Plant litter decomposition in global drylands is better predicted by precipitation seasonality and temperature than by aridity

Abstract

Understanding the global carbon (C) balance in terrestrial ecosystems is crucial for predicting their current and future roles as C sources or sinks in the context of global change. Drylands, covering nearly 45% of Earths land surface, contribute significantly to net primary production (NPP) and influence the interannual variability of the terrestrial C sink. However, the controls on plant litter decomposition, a major pathway of C release, remain unclear in these ecosystems. Here, we present a global analysis of plant litter decomposition in drylands, using a dataset from 116 sites across five continents spanning diverse climates and ecosystems. We found that litter decomposition does not correlate with mean annual precipitation (MAP) at the global scale, challenging the paradigm that water availability is the primary constraint on ecological processes in drylands. Instead, our analysis identifies mean annual temperature (MAT), precipitation-temperature synchrony, precipitation variability, and cloud cover frequency as key drivers. Specifically, our model predicted faster decomposition rates for warmer and more monsoonal ecosystems, but vary independently of MAP. Additionally, decomposition correlated positively with both lignin and nitrogen content, in contrast to the negative lignin-decomposition relationship commonly observed in mesic ecosystems. These findings suggest a fundamental mismatch between aridity and its expected effects on decomposition rates in terrestrial ecosystems. Given the ongoing expansion of drylands, rising temperatures and changes in precipitation variability under climate change; our results underscore the need to refine decomposition models beyond traditional aridity frameworks. Such refinement is essential for accurately predicting dryland contributions to the global C balance.