Hydrology laboratory research modeling system (HL-RMS) of the US national weather service

Abstract

This study investigates an approach that combines physically-based and conceptual model features in two stages of distributed modeling: model structure development and estimation of spatially variable parameters. The approach adds more practicality to the process of model parameterization, and facilitates an easier transition from current lumped model-based operational systems to more powerful distributed systems. This combination of physically-based and conceptual model features is implemented within the Hydrology Laboratory Research Modeling System (HL-RMS). HL-RMS consists of a well-tested conceptual water balance model applied on a regular spatial grid linked to physically-based kinematic hillslope and channel routing models. Parameter estimation procedures that combine spatially distributed and 'integrated' basin-outlet properties have been developed for the water balance and routing components. High-resolution radar-based precipitation data over a large region are used in testing HL-RMS. Initial tests show that HL-RMS yields results comparable to well-calibrated lumped model simulations in several headwater basins, and it outperforms a lumped model in basins where spatial rainfall variability effects are significant. It is important to note that simulations for two nested basins (not calibrated directly, but parameters from the calibration of the parent basin were applied instead) outperformed lumped simulations even more consistently, which means that HL-RMS has the potential to improve the accuracy and resolution of river runoff forecasts. Published by Elsevier B.V.

State updating of a distributed hydrological model with Ensemble Kalman Filtering: effects of updating frequency and observation network density on forecast accuracy

“…Such a grid-based approach is a popular concept in applied hydrology (e.g. Koren et al, 2004;Blöschl et al, 2008;Cole and Moore, 2009;Thielen et al, 2009). For each 1 km × 1 km grid cell of the Upper Ourthe the HBV-96 model was implemented and is used as a benchmark case.…”

“…Currently, most operational hydrological forecasting systems employ lumped hydrological models (with deterministic or manual state updating), but there is a clear tendency to move towards distributed models combined with hydrological ensemble forecasts, (e.g. Koren et al, 2004;Cole and Moore, 2009;Weerts et al, 2012). The main advantage of spatially distributed models is the possibility to force them with spatially measured data, which nowadays become more readily available due to rapid developments in telemetry.…”

Sensitivity of hydrological models to uncertainty in rainfall input

“…For distributed approaches we do not calibrate the parameters for each pixel (i). We imposed a dimensionless spatial repartition of these parameters and calibrated a weighting coefficient, as applied in the Hydrology Laboratory Research Modeling System (HL-RMS) (Koren et al, 2004;Reed et al, 2007). The spatial repartition of S(0)/A was imposed by a regionalization study, and the spatial repartition of B was imposed by the distance between the pixel and the catchment outlet.…”

Comparing sensitivity analysis methods to advance lumped watershed model identification and evaluation

“…The SNOW-17 (Anderson, 1973) and the Sacramento soil moisture accounting (SAC-SMA) models (Burnash, 1995) are popular and the United States National Weather Service (US NWS) uses them for river forecasting (Moreda et al, 2006;Koren et al, 2004;Smith et al, 2004;Reed et al, 2004). In this study, lumped versions of these models have PLWHC-Percent (decimal) liquid-water holding capacity.…”

“…Hydrologic models are evolving from single purpose tools to complex decision support systems that can perform all (or at least many) of the tasks mentioned above in a single software package. Hydrological models vary in complexity from lumped conceptual models to distributed models that include close coupling of surface and groundwater flow processes, feedbacks with the atmosphere, transport of water and solutes, and spatially explicit representations of system characteristics and states (e.g., Duffy, 1996Duffy, , 2004Koren et al, 2004;Panday and Huyakorn, 2004). In integrated assessment applications models may even include socioeconomic components to integrate human behavior (Wagener et al, 2005).…”