Parameterizing MTCLIM for the diverse meteorological and topographical environments across the Hawaiian islands

Abstract

Continuous and high resolution climatologies are important inputs in determining future scenarios for land-surface processes. In Hawaiʻi, a lack of continuous meteorological data has been a problem for both ecological and hydrological research of land-surface processes at daily time scales. For climate variables of downward shortwave radiation (SWdown) and relative humidity (RH), the limited number of surface stations which record daily values are situated at city airports or in convenient locations leaving large sections of the islands remain poorly represented. The aim of this study is to evaluate the rationale behind using the mountain microclimate simulator MTCLIM to obtain observation-based climate estimates of SWdown and RH data at a daily increment for the period of 1990–2014 for the main Hawaiian Islands. Results testing model output, after parameter calibration with observed daily SWdown data, show a mean bias error (MBE) of 3.17 Watts per meters squared (W/m2; 1.15%), mean absolute error (MAE) of 49.12 W/m2 (32.83%), and root mean square error (RMSE) of 60.04 W/m2 (40.73%). This is reduced from a MBE of 107.63 W/m2 (66.47%), a MAE of 127.80 W/m2 (82.55%), and a RMSE of 152.11 W/m2 (102.31%) for observed SWdown data processed through MTCLIM with default parameter settings. Optimized RH results for all stations combined showed a MBE of -0.8%, a MAE of 14.96%, and a RMSE 17.77%. Focusing on stations below the trade-wind inversion (TWI), optimized RH results for show a MBE of -1.03%, a MAE of 9.86%, and a RMSE of 11.75%. Results above the TWI show a MBE of 11.31%, a MAE of 21.71%, and a RMSE of 26.39%. Before calibration, results for all RH stations showed a MBE -6.0%, a MAE of 15.29%, and a RMSE 17.61%. These results indicate that MTCLIM performs well at simulating daily SWdown. However, the model performed poorly at simulating daily RH data at the stations analyzed in this study. The poor results for RH occurred due to MTCLIM’s algorithms inability to consider the strong influence of the oceanic airmasses on the humidity measurements and the presence of a persistent TWI around 7100 ft (2164 m) occurring throughout the islands. Additionally, the unique topography and climate gradients further complicate the issue and make predicting accurate RH estimates difficult for the Hawaiian Islands. Future versions of MTCLIM may need to be re-evaluated and updated to include options to calculate RH at coastal sites that are influenced by moist oceanic air masses. It would also be advisable for future updates of the model’s algorithms to include an option that could account for an inversion layer. Otherwise, MTCLIM should only be used in a limited capacity for locations that experience an inversion or are strongly influenced by nearby humid air masses.