Phenology and seasonal forecasting in the Pacific: combining traditional knowledge with statistical and dynamical methods to increase community resilience

Many indigenous people in the Pacific forecast seasonal climate conditions through observation and monitoring of meteorological, astronomical and biological indicators (e.g. phenology). Built over many generations, these knowledge systems are adapted to local conditions to cope with a highly variable and vulnerable environment. Traditional climate knowledge systems continue to influence all aspects of modern Pacific livelihoods from agricultural productivity to disaster response and recovery. In recent years, alternative forecasting methods have been promoted by national meteorological services based on statistical and dynamical modeling of the climate system. However, in some locations, uptake of these ‘new' methods is low, with locals continuing to use traditional forecasts for many reasons including inadequate access to the new forecasts, insufficient trust in new forecasting methods or historical usefulness of traditional forecasts. Enabling adaptation to changing climatic conditions in the Pacific requires improved understanding of how traditional forecasting methods compare to those based on climate models. Building on global experiences with traditional seasonal forecasting, we develop a methodology for the Pacific region that includes:

1. Documenting traditional indicators used for seasonal climate forecasting;

2. The development of a national monitoring network, based on the traditional indicators identified, including phenology;

3. Comparing the outcomes of traditional forecasts with those based on statistical and dynamical modeling of the climate system;

4. Development of methods for optimally combining traditional and modeling-based forecasts.

This integrated approach to forecasting, illustrated with a case study from Vanuatu, has the potential to improve the accuracy and utility of local forecasts as well as ensuring the communication of climate information is in a locally relevant context.