Model Regresi Untuk Prediksi Muka Air Tanah Di Lahan Gambut

Abstract

Peatland fires in Indonesia occur almost every year with an ever-increasing area. Factors affecting peat drought and fire are the high and low ground water level of the peat, for this reason it is necessary to review how high the peat water level can cause a fire so that it can be used as an early warning before peatland fires occur. This study created a hydrological model using regression analysis to predict groundwater levels in peatlands which were then verified using correlation coefficients. The data used were obtained from the SESAME tool namely rainfall data and water surface data recorded in time series per 10 minutes. The hydrological model is carried out with several variations of data usage periods to get the best value. The results showed the use of data for the period July 2017 - December 2017 got the best correlation coefficient value of 0.86, while the use of data for the period February 2018 - July 2018 obtained the lowest correlation coefficient value of 0.15.
Key words: Fires, hydrological models, regression analysis, peatland