Heat Stress in a Climate Setting: A Framework for Reanalyses

The proliferation of reanalysis models for the atmosphere in recent decades has allowed researchers to study Earth's past climate in great detail. While much work has gone into understanding key climate indicators such as surface temperature and precipitation trends, there have been few studies dealing with heat stress. As climate change grows increasingly exigent, it is becoming vitally important to understand the thermal impacts on biological systems. This study analyzed data from five reanalysis models (20CRv2, NARR, NNRA 1, NCEP DOE 2, and ERA-I) and found agreement in average surface temperature increases of 0.2-0.6°C per decade across the U.S. west coast and east coast since 1979. These trends were consistent with previous studies. Less agreement was found for the central U.S. The Temperature Humidity Index and the Heat Index were found to generally follow the temperature trends. An analysis of the role of moisture indicated that the effect of specific humidity on heat stress is dependent on climatology. Trends of heat stress over arid regions such as the desert southwest were found to be much more influenced by temperature trends than by moisture trends. In contrast, moisture seemed to play a stronger role in the more humid southeast. There appeared to be a more equal effect of temperature and moisture on heat stress in the northeast and Great Lake states. Perhaps equally as important, the study provides a framework to reduce computational time but allows for more rigorous statistical methods that are not available in the typical suite of software and programming languages to analyze climate data. Functionality was developed to infer daily extrema from six-hourly reanalysis data. A shapefile was used to aggregate the data according to prescribed geographic boundaries and reduce the load of data for statistical analysis. Time series decomposition was performed on the aggregated daily data to determine linear trends which were then mapped out to visualize their spatial features.