GROUNDWATER LEVEL FORECASTING MODEL IN TROPICAL PEATLAND USING ARTIFICIAL NEURAL NETWORK
Imam Suprayogi, Ari Sandhyavitri, Nurdin Anon, Joleha Anon, Wawan Anon, Azmeri Anon · INTERNATIONAL JOURNAL OF CIVIL ENGINEERING AND TECHNOLOGY (IJCIET) · 2020
One of the key parameters in peat land management is water, which is expressed in the water level of peat land.The water level fluctuations of a peat land are closely related to the decomposition of the peat constituent material, its cover and hydrological conditions.When the water level drops, the peat decomposition increases and will release the carbon into the atmosphere.In addition, the condition of peat will be dry so that the area becomes prone to fire.The main purpose of the research is to develop a groundwater level forecasting model to monitor the dynamics of land water fluctuations in tropical peatland in order to comply with government regulation No. 57 of 2016 on the protection and management of peat ecosystems, especially the necessity to maintain water level at a rate of 40 cm.The method of research approach used is using ANN as one branch of soft computing.The location of research on Imam Suprayogi, Ari Sandhyavitri, Nurdin, Joleha, Wawan, Azmeri http://iaeme.com/Home/journal/IJCIET69 [email protected] of PT Meskom Pulau Bengkalis, Riau province.Data that is used to build a model of groundwater level forecasting in tropical peatland sourced from Hobo Water Logger measuring device that record water table in 2014.The main results of the research proved that the implementation of a groundwater level forecasting model on the tropical peatlands in Bengkalis using the ANN method approach for the next day (t + 24) has a very strong classification tested using statistical parameters coefficient of correlation (R) and Mean Square Error (MSE) respectively 0.995929 and 0.0003026 so that the model can be applied on tropical peatlands.