Spatial Patterns of Global Precipitation Change and Variability during 1901–2010

Abstract

his study examines global precipitation changes/variations during 1901?2010 by using the long-record GPCC land precipitation analysis, the NOAA/Cooperative Institute for Climate and Satellites (CICS) reconstructed (RECONS) precipitation analysis, and the CMIP5 outputs. In particular, spatial features of long-term precipitation changes and trends and decadal/interdecadal precipitation variations are explored by focusing on the effects of various physical mechanisms such as the anthropogenic greenhouse gas (GHG) and aerosol forcings and certain internal oscillations including the Pacific decadal variability (PDV) and Atlantic multidecadal oscillation (AMO).Precipitation increases in the Northern Hemisphere (NH) mid- to high-latitude lands observed in GPCC can also be found in RECONS and model simulations. Over tropical/subtropical land areas, precipitation reductions generally appear in all products, but with large discrepancies on regional scales. Over ocean, consistent spatial structures of precipitation change also exist between RECONS and models. It is further found that these long-term changes/trends might be due to both anthropogenic GHG and aerosols. The aerosol effect estimated from CMIP5 historical simulations is then removed from the GPCC, RECONS, and AMIP simulations. These isolated GHG-related changes/trends have many similar spatial features when compared to the CMIP5 GHG-only simulations, especially in the zonal-mean context.Both PDV and AMO have influence on spatial patterns of precipitation variations during the past century. In the NH middle to high latitudes, PDV and AMO have played an important role on interdecadal/multidecadal time scales. In several tropical/subtropical regions, their impacts may even become dominant for certain time spans including the recent past two decades. Therefore, these two internal mechanisms make the estimations of GHG and aerosol effects on precipitation on decadal/interdecadal time scales very challenging, especially on regional scales.