Predicting soil respiration for the Qinghai-Tibet Plateau: An empirical comparison of regression models

Abstract

Alpine ecosystems like the Qinghai-Tibet Plateau strongly respond to global warming. Their soils, containing large carbon stocks, release more carbon dioxide as a possible consequence. Reciprocally, this may intensify climate warming. The Qinghai-Tibet plateaús large and almost inaccessible terrain results in a general data scarcity for this area making the quantification of soil carbon dynamics challenging. The current study provides an area-wide estimation of soil respiration for the Qinghai-Tibet Plateau, which is a key region for climate change studies due to its size and sensitivity. We compared the ability of six regression models to predict soil respiration that were developed within different studies and are based on mean annual air temperature, mean annual precipitation and belowground biomass. We used the WorldClim data sets to approximate annual soil respiration on a regional scale. Compared to field measurements of soil respiration at single spots in different vegetation zones on the Qinghai-Tibet Plateau (max. 1876.63gCm−2year−1), our predicted results (max. 1765.13gCm−2year−1) appear to be consistent. The basic difference between grasslands and forests in soil respiration is indicated by all regression models, however, a more precise differentiation between vegetation types is only exhibited by the regression model based on mean annual precipitation. Overall, this model performs best for most and the largest vegetation zones. Nevertheless, the approximations of the model based on mean annual temperature by Raich and Schlesinger (1992) with a lower constant better represent the vegetation zone of the alpine steppe. With this spatial estimation of soil respiration at a regional scale, a basis for assessing an area-specific potential of greenhouse gas emissions on the Qinghai-Tibet Plateau is provided. Moreover, we quantify a complex soil ecological process for this data-scarce area.