Authors: Aline Falck, Giuseppe Torri, Yinphan Tsang, Jan Hafner, Keri Kodama, Yu-Fen Huang, Thomas Giambelluca – University of Hawai’i at Mãnoa, Andy Wood – Colorado School of Mines, Kevin Kodama – NOAA
Title: Assessing the National Water Model for flood forecasting in tropical islands
Abstract: The Hawaiian islands are affected by multiple flood events every year, with relatively small watersheds characterized by intermittent flow and high topography slopes. This makes a challenge for flood forecasting, and its accuracy and reliability are extremely important due to the significant impact in human lifes. In this study we briefly evaluate the performance of the streamflow forecasts from the operational NOAA National Water Model (NWM) to provide flood forecast guidance to emergency management communities over the islands of Hawai’i. We evaluate the NWM streamflow forecasts over Hawai’i using streamflow observations available from the U.S. Geological Survey, as well as the NWM analyses used to initialize the model. We also evaluate the precipitation forecasts used as input to the NWM. Different evaluation versions of the NWM over Hawaiian Islands were considered to account for the impact of introducing Multi-Radar Multi-Sensor (MRMS) model forcings during NWM update cycles. Preliminary results show that the NWM forecast systematically overestimates the streamflow over all the islands. The discrepancies in the predicted streamflow simulations are particularly evident during maximum streamflow, especially during the period with the presence of the MRMS, reducing the biases in streamflow forecasts.