Characterizing the Increase in U.S. Precipitation

U.S. annual precipitation has increased in recent decades, especially over the Northeast where totals have risen about 15% since 1901. Accompanying this increase in total precipitation have been increases in the frequency and intensity of very heavy rain events. Is this regional wetness a new normal, and furthermore does the upward trend portend even wetter conditions for upcoming decades? One view on the mechanism for these U.S. increases in precipitation is that they are consistent with increasing water vapor in a warmed climate (U.S. National Climate Assessment; Melilo et al. 2014). It remains to be determined if alternate mechanisms could also be operating. To what extent, for instance, could multi-decadal changes in regional U.S. precipitation arise naturally, either via coupled modes of atmosphere-ocean variability or from purely atmosphere-land low frequency variability? Could the sign of recent precipitation change over the Northeast U.S. be consistent with a signal due to global warming, with the magnitude of the increased wetness reflecting mainly a particular phase of internal variability? The issue of fully characterizing the increase in regional U.S. precipitation thus hinges critically on knowing not just the anthropogenic climate change signal but also knowing the statistics of internal variability. Here we use very long integrations of atmosphere and coupled atmosphere -ocean models to diagnose statistics in the variability of 20-yr averaged precipitation over a region of the Northeast U.S. that has witnessed substantial recent increases in precipitation. Parallel 2000-yr long equilibrium integrations of the coupled model are diagnosed, one that was subjected to radiative forcing of the late 19th century and the other to radiative forcing of the early 21st century. A large number of samples of 20-yr averaged regional precipitation are analyzed from each to 1) characterize the internally driven decadal-scale climate variability, and 2) to compare the magnitude of that variability to the signal of externally driven mean change. The importance of atmosphere-land only interactions is assessed by comparing with the statistics of 20-yr precipitation variations occurring in a parallel integration of an atmospheric model subjected to the fixed mean SSTs, sea ice, and radiative forcing.