This study summarizes an assessment of the wind resource at selected locations in Fiji for the potential of future utility-scale wind-power development. We use 2 -8 years of near surface wind observations (2011 -2018) from thirty automatic weather stations. The standard windindustry software, WAsP is used to simulate the wind resource in terms of mean wind speed, dominant wind direction, power density and annual energy production (AEP) using a Vergnet 275-kW wind turbine. Our analysis identifies three sites: Rakiraki, Nabouwalu and Udu, which should be considered for a future comprehensive resource assessment for utility-scale windpower development once further wind resource data is available. High-resolution resource maps for each wind resource parameter at a horizontal resolution of 50 m are produced for 6 km × 6 km domains around these sites. Rakiraki, Nabouwalu and Udu have average wind speeds of 7.6 m/s, 7.1 m/s and 7.0 m/s, with an average power density of 401 W/m 2 , 512 W/m 2 and 294 W/m 2 , and a potential average AEP of 0.91 GWh, 0.80 GWh and 0.72 GWh at 55 m AGL, respectively. The dominant wind direction is southeasterly. Modelling a 10 MW wind farm at each site yields a net AEP of 43 GWh, 42 GWh and 37 GWh for Rakiraki, Nabouwalu and Udu, respectively with capacity factors of 0.42 -0.48 and wind farm efficiencies of 97 -98 %.

This study presents a high-resolution mesoscale wind-resource assessment of the small island developing state (SIDS) of Fiji using a 10-year simulation of the Weather Research and Forecasting (WRF) model with convection-permitting resolution. Our analysis evaluates the wind speed and Weibull distributions, diurnal and seasonal wind speed patterns, resource maps of annual and seasonal wind speed, power density, model statistical analysis and interannual wind speed variability. The results reveal that the WRF-model simulated wind resource parameters are in good agreement with observations at 24 existing weather stations. At 55 m above ground, the annual mean wind speed and wind power density varies from 1.5 m/s to 8 m/s and 50 W/m 2 to 300 W/m 2 , respectively, for onshore land areas. Higher wind speeds are observed during austral winter than in austral summer. Forty high wind-resource areas are identified in this study, which were previously unknown. This indicates that there is potential for utility-scale wind power generation at selected locations with wind speed and power density greater than 6.4 m/s and 300 W/m 2 (NREL, Wind Power Class 3). An estimated 1000 MW theoretical potential installed capacity is available for utility-scale wind power applications on Viti Levu and Vanua Levu.

Evaluation of the performance of the WRF model is carried out for simulating the surface winds and the diurnal cycle of wind speed for the small island developing state of Fiji at a 1.33 km by 1.33 km grid resolution using 1deg gridded data from NCEP-FNL. Simulations are performed for an austral summer (January 2017) and an austral winter (July 2017) month using the dynamical downscaling and the two-way nested approach. A set of physics parameterization schemes together with topo_wind = 1, 2 and ysu_topdown_pblmix = 1 physics settings associated with YSU PBL scheme are used to correct the surface winds and the diurnal cycle of wind speed. The results reveal that the WRF model is able to capture the surface winds and the diurnal cycle of wind speed on the windward side. Surface winds on the leeward side and the outer islands, show positive bias especially at nighttime for January and at both the day and night time for July. The statistical evaluation of all stations for January (July) showed a bias of 1.16 m/s (1.89 m/s), RMSE of 2.40 m/s (3.14 m/s), STDE of 1.88 m/s (2.08 m/s) and diurnal cycle correlation of 0.74 (0.68) using topo_wind = 2 and ysu_topdown_pblmix = 1.

A grid sensitivity study of the Weather Research and Forecasting (WRF) model was conducted for simulating the surface winds and the diurnal cycle over the small island developing state (SIDS) of Fiji. Two different sets of grid resolutions: 20 km -4 km -1.33 km and 15 km -5 km -1 km are used with the two-way nested approach and 1deg gridded input data from the National Centers for Environmental Prediction -(Final) Operational Global Analysis (NCEP-FNL) as initial and boundary conditions. Simulations are performed for an austral summer (January 2017) and austral winter (July 2017) month using the dynamical-downscaling approach and the tropical suite of the physics parameterization scheme to simulate the surface winds and the diurnal cycle of wind speed for both grid set-ups. The results revealed that the WRF model is able to capture the surface winds and the diurnal cycle of wind speed more accurately for the higher grid resolution of 1 km x 1 km in comparison with the 1.33 km x 1.33 km, indicating that the topographical representation is better in the higher grid resolution. For January (July) the bias reduced from 0.85 m/s (1.04 m/s) to 0.50 m/s (0.83 m/s), which is a reduction of 20.78 % (15.44 %). The Root Mean Square Error (RMSE) reduced from 2.08 m/s (2.55 m/s) to 1.96 m/s (2.50 m/s). The Standard Deviation Error (STDE) is 1.65 m/s (1.92 m/s), which is almost the same for both grids. The diurnal cycle correlation also improved from 0.69 (0.60) to 0.81 (0.75).

This study carries out an analysis of the 10 MW Butoni wind farm in the tropical southwest Pacific island of Fiji using 6 years of uninterrupted near-surface wind observations (2013–2018). The standard wind-industry software, WAsP is used to analyse and evaluate the wind characteristics of the wind farm and the surrounding areas. The modelled and operational AEP are discussed with the related economic analysis together with the main causes for the under-performance of the wind farm. The results revealed that the mean wind speed, power density and the AEP at the Butoni wind farm are below the utility-scale standard of 6.4 m/s, 300 W/m2 and 500 MWh/year/turbine respectively, at 55 m above ground level (AGL). The main reason for the under-performance of the wind farm is that it was commissioned for a low mean wind speed regime of Wind Power Class 1. The wind farm has a lower-than-expected capacity factor of 5.4% and a higher wind shear coefficient of 0.35. An economic analysis revealed that the payback time is 24.5 years, and the cost of energy generation is FJD $ 0.55/kWh.