Ecological networks, formed by species interactions, are crucial for understanding biodiversity and ecological functions. While climate and land use significantly impact species richness and composition, their effects on ecological networks remain understudied, particularly regarding multi-trophic food webs across different biomes. Previous research often focused on simplified systems or limited taxonomic groups, hindering a comprehensive understanding of real-world food web responses to environmental change. Aquatic and terrestrial food webs, while parts of the same landscape, have largely been studied separately. Aquatic systems tend to have long food chains and nested structures, while terrestrial systems often exhibit shorter chains and modular structures. However, systematic comparisons of their responses to environmental drivers are lacking. This study addresses this gap by investigating how multi-trophic food webs are influenced by environmental drivers (elevation and land use) and whether aquatic and terrestrial food webs respond differently. The large-scale, comparative approach allows for an unbiased assessment of food web structural differences resulting from compositional differences in local communities, offering crucial insights into biodiversity management in a changing world.

The literature highlights the importance of understanding ecological networks to effectively manage biodiversity in a changing world. Studies examining the structural changes of ecological networks along environmental gradients and their effects on species persistence and ecosystem functioning are beginning to emerge, but mostly focus on interactions between two distinct taxonomic groups (e.g., herbivore-plant or host-parasitoid networks) or simplified experimental systems. The challenges in measuring interactions across various taxonomic groups and environmental gradients have limited research on real-world multi-trophic food webs' responses to environmental change. While aquatic and terrestrial food webs are integral parts of the same landscape, they have often been studied in isolation. Known structural differences exist, with aquatic systems characterized by long food chains and nested structures, and terrestrial systems showing shorter chains and modularity. This research gap necessitates a landscape-scale comparison of blue and green food webs' responses to shared environmental drivers.

This study employed a metaweb approach combined with empirical species co-occurrence data to infer local food webs for 462 terrestrial and 465 aquatic sites across Switzerland, spanning a wide elevational range and various land-use types. A metaweb, representing the regional food web with trophic interactions among species, was constructed using literature, published datasets, and expert knowledge. Local food webs were inferred by identifying trophic relationships among co-occurring species. The study focused on five key food-web metrics: number of nodes (species richness), connectance, nestedness, modularity, and consumer diet niche overlap. These metrics were analyzed in relation to elevation and five major land-use types (forests, scrubs, open spaces, farmlands, and urban areas). Elevation was chosen as a proxy for temperature, as the two variables were highly correlated, and analyses showed that elevation encapsulated most of the temperature's effects on food web structure. Piecewise structural equation modeling (SEM) was used to analyze the associations between food-web properties and environmental drivers, considering the mutual dependence among food-web metrics. Generalized additive models and linear models were employed to investigate nonlinear elevational patterns and responses among land-use types. Randomization tests were conducted to determine whether observed patterns resulted from changes in food-web size and connectance or from changes in species composition along elevation. The data for the study came from several sources including the Biodiversity Monitoring Switzerland programme and the Swiss database on faunistic records (info fauna). Taxa were resolved to the species level whenever possible, but some higher-level taxonomic groupings were used for some groups.

Principal component analysis revealed distinct structural differences between terrestrial and aquatic food webs. Aquatic webs were generally smaller, more connected, and less modular than terrestrial webs. Piecewise structural equation modeling showed significant associations between elevation and various food-web properties in both systems, but with contrasting relationships. In terrestrial food webs, modularity increased and niche overlap decreased with elevation. In aquatic food webs, the opposite pattern was observed. These contrasting relationships were particularly pronounced in farmlands. Generalized additive models revealed nonlinear elevational patterns in both systems. Terrestrial food webs showed an increase in the number of nodes with elevation up to a certain point (around 1500–2000 m a.s.l.), followed by a decrease, likely due to the tree line effect. Aquatic food webs showed a consistent linear decrease in the number of nodes with elevation. Connectance decreased linearly with elevation in terrestrial webs, while in aquatic webs it initially decreased and then mildly increased above 1000 m a.s.l., reflecting the loss of fish species and the dominance of generalist invertebrates at higher elevations. Randomization tests revealed that terrestrial webs were more nested and modular (below 2500 m a.s.l.) than their randomized counterparts, while aquatic webs were more nested but less modular. Both types showed decreasing nestedness with elevation, and the pattern in aquatic webs was linked to the decline in fish species richness. Consumer diet niche overlap was higher in inferred food webs than in randomized ones, with terrestrial webs exhibiting less overlap and aquatic webs more overlap than expected by chance. Significant differences in the regression slopes of terrestrial and aquatic food webs were observed for all five metrics in farmlands, with contrasting relationships between elevation and modularity/niche overlap in the two systems.

This study demonstrates that the structure and ecological properties of local food webs vary significantly with elevation and land-use type. Importantly, terrestrial and aquatic food webs exhibit contrasting responses to these environmental factors. The nonlinear elevational patterns observed are likely driven by biogeographic boundaries (e.g., tree line) and land-use type turnover along elevation. Elevation serves as a useful proxy for climate change effects. Terrestrial food web patterns likely reflect the influence of temperature variation on species distribution and community composition, while aquatic patterns are influenced by a combination of temperature and topographical factors (e.g., stream structure and hydrology). The nested structure of aquatic food webs reflects the broad feeding habits of aquatic consumers constrained by body size, whereas the modular structure of terrestrial food webs indicates a high proportion of specialized consumer-resource interactions. Elevation had contrasting effects on diet specialization in the two systems, especially in farmlands. Farmlands stand out as a critical land-use type where blue and green food webs showed qualitatively different responses to elevation, suggesting the importance of considering this factor when managing local biodiversity. Anthropogenic land-use significantly affects both terrestrial and aquatic food webs. Agricultural practices simplify vegetation and favor generalists over specialists, leading to differences in food web structure compared to natural habitats. The interactive effects of land use and climate change on food webs are evident in the contrasting elevational patterns observed in forests and farmlands.

This large-scale analysis provides strong evidence that aquatic and terrestrial food webs in the same landscape respond differently to elevation and land use. The observed patterns reflect the combined effects of evolutionary and ecological processes. The study’s findings highlight the potential impacts of climate change on biodiversity and suggest that management strategies should consider the distinct responses of aquatic and terrestrial systems to environmental changes. Future research could focus on finer temporal resolution studies of interconnected blue-green systems and broader taxonomic coverage.

While the study used a comprehensive dataset, some limitations exist. The metaweb approach might overestimate the number of trophic links by assuming fixed diets for species, ignoring intraspecific diet variation. The simplified representation of basal resources in aquatic systems could also influence the results. The study focused on a specific region (Switzerland) and might not be generalizable to all ecosystems. Further, while the study considers the interaction between land use and elevation, the influence of other environmental factors is not fully explored.