Semestral Journal of Centro Argentino de Meteorólogos, which is published  since 1970 and serves on the Core of Argentine Scientific Journals since 2005. Meteorologica publishes original papers in the field of atmospheric sciences and oceanography.

Registration number of intellectual property: RL-2018-42420861-APN-DNDA#MJ

ISSN 1850-468X

Volume 39 – N°2 MORE NOTES FROM THIS ISSUE

APPLICATION OF MET (MODEL EVALUATION TOOL) FOR THE VERIFICATION OF THE WRF-ARW/SHN-SMN MODEL FORECASTS DURING THE 2011 SPRING

Gisela D. Charó, Estela A. Collini y M. Eugenia Dillon

Servicio Meteorológico Nacional. CABA, Argentina
Servicio de Hidrografía Naval. CABA, Argentina
CONICET. CABA, Argentina
Departamento de Matemática. FCEyN. UBA. CABA, Argentina

Manuscritpt received on 2 July 2013, in final form on 4 December 2013

ABSTRACT

In this article we evaluate the MET v3.1 (Model Evaluation Tool) used for verification of the forecasts of the WRF-ARW (Weather Research and Forecasting – Advanced Research Weather) v 3.1.1 (WRF-ARW/SHN-SMN) during the spring of 2011. Among MET’s tools, we implemented Point-Stat, Stat-Analysis and MODE. We have also studied its mathematical aspects and capabilities. One of the main reasons for the application of the MET is its efficacy in handling upper air data for comparison purposes and precipitation verification object based. PointStat provides verification statistics for forecasts at observation points, meanwhile Stat-Analysis uses Point-Stat’s results to compute statistics in time periods like weeks, months and even years. They were used to compare the WRF-ARWforecasts with the upper air observations made at Córdoba, Resistencia, Mendoza, Ezeiza, and Santa Rosa stations for the spring of 2011. The evaluation was applied to temperature and winds at standard levels. Scatter plots between observations and forecasts of temperature and winds at standard levels were made. In addition the frequency distributions of the forecasts errors were calculated, which were plotted in histograms and boxplots. Finally, a comparison between the means absolute errors from ETA and WRF-ARW models was made for the same period to show the differences between them. In addition, the MODE, a verification tool based on objects, was implemented. It was applied to an intense precipitation event that took place on 24 October of 2011 over large part of the northeast of Argentina. This particular comparison using TRMM data was a first approach to understand MODE success and failures. The MET package includes traditional statistical verification methods, as well as advanced methods using fuzzy logic tools. In this article we show some examples of its usage to evaluate the WRF-ARW SHN-SMN forecasts on experimental basis.

Keywords: VERIFICATION, MET, POINT-STAT, STAT-ANALYSIS, MODE