Improving Operational Flood Forecast
The National Weather Service provides real-time flood forecasts in the United States. Thirteen River Forecast Centers (RFCs) are responsible for providing the 5-day streamflow forecasts for rivers throughout the country. The operational hydrologists at the RFCs, apply their expert judgment to adjust and modify the short-range streamflow forecasts as new information becomes available in real-time.
In this project, we developed a process to identify the relative importance of uncertain factors in Ohio River Forecast Center (OHRFC) flood forecasts. The goal of this project was to empower the forecasters to quickly reduce forecast errors through intelligent adjustment of model input data and state variables. We used ANOVA to identify the relative contribution to forecast uncertainty of each modifier, Heat map, and bar chart to illustrate the results. We found that the effect of precipitation, soil moisture, and unit hydrograph shape, are most likely to provide improvement in flood forecast accuracy by studying the flood events sampled from 1990 to 2018 in the Ohio River.
The findings of this project was presented in the AGU Fall meeting 2020 (Abstract). This work was published in the Journal of Hydrology X in 2021 (Full Paper) .
References:
- Ray, P. A., Wasti, A., Schade, T., & Zhu, Z. (2020, December). Evaluation of the Effectiveness of the Modifier Process in National Weather Service Flood Forecasts. In AGU Fall Meeting Abstracts (Vol. 2020, pp. GC072-0006).
- Zhu, Z., Wasti, A., Schade, T., & Ray, P. A. (2021). Techniques to evaluate the modifier process of National Weather Service flood forecasts. Journal of Hydrology X, 11, 100073.