Metodologia para análise de desempenho de simulações de sistemas convectivos na região metropolitana de São Paulo com o modelo ARPS: sensibilidade a variações com os esquemas de advecção e assimilação de dados

Abstract

RESUMOTempestades seguidas de enchentes e alagamentos em pontos da Região Metropolitana de São Paulo são eventos recorrentes nas estações chuvosas. O desempenho do modelo ARPS nas simulações numéricas de alta resolução espacial, para o evento de 04 de fevereiro de 2004, é avaliado por meio de erros estatísticos e índices de acurácia, com base no confronto entre a distribuição espacial da precipitação acumulada simulada e a estimada com o radar meteorológico de São Paulo. Os resultados quantificam a influência dos termos de advecção das equações primitivas nos erros de fase dos sistemas convectivos simulados, para o tipo de configuração atmosférica dominante no caso estudado. A assimilação incremental de medidas obtidas em estações de superfície durante a execução do modelo, corrige significativamente a fase das células convectivas profundas na região de influência das medidas, a qual é definida durante o processo de análise objetiva. Um novo índice de destreza para modelos de mesoescala é proposto (D PIELKE ). As análises preliminares para o desempenho deste índice indicam adequação de sua formulação para o tipo de simulação efetuada. Palavras-chave: tempestades, simulações numéricas de mesoescala, erros de fase e amplitude, índices de acurácia. ABSTRACT: METODOLOGY FOR PERFORMANCE ANALYSIS OF SIMULATIONS OF CONVECTIVE SYSTEMS IN THE METROPOLITAN AREA OF SÃO PAULO WITH THE ARPS MODEL: SENSITIVITY TO VARIATIONS WITH THE ADVECTION AND THE DATA ASSIMILATION SCHEMESThunderstorms followed by flash flood events are recurrent in the Metropolitan Area of São Paulo during the wet seasons. The performance of several numerical simulations of the February 04, 2004 event carried out with the ARPS system in high horizontal resolution is evaluated through statistical errors and indices of accuracy. The evaluation is based on direct confrontation of the accumulated precipitation fields simulated by the model and measured by the São Paulo weather radar. The results quantify the influence of the advective terms of the primitive equations on phase errors of the simulated convective systems. The incremental assimilation of surface station data done during the model execution improved its performance in the area limited by the influence radius, which is defined during the objective analysis process. A new skill index formulation is presented (D PIELKE ). The preliminary results indicate that the D PIELKE index is adequate for assessing the simulation skill of small-scale convective systems.

Implementation of Observed Sky-View Factor in a Mesoscale Model for Sensitivity Studies of the Urban Meteorology

“…Table 3 lists all surface stations used to evaluate the model. Temperature and specific humidity taken at 2 m agl were used for sake of validation, using root mean square error (RMSE), bias [45] and Pielke index (D Pielke ) [46], which considers the Pielke model skill definition [47] as presented in Equation (4):…”

Contributions of mobile, stationary and biogenic sources to air pollution in the Amazon rainforest: a numerical study with the WRF-Chem model

“…1). The analysis of meteorological variables has been conducted based on Pielke's skill score (S pielke ) (Pielke, 2002;Hallak and Perreira Filho, 2011), Pearson's correlation coefficient (r), and mean bias (MB). For the air pollutants, in addition to r and MB, two statistical indexes were used, the mean normalized bias error (MNBE) and the mean normalized gross error (MNGE).…”

The effects of an osmoregulator, carbohydrates and polyol on maturation and germination of ‘Golden THB’ papaya somatic embryos

“…The root-mean-square error (RMSE) informs the actual value of the error produced by the model (Souza et al 2011). It usually expresses the accuracy of numerical results with the advantage that RMSE has errors in the same dimensions of the analyzed variable (Hallak and Pereira Filho 2011). The lower this measure, the better the model performance in making estimates (Souza and Escobedo 2013).…”