Mesoscale Covariation in the Plant–Soil Neighbourhood of a South African Montane Grassland

Abstract

ABSTRACT Plant–soil relationships are the foundation of terrestrial life, sustaining global ecosystems through biomass and food production, nutrient cycling and microbial activity, and ecosystem regulation. To better understand plant–soil relationships in a changing world, we investigated mesoscale covariation in the plant–soil neighbourhood of a South African flammable montane grassland. In a randomized block experiment of 30 plots, we assessed variation in plant species composition and cover, and on three occasions within a single growing season (spring–summer–autumn), we determined above‐ground plant biomass and nutrient concentrations by plant functional type (PFT). We also assessed variation in soil depth, and soil chemical and physical properties. Some 10% of the variation in plant species composition was explained by plot identity. Forbs accounted for 68% of species richness, whereas graminoids (almost exclusively grasses) contributed 93% of biomass. The C 4 grasses Themeda triandra , Tristachya leucothrix , Heteropogon contortus and Harpochloa falx, and the C 3 grass Koeleria capensis , dominated cover and biomass. Graminoid biomass was positively correlated with soil depth, whereas forb biomass was negatively correlated with soil total carbon and nitrogen. Plant diversity (using Shannon‐Weiner and Simpson's Indices of Diversity) did not share a significant relationship with biomass. PFT, season, and their interaction influenced biomass. Results suggest that graminoids and forbs exhibit highly divergent resource strategies presumably as an adaptation to fire and grazing. Graminoids appeared to show a trade‐off prioritizing biomass over nutrients to minimize nutrient losses in flammable biomass. Forbs, by contrast, invested more heavily in nutrients. Mobile macronutrient concentrations decreased in autumn due to translocation to below‐ground storage organs prior to senescence. Additionally, microbially mediated soil mineralization, activated by warming and precipitation, resulted in higher N and P herbage concentrations in spring. Although soil variables contributed to biotic variation, we speculate that historical land management legacies have had a confounding influence on site variation. Our study showed that the plant–soil neighbourhood of a montane grassland is tightly coupled at a small spatial scale. Such focused studies are essential for understanding how plant–soil relationships may be affected by a changing world. This has implications for the management and conservation of montane grasslands.