Application of soft computing techniques and Arc-GIS in a holistic sense for the development of future rainfall maps: case study the island of St Lucia

Changes in rainfall patterns and their adverse environmental impacts are major concerns for Small Island Developing States (SIDS) in particular the islands of the Caribbean. In this paper, rainfall data analysis for the island of St. Lucia was performed to identify key hydrological years as well as to investigate the inter-annual behavior of the rainfall variability during the period 1961–2020. Data analysis using the Gene Expression Programming (GEP) soft computing technique was applied to investigate existing rainfall patterns from available rainfall datasets and to make predictions of future annual rainfall volumes for areas on the island. In addition, ArcGIS a geographic information system was used to support the spatial development and descriptive presentation of past, present and future rainfall predictive maps. The detailed analysis indicated that there was no significant decreasing trend of rainfall on the island over the last 3 decades. However, the results of the predictive future rainfall maps indicated that during the next 10 years the country will experience drought-like instances of low rainfall volumes. The results further suggest that the proposed procedure for coupling predictive soft computing learning algorithms and GIS systems can be an effective tool for precipitation monitoring and can be easily transferred to other similar applications. Additionally, the yearly map-to-map outputs give the user a powerful mapping and real-time annual rainfall forecasting tool.