10:40 〜 11:00
[E2-01] Seasonal ARIMA Prediction of Streamflow: Sobat River Tributary of the White Nile River
キーワード：Floods, Timeseries, Prediction, Forecasting, Streamflow
From 2019 to 2022 South Sudan has witnessed annual floods consecutively from the White Nile River and its tributaries, resulting in flood disasters and dire humanitarian conditions.
Understanding the patterns and behavior of its flow is essential for effective optimization of water resources, and forecasting for flood management which also plays a crucial role in mitigating the risks associated with flood disasters.
In this study, we utilize Seasonal Autoregressive Differencing and Moving Average Model to perform Streamflow forecasting. We use the Seasonal Autoregressive, Differencing, and Moving Average Model (SARIMA) for modeling and forecasting. The data provided consists of stream-flow volume in m3 collected from the Sobat River, which is the main tributary of the White Nile in South Sudan. The data was collected between 1915 and 1944. A Unit Test using the Ad fuller test indicated that the stream-flow data is stationary (p-value is less than α=0.05), with seasonal differencing of D=1. An optimal model of ARIMA(2,0,1)(2,1,0)12 produced a coefficient of determination greater than 0.9.
Understanding the patterns and behavior of its flow is essential for effective optimization of water resources, and forecasting for flood management which also plays a crucial role in mitigating the risks associated with flood disasters.
In this study, we utilize Seasonal Autoregressive Differencing and Moving Average Model to perform Streamflow forecasting. We use the Seasonal Autoregressive, Differencing, and Moving Average Model (SARIMA) for modeling and forecasting. The data provided consists of stream-flow volume in m3 collected from the Sobat River, which is the main tributary of the White Nile in South Sudan. The data was collected between 1915 and 1944. A Unit Test using the Ad fuller test indicated that the stream-flow data is stationary (p-value is less than α=0.05), with seasonal differencing of D=1. An optimal model of ARIMA(2,0,1)(2,1,0)12 produced a coefficient of determination greater than 0.9.