Numerical Simulation of the 2004 North American Monsoon sensitivity to surface data

We use global ensemble simulations of the 2004 warm season to study the North American monsoon, and its sensitivity to surface conditions and data. A control simulation will be performed, using NCEP analysis initial conditions in the atmosphere. The land initial conditions are spun up from a pure model simulation, with prescribed time dependent Reynolds SST. The spatial resolution of the model is ¼ degree, so the Baja Peninsula and Gulf of California are apparent in the land-sea mask. For this initial presentation, we will evaluate the control ensemble seasonal simulations, to ascertain the model's ability to represent both the large-scale environment and the local circulations around the monsoon. We will compare the simulations with observations and the North American Regional Reanalysis (NARR). Data from the North American Monsoon Experiment (NAME) will also provide a crucial benchmark for the simulations. The control simulation will be the baseline by which we test the impact of surface observations on the monsoon circulation. Global Land Data Assimilation System (GLDAS) data integrates numerous conventional and remotely sensed observations of the land surface with model fields. A second experiment has been run where GLDAS provides prescribed land states in the seasonal simulation to assess the impact of improved land conditions on the ensemble simulation.