Impact of Soil Moisture Initializations on WRF‐Simulated North American Monsoon System

Abstract

AbstractThis study investigates the influence of soil moisture initialization on the North American Monsoon System (NAMS) using the Weather Research and Forecasting (WRF) model coupled with Noah‐MP Land Surface Model. Five sets of experiments using different soil moisture initializations were conducted from May 1 to October 1 in 2015–2017. The model forecast for each set of experiments consists of five ensemble members. The simulation using North American Regional Reanalysis is the control run (WRF‐Ctrl). Four sets of sensitivity experiments were conducted with extremely wet (WRF‐Wet) and dry (WRF‐dry) initial conditions and using different soil moisture products from the Global Land Data Assimilation System (WRF‐GLDAS) and NASA Soil Moisture Active Passive L4 (WRF‐SMAP). Results show that the WRF model can capture the key features of the NAMS, but it exhibits a wet bias in the higher elevation regions and a dry bias in the lower elevation regions. Our analysis of extremely wet and dry cases reveals that initial soil moisture conditions can affect precipitation through both modulating surface energy partitioning and influencing large‐scale pressure and wind patterns. When different soil moisture products are used to initialize the model, it can lead to substantial changes in spatial pattern of the NAM precipitation. Initializing the model with SMAP and GLDAS can reduce the overestimation of precipitation over high terrain areas. The results indicate that high‐quality soil moisture products have the potential to improve model representation of the NAMS, especially with future improvements in model convective parameterization.