Forecasting Rainfall and Temperature Trends in Saint Lucia Using Nonparametric Statistics and Soft Computing Techniques

Abstract

Climate change and its adverse environmental impacts are major concerns for Small Island Developing States (SIDS), such as Saint Lucia. In particular, the annual atmospheric temperature and rainfall continue to receive a great deal of attention in the Caribbean and worldwide. Subsequently, looking at the spatiotemporal elements of meteorological factors with regard to the evolving environment, especially in countries where tourism and agriculture activities are dominating, is essential to evaluate climate prompted changes and propose feasible adaptation techniques. The present study attempts to examine and provide findings on the long-term deviations and fluctuation in atmospheric temperature and rainfall over the island of St. Lucia using data for the last three decades . The problem was analyzed using soft computing techniques of Gene Expression Programming (GEP) and a nonparametric statistical trend analysis technique of Mann Kendall (MK) and Sen's slope (Q) tests. The results showed the descriptive capabilities of the GEP output and significance as a prediction tool when compared to MK and Q tests. The results of the detailed analyses indicated that there was no significant trend of rainfall on the island over the last 30 years. However, the GEP analysis indicated that during the next 10 years, the country will experience drought-like instances of low rainfall volumes. Further analysis showed that the island's mean temperature will increase at a rate of 0.032°C per year. Subsequently, due to the island becoming warmer with little to no increase in rainfall, the island would require in-depth planning in the area of energy consumption as well as mitigation measures due to the potential impact of increased storm activities that can be correlated with these climatic variabilities.