Simulating saturation‐excess surface run‐off in a semi‐distributed hydrological model

Abstract

The level of complexity, and the number of parameters, to include in a hydrological model is a relatively contentious issue in hydrological modelling. However, it can be argued that explicitly representing important run‐off generation processes can improve the practical value of a model's outputs. This paper explores the benefits of including a new function into an existing semi‐distributed hydrological model (the Pitman model) that is widely used in the sub‐Saharan Africa region. The new function was designed to represent saturation‐excess surface run‐off processes at subcatchment scales and was motivated by the evidence of dambo (low topography riparian areas) type features in many sub‐Saharan river basins. The results for uncertainty versions of the model, with and without the new function, were compared for 25 catchments, which were divided up into those where evidence of dambos exists and those where there is no such evidence. The results suggest that the new function certainly improves the model results for the catchments where dambos exist, but not in situations where saturation‐excess surface run‐off is not expected to occur. The overall conclusion is therefore that the addition of the new function is justified.

Assessing development and climate variability impacts on water resources in the Zambezi River basin: Initial model calibration, uncertainty issues and performance

“…The WEAP model results are very similar as the parameterisation was guided by the parameter sets used for the Pitman model. In general terms, the partitioning shown is consistent with what we expect in the different parts of the Zambezi, as well as with previously reported results for similar areas ( Winsemius et al, 2006 ; Hughes and Mazibuko, 2018 ). KAF4 on the Lunga River is quite typical of many of the sub-basins of the Kafue and Luangwa river systems, where there is an almost equal split between the three runoff components and the surface runoff is dominated by saturation-excess runoff due to the presence of Dambos.…”

“…A key geomorphological feature of the western parts of the Luangwa River basin is the presence of Dambos, which are relatively small, but frequent, valley bottom wetlands that have the potential to generate saturated surface runoff during the early part of the wet season ( Hughes and Mazibuko, 2018 ). The other key feature is the Luangwa River floodplain, although recent work suggests that the downstream impacts (attenuation and delay) are quite minimal at the monthly time scale ( Makungu, 2019 ) and no wetland effects were included in the WEAP model.…”

“…While other models have been previously applied in some parts of the basin ( Winsemius et al, 2006 ; Kling et al, 2014 ; Jury, 2017 ), the two models used in this study are the Water Evaluation And Planning (WEAP) model ( https://www.weap21.org/: accessed November 2019) and the Rhodes University version of the Pitman model ( Pitman, 1973 ; Hughes, 2013 ). Both are semi-distributed conceptual type models that operate at a monthly time step and both have been previously applied in parts of the Zambezi River basin, as well as other areas of sub-Saharan Africa ( Cervigni et al, 2015 ; Hughes, 2016 ; Spalding-Fecher et al, 2017 ; Hughes and Mazibuko, 2018 ). The choice of these models was based on the authors existing calibration experience, the frequency of use of the models in sub-Saharan Africa and because both represent compromises between detailed and simple models, and either explicitly or implicitly include the key runoff generation processes expected to occur within the Zambezi.…”

“…Without further site specific information about real processes it may not be possible to resolve such equifinalities, but it is useful to know that they exist. This version of the model has been previously applied in several areas of the basin ( Hughes, 2013 ; Hughes and Mazibuko, 2018 ) and the results were very useful in guiding the manual calibration process. Several additional runs were made during the current calibration study to further explore the parameter space for parameter sets that yield satisfactory model performance.…”

Unpacking some of the linkages between uncertainties in observational data and the simulation of different hydrological processes using the Pitman model in the data scarce Zambezi River basin

“…Most of the original structure (Pitman, 1973), as well as more recent additions (Hughes, 2004;Hughes and Mazibuko, 2018) to the model have been designed to represent processes explicitly, albeit at the sub-basin scale, using approaches that are similar to the probability distributed principle of Moore (1985). Figure 1 provides an illustration of the model structure, and includes the equations, parameter symbols and simulated state variables referred to in the following sub-sections (more details about some of the other model components can be found in Pitman (1973), Hughes (2004) and Hughes and Mazibuko (2018)).…”