Enhanced Arid And Semi-arid Warming Over Mid-high Latitudes Of Northern Hemesphere

By analyzing the high resolution data from Climate Research Unit (monthly precipitation and mean surface air temperature) version TS3.1, we examined the trends of annual mean surface air temperature as function of climatological mean precipitation. It’s found that the warming was particularly enhanced over the mid-high latitude semi-arid areas of Northern Hemisphere, especially a warming of the winter periods. The warming over semi-arid areas is faster than any other regions and contributes most to the global warming. From this perspective, we define a new trend by the spatial-temporally multidimensional ensemble empirical mode decomposition (MEEMD) method, diagnose the evolution of global land surface air temperature trend in the past century, analyze the temporal and spatial characteristics of dry/wet variation at the surface of North America and Eurasia, and their correlations with the enhanced warming. At the end of this study, we define a Modulated Annual Cycle (MAC) of land surface air temperature at each grid point, then we study the evolution of MAC trend, and characteristics of the changes of MAC under different timescales, on which basis we evaluate the correlations between MAC changes and the enhanced arid and semi-arid warming over the mid-high latituds of Northern Hemisphere. The main results are summarized:(1) The entire globe is warming from 1900 onward. The noticeable warming first started around the regions circling the Arctic and subtropical regions in both hemispheres. But the largest warming to date has been at the northern mid-high latitudes which include most of the arid and semi-arid regions, Asia, Europe and North America have been experiencing the largest warming since 1900. Before 1950, there were both moderate warming and weak cooling regions. The cooling regions shrank and most of them turned into warming regions with an accelerated pace of warming over the next three decades. By 1980, except for the weak cooling in the northern tip of Greenland and in the vicinity of the Andes, almost all the global land had been warming. The warming rates over the global land have changed little since. The fastest warming in recent decades (>0.4 K per decade) occurred in arid and semi-arid regions of northern mid-latitudes.(2) By analyzing the PDSI, it is found that there was no clear trend of precipitation in North America and Eurasia during winter seasons in1950-2008, so the dry/wet variation was mainly affected by the change of temperature, especially in the places where winter warming was enhanced. It’s also found that most of the global land is drying by analyzing the aridity index (AI), and there are large regional differences in AI trends:at hemisphere scale, the eastern hemisphere is drying while the western hemisphere is wetting. At continental scale, North America becomes wet before it’s drying, the wetting-drying shift of 50°N and it’s North takes place around 70s while it’s South starts around late 90s; for Euroasia, drying of the entire continent is much severe, which starts at high latitudes and extends toward lower latidudes.(3) There are also large regional differences in trends of the amplitudes of land surface air temperature annual cycle. Amplitudes of annual cycle are increasing in most parts of Southern Hemisphere. As for Northern Hemisphere, most parts are increasing first, then decreasing, while there are long-term decreasing trends in the high latitudes. From the zonally averaged trend (over the land area only), one could find that amplitudes of the annual cycle experienced the most dramatic changes are located between 30°N and 60°N, which are the mid-latitudes. Before 1940, amplitudes of annual cycle in mid-latitudes are increasing and decrease fast since then, while the high-latitudes have experienced the significant decreasing, and the decreasing trends extend towards mid-latitudes latterly, which coincides with the enhanced arid and semi-arid cold season warming. At the same time, the characteristics of annual cycle amplitudes at different time scales are studied by EOF analyses.