PREDICTION OF WATER HYACINTH COVERAGE ON THE HARTBEESPOORT DAM

• Claudia Camacho De Gouveia School of Mechanical, Industrial and Aeronautical Engineering, University of Witwatersrand, South Africa https://orcid.org/0000-0001-9176-7677
• Jacoba Buhrmann School of Industrial Engineering, North-West University, South Africa https://orcid.org/0000-0003-0657-9933

DOI:

https://doi.org/10.7166/35-3-3090

Abstract

Water hyacinth is an invasive weed that contributes to the Hartbeespoort Dam’s poor water quality. Although biological control is the most effective and sustainable method of controlling water hyacinth, a prediction model to plan the biological controls is essential for successful intervention. The literature shows that mathematical models and remote sensing have been used successfully in the past to estimate plant growth rates in similar applications. This study presents various machine-learning models that were investigated to predict water hyacinth coverage.

The complex relationships of water hyacinth growth were simplified to focus on the most influential factors: temperature and nutrients. Missing data were imputed using the multiple k-nearest neighbours imputation. The nutrient datasets were extrapolated to the correct timeline using Monte Carlo simulations and seasonal patterns. Ensemble learning, decision trees, artificial neural networks, and support vector machine models were developed, with ensemble learning (bag algorithm) resulting in the best predictions.