Observational data showed heavy snow in North-West US and adjacent Canadian area during the winter season correlate with cold spring temperatures in the western US, and relatively weak summer precipitation in the eastern US. Observational data indicate there are a negative correlation between winter snow and surface temperature and a positive correlation between surface temperature in later spring and summer precipitation over the southeastern U.S.

Modeling study is conducted to test this relationship and elucidate mechanisms that cause this correlation. As a first step, we test the impact of deep soil temperature in the western US during the late spring on the summer US precipitation. A coupled Eta/SSiB model has been used to test for short-term weather prediction (Xue et al., 2000). In this study, a climate version of the coupled Eta/SSiB model is used for this study.

The model is integrated for three months from May 1, 1998 through July 31, 1998 with two different initial deep soil temperatures over the western US, one corresponds with warm years and the other corresponds with cold years. Each case consists of five different initial atmospheric conditions. The ensemble means are taken for analyses. The Eta/SSiB produces a reasonable monthly mean climate. The experiments with warm initial deep soil temperature in the western US produce stronger monthly mean precipitation over the southern U.S. Meanwhile, precipitation is reduced to the north. The impact mainly occurs in June, which is consistent with observed anomaly. The temperature anomalies induce the variations in wind field, which in turn change the amount of moisture flowing into the continent, resulting in the change in rainfall.

In addition, different initial deep soil anomalies are also specified to test the model response. In one experiment, deep soil temperature anomalies that are greater than the previous experiment are specified, which produces a much larger anomalies in precipitation. In another experiment, observed May deep soil temperature anomalies between 1992 (a warm year) and 1998 (a cold year) are specified in the Eta model. The simulated precipitation anomalies patterns are not only consistent with the above-mentioned experiments, but are also closer to the observed precipitation difference between 1998 and 1982.

This preliminary study indicates that deep soil temperature over the western U.S. in late spring has impact on the summer U.S. precipitation, which could have a substantial implication for the North American snow-monsoon interactions.