How sensitive is the North American Monsoon to land surface characteristics? An examination of land surface-atmosphere interactions using a mesoscale model for the 2004 North American Monsoon season

Initiation and maintenance of convection has been shown to be sensitive to land surface properties such as soil moisture, soil temperature, topography, and vegetation. A hotspot of land surface-atmosphere coupling is the semiarid region of western Mexico and the southwestern United States, which is the domain of the North American Monsoon (NAM). Many studies have observed a pattern of diurnal convection within the NAM and intensive modeling efforts are underway to improve the simulation the diurnal cycle of convection in global and regional models.

The current study examines the diurnal cycle on a seasonal basis, isolating the relative importance of land surface characteristics on local circulations within the NAM by running contrasting high-resolution, convection-allowing Weather Research and Forecasting model (WRF) simulations side-by-side. Two WRF simulations were performed over the period of July 1, 2004 through August 15, 2004, varying only the factors of soil moisture and soil temperature. Land surface conditions for WRF initialization were given by the enhanced North American Regional Reanalysis (NARR) or a 2-year spinup using the Noah Land Surface Model (Noah LSM) within the Land Information System (LIS). Initial soil moisture and soil temperature differences were found to impact the partitioning of latent and sensible heat, low-level atmospheric moisture and temperature, strength of the monsoon circulation, and distribution and intensity of precipitation, though differences between WRF simulations diminished with model run-time. To determine which land surface initialization is more representative of observed states, initialization and model output will be compared against existing surface-based observations, satellite and ground-based radar precipitation estimates, and other datasets such as the NCEP/NCAR reanalysis and GLDAS. We also plan to explore possible explanations for observed NARR overestimates of soil moisture over the coastal plain of western Mexico.