Relative contributions of global warming, AMO and IPO to the land precipitation variabilities since 1930s

Abstract

The relative contributions of ocean modes to the JJA and DJF land precipitation variabilities during 1934–2015 are investigated using a variety of statistical and dynamical system methods, i.e., singular value decomposition (SVD), multivariate linear regression, and information flow analysis. Through SVD analysis for the tropical land precipitation and sea surface temperature (SST), three ocean modes are found to most affect the trend and interdecadal variation of the land precipitation. They are the global warming (GW) mode, Atlantic Multidecadal Oscillation (AMO) and Interdecadal Pacific Oscillation (IPO). GW contributes dominantly to the tropical land rainfall variability in both the JJA and DJF seasons. In JJA (DJF), AMO (IPO) plays a role only secondary to GW. Locally, within the thin latitude bands 10° S–10° N, 50° N–60° N and 40° S–50° S, GW, AMO and IPO are of equal importance in JJA; outside these bands, in the same season the first two dominate. In the band 10° N–40° N, IPO is the primary contributor in DJF, but outside it, GW dominates. Also, these contributions differ geographically from continent to continent. These results have been substantiated in the application of information flow analysis, a recently developed method in physics for the inference of causality between dynamical events. In terms of information flow, we have presented the regions of sensitivity to the three modes. Also presented are a number of ECHAM model experiments, which, besides verifying the above results, show for the first time that the Indian Ocean is pivotal in having AMO and IPO in effect in causing the precipitation variabilities.