Role of Natural Climate Variability in Regional Climate Change and its Application to Water Resources

Abstract

We have proposed a new methodology to reduce uncertainties in regional climate hindcasts and projections. The fundamental premise of this work is observed regional climate trends e.g. temperature trend is a combination of two major components: (1) natural climate variability signal, and (2) global climate change signal in the region concerned. The new methodology combines regional natural (low frequency) climate variability with the global change signal from Coupled Model Inter-comparison projects phase 5 (CMIP5) climate models. While most parts of the World have shown a warming trend during the 20 th century, the eastern United States has shown a cooling trend or lack of warming trend particularly after 1930’s. Using this anomalous temperature trends’ example, we have first identified relevant natural climate variability signal. We found that the North Atlantic-Multi-decadal oscillation (AMO) explains approximately 80% variance in observed trend variability in the Eastern United States. We employed AMO as natural climate variability signal and CMIP5 median trend as global climate change signal in the study. We found the proposed methodology provides a better model than all individual CMIP5 models, and the natural climate variability only model for the 20 th century temperature trends hindcasts in the Eastern United States. For 21 st century projections, we developed a statistical model for AMO variability (65-to 70 years oscillation cycle) using past 155 years of observations. We relied on AMO statistical model because: (1) CMIP5 models have limited AMO simulation skill, and (2) it is rather difficult to extract AMO signal from CMIP5 simulations. We combined AMO simulations with the CMIP5 median trend for the 21 st century temperature trends projections in the Eastern United States. The proposed new methodology shows significant potential in reducing uncertainties in the Eastern United States temperature trends’ projections.