Many small island states face significant, even existential, threats due to climate change. Many of these islands are
not represented by Global Climate Models (GCMs) due to their coarse resolution and are outside current CORDEX
domains so are not covered by regional climate projection ensembles. As such the climate change projections for
many small islands are based on GCM climate projections for the open ocean and do not reflect the actual island
climate. Beyond sea level rise the islands are impacted by changes to various climate extremes (tropical cyclones,
other extreme rain, extreme heat, drought,...) but future changes of these hazards have low confidence according
to the recent IPCC report.
This project aims to produce useful climate change projections for a region containing many Pacific island states and territories. In order to capture important island climate processes these projections need to be high resolution (km scale), and in order to quantify confidence levels (and uncertainty) a multi-model ensemble is required. The use of empirical-statistical downscaling approaches including those based on machine learning techniques in the context of island climates will also be included. In doing this we will investigate the ability of convection permitting RCMs to simulate island climates including an evaluation against a variety of observations. Local/regional organisations such as the Secretariat of the Pacific Regional Environment Programme and the Pacific Community are part of discussions and will provide further connections to stakeholders in various countries in the region. The work will also improve the assessment of island climates within IPCC reports. Participants from a variety of countries will take part in this project.
It is intended that the experiment design developed in this region could then be applied to other island regions
globally.
