Application of the Simple Verification Method to Estimate the Weather at Makassar Maritime Station, Indonesia

Abstract

Verification is used to measure the quality of a weather prediction, improve process performance, and measure the value of weather estimation. Initially, weather verification developed after Finley published his paper on the verification of tornado events. The type of data, objectives, and scale can make a different method in using weather verification. If there are some parameters that can be predicted, a simple question is consequently often asked by the public: how accurate are weather forecasts? Nowadays, the public wants a simple answer in 1 value that is presented quantitatively. The aim of the research is to develop a simple method that can answer the accuracy of weather prediction in a value that is easily understood by the public. Practically, validation comparing between prediction and observation parameters is divided into 2, namely dichotomous and comparing the values. This research tries to combine all weather prediction variables into a dichotomous variable with a threshold. Moreover, this technique is tested on weather predictions for the port of Makassar over a year. The results show that a certain threshold can be used to change the weather variable to be dichotomous. With the application of this method, forecast accuracy and suitability between the predicted parameters can be obtained. Moreover, the weather forecast issued by the Makassar Maritime Station shows the average true value of the forecast to be 69.1 %, and then the capabilities vary by forecasters, which range from 61 to 79 %. HIGHLIGHTS Weather forecast verification is used to measure the quality of a weather forecast, improve process performance, and the value of weather forecasts The character of the weather variables and their predictions is unique and influences the type of evaluation method To facilitate the public's assessment of the accuracy of weather predictions, it is necessary to combine weather prediction evaluation methods in one value Using the tolerance threshold whether a deviant prediction is used to combine various weather predictive variables Average true value of the forecast is 69.1 % and the different capabilities of each forecaster, which range from 61 to 79 % GRAPHICAL ABSTRACT