Viable forecasting monthly weather data using time series methods

• Published in: Ecological questions 2023-01, Vol.34 (1), p.1-16
• Main Authors: Vambol, Sergij, Soomro, Ramzan, Ghauri, Saghir Pervaiz, Marri, Azhar Ali, Dung, Hoang Thi, Manzoor, Nazish, Bano, Shella, Shahid, Sana, Asadullah, Farooq, Ahmed, Lutsenko, Yurii
• Format: Article
• Language: English
• ISSN: 1644-7298
• DOI: 10.12775/EQ.2023.003

The main object of the research was to assess the forecast values of the weather parameters by using three-time series methods such as Decomposition of time series, Autoregressive (AR) model with seasonal dummies and Autoregressive moving average (ARMA) /Autoregressive Integrated moving average (ARIMA) model. A recent phenomenon in weather changing has disturbed the world in general and Pakistan in particular. In Pakistan due to climate change, flood and heat stroke have taken many lives. Stationarity was measured through the Augmented Dickey-Fuller test; results showed that some variables are I(0) and some are I(1). The reliability of the forecast results was examined through the goodness of fit test. For finding the best fit model, the performance measures of various models: Root Mean Squire Error, Mean Absolute Error and Mean Absolute Percentage Error were considered. The model in which the above statistics are the minimum was chosen as the appropriate model. After model analysis and validation, it was observed that AR-model with seasonal dummies was found to be the best fit model between the three models. Meanwhile, the forecasting for the period Jan.2018 to Dec.2018 was made based on the best fit model. Given the future forecasting results, the temperature will be normal at selected stations. The wind and rainfall will also be present. Overall, it was suggested that the obtained findings of meteorological stations' weather might be normal for the coming few months over there, and no chance of heatstroke and flood might be expected. Future studies must be carried out to provide the awareness to well-being regarding ecological hazardous to minimize their economic loss through mass media.