Estimating future bathymetric surface of Kainji Reservoir using Markov Chains and Cellular Automata algorithms

Abstract

The menace of sedimentation to reservoirs has a significant implication for water quality, storage capacity and reservoir lifetime. Rainfall patterns and other anthropogenic and environmental impacts alter the erosion rate and, by extension, directly affect sedimentation rates if left unchecked. This research focused on using the integration of Markov Chains and Cellular Automata (MC – CA) models to estimate and forecast the future bathymetric surface of the Kainji reservoir in Nigeria for the year 2050. The bathymetric datasets used for this research comprise two different epochs (1990 and 2020). The datasets were acquired using a Single Beam Echosounder at Low and High frequencies of 20 kHz and 200 kHz. The preliminary investigation revealed that sedimentation is exacerbating a greater danger to the reservoir functionality. The results show that the maximum observed depth is 71.2 m, indicating a 7.53% loss in depth from the 1990 archived data and a 16.24% depth loss to sedimentation from 1968 to 2020 and 22.35% depth loss in the year 2050 as shown on the projected surface. Consequently, the integrated model (MC and CA) efficiently predicted the future bathymetric surface of the Kainji reservoir for the year 2050 based on the data characteristics. However, the proven techniques for analysing spatial data, such as the Markov Chain and Cellular Automata, best suited for analysing categorical transition data, show some artefacts (black spots) on the projected generated map which is subject to further investigation.