No description available.

Pluvial flood is a natural hazard that severely threatens people’s property and safety. With the development of algorithms and computer technologies, numerical modeling has emerged as an effective tool for predicting the impacts of floods. Despite being one of the most costly types of floods in West Africa, pluvial flooding has not been studied as extensively as riverine flooding, probably because modeling runoff across urban areas remains a challenge. Recently, a module based on the SCS Curve Number Method is incorporated in the open-source software TELEAMC-2D, which provides a possibility to model the infiltration process dynamically. TELEMAC-2D is one of the first hydraulic models to consider hydrologic parameters. Although the update is expected to increase the suitability of TELAMC-2D in pluvial flood modeling, the infiltration routine has not yet been tested in a real situation in a semi-arid area. This study aims to investigate the capability of TELEMAC-2D in simulating the rainfall-runoff process during a pluvial flood event in a semi-arid urban area, Niamey city in west Africa. Due to the lack of calibration data, a hydrological model SWAT is used to evaluate the performance of TELEMAC-2D. Through the comparison between the runoffs generated by the two models, it is found that TELEMAC-2D has a similar trend with SWAT in runoff simulation. However, TELEMAC-2D significantly overestimates the runoff magnitude despite having the same SCS values as SWAT. The reason for the overestimation is TELEMAC-2D that does not properly consider evaporation. Two suggestions are made to improve pluvial floods simulations using TELEMAC-2D in semi-arid areas: 1) couple TELEMAC-2D with a hydrologic model, and use net rainfall generated by the hydrologic model as precipitation input; 2) provide functions in infiltration subroutine that calculate rainfall abstractions by other hydrologic phenomena in addition to the infiltration process.

När klimatet i Sverige blir varmare ökar både förekomsten och omfattningen av intensiva skyfall. Dessa kan leda till stora konsekvenser lokalt för drabbade samhällen. Omfattande översvämningar, skadade fastigheter, erosionsskador och hindrad framkomlighet är några av de potentiella följderna. Ett första steg för att öka beredskapen för skyfall är att göra en skyfallskartering där ett nederbördstillfälle med skyfallskaraktär läggs in i en hydraulisk modell över ett område. Modellen visar sedan områden som riskerar att drabbas av översvämningar samt vilken väg som ytavrinningen tar under förloppet. I detta arbete har en sådan modell gjorts över Sala stad. Skyfallsmodellen skapades i programmet HEC-RAS och för att simulera infiltrationen användes metoden SCS curve number. SCS curve number metoden uppskattar infiltration baserat på markanvändning och jordart. Typregnet valdes till 100 års återkomsttid och modellerades som ett CDS-regn med varaktighet sex timmar och klimatfaktor 1,3. Ett avdrag på nederbördsmängden gjordes även för dagvattennätets kapacitet. Resultatet blev att flertalet områden i Sala kan drabbas av potentiellt stora vattendjup. De största problemområdena identifierades till bebyggelsen mellan Ringgatan och Östra Tulegatan söder om centrala Sala samt området kring Pråmån uppströms Jakob Mats kvarn. Då infiltrationsmetoden SCS curve number är baserad på empiriska data gjordes en känslighetsanalys på infiltrationsparametrarna. Det som undersöktes var påverkan på total andel infiltrerad nederbörd och översvämningarnas utbredning. Tre olika scenarion baserat på osäkerheten i curve number-talet användes samt ett scenario utan infiltration i modellen. Skillnaden i total översvämmad yta med ett djup större än 0,1 m mellan scenariot med högst infiltration och scenariot med lägst infiltration var i modellen totalt 9,1 procentenheter. I det scenario där ingen infiltration modellerades drabbades flertalet nya fastigheter och infrastruktur av översvämning vilket visar på infiltrationens betydelse för riskbedömningen vid en skyfallskartering. Totalt infiltrerade 30 % av nederbörden i modellen vid scenariot med lägst infiltration, 49 % av nederbörden vid scenariot med direkta litteraturvärden för curve number- talet och 66 % av nederbörden i scenariot med högst infiltration. Eftersom infiltrationen mellan scenarierna til hög grad berodde på hur jordarterna klassades i modellen är det dock svårt att dra en generell slutsats om vilket infiltrationsscenario som är bäst lämpat att använda vid skyfall. / As the climate becomes warmer in Sweden the frequency and extent of cloudbursts are expected to rise. When these rain events occurs in populated areas the consequences can be severe for the local community with extensive flooding leading to for example damaged properties, erosion and obstructions of traffic. One step to increase the awareness of these risks is to do a cloudbust mapping were a rain with cloudburst extent is modelled hydraulically. Areas that are at risk of flooding during the rain event and flow paths for the surface runoff can be assessed from the model. In this thesis such a model was created for the town of Sala. The cloudbust model was made using the software HEC-RAS and to simulate infiltration the method SCS curve number was implemented. The rain event was modeled as a Chicago design storm with a return period of 100 years, total duration of six hours and a climate factor of 1,3. The results were that several areas in Sala were at risk of flooding. Two main areas with risk of major flooding extents were located to just south of central Sala between Ringgatan and Östra Tulegatan and adjacent to Pråmån upstream of Jakob Mats kvarn. Because of uncertainties in the SCS curve number model and due to the fact that it is based on empirical data a sensitivity analysis was also done on the infiltration parameters. For the sensitivity analysis four different scenarios were used. One scenario had no modeled infiltration and the three other scenarios were based of error estimations of the curve number parameter called antecedent runoff conditions. The difference in flooding extent with a depth greater than 0,1 m in the model between the scenario with high infiltration and the scenario with low infiltration was 9,1 percentage points. Several new properties and infrastructure were affected in the scenario when no infiltration was used. In total 30% of the precipitation infiltrated in the model in the scenario with low infiltration, 49% infiltrated in the scenario with literature values for the curve number value, and 66% of the precipitation infiltrated in the scenario with high infiltration. Since the infiltration is highly dependent on the classification of the soils in the model no overall conclusion could be drawn on which infiltration scenario that is best suited for a cloudburst model in general when the SCS curve number method is used to model infiltration.