Sensitivity of Weather Forecasts to Soil Moisture Perturbations in Coastal Regions

Land-atmosphere interactions in coastal regions remain a significant challenge to understand and represent in models. In this study we focus on two coastal regions of the United States with different land surface conditions: 1) a moist and vegetated area in the Mid-Atlantic, and 2) an arid and mountainous area in the southwestern United States. The terrestrial coupling index (TCI), for which larger positive values indicate greater sensitivity of surface latent heat flux to soil moisture variability, is calculated for each region and month during 2000–2020 using the 0.1° ERA5-Land dataset. The largest positive values of TCI are found during March and July­–September in the West Coast region and during July­–September in the East Coast region. The sensitivity of weather forecasts to soil moisture perturbations are examined for selected times during the months of largest positive values of TCI for each region using the Coupled Ocean–Atmosphere Mesoscale Prediction System (COAMPS). Specifically, we run pairs of COAMPS simulations out to 120 h at convection permitting horizontal resolutions, in which one simulation includes soil moisture perturbations at forecast initialization and the other simulation does not. Differences in forecast quantities, including surface heat fluxes and lower-tropospheric temperature and humidity, between the simulations in each pair are examined across the regions and adjacent offshore areas. Simulations performed for an offshore low-level flow event in the East Coast region during August 2020 indicate that soil moisture perturbations can have a notable impact on surface fluxes and atmospheric quantities over land and over offshore areas. The simulations for the East Coast region are compared to simulations of an offshore low-level flow event for the West Coast region for a similar time of the year to examine regional variability of the sensitivity of weather forecasts to soil moisture perturbations.