COMPARATIVE ANALYSIS OF DISCHARGE FORCASTING METHODS USING F. J. MOCK, MARKOV, ARIMA, AND ARTIFICIAL NEURAL NETWORKS (CASE STUDY: GUNG WATERSHED, PEMALI-COMAL RIVER BASIN, CENTRAL JAVA)

• Main Author: Aulia Oktarini, Ahya
• Format: Theses
• Language: Indonesia
• Subjects: Teknik saniter dan perkotaan; teknik perlindungan lingkungan
• Online Access: https://digilib.itb.ac.id/gdl/view/81706
• Institution: Institut Teknologi Bandung

Water availability is a critical need that continues to grow as the population increases. Challenges related to water availability and potential disasters due to suboptimal water infrastructure management are significant concerns. Gung watershed, part of the Pemali-Comal River Basin (RB), is one of the primary water sources in Tegal Regency, facing water scarcity and flooding due to land cover degradation. Additionally, no prior research has analyzed discharge forecasting in this region. This study aims to identify the best model for managing water resources in Gung watershed through a comparative analysis of various discharge forecasting models. The study investigates key hydrological components, including rainfall patterns and discharge, and performs a comparative analysis of several discharge forecasting models in Gung watershed. The models compared in this study include F. J. Mock, Markov A & B, ARIMA, and Artificial Neural Networks (ANN). These models are validated using statistical analysis of correlation coefficients (r), root mean square error (RMSE), and Nash-Sutcliffe efficiency (NSE) with actual discharge data. Secondary data collection in this study encompasses hydrological and climatological data from four rainfall gauge stations (Stations P1 Dukuhwringin, P2 Procot, P3 Pagongan, and P4 Slerok), one discharege recording station (Q1 Pesayangan), and one climate monitoring station around Gung watershed provided by Pemali-Comal River Basin Water Resources Management Agency (BPSDA) and Meteorological, Climatological, and Geophysical Agency (BMKG) Maritime Tegal. This study consists of two interconnected phases. The first phase involves identifying trends in rainfall patterns, discharge, and climatic variation in Gung watershed. The second phase includes a comparative analysis of five water source models that are statistically tested to identify the best model for discharge forecasting in Gung watershed.Hydroclimatological trend analysis in Gung watershed uses 20 years of historical data (2003–2022) to study continuity and consistency of rainfall, discharge, and climate data. The methods involved include filing nonexisting data, consistency testing, homogeneity testing, calculation of regional precipitation averages, and determination of climate type of Gung watershed. The complete discharge dataset enables immediate analysis of trends and modeling. Climate data from BMKG Maritime Tegal encompasses temperature, humidity, wind speed, and sunlight duration over 20 years (2003–2022). Gung watershed exhibits a monsoon climate with a peak rainy season in February and a dry season in August. This rainfall trend aligns with discharge patterns, with the highest discharge occurring in the rainy season and the lowest in the dry season. Climatological trends indicate a significant increase in temperature throughout the year, while humidity and wind speed have seen slight increases, leading to increased dry discharge extremes throughout the year. Validation of water source models in Gung watershed considered statistical measures of r, RMSE, and NSE against actual discharge data. The Markov B model (5 classes) achieved the best results with the highest correlation coefficient (r = 0.9331), the lowest RMSE (5,2555), and the highest NSE (0.8593). It was followed by the ARIMA with r = 0.8214, RMSE =8,0915, and NSE = 0.6665; ANN model with r = 0.8034, RMSE =8,3593, and NSE = 0.6440; Markov A with r = 0,8334, RMSE = 8,8073, and NSE = 0,6049, and F. J. Mock model with r = 0.7906, RMSE =12,2695, and NSE = 0.5281. This study provides information on the dynamics of discharge and hydroclimatology patterns around Gung watershed. It also serves as a reference for water resource managments in Gung watershed for optimal water resource planning, given many challenges in the region, primarily landslides, flooding, and land-use changes. The accuracy of discharge forecasting is crucial in disaster risk mitigation and meeting water needs. This study contributes significantly, as there has been no prior research on discharge forecasting in Gung watershed.