Assessing the Impact of Climate Change on Volume of Water Consumed by Poor Urban Households Using Geospatial Technology and Machine Learning Models

Abstract

The United Nations declares that climate change is primarily a water crisis. The impact is felt more by poor urban households who have to spend more time and scarce resources to access clean water. This study assesses the impact of climate change temperature variability on volume of water consumed by poor urban households. Spatial and seasonal variations of land surface temperature as a result of climate change was mapped in relation to the location of poor urban households using geospatial technology. The mapped pattern was fed into four machine learning models for modeling and prediction. The machine learning models predicted a difference of 0.1 liter between volume of water consumed by poor urban households with and without climate change data. Significance test shows that there is no significant difference in model performance when climate data was included and when climate data was not included in the modelling. In addition, weak correlation of 0.04 was found between LST and volume of water consumed, which conforms with the fact that there is no significant difference in volume of water consumed with or without temperature variability. That is, climate change temperature variability does not influence the volume of water consumed by poor urban households in the study area.