Soil is a non-renewable resource that is globally being lost to erosion many times faster than formed through pedogenesis. Soil aggregate stability is a key property conferring erosion resistance and a key indicator of soil health. The classic view holds that soil aggregate stability is determined by vascular plants (and their symbionts, exudates and residues) that generate soil organic matter, alongside soil texture and chemistry. Biological soil crust (biocrust) communities are the missing element in this view. Biocrusts are of particular importance in drylands where soils generally have low organic matter content and plant cover, and they may cover sizeable interspaces of exposed soil between plants. Interspace soil surfaces are often colonized by biocrusts. We conducted a synthesis comparing the contributions of plants, biocrusts, and abiotic soil properties to surface aggregate stability. We gathered nearly 100 original datasets from US and international drylands, spanning an aridity gradient. We fitted them with similar structural equation models and extracted coefficients for use in a subsequent meta-analysis. We asked: 1. Which influence – biocrusts, vascular plants, or abiotic soil properties– most strongly affects surface soil aggregation? and 2. How does aridity alter the magnitude of these effects?
Results/Conclusions
On average, our models explained 44% of the variation in soil aggregate stability with standardized direct effects as follows: biocrusts = |0.46|, vascular plants = |0.35|, abiotic soil properties = |0.23|. Across all models, the biocrust influence was stronger than all other effects, but plant and abiotic effects were not clearly distinguishable in magnitude. Biocrusts effects were the strongest influence on soil aggregate stability within 58% of the models in which they could be compared to plant effects, abiotic soil effects or both. Plant effects were the strongest influence within 36% of the models in which they could be compared to biocrust or abiotic effects or both. Direct effects of both biocrusts and plants on aggregate stability were prevailingly positive; however, plants also exerted substantial indirect effects through their effect on biocrusts (|0.36|), and these were a mixture of positive and negative effects. Further we found that the influence of biocrusts increases as sites become more arid (R2 = 0.09, P = 0.01), but the influence of the other effects was not sensitive to aridity. These results suggest that the generation of aggregate stability follows a different set of rules in drylands than the classic view based on mesic ecosystems.