Prediction of land degradation by Machine Learning Methods

Abstract

In this study, three models were used to monitor and predict the GWL and the land degradation index via the IMDPA method. In all models, 70% of the data was applied for training, while 30% of data were employed for testing and validation. Monthly rainfall, TWI index, the distance of the river, Geographic location was the inputs and the level of groundwater was the output of each method. we found that ANN has the highest efficiency, which agrees with other findings. We combined the results of ANN with Ordinary Kriging and produced a groundwater condition map. According to the potential desertification map and groundwater level index, the potential of desertification had become severe since 2002 and was at a rate of 60% of land area, which, due to incorrect land management in 2016, increased to almost 98% of the land surface in the study area. Using ANN, we predicted that around 99% of the area was severely degraded for 2017. We also used latitude and longitude as input variables which improved the model. In addition to the target variable, latitude and longitude play important roles in Ordinary Kriging and decreased the total error of two combined models.

Towards Artificial Intelligence in Sustainable Environmental Development

“…The application of machine learning and Artificial Neural Networks (ANN) in different environmental components is outlined below. Commonly, machine learning in ANN's form is utilized for problem evaluation and resolution in a variety of environmental sustainability components, such as plastic pollution [14], soil [15], and water quality [16], waste management [17], land degradation [18], and agriculture [19]. These issues may fall under larger SDG categories, such as "life on land" and "life below water", but they affect various social and economic sectors, including partnership objectives, sustainable cities, infrastructure development, poverty, and hunger [20,21].…”