Complex, deterministic hydrological modelling towards decision support for urban catchment management

[摘要] ENGLISH ABSTRACT: Historically, urban waterresources have too often been managed without recognition that theflow in a river integrates many landscape and biological features. This has often resulted inthe elimination of natural processes and their replacement by man-made streamlinedstructures with the effects of increased urbanisation being primarily addressed from anengineering and economics point of view to the detriment of environmental and social issues.Catchment Management, as legislated in the Water Act, No. 36 of 1998, is a managementapproach to address the negative consequences of an urban stormwater design philosophyrestricted to flood restriction. It is a systems approach that integrates engineering andscientific skills, socio-economic concerns, and environmental constraints within a new multidisciplinarydecision-making process that recognises the different components of thehydrological and aquatic cycles are linked, and each component is affected by changes inevery other component.In order to make effective management decisions, catchment managers require tools toprovide reliable information about the performance of alternative arrangements of stormwatermanagement facilities and to quantify the effects of possible management decisions on thewater environment. A deterministic hydrological model is such a tool, which provides thelink between the conceptual understanding of the physical catchment characteristics and theempirical quantification of the hydrological, water quality and ecological response.In order to provide effective computer based decision support, the hydrological model mustbe part of an integrated software application in which a collection of data manipulation,analysis, modelling and interpretation tools, including GIS, can be efficiently used together tomanage a large potion of the overall decision process. This decision support system musthave a simple and intuitive user interface able to produce easily interpreted output. It musthave powerful graphical presentation capabilities promoting effective communication and bedesigned to solve ill-structured problems by flexibly combining statistical analysis, modelsand data.The Great Lotus River canal, situated on the Cape Flats, Cape Town, has been designed andcontrolled through extensive canalisation and the construction of detention pond facilities toavoid the flooding of urban areas of the catchment. This approach has resulted in thesechannels becoming stormwater drains, transporting waste and nutrients in dissolved andparticulate forms, and reducing their assimilatory capacity for water quality improvement.In order to investigate the use of hydrological modelling in decision support for CatchmentManagement, the semi-distributed, physically based model, SWMM, was applied to the GreatLotus River canal. SWMM consists of a number of independent modules allowing thehydrological and hydraulic simulations of urban catchments and their conveyance networkson an event or continuous basis.In order to ease the application of the Fortran based SWMM model, the GUl, PCSWMM98,was developed by Computational Hydraulics Inc (CH!). This provides decision support forSWMM through large array of tools for file management, data file creation, outputvisualisation and interpretation, model calibration and error analysis and storm dynamicanalysis thus easing any simulations with SWMM. In addition, PCSWMM was developedwith a GIS functionality for graphically creating, editing and/or querying SWMM modelentities and attributes, displaying these SWMM layers with background layers and dynamicmodel results, and exporting data to SWMM input files thus providing an interface between aGIS and SWMM.In terms of Catchment Management, the above DSS can be used effectively to assist decisionmaking.This is to address tensions between the fundamental catchment managementconsiderations of physical development, social considerations and maintaining ecologicalsustainability. It is at the stages of Assessment and Planning that the model can play the mostsignificant role in providing decision support to the Catchment Management process.Assessment in the Catchment Management process refers to the collection, storage, modellingand interpretation of catchment information. It is in this quantification, interpretation andassessment of catchment information that a hydrological model contributes to an increase inknowledge in the Catchment Management process. In identifying and quantifying, at asufficient temporal and spatial scale, the dominant cause and effect relationships in the urbanphysical environment, a hydrological model is able to highlight the main contributing factorsto an issue. This is used in the Planning stage of the Catchment Management process andwhen combining these contributing factors with assessments of the socio-economic andadministrative environments, enables the prioritisation of the principal issues requiringattention in a Catchment Management Strategy.It is possible to link the multiple decision-making requirements of Catchment Managementwith the abilities of a hydrological model to provide information on these requirements in aconceptual framework. This framework consists of the fundamental catchment considerationsof Physical Development, Environmental Management and Social Development and resolvesthese considerations into the various management issues associated with each consideration ~swell as its management solution. The management solutions are linked to the model throughformulating the solution in terms of the model parameters and perturbing the affectedparameters in ways to simulate the management solution. This results in model output andgraphical interpretation of the effects of the suggested management solution. A comparisonbetween the simulated effects of each management solution allows the CatchmentManagement body to identify optimal management solutions for the various managementIssues.The present model of the Great Lotus River catchment is sufficient to simulate the overlandand subsurface flows from individual parts of the catchment and to route these flows andassociated pollutant loadings to the catchment outlet. At its present level of complexity, thefinely discretised model subcatchment and conveyance network provides decision support forCatchment Management through the simulation, at a pre-feasibility stage, of variousCatchment Management issues and their proposed solutions.Given more detailed canal and drainage network dimensions and water quality data, it ispossible for the model to incorporate hydraulic calculation routines to assess the implicationsof alternative river rehabilitation techniques and waste management strategies. This wouldallow greater capability in assessing the role of the various BMPs in ameliorating stormwaterimpacts and pollutant loading. In addition, a detailed level survey of the stormwater pipe andcanal network could result in hydrological modelling being utilised to identify critical areaswhere stormwater upgrading would be necessary.In order to facilitate future complex, finely discretised catchment hydrological models, it isimperative that complete and detailed drainage patterns and stormwater networkcharacteristics are available. In addition, to minimise model generation costs and time ofmodel setup, this spatially representative data must be captured in a GIS for rapid inclusioninto the model. Furthermore, complete spatially representative precipitation datasets arenecessary to ensure that model error is reduced. These two issues of available spatial data andcomprehensive precipitation records are crucial for the generated models to function aseffective decision support systems for Catchment Management.

[发布机构] Stellenbosch University