Improving Hurricane Harvey rainfall simulation with the Local Climate Zone urban classification

The Local Climate Zone (LCZ) classification scheme categorizes urban areas into ten different classes to capture the urban heterogeneity for urban studies. Numerous studies have shown that incorporating LCZ in the Weather Research and Forecasting (WRF) model can lower the errors of simulated temperature, humidity, wind, and heavy precipitation. However, the impact of employing LCZ on WRF-Urban simulated hurricane events has yet to be explored. As a case to demonstrate the LCZ effects on simulating hurricane rainfall, this study selects Hurricane Harvey, a category 4 hurricane that landed on Texas in 2017 and caused 68 deaths and $125 billion in economic loss.

This study investigates the urban heterogeneity effect by simulating Hurricane Harvey with two land use land cover experiments: MODIS only (MOD2018) and MODIS with LCZ (MOD2018.LCZ). Eight ensemble members with different combinations of model physics are evaluated. The results from the ensemble mean show that MOD2018.LCZ has reduced errors in tracks (11 km; 6.9 %), landfall location (4.3 km; 8.7 %), landfall time (0.125 hr; 33.3 %), and urban cumulative precipitation (15mm; 6.5 %). We further evaluate the rainfall clusters and show that clusters with intensity > 30 mm/hr have improved the location and coverage.