Assessment of Relative Accuracy of TxRR and NWM in Simulating Streamflow for a Cluster of Watersheds along the south Texas Coast

Abstract

Study Region -This study is conducted for a cluster of watersheds within the Matagorda Basin along the south Texas coast in the United States. Study Focus - Retrospective streamflow simulations of two hydrologic models of contrasting formulations and complexity, namely the TxRR, a simple analog model, and the National Water Model (NWM), a land surface-hydrologic model that explicitly accounts for surface energy balance in calculating water budget, and whose output. The comparison was motivated by a) the need for improving the modeling of runoff dynamics for watersheds along central Texas coast through the introduction of NWM to inform freshwater inflow monitoring and flood planning, and b) the need to better understand the potential ability of the NWM to address shortcomings of TxRR related to the latter’s mechanistic deficiencies in a region known for hydroclimatic extremes. The study focuses on the temporal scale and climate dependence in the relative performance of the two models, the behaviors of models during extreme floods and droughts, and the relative efficacy of parameter transfer schemes. It further seeks to relate the performance differentials to model physics and calibration approaches. New Hydrologic Insights – NWM, with sophisticated representations of energe closure, is advantageous to TxRR in capturing low-frequency (interannual) variations in streamflow modulated by the surface energy balance, as evidenced by the superior correlation of its streamflow simulations at the interannual range. Yet, its overall performance trails behind TxRR at daily scale, and it tends to underpredict runoff volumes during two major floods. These features underscore the tendency of the model to suppress soil moisture at the onset of these events. On the other hand, TxRR outperforms in reproducing the volumes for the flooding events, but it overpredicts runoff during the extreme drought episode of 2011, likely an outcome of inadequate representation of the impacts of root zone soil moisture in regulating runoff. TxRR’s parameteralization scheme proves more effective for adaptation across watersheds due to the presence of steep gradation in soil properties .

Quantifying patterns of streamflow peaks over the southeastern United States using a long‐term retrospective data set

“…This includes 715 USGS gauges within the southeast US, which were used as streamflow benchmark locations for the NWM model (Foks et al, 2022;Henderson et al, 2020). Validation studies indicate a good performance of the model within higher flow ranges, with some evidence of lower performance for low flows (Garousi-Nejad & Tarboton, 2022;Johnson & Kim, 2019;Rahman et al, 2022;Towler et al, 2022;Wan et al, 2022). For additional verification of the data specifically used in this study, the daily, monthly, and annually aggregated data series available in NWM nodes were compared to USGS observation data over the same aggregated time periods.…”