.\u0026rdquo; Ecosystems 16(5):894\u0026ndash;908. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10021-013-9647-2\u003c/span\u003e\u003cspan address=\"10.1007/s10021-013-9647-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNguyen, Lan H., Samuel V. J. Robinson, and Paul Galpern. 2022. \u0026ldquo;Effects of Landscape Complexity on Crop Productivity: An Assessment from Space.\u0026rdquo; Agriculture, Ecosystems \u0026amp; Environment 328:107849. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.agee.2021.107849\u003c/span\u003e\u003cspan address=\"10.1016/j.agee.2021.107849\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eOstrom, Elinor. 2009. \u0026ldquo;A General Framework for Analyzing Sustainability of Social-Ecological Systems.\u0026rdquo; Science 325(5939):419\u0026ndash;22.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePashanejad, Ehsan. 2025. Balancing Agriculture and Conservation in the Canadian Prairies. Ottawa, Ontario.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePashanejad, Ehsan, Ali Kharrazi, Zuelclady M. F. Araujo-Gutierrez, Brian E. Robinson, Brian D. Fath, and Lael Parrott. 2024. \u0026ldquo;A Functional Connectivity Approach for Exploring Interactions of Multiple Ecosystem Services in the Context of Agricultural Landscapes in the Canadian Prairies.\u0026rdquo; Ecosystem Services 68:101639. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ecoser.2024.101639\u003c/span\u003e\u003cspan address=\"10.1016/j.ecoser.2024.101639\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eQiu, Jiangxiao, and Matthew Mitchell. 2024. \u0026ldquo;Understanding Biodiversity \u0026ndash; Ecosystem Service Linkages in Real Landscapes.\u0026rdquo; Landscape Ecology 39(11):188. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10980-024-01980-3\u003c/span\u003e\u003cspan address=\"10.1007/s10980-024-01980-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eQualman, Darrin. 2019. \u0026ldquo;Tackling the Farm Crisis and the Climate Crisis: A Transformative Strategy for Canadian Farms and Food Systems.\u0026rdquo; 102.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eQualman, Darrin, Annette Aur\u0026eacute;lie Desmarais, Andr\u0026eacute; Magnan, and Mengistu Wendimu. 2020. \u0026ldquo;Concentration Matters: Farmland Inequality on the Prairies.\u0026rdquo;\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRehman, Abdul, Muhammad Farooq, Dong-Jin Lee, and Kadambot H. M. Siddique. 2022. \u0026ldquo;Sustainable Agricultural Practices for Food Security and Ecosystem Services.\u0026rdquo; Environmental Science and Pollution Research 29(56):84076\u0026ndash;95. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11356-022-23635-z\u003c/span\u003e\u003cspan address=\"10.1007/s11356-022-23635-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRichter, Franziska J., Matthias Suter, Andreas L\u0026uuml;scher, Nina Buchmann, Nadja El Benni, Rafaela Feola Conz, Martin Hartmann, Pierrick Jan, and Valentin H. Klaus. 2024. \u0026ldquo;Effects of Management Practices on the Ecosystem-Service Multifunctionality of Temperate Grasslands.\u0026rdquo; Nature Communications 15(1):3829. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41467-024-48049-y\u003c/span\u003e\u003cspan address=\"10.1038/s41467-024-48049-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRieb, Jesse T., and Elena M. Bennett. 2020. \u0026ldquo;Landscape Structure as a Mediator of Ecosystem Service Interactions.\u0026rdquo; Landscape Ecology 35(12):2863\u0026ndash;80. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10980-020-01117-2\u003c/span\u003e\u003cspan address=\"10.1007/s10980-020-01117-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRies, Leslie, Robert J. Fletcher, James Battin, and Thomas D. Sisk. 2004. \u0026ldquo;Ecological Responses to Habitat Edges: Mechanisms, Models, and Variability Explained.\u0026rdquo; Annual Review of Ecology, Evolution, and Systematics 35(1):491\u0026ndash;522. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1146/annurev.ecolsys.35.112202.130148\u003c/span\u003e\u003cspan address=\"10.1146/annurev.ecolsys.35.112202.130148\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTamburini, Giovanni, Riccardo Bommarco, Thomas Cherico Wanger, Claire Kremen, Marcel G. A. van der Heijden, Matt Liebman, and Sara Hallin. 2020. \u0026ldquo;Agricultural Diversification Promotes Multiple Ecosystem Services without Compromising Yield.\u0026rdquo; Science Advances 6(45). doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1126/sciadv.aba1715\u003c/span\u003e\u003cspan address=\"10.1126/sciadv.aba1715\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTscharntke, Teja, Alexandra M. Klein, Andreas Kruess, Ingolf Steffan-Dewenter, and Carsten Thies. 2005. \u0026ldquo;Landscape Perspectives on Agricultural Intensification and Biodiversity \u0026ndash; Ecosystem Service Management.\u0026rdquo; Ecology Letters 8(8):857\u0026ndash;74. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/j.1461-0248.2005.00782.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1461-0248.2005.00782.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTscharntke, Teja, Jason M. Tylianakis, Tatyana A. Rand, Raphael K. Didham, Lenore Fahrig, P\u0026eacute;ter Bat\u0026aacute;ry, Janne Bengtsson, Yann Clough, Thomas O. Crist, Carsten F. Dormann, Robert M. Ewers, Jochen Fr\u0026uuml;nd, Robert D. Holt, Andrea Holzschuh, Alexandra M. Klein, David Kleijn, Claire Kremen, Doug A. Landis, William Laurance, David Lindenmayer, Christoph Scherber, Navjot Sodhi, Ingolf Steffan-Dewenter, Carsten Thies, Wim H. van der Putten, and Catrin Westphal. 2012. \u0026ldquo;Landscape Moderation of Biodiversity Patterns and Processes ‐ Eight Hypotheses.\u0026rdquo; Biological Reviews 87(3):661\u0026ndash;85. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/j.1469-185X.2011.00216.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1469-185X.2011.00216.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang, Tongli, Andreas Hamann, Dave Spittlehouse, and Carlos Carroll. 2016. \u0026ldquo;Locally Downscaled and Spatially Customizable Climate Data for Historical and Future Periods for North America\u0026rdquo; edited by I. \u0026Aacute;lvarez. PLOS ONE 11(6):e0156720. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1371/journal.pone.0156720\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0156720\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWood, Sylvia L. R., Kyle T. Martins, V\u0026eacute;ronique Dumais-Lalonde, Olivier Tanguy, Fanny Maure, Annick St-Denis, Bronwyn Rayfield, Amanda E. Martin, and Andrew Gonzalez. 2022. \u0026ldquo;Missing Interactions: The Current State of Multispecies Connectivity Analysis.\u0026rdquo; Frontiers in Ecology and Evolution 10. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fevo.2022.830822\u003c/span\u003e\u003cspan address=\"10.3389/fevo.2022.830822\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"landscape-ecology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"land","sideBox":"Learn more about [Landscape Ecology](https://www.springer.com/journal/10980)","snPcode":"10980","submissionUrl":"https://submission.nature.com/new-submission/10980/3","title":"Landscape Ecology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"agroecosystem, landscape structure, multifunctionality, connectivity, diversity, crop yield, resilience","lastPublishedDoi":"10.21203/rs.3.rs-7467778/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7467778/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eContext.\u003c/h2\u003e\u003cp\u003eSustaining agricultural productivity while maintaining ecological integrity requires understanding the spatial dynamics of ecosystem services (ES). In the Canadian prairies\u0026mdash;an intensively modified agricultural landscape\u0026mdash;the degradation of natural habitats has impacted ES flows crucial for food security.\u003c/p\u003e\u003ch2\u003eObjectives.\u003c/h2\u003e\u003cp\u003eWe investigated how internal ES flows, mediated by landscape structure, influence crop yield at the Soil Landscape of Canada (SLC) scale, an ecologically meaningful delineation based on natural features. Our primary objective was to determine the relative importance of landscape composition versus configuration in predicting agricultural productivity.\u003c/p\u003e\u003ch2\u003eMethod.\u003c/h2\u003e\u003cp\u003eWe conducted a biophysical assessment of key ES (pollination, carbon storage, habitat quality, soil erosion control) for the year 2020. We quantified landscape composition and configuration metrics at the SLC scale to represent ES flow pathways. Generalized additive models (GAMs) were used to analyze the non-linear effects of these variables on a composite crop yield index.\u003c/p\u003e\u003ch2\u003eResults.\u003c/h2\u003e\u003cp\u003eOur findings reveal that landscape configuration\u0026mdash;notably connectivity (positive linear effect) and crop diversity (complex non-linear effect)\u0026mdash;significantly predicts crop yield, often exerting greater influence than the mere amount of natural habitat. A secondary analysis showed that yield in specific crops like canola, which depends on pollination, responded positively to natural habitat extent. The models explained a substantial portion of yield variance (Adjusted R\u0026sup2; \u0026asymp; 0.66\u0026ndash;0.67).\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e\u003cp\u003e. Our analysis highlights that agricultural output is not solely a function of field-level inputs but is deeply embedded within, and responsive to the landscape matrix at SLC scale and the ecological processes it mediates. Strategically enhancing landscape cohesion and crop diversity may therefore offer greater yield benefits than focusing on increasing isolated natural habitat, guiding a shift towards spatially explicit, multifunctional landscape planning.\u003c/p\u003e","manuscriptTitle":"Pathways to productivity: how functional ecosystem service flow determines crop yield in a prairie landscape","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-12 13:18:42","doi":"10.21203/rs.3.rs-7467778/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-10-27T12:18:38+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-10T04:36:53+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-02T15:49:47+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"296313751984399471520625714924138741478","date":"2025-09-10T14:53:04+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"5107308043967795604761172276080946850","date":"2025-09-08T12:14:56+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-07T13:57:37+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-08-29T07:41:11+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-08-29T07:41:06+00:00","index":"","fulltext":""},{"type":"submitted","content":"Landscape Ecology","date":"2025-08-27T04:40:41+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"landscape-ecology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"land","sideBox":"Learn more about [Landscape Ecology](https://www.springer.com/journal/10980)","snPcode":"10980","submissionUrl":"https://submission.nature.com/new-submission/10980/3","title":"Landscape Ecology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"6004ffdc-5f5c-4bc8-b375-1da04e676738","owner":[],"postedDate":"September 12th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2026-03-23T16:00:21+00:00","versionOfRecord":{"articleIdentity":"rs-7467778","link":"https://doi.org/10.1007/s10980-026-02333-y","journal":{"identity":"landscape-ecology","isVorOnly":false,"title":"Landscape Ecology"},"publishedOn":"2026-03-21 15:57:41","publishedOnDateReadable":"March 21st, 2026"},"versionCreatedAt":"2025-09-12 13:18:42","video":"","vorDoi":"10.1007/s10980-026-02333-y","vorDoiUrl":"https://doi.org/10.1007/s10980-026-02333-y","workflowStages":[]},"version":"v1","identity":"rs-7467778","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7467778","identity":"rs-7467778","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}