Abstract
The aim of this study is to evaluate the performance of simulations generated by the ensemble of the Eta-5km model from the National Institute for Space Research. The evaluation was performed over the Paraíba do Sul River Basin during an extreme rainfall event, occurred early in January 2000, in a South Atlantic Convergence Zone event. The 5-member ensemble is formed with different convective (Betts-Miller-Jancic - BMJ or Kain-Fritsch - KF) and microphysics (Ferrier or Zhao) parameterizations and are nominated respectively as: CNTRL (Betts-Miller-Janjic/Ferrier); KF (Kain-Fritsch/Ferrier); KFMX (Kain-Fritsch with inclusion of momentum flux/Ferrier); KFMXP (Kain-Fritsch with inclusion of momentum flux and force of the perturbation of pressure gradient/Ferrier) and ZHAO (Betts-Miller-Janjic/Zhao). The simulations are integrated up to 120 h and the initial and boundary conditions were taken from the Climate Forecast System Reanalysis. The results indicate that all members were able to position the rain band in about the same observed location. Although the model underestimates the highest daily precipitation amount, it simulates higher frequency of heavy rainfall events than it really occurs. In general, the ensemble mean simulation presents higher performance than each individual members and the 48-h simulations are better than the other forecast lead times.
Keywords:
simulations; rainfall; Eta model; high resolution; convective and microphysics parameterizations
