Unraveling ecosystem function responses to grazing intensity through UAV-based hyperspectral analysis

Hyperspectral remote sensing conducted with unmanned aerial vehicles (UAVs) is revolutionizing ecological research by enabling fine-scale monitoring of ecosystem functions. This study investigates how grazing intensity mediates the relationship between functional trait composition and aboveground biomass in a grassland using hyperspectral imagery from the Inner Mongolian Steppe. High-resolution spectral data were utilized to retrieve aboveground biomass and 9 key functional traits with reliable accuracy (nRMSE < 0.2). These traits include leaf mass per area (LMA), leaf thickness, leaf dry matter content (LDMC), leaf carbon concentration (leaf [C]), leaf nitrogen concentration (leaf [N]), leaf phosphorus concentration (leaf [P]), leaf potassium concentration (leaf [K]), leaf calcium concentration (leaf [Ca]), and leaf magnesium concentration (leaf [Mg]). Functional traits were analyzed through community-level means, diversity indices, and plant trait networks to quantify their contributions to biomass under varying grazing pressures. Our results suggest that the impact of dominated species on community biomass became more pronounced under conditions of heavy grazing. Functional diversity exhibited increasingly positive effects on biomass with higher grazing intensity, emphasizing stronger complementarity effects within communities. Furthermore, communities with tightly integrated trait networks, characterized by lower modularity and shorter average path lengths, supported greater biomass under stressful grazing conditions. This study highlights the capability of UAV-mounted hyperspectral remote sensing to capture detailed trait variations and ecosystem functions in grazing systems. Our findings provide valuable insights for advancing remote sensing applications in grassland management and conservation, particularly under intensifying grazing pressures.