Soil C and N as causal factors of spatial variation in extracellular enzyme activity across grassland-woodland ecotones

•Kriging showed decomposition hotspots with co-occurrence of high enzyme activities.•Spatial range of enzymes closely resembled soil fertility level.•Soil C and N levels and not pH were the consistent drivers of enzymes activities. Soil enzymes are a ubiquitous component of ecosystem function and play a pivotal role in key biogeochemical processes including soil organic matter (SOM) decomposition and nutrient cycling. Enzyme activity has therefore been used as a surrogate for microbial activities and an indicator of soil productivity. Although numerous studies have assessed enzyme activity individually in different ecosystems, little is known about soil factors that determine their spatial structure and associated variation in activities across ecotones. Ecotones provide unique opportunities to study how soil properties vary spatially within and across adjacent land-uses. This study used multiple approaches, including geostatistical kriging, principal component analysis (PCA) and structural equation modelling (SEM) to examine the determinants and spatial distribution patterns of six key soil enzymes across two grassland-woodland ecotones in south-eastern Australia. Our results showed significant spatial dependencies for the activities of the enzymes across ecotones, and highlighted that spatial autocorrelation patterns were controlled by soil fertility factors regardless of enzyme type. Kriging maps revealed “decomposition hotspots” across ecotones with spatial co-occurrence of high enzyme activities. The spatial range of enzymes closely resembled soil fertility level as measured by a range of soil chemical properties, indicating a high degree of spatial co-dependency. Consistent with this, the SEM analysis indicated that soil resources such as C and N levels were the causal factors of variation in enzyme activities at field-scale. Soil pH showed little variation across the landscape at both sites and thus, had no major effect on enzyme activities. Overall, this study demonstrates spatial dependency and edaphic determinants of soil enzyme activities are consistent across grassland-woodland ecotones.