Weather Forecast Based on Daily Temperature Data Using Seasonal Auto-regressive Integrated Moving Average (SARIMA) Method: Surabaya – BMKG Juanda

Abstract

Indonesia is a country that has two seasons in one period, namely the dry season and the rainy season. Problems often arise in the field of climatology, is the presence of missing data, which affects the quality of the prediction results. This study uses time-series results, which show the characteristics of statistical values ??with observed data and behave according to the probabilistic concept. In this research case study, we took data from Kaggle which contained Global Climate Change data for the Surabaya area from 1900 to 2012. In addition, we took BMKG online data from 2013 to 21 April 2021 based on observations from the Juanda Meteorological station for the Surabaya area, and surrounding. Specifically for this research, we focus on the rainfall temperature which is modeled using the ARIMA Seasonal model, to predict the weather in the Surabaya area and its surroundings. Therefore, we want to analyze the weather prediction using the ARIMA Seasonal time series model for a case study in the Surabaya area. It can be simplified that the temperature prediction in the Surabaya and surrounding areas has an accurate value with the selection of the best model SARIMA (3,0,0) (0,1,1,12).