Improving predictions of snow resources using midlatitude SSTs with convergent cross mapping

Abstract Effective water resource management in the western United States (WUS) is possible only with accurate monitoring and forecasting of seasonal snowpacks. Seasonal snowpack, a major water source for the WUS, is declining due to anthropogenic climate change. Overprinted on this trend is year-to-year variance in snowpack extent and mass due to influences from teleconnections related to the El Niño Southern Oscillation and the Pacific Decadal Oscillation. Recently in the 2015 and 2016 winters, extreme droughts in the coastal WUS, mainly the Pacific Northwest (PNW) states of Washington and Oregon were linked with anomalously warm sea surface temperatures (SST) in northeastern Pacific Ocean. Here, we use convergent cross maps (CCMs) to analyze time series of SSTs and snow water equivalent (SWE) in the PNW. For some ecoregions, we show that extratropical SSTs may have a stronger influence on snowfall and snow accumulation in the PNW compared to tropical indices of climatic variability. Cold (warm) SSTs in the northeast Pacific lead to high (low) snow years. CCMs also performed better in recreating SWE anomalies compared to linear regressions with lagged predictor variables. Accounting for the influence of SSTs may help water resource managers to better predict and prepare for extreme snow events in the future.