Aims

Biocrusts, communities dominated by mosses, lichens, cyanobacteria, algae, and fungi living on the soil surface, constitute a vital biotic component of dryland ecosystems that play critical roles in maintaining their structure and functioning. However, there are substantial knowledge gaps regarding the global distribution of biocrusts, which has primarily been estimated using statistical methods. Given that water is a key limiting factor for dryland biota, it holds significant potential as a key driver of the global distribution of biocrusts.

Methods

Here we utilized a novel global biocrust dataset comprising 2348 data points from six continents (excluding Antarctica) and adapted an existing ecohydrological model to predict biocrust cover across global drylands.

Results

We found that in the regions with sufficient data, such as the Loess Plateau in China, the ecohydrological model reached good correlation coefficients between the simulated and observed biocrust cover (average correlation coefficient of 100 simulations was 0.727, with a corresponding RMSE of 0.311). However, at a global scale the model was unable to reproduce observed biocrust cover (correlation coefficient < 0.352, RMSE > 0.331). The high variability observed in biocrust cover globally may contribute to the low performance of the model at this scale.

Conclusions

Our findings indicate that ecohydrological models show promise to map biocrust distribution at the regional scale when accompanied with region-specific detailed information about biocrusts and improved environmental data. They also provide a stepping-stone for future research to advance towards a mechanism-based prediction of global biocrust distribution.