Dynamical downscaling of seasonal climate prediction over northern South America with NCEP's Regional Spectral Model at IRI

Dynamical downscaling of seasonal climate prediction over northern

South America with NCEP's Regional Spectral Model at IRI.

 

Paulo Nobre^*, Antonio D. Moura, Liqiang Sun

 

International Research Institute for climate prediction

Lamont-Doherty Earth Observatory of Columbia University

P.O.Box 1000, Palisades, NY, 10964, USA

 

 ABSTRACT

This study presents an evaluation of seasonal climate forecasts done with the IRI's dynamical forecast system (regional model nested into an AGCM) for the northern South America region for 1998 and 1999 during the rainy period of February to April. An ensemble of three realizations was done for the 1999 experiment, while one integration was done for an earlier forecast experiment in 1998. The model suite consists of the hydrostatic version of the Regional Spectral Model (RSM) from NCEP and implemented at the IRI with 80 km resolution, nested on the ECHAM-3 atmospheric general circulation model (AGCM) with horizontal spectral resolution T42 and 18 levels. Lateral boundary conditions and the mean state in the interior of the RSM integration domain are given at 6 hours intervals, and are extracted from the AGCM forecasts. Lower boundary conditions for the AGCM runs are forecast SSTs over the Tropical oceans during 1999 and a blend of forecast (Pacific) and persisted SST for the 1998 experiment.

The rainfall forecasts generated by the regional model are compared with those of the AGCM and with observations. For the cases analyzed, it is shown that the regional model can improve the rainfall forecast over some areas, noticeably over the Brazilian Nordeste, but it is generally dependent of the AGCM forecast skill. The position and width of the Intertropical Convergence Zone (ITCZ) over the Tropical Atlantic is better captured by the RSM than by the AGCM, resulting in a better rainfall forecast over the Nordeste done by the RSM. It is shown also that the statistics of synoptic scale weather variability on seasonal time scales can be forecast with the RSM model over the Nordeste. Diagrams of forecast and observed daily rainfall shows that the RSM ensemble forecast captured the onset of the 1999 Nordeste rainy season reasonably well.

On the other hand, the RSM forecasts generated spurious rainfall anomalies along major mountain barriers (e.g. along the Andes) and along the East coast of South America. The RSM has also shown a tendency for a dry bias over part of the Amazon region and the Brazilian central plateau.

 

* On leave of absence from Centro de Previsão de Tempo e Estudos Climáticos - CPTEC/INPE. Cachoeira Paulista, São Paulo 12630-000, Brazil.