Abstract:
Catchment management is a complex task that, over the past decade, has become increasingly important to the community. Information for this management increasingly is being obtained from implementation ofcatchmentmodellingsystems. Implementation ofa catchmentmodelling systemrequiresthree steps,whichare the calibrationofthe system, the validationofthe calibration, andthe extrapolation ofthe system to different hydrologic events and catchment conditions. The robustness of the simulations when this extrapolation is undertaken is relatedto the calibrationandvalidationof the catchmentmodellingsystem,which inturn are impactedbythe qualityandreliabilityofthe inputdata. Acatchment modellingsystemfor simulation of catchment processes can be considered to consist of four conceptual components. Within each of these components, there are manyalternativetransformations availablewith each of these transformations beingthe resultof differentcombinations of processmodels. Theconceptof systemcalibrationis to selectthe appropriate transformations andinputinformation that best represents the catchment beingsimulated. Withinthe generation componentan importantaspect is the model used to transformthe point rainfallmeasurements into a spatially distributed rainfall-over the catchment. There have been many alternative models proposed for this transformation. Presentedherein is an analysisof the influence of alternative rainfall models on the simulated hydrographandhencethe influenceonthe systemcalibration. This analysis willbebasedonreal eventsrecorded at the outletsof the Centennial Park catchment(LBkrrr') andthe UpperParramattaRiver catchment(110km2 ) in Sydney, Australia. It wasfoundthat an increaseinthe robustness ofthe predictions obtainedwasrelateddirectly to the storm variability as definedby the spatial and temporal semi-variograms developed duringthe study.
