资源环境科技发展态势分析平台

Detecting and attributing a human influence on observed rainfall trends is a major challenge due to the presence of large amplitude internal variability on all time scales and by limited temporal and spatial data coverage. Here we apply a dynamical adjustment methodology to a gridded archive of monthly precipitation to estimate an anthropogenic influence on long-term (1920-2015) trends over North America and Eurasia during winter (November-March). This empirical approach aims to remove atmospheric circulation influences from precipitation variability and trends, thereby revealing the thermodynamically induced component as a residual. The geographical pattern and amplitude of this observed thermodynamic residual precipitation trend are in good agreement with anthropogenically forced trends obtained from ensembles of historical climate model simulations. Such consistency helps to reconcile observations and models and provides compelling evidence for a human influence on century-scale precipitation trends over North America and Eurasia during the cold season.

Plain Language Summary It is difficult to isolate the anthropogenic influence on long-term precipitation trends due to confounding effects from internal variability. Here we remove the influence of atmospheric circulation variability, which is primarily unforced, from observed precipitation trends using an empirical approach called "dynamical adjustment." This removal isolates the thermodynamic component of observed precipitation trends as a residual. We find that this thermodynamic component is in good agreement with the anthropogenic component determined from historical simulations from climate models. Thus, we conclude that we are able to identify a human influence on observed century-scale precipitation trends over North America and Eurasia.