In the heart of Johor, Malaysia, the Muar River flows, its water levels a critical factor for local agriculture, industry, and energy production. Predicting these levels with accuracy has long been a challenge, but a recent study published in the journal *Scientific Reports* (translated as *Nature Research: Scientific Reports*) offers a promising solution. Led by Chaitanya Baliram Pande from the Institute of Energy Infrastructure at Universiti Tenaga Nasional, the research introduces a novel approach to river water level prediction that could revolutionize water resource management and energy production.
Pande and his team developed a hybrid modeling approach that combines complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and machine learning algorithms like support vector machine (SVM) and random forest (RF). The goal? To accurately predict the monthly water levels of the Muar River at the Buloh Kasap station from 2014 to 2023.
“The hybrid models outperformed standalone models in predicting river water levels,” Pande explained. “The CEEMDAN-RF model, in particular, showed exceptional performance with the highest coefficient of determination (R2) of 0.94 and the lowest errors in root mean square error (RMSE) and mean square error (MSE).”
So, why does this matter for the energy sector? Accurate water level predictions are crucial for hydropower generation, which relies on consistent water flow. “By improving the prediction of river water levels, we can optimize hydropower generation, ensuring a more reliable and sustainable energy supply,” Pande added.
The study also highlights the importance of data decomposition techniques in understanding complex data sets. By breaking down the data into various sub-frequencies, the CEEMDAN technique allows for a better understanding of trends, seasonality, and fluctuations in river water levels. This understanding can lead to more informed decision-making in water resource management and energy production.
The implications of this research extend beyond the Muar River. As Pande noted, “The CEEMDAN-based novel hybrid modeling is effective for complex fields utilized in the sustainable and optimized utilization of water resources for sustainable development goals (SDGs).” This could pave the way for similar applications in other rivers and regions, ultimately contributing to global efforts in sustainable development.
In the quest for sustainable energy and efficient water resource management, Pande’s research offers a significant step forward. By harnessing the power of artificial intelligence and data decomposition, we can better predict and manage our natural resources, ensuring a more sustainable future for all.