Evaluation of the SDF method using a customised design flood estimation tool

Gericke, Ockert Jacobus (2010-12)

Thesis (MScEng (Civil Engineering))--University of Stellenbosch, 2010.

Thesis

ENGLISH ABSTRACT: The primary aim of this study was to evaluate, calibrate and verify the SDF run-off coefficients at a quaternary catchment level in the C5 secondary drainage region (SDF basin 9) and other selected SDF basins in South Africa by establishing the catchment parameters and SDF/probability distribution-ratios. The probability distribution-ratios were based on the comparison between the flood peaks estimated by the SDF method and statistical analyses of observed flow data. These quaternary run-off coefficients were then compared with the existing regional SDF run-off coefficients, whilst the run-off coefficient adjustment factors as proposed by Van Bladeren (2005) were also evaluated. It was evident from this study that the calibrated run-off coefficients obtained are spread around those of Alexander (2003), but were generally lower in magnitude. The adjusted run-off coefficients (Van Bladeren, 2005) had a tendency to decrease in magnitude with increasing recurrence interval, whilst some of the adjusted run-off coefficients exceeded unity. The extent to which the original SDF method overestimated the magnitude and frequency of flood peaks varied form basin to basin, with the SDF/probability distribution-ratios the highest in the Highveld and southern coastal regions with summer convective precipitation. In these regions the flood peak-ratios were occasionally different by up to a factor of 3 or even more. The southern coastal regions with winter orographic/frontal precipitation demonstrated the best flood peak-ratios, varying from 0.78 to 1.63. The adjusted SDF method results (Van Bladeren, 2005) were only better in 26% of all the basins under consideration when compared to those estimated by the original SDF method. On average, the adjusted SDF/probability distribution-ratios varied between 0.30 and 6.58, which is unacceptable. The calibrated version of the SDF method proved to be the most accurate in all the basins under consideration. On average, the calibrated SDF/probability distribution-ratios varied between 0.85 and 1.15, whilst at some basins and individual return periods, less accurate results were evident. Verification tests were conducted in catchments not considered during the calibration process with a view to establish whether the calibrated run-off coefficients are predictable and to confirm that the method is reliable. The verification results showed that the calibrated/verified SDF method is the most accurate and similar trends were evident in all the basins under consideration. On average, the verified SDF/probability distribution-ratios varied between 0.82 and 1.19, except in SDF basins 6 and 21 where the 5 to 20-year return period flood peaks were overestimated by 41% and 56% respectively, which is still conservative. The secondary aim of this study was to develop a customised, user-friendly Design Flood Estimation Tool (DFET) in a Microsoft Office Excel/Visual Basic for Applications environment in order to assess the use and applicability of the various design flood estimation methods. The developed DFET will provide designers with a software tool for the rapid investigation and evaluation of alternative design flood estimation methods either at a regional or site specific scale. The focus user group of the application will comprises of engineering technicians, engineering technologist and engineers employed at civil engineering consultants, not necessarily specialists in the field of flood hydrology. The DFET processed all the catchment, meteorological (precipitation) and hydrological (observed flows) data used as input for the various design flood estimation methods.

AFRIKAANSE OPSOMMING: Die primêre doelwit van die studie was om die SDF-afloopkoëffisiënte op ‘n kwartinêre opvangsgebiedvlak in die C5-sekondêre dreineringsgebied (SDFopvangsgebied 9) en ander gekose SDF-opvangsgebiede in Suid-Afrika te evalueer, te kalibreer en te verifieer deur die opvangsgebiedparameters en SDF/waarskynlikheidsverspreiding-verhoudings vas te stel. Dié waarskynlikheidsverspreiding-verhoudings was gebaseer op die vergelyking tussen die vloedpieke soos beraam deur die SDF-metode en statistiese analises van waargenome vloeidata. Dié kwartinêre afloopkoëffisiënte is met die bestaande streeksgebonde SDF-afloopkoëffisiënte vergelyk, terwyl die afloopkoëffisiënt-aanpassingsfaktore soos voorgestel deur Van Bladeren (2005) ook geëvalueer is. Dit het duidelik uit die studie geblyk dat die gekalibreerde afloopkoëffisiënte verspreid rondom die van Alexander (2003) is, maar in die algemeen laer in omvang. Die aangepaste afloopkoëffisiënte (Van Bladeren, 2005) was geneig om af te neem in grootte met ‘n toename in die herhalingsperiode, terwyl sommige afloopkoëffisiënte ‘n waarde van 1 oorskry het. Die omvang waartoe die oorspronklike SDF metode die grootte en herhaalperiode van vloedpieke oorskat het, wissel van opvangsgebied tot opvangsgebied, met die SDF/waarskynlikheidsverspreiding-verhoudings die hoogste in die Hoëveld en suidelike kusstreke gekenmerk deur konveksie-somerreënval. In hierdie streke het die vloedpiekverhoudings gereeld verskil tot en met ‘n faktor van 3 of selfs meer. Die suidelike kusstreke met kenmerkende ortografiese/frontale winterreënval het oor die beste vloedpiekverhoudings beskik wat gewissel het tussen 0.78 en 1.63. Die resultate van die aangepaste SDF-metode (Van Bladeren, 2005) was slegs in 26% van al die opvangsgebiede beter as die beramings van die oorspronklike SDF-metode. Die aangepaste SDF/waarskynlikheidsverspreiding-verhoudings het, met verwysing na gemiddeldes, tussen 0.30 en 6.58 gewissel, wat onaanvaarbaar is. Die gekalibreerde weergawe van die SDF-metode was die mees akkurate metode in al die opvangsgebiede van belang. Die gekalibreerde SDF/waarskynlikheidsverspreiding-verhoudings het, met verwysing na gemiddeldes, tussen 0.85 en 1.15 gewissel, terwyl die resultate van sommige opvangsgebiede en individuele herhalingsperiodes minder akkuraat was. Verifikasietoetse is uitgevoer in die opvangsgebiede wat nie tydens die kalibrasieproses gebruik was nie om vas te stel of die gekalibreerde afloopkoëffisiënte voorspelbaar is en om te bevestig dat die metode betroubaar is. Die verifikasieresultate het getoon dat die gekalibreerde/geverifieerde SDFmetode die mees akkurate metode is en dat soortgelyke tendense duidelik was in al die relevante opvangsgebiede. Die geverifieerde SDF/waarskynlikheidsverspreiding-verhoudings het, met verwysing na gemiddeldes, tussen 0.82 en 1.19 gewissel, behalwe in SDF-opvangsgebiede 6 en 21 waar die 5- en 20-jaar herhalingsperiode-vloedpieke onderskeidelik met 41% en 56% oorskat is, wat steeds konserwatief is. Die sekondêre doelwit van die studie was om ‘n gebruikersvriendelike “Design Flood Estimation Tool” (DFET) in ‘n Microsoft Office Excel/Visual Basic for Applications omgewing te ontwikkel om die gebruik en toepaslikheid van die verskeie ontwerpvloedberamingsmetodes te bepaal. Die DFET sal ontwerpers voorsien van ‘n sagtewareprogram om alternatiewe ontwerpvloedberamingsmetodes op streek- of plaaslike skaal te ondersoek en te evalueer. Die fokus-gebruikersgroep vir die toepassing van die program sal bestaan uit ingenieurstegnici, ingenieurstegnoloë en ingenieurs werksaam by raadgewende siviele ingenieurs, nie noodwendig vakkundiges in die veld van hidrologie nie. Die DFET was gebruik om al die opvangsgebied-, meteorologiese (reënval) en hidrologiese (waargenome vloeie) data vir die verskeie ontwerpvloedberamingsmetodes te verwerk.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/5340
This item appears in the following collections: