Productivity at a price: mixed pastures show higher ecosystem respiration vulnerability to drought–grazing stress than grass-only patches

Pastures contain a fine-scale mosaic of grass-only (GP) and mixed (grass–legume–forb; MP) patches, whose functional traits shape belowground C inputs and ecosystem respiration (ER). These traits make ER sensitive to drought and frequent grazing, yet their combined effects across patch types remain poorly understood.

To address this, we conducted a 2×2 rainfall × grazing factorial in GP and MP in a field-based temperate pasture climate-manipulation experiment, quantifying the effects of drought, frequent grazing, and their combination on ER, its temperature and moisture sensitivity, and plant C-use efficiency (AGB/ER). Rainfall was based on 30-year records, and grazing simulated via one or three harvests per season. ER was measured during spring, summer, and autumn 2023–2024, and structural equation modelling identified the key pathways by which biophysical factors regulate ER in each patch type.

Highly productive MP, compared to GP, consistently exhibited higher ER (spring: 0.17 vs. 0.11; summer: 0.32 vs. 0.17; autumn: 0.41 vs. 0.12 g C m^-2 hr^-1), greater C-use efficiency (3.2 ± 0.63 vs. 0.09 ± 0.02), higher apparent temperature sensitivity (Q₁₀: 1.46 vs. 1.22), and weaker moisture constraints. However, this higher functioning came with greater vulnerability: under combined drought and frequent grazing, ER declined non-additively and more sharply in MP (–14.5%, –42.8%, –67.3%) than in GP (–4.6%, –34.2%, –11.2%). C-use efficiency dropped by 80% in MP but remained stable in GP, accompanied by larger reductions in AGB and Q₁₀. Mechanistically, ER in MP was plant-biomass driven, whereas in GP it was microbial-substrate driven, with both indirectly constrained by moisture and temperature-induced soil drying.

These results show that the higher productivity of MP comes at the cost of greater ER vulnerability to drought–grazing stress, offering guidance for grazing management and strengthening predictions of pasture C–climate feedbacks.