Seasonal Forecasts

A Seasonal Forecast of Swell activity has been developed.
The model developed for forecasting the wave activity is based on finding the relationship between a large-scale predictor and a local predictand. In this case, as the waves are the result of storm situations characterized by anomalous sea level pressure fields and winds, we have used the SLP fields and its gradients (proportional to the wind) to build a tailor-made predictor that includes the information of the travel time of waves in the Pacific.

As the waves in the Pacific have a strong multimodality and are characterized by a number of different swells and a sea concurrent at the same time, instead of predicting the bulk parameters of the waves (Hs, Tp and Dir), we have decided to incorporate the use of the complete wave spectra, in order to have all the possible wave families reaching the study site at any time. Thus, our predictand variable is the daily wave spectra.

Below, an sketch of the general methodology for constructing the model and forecasting seasonal wave spectra in the following 9 months is presented.

A Seasonal Forecast Outlook of Tropical Cyclone Activity has been developed by means of climate-based statistical techniques, for an area covering both Samoa and Tonga. This area is big enough to identify the generation areas as well as local patterns. The model is based on finding the relationship between the sea surface temperature (SST) and the mixed layer depth (MLD) as predictor variables, and the TCs occurrence in the target area. A statistical downscaling method that identifies the relationship between the predictor and predictand, based on weather typing, is developed. This relationship is then used with forecast data to predict TC activity in the area.

Below, an sketch of the general methodology for constructing the model and forecasting the number of expected TCs in the following season is presented.