Jose L. Hernandez hernandezfj@ornl.gov

Carlos Perez cperez@snet.gob.sv
David Erickson IIIericksondj@ornl.gov
Bob Oglesby Robert.J.Oglesby@nasa.gov
 

ABSTRACT

Diagnose is a fundamental step in any kind of modeling studies. At finer time and spatial scales in geosciences, it is a challenge to have good model performance since we need to consider local interactions that impose their variability on signals from larger neighbor areas and longer periods. Ultimately, an observed signal of any climate parameter is a combination of a wide spectrum of space and time variability. We focus our attention on the MM5 regional model and its performance to describe the regional climate in Central America. Several test simulations were carried out to evaluate this model. In the first test ( experiment1_part_1 and experiment1_part_2 ), at 3 Km and 9 Km, MM5 can describe with very good performance the diurnal cycle of temperature and wind, however the model underestimate precipitation, yielding a 10% of observed amounts. In a second test , two configurations were used under 3 Km, 9 Km and 27 Km. The physics options considered were: experiment 2 (IMPHYS=4; ICUPA=3); and experiment 3 (IMPHYS=6; ICUPA=6). In the experiment 3, MM5 shows improvements in precipitation and still present excellent agreement with observations for temperature and wind. Here, coastal areas show a diurnal change in the temperature gradient and display well-defined land and sea breeze effects. The experiments are summarized in three presentations by Carlos Perez, Jose Hernandez and David Erickson in the Regional Climate Modeling Workshop for Central America, October 18-20, 2005, in the Republic of Panama. This workshop is part of the annual activities from SERVIR, a project primarily funded by NASA and the United States Agency for International Development (USAID).