Toward improved streamflow forecasts: value of semidistributed modeling

Abstract

Abstract. The focus of this study is to assess the performance improvements of semidistributed applications of the U.S. National Weather Service Sacramento Soil Moisture Accounting model on a watershed using radar-based remotely sensed precipitation data. Specifically, performance comparisons are made within an automated multicriteria calibration framework to evaluate the benefit of "spatial distribution" of the model input (precipitation), structural components (soil moisture and streamflow routing computations), and surface characteristics (parameters). A comparison of these results is made with those obtained through manual calibration. Results indicate that for the study watershed, there are performance improvements associated with semidistributed model applications when the watershed is partitioned into three subwatersheds; however, no additional benefit is gained from increasing the number of subwatersheds from three to eight. Improvements in model performance are demonstrably related to the spatial distribution of the model input and streamflow routing. Surprisingly, there is no improvement associated with the distribution of the surface characteristics (model parameters).

Advances in the use of observed spatial patterns of catchment hydrological response

“…Many researchers have noted that it is common for a large number of parameter sets to give similar fits to observed data. Gupta et al [48] propose methods that consider the multiobjective nature of calibration, and that allow for explicit consideration of model error, with a recent application provided in Boyle et al [23]. Mroczkowski et al [79] have illustrated the value of different types of time-series data for model testing and calibration.…”

“…For example Wagener et al [106,107] show the value of different parts of a hydrograph in constraining uncertainty in predictions using their DYNIA (DYNamic Identifiability Analysis) approach, and Boyle et al [23] split the hydrograph into different sections with separate objective functions for each. The same argument applies to the use of spatial patterns.…”

A constraint-based search algorithm for parameter identification of environmental models

“…The multi-objective approach seeks to identify parameter sets that simultaneously provide optimal performance for different aspects of system response (Gupta et al, 1998;Boyle et al, 2000Boyle et al, , 2001. This can include constraining the model to reproduce multiple system fluxes and state variables such as runoff, evaporation, groundwater levels or tracer concentrations (e.g., Gupta et al, 1999;Bastidas et al, 1999;Freer et al, 2002;McDonnell, 2002, 2013;Khu and Madsen, 2005;Fenicia et al, 2008;Winsemius et al, 2008;Birkel et al, 2011;Hrachowitz et al, 2013).…”

Comparison of point forecast accuracy of model averaging methods in hydrologic applications

“…In a previous study Vrugt and Robinson (2007) generated a 36-year ensemble of daily streamflow forecasts using eight different conceptual watershed models involving the ABC (3) (Fiering 1967;Kuczera and Parent 1998), GR4J (4) (Perrin et al 2003), HYMOD (5) (Boyle et al 2001;Vrugt et al 2002;Vrugt et al 2005), TOPMO (8) (Oudin et al 2005), AWBM (8) (Boughton 1993;Marshall et al 2005), NAM (9) (Nielsen and Hansen 1973), HBV (9) (Bergström 1995), and SAC-SMA (13) (Burnash et al 1973). These eight models are listed in order of increasing complexity, and the number of user-specified parameters is indicated in parentheses.…”