Improvement And Evaluation Of The Iowa Flood Center Hillslope Link Model (HLM) By Calibration-free Approach

Abstract

This study evaluates the performance of Iowa Flood Center's real-time distributed hydrologic model, Hillslope-Link Model (HLM). The HLM provides information about current and future streamflow conditions for over 1000 locations in Iowa, including small communities and stream gauge locations. The HLM uses a calibration-free approach in which a system of ordinary nonlinear differential equations represents the processes of rainfall to runoff translation as well as the transport of water in the river network. The authors developed a comprehensive evaluation framework to improve the model performance through modification of its structure, and not by parameter calibration. The authors explored modifications to the runoff generation modeling, and to the approach for routing water in channels. Model performance was assessed using different metrics: Nash Sutcliffe efficiency (NSE); Kling Gupta efficiency (KGE); mean absolute error of river stages (MAE); percent bias (PB); annual peak difference; ratio of hits, misses and false alarms, and timing difference. The inclusion of an interflow component alleviated problems of volume underestimation. Timing of model simulations was found to be more accurate when the routing component considered the channel stream order. In the study, effect of rainfall input on the model performance was verified with Stage IV, which includes bias adjustment using gauge information, and a radar rainfall product created in real-time by IFC without bias correction. Open-loop simulations obtained with Stage IV and IFC rainfall show similar performance. Out of all the simulations obtained with Stage IV, 88% have NSE > 0; 95% have KGE > 0; 65% have MAE < 1 foot; 83% have |PB| < 40; 35% have hit rate >40%; average peak estimation errors are below 5% at large basins and below 20% at small basins; and the average timing difference is smaller than three hours. The characteristics of landforms and the impact of agricultural practices in the region are related to the spatial distribution of model performance. The study provides the basis for future improvements of the model.

Department(s)

Civil, Architectural and Environmental Engineering

Comments

University Corporation for Atmospheric Research, Grant None

Keywords and Phrases

Floods; HLM; Interflow; Iowa; Performance; Routing; Stage IV

International Standard Serial Number (ISSN)

0022-1694

Document Type

Article - Journal

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2023 Elsevier, All rights reserved.

Publication Date

01 May 2020

Link to Full Text

Share

 
COinS