MODELING OF TURBIDITY AND WATER CLARITY BASED ON REMOTE SENSING SATELLITE LANDSAT IN SAGULING RESERVOIR, WEST JAVA

• Main Author: Ritka May, Annisa
• Format: Theses
• Language: Indonesia
• Subjects: Teknik saniter dan perkotaan; teknik perlindungan lingkungan
• Online Access: https://digilib.itb.ac.id/gdl/view/80331
• Institution: Institut Teknologi Bandung

Indonesia faced several challenges regarding water quality such as water exploitation and contamination caused by human activities. A comprehensive and sustainable water management is required to ensure the availability for the society. Ecosystem quality monitoring is needed to make sure the availability of water resource all year round by using modelling to assist. This paper presented an application of Artificial Neural Network (ANN) method using multilayer perception model with a backpropagation algorithm to predict water clarity, turbidity and TSS in Saguling Reservoir provided by PT Indonesia Power throughout 2013 until 2022. Evaluation of regression analysis (R2 ), Mean Absolute Error (MAE) and Mean Square Error (MSE) are applied to determine ANN performance of predicting water quality level. Based on the results obtained, the transparency data training model has R2 = 0.63 with the resulting model error rate MAE = 0.267 and MSE = 0.239. Transparency data testing has R2 = 0.79 with the resulting model error rate MAE = 0.154 and MSE = 0.165. For the training data turbidity parameter, it has an R2 of 0.8 with an MAE value of 0.149 and an MSE of 0.124. The testing data turbidity parameters have R2 = 0.99, MAE is 0.0007 and MSE is 0.0009. ANN prediction model is expected to be able to demonstrate a good prediction capability when the parameter level changed. Thus, the model developed is proved to be an efficient tool in classifying the water quality level and is beneficial to Saguling Reservoir quantity and quality integrated maintenance.