The Role of Resolution in Characterizing the North American Monsoon Variability

Abstract:

The North American Monsoon (NAM) is characterized by a seasonal maximum in precipitation over the SouthWestern US. The combination of extreme precipitation events and significant year-to-year variability makes NAM difficult to simulate in climate models, increasing the challenges for predicting future changes and employing robust mitigation efforts. Given the complex terrain of the NAM region and the localized nature of extreme events, we postulate that resolution may play a role in the reliability of any climate change signal. We use the National Center for Atmospheric Research (NCAR)’s Community Earth System Model 1 (CESM1) Large Ensemble (LENS) for both historical (1920-2005) and RCP8.5 future (2006-2100) climate projections. The study examines baseline LENS members at 100 km (1 deg, LR) resolution and a companion subset of high resolution simulations at 25 km (0.25 deg, HR) atmosphere.

Both resolutions are in agreement that global temperature will increase by around 4.5K by 2100. However, the resolution captured varying predictable increases. During summertime, the NAM region is projected to experience a 3.0% decrease in the LR simulations and 2.9% increase in the HR simulations. During the wintertime, however, the NAM region is projected to experience a 0.8% decrease in LR and 2.3% increase in HR. Given the large interdecadal, internal variability in precipitation over the NAM region in the summertime, LENS provides a more reliable future climate signal. The study finds that the HR signal is distinctly different in the future, where global warming and reduced precipitation is more pronounced compared to that of LR. We analyze the role of resolution in NAM, but contrast this with the resolution dependence in other seasons and climate regimes such as the Pacific NorthWest. There can clearly be fundamental differences when using higher resolution and we explore the tradeoffs between higher computational costs and improved signal reliability