| || || Rain and rainfall -- Fiji|
| || || Impacts of ENSO on rainfall and temperature in Fiji : evaluation of CCAM downscaling at 60 and 8 km resolutions|
Author: Kumar, Vandhna D.
Institution: University of the South Pacific.
Subject: Rain and rainfall -- Fiji, Climatic changes
Call No.: Pac QC 925 .8 .F5 K86 2014
Copyright:20-40% of this thesis may be copied without the authors written permission
Abstract: In Fiji, the high mountain ranges on the main island of Viti Levu, together with the influence of the predominant southeasterly trade winds, leads to a much drier and hotter western division compared to the eastern division. Different effects of the El Niño Southern Oscillation (ENSO) are also observed between the windward and leeward sides of the island. With resolutions typically ranging from 100-400 km, general circulation models (GCMs) are too coarse to resolve such fine level details on small islands, and numerous efforts have been made to successfully downscale low resolution climate data to deliver reliable information for small island nations. Amongst these, the CSIRO’s stretched grid variable resolution Conformal Cubic Atmospheric Model (CCAM) has been used to perform 60 km global and 8 km regional simulations for selected GCMs. The skill of CCAM in resolving the orographic differences in the climatologies of precipitation and temperature between the wet and dry sides of Viti Levu, and the variation in the impacts of ENSO on these two parameters have been investigated for the 60km and 8km resolution experiments performed over Fiji for the models GDFL-CM2.1, UKMO-HadCM3 and ECHAM5/MPI-OM. The CCAM simulations are evaluated against the station-based observations from study sites Suva and Nadi, each selected from the wet and dry sides of the island respectively. It is found that while CCAM exhibits good skill in reproducing the different climatologies of precipitation and temperature at the two sites, it is not able to successfully capture the response to ENSO at the individual locations. Based on the evaluation results, model projections for the aspects studied over the 21st century, for the SRES A2 scenario, are also generated. The results of the study show that while further improvement is needed in CCAM downscaling to correctly simulate the ENSO signal at individual locations, the model provides high improvement upon GCM simulations in terms of topography and associated local climate effects on small islands, and can be applied in generating better information about the current and future climate from lower resolution datasets for risk assessment and adaptation planning purposes.