Forecasting Sea Surface Temperature in the Kiribati Region

A slight change in sea surface temperature (SST) is a critical condition as too high of SST could cause coral bleaching, which could results in the declining numbers of fish individuals and species on coral reefs. The people of Kiribati, one of the most affected countries by climate change, depend on the reef and ocean for food and economics, thus advance knowledge of SST in the region could benefit the country. In this paper, a multiple linear regression (MLR) is developed for forecasting the sea surface temperature anomaly (SSTA) of the Kiribati Region (7^0 N-15^0S, 150^0 W-170^0 E) using the Sea Level Pressure Anomaly (SLPA), Air Temperature Anomaly (ATA), Total Cloudiness Anomaly (TCA), Relative Humidity Anomaly (RHA), Wind Eastward component Anomaly (WECA), Wind Northward component Anomaly (WNCA) and Wind Scalar Anomaly (WSA) as predictors. We validate the proposed model and determine which predictors should we include, and to what extend does this model predict the SSTA in the Kiribati region. The proposed model is compared with the Naïve Method by various error functions such as the Root Square Mean Error (RSME), Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE). We found that ATA, TCA, WNCA, WECA and WSA are the best predictor variables in the forecasting model, which satisfy all the MLR assumptions and performing better than the Naïve method, hence, it may be useful for forecasting SST accurately.