Evaluation of Eta Model Rainfall Forecasts at São Francisco River Basin in Minas Gerais, Brazil

Raphael Moreira Gomes Vieira, Claudine Pereira Dereczynski, Sin Chan Chou, Jorge Luis Gomes, Arthur Chaves de Paiva Neto


In this work, the CPTEC/INPE's Eta model (8 km and 40 km) forecasts, with 72 hours in advance, are evaluate for 25 intense rainfall cases, occurred in the period 2005-2012, at São Francisco River Basin, upstream Três Marias Hydroelectric Power Plant, in Minas Gerais. The objective is to contribute to the improvement of intense rainfall forecasts in the studied area. Evaluation of rainfall forecasts is made objectively and subjectively, comparing them with observational data. Moreover, synoptic fields forecasts for several variables are compared with the same fields of ERA-Interim Reanalysis (ECMWF). The Eta model forecasts for 25 intense rainfall cases were generated initially nesting Eta-40km model to the CPTEC/INPE's Global model (T126L28). Finally, the Eta-8km model is integrated using as initial conditions the NCEP analysis and as lateral boundary conditions the Eta-40km model forecasts, updated every 6 hours. Considering the 25 intense rainfall cases, evaluating objetively the rain, Eta-8km has better (worst) performance compared to Eta-40km, to predict daily moderate to strong (weak) precipitation. Based on the subjective evaluation, it is noted that when CPTEC/INPE's Global model fails in positioning the active meteorological system, both versions of Eta model are affected. In cases where meteorological systems are more pronounced, as in the 25 case, it increases the accuracy of both Eta model versions, with Eta-8km showing better results than Eta-40km.


Evaluation of precipitation; Eta model; Minas Gerais

DOI: https://doi.org/10.11137/2015_2_15_23


  • There are currently no refbacks.

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Indexers and Bibliographic Databases

Social Media

SCImago Journal & Country Rank
22nd percentile
Powered by  Scopus
Google Scholar
 Except where otherwise noted, content on this site is licensed under a Creative Commons Attribution 4.0 International license.