University of Florida Homepage
College of Liberal Arts and Sciences Liberal Arts and Sciences Liberal Arts and Sciences

Geography
Make a Gift
Make a Gift

A Hybrid Framework for Rapidly Locating Transition Zones: A Comparison of Event- and Response-Based Return Water Levels in the Suwannee River FL

Top left shows a map of Georgia and Florida with colored lines showing the tributaries of the Suwannee River. The middle panel shows a map of the lowest part of the Suwannee River, with blue circles showing stream gauge sites and a red triangle showing a tide gauge. Bottom left is the legend for both maps. The right panel shows bars to indicate the extent of observations that exist for each location.
Figure 1. Case study site. (a) Location of tide and stream gauges in the Suwannee River and its main tributaries. (b) Gray box in (a), solid blue line denotes the length of the lower portion of the Suwannee River included in the HEC-RAS model domain. (c) Completeness of the observation record.

TIMMERS, SERAFIN – A Hybrid Framework for Rapidly Locating Transition Zones: A Comparison of Event- and Response-Based Return Water Levels in the Suwannee River FL

R. A. Jane, V. Malagón-Santos, M. M. Rashid, L. Doebele, T. Wahl, S. R. Timmers, K. A. Serafin, L. Schmied, C. Lindemer

Article first published online: 15 November 2022

DOI: https://doi.org/10.1029/2022WR032481

ABSTRACT: Flood risk assessments commonly use event-based approaches to reduce the number of scenarios required to be run through computationally intensive physical process models. Often the return period of the response variable (e.g., a fluvial water level or overtopping discharge) generated by an event (e.g., upstream/downstream water level or set of sea state variables) does not match that of the event itself; a limitation of event-based approaches which can lead to the misspecification of flood risk. We present a transferable hybrid statistical-hydraulic modeling framework for rapidly locating transition zones; river reaches where extreme water levels are driven by both upstream riverine discharge and downstream sea level. Instead of an event-based approach the framework utilizes a surrogate model to reduce computational expense of the hydraulic model. The surrogate-based approach allows the empirical estimation of response-based along-river return levels from a large number of plausible discharge–coastal still water level events simulated from the statistical model. We assess the robustness of the event-based approach by comparing the associated return levels with the response-based return levels. The framework is applied to the Suwannee River in Florida (United States). Three surrogate models are evaluated, highlighting the enhanced ability of non-linear models to accurately capture discharge-sea level interactions along the river. The along-river return levels of the “most-likely” design event are found to lie within the range of variability of the response-based return levels for most of the transition zone.

Read the full publication at Water Resources Research