Using ARIMA and ETS models for forecasting water level changes for sustainable environmental management

Area of Science:

• Hydrology
• Environmental Science
• Data Science

Background:

• Accurate hydrological forecasts are essential for water management, flood control, and drought mitigation.
• Predicting river water levels is crucial for sustainable environmental management and public safety.

Purpose of the Study:

• To evaluate the effectiveness of Autoregressive Integrated Moving Average (ARIMA) and Exponential Smoothing (ETS) models for river water level forecasting.
• To identify the most suitable model for hydrological time series prediction using historical data.
• To provide insights for water resource management and hazard control in the Morava e Binçës river basin.

Main Methods:

• Utilized nine years of monthly water level data (2014-2022) from the Vitia station on the Morava e Binçës river.
• Applied Autoregressive Integrated Moving Average (ARIMA) and Error, Trend, and Seasonality (ETS) models for time series forecasting.
• Validated model performance using Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) metrics via the R package.

Main Results:

• Both ARIMA and ETS models demonstrated applicability in predicting river water levels.
• The models successfully identified distinct periods of high and low water levels for 2022-2024.
• Model performance was quantitatively assessed using RMSE and MAE, indicating their predictive accuracy.

Conclusions:

• ARIMA and ETS models are effective tools for hydrological forecasting of river water levels.
• Accurate water level predictions support crucial water resource management and flood hazard control strategies.
• The findings are vital for areas prone to frequent flood events, like the studied region.