Soil–litter mixing and microbial activity mediate decomposition and soil aggregate formation in a sandy shrub-invaded Chihuahuan Desert grassland

Drylands account globally for 30% of terrestrial net primary production and 20% of soil organic carbon. Present ecosystem models under predict litter decay in drylands, limiting assessments of biogeochemical cycling at multiple scales. Overlooked decomposition drivers, such as soil–litter mixing (SLM), may account for part of this model-measurement disconnect. We documented SLM and decomposition in relation to the formation of soil-microbial films and microbial extracellular enzyme activity (EEA) in the North American Chihuahuan Desert by placing mesh bags containing shrub (Prosopis glandulosa) foliar litter on the soil surface within contrasting vegetation microsites. Mass loss (in terms of k, the decay constant) was best described by the degree of SLM and soil-microbial film cover. EEA was greatest during periods of rapid litter decomposition and associated SLM. Soil-microbial film cover on litter surfaces increased over time and was greater in bare ground microsites (50% litter surface area covered) compared to shrub and grass microsites (37 and 33% covered, respectively). Soil aggregates that formed in association with decomposing leaf material had organic C and N concentrations 1.5–2× that of local surface soils. Micrographs of soil aggregates revealed a strong biotic component in their structure, suggesting that microbial decomposition facilitates aggregate formation and their C and N content. Decomposition drivers in arid lands fall into two major categories, abiotic and biotic, and it is challenging to ascertain their relative importance. The temporal synchrony between surface litter mass loss, EEA, biotic film development, and aggregate formation observed in this study supports the hypothesis that SLM enhances decomposition on detached litter by promoting conditions favorable for microbial processes. Inclusion of interactions between SLM and biological drivers will improve the ability of ecosystem models to predict decomposition rates and dynamics in drylands.