INTERNATIONALJOURNALOF GEOMATICSANDGEOSCIENCES Volume1,No 3,2010

Copyright2010AllrightsreservedIntegratedPublishingservices

Researcharticle

ISSN0976 4380

1 1 2 BhankaurallyMustaphaYahya NowbuthMantaDevi andBhavanaUmrikar 1.DepartmentofCivilEngineering,FacultyofEngineering,Universityof Mauritius 2.DepartmentofGeology,UniversityofPune,Pune,Maharashtra,India bnumrikar@gmail.com

FloodHazardMappingbyIntegratedGISSCSModel

ABSTRACT Naturalhazardshaveexistedsincethedawnofcivilization.Withtime,theyhaveslowly changedintodisasterswherethemainfactorgearingtheamplitudeoftheireffectsisthe extent of urbanization within the critical regions of occurrence. Thus, with the development,thesedisastershavebecomethecauseofsocioeconomiclosses.Amongst all, disasters of hydrometeorological origins are most frequent and seriously damaging nature. Small countries and islands suffer much from such events, because of the precocity of their economies and lack of appropriate infrastructure. Mauritius, given its geographiclocationandclimate,isquitepronetosuchextremeevents,eveniffloodsare themostrecurrentones.Theaimofthisresearchstudyistoexplorethepossibilityofthe relevance of an adapted flood study model to the island, and how it could better help understandingthesourceoftheproblem.TheintegratedGISSCShydrologicalmodelis utilized to highlight the flooding problem in Mauritius. Flood sensitive zones were identifiedusingDigitalElevationModel(DEM)andfloodhydrographswerederivedfor twocatchmentsnamelyJandM.Theresearcharticlefurtherdiscusseshowcouplingof thesetwotechniquescouldhelptoidentifyeventhemostcriticalfloodproneareaswithin asubcatchment. Keywords: GIS SCSmodel,Floodhazard,subcatchments,Mauritius. 1.Introduction Sincethebeginningofcivilizationnaturalhazardshaveinfluencedhumanactivities.The frequencyofsucheventshas increasedoverthe years.Usually,regionsaffectedrequire much time and resources for reconstruction (Scaruffi 2008 Crossley 2005). The most common natural hazards are either of geophysical nature Earthquakes, landslides, volcanoes,mudflows,avalanchesorofHydrometeorologicalnatureFloods,droughts, storm surges, cyclones (windstorms), cloudburst, and heat and cold waves. Floods are probably the most recurring, widespread, disastrous and frequent natural hazards of the world.Withtime,peopleandtheenvironmentarebeingconstantlyputunderstressfrom theeffectsofnaturaldisasters. Mauritiusisexperiencingextremeeventssuchaslongspellsofdryseasons,cyclonesand longdurationhighintensityrainstormsetc.Thesemaysubsequentlybringmajorflooding problemsthathavebeenunderminingtheislandforthepastdecades,andthesituationis gettingevenworstovertime.Inthepasttheislandusedtoreceivemorethan60%ofits totalrainfallduringthewetsummerseasons,i.e.fromNovembertoApril,withtheflashy

489

INTERNATIONALJOURNALOF GEOMATICSANDGEOSCIENCES Volume1,No 3,2010

Copyright2010AllrightsreservedIntegratedPublishingservices

Researcharticle

ISSN0976 4380

rainfall patterns lasting for a maximum of two hours. Recently the island has been recording heavy intensity rainfall over long time periods which give rise to flood conditions. Aftermathof floods can also be very incrementaltothe primary effects, asthe affected regiontakesmuchtimetorecoverwherediseasesandepidemicsoutburstsarefrequent. Floods are usually recurrent and have massive impactsonagriculture based economies, whichisthecaseofmostsmallislands.Thussmallcountriesandislandssufferaheavy burdenfromsucheventsontheeconomicaspect.Thus,floodstudiescanbeveryuseful inalleviatingthealreadystressedeconomiesoftheinsularcountries.Thestudyoffloods helpstounderstandthecharacteristicsofthefloods,theircauseandthusproviderelevant data in assessment of these events. Identify the source of the problem can thus help in decision making and contingency planning. Hence the need to delineate flood prone areas is nowadays a priority for the island. In the Flood Risk Management of the European Commission (2005), several issues have been proposed and discussed, pertaining to flood hazard management, and integrated development planning. Eastman (1995)haddiscussedrasterproceduresformulticriteriaanalysis.Intheresearcharticle byYalnandAkyrek(2002)theGeographicalInformationSystem(GIS)isintegrated withMulticriteriaDecisionAnalysis(MCDA)foranalysisofthefloodvulnerableareas. The Decision Support System for flood warning has also reported by few (Abebe et al 2001Catellietal1998Haimesetal1990). 2.Methodology The study was carried out in two stages. The first stage consist of GIS based hydrological model (HEC1) and Digital Elevation Model (DEM), to produce a map showing flood prone areas. In this stage DEM was used to identify the low elevation regions over the island. This map illustrated the vulnerable zones where runoff would accumulateoverthecatchmentorsubcatchmentduringfloodpronerainfallevents.This map also gave more detailed information about the physical characteristics of the catchmentsoverasmallerarea. Thesecondstageofthisstudywastopredictthefloodhydrographslikelytobeobtained during flood type rainfall. A unit hydrograph was derived using the SCS method followedby floodtyperainfalldata,thepeakdischargewassimulatedatthevulnerable subcatchments.Thecontributionofrunofffromvarioussubcatchmentswithinamain catchmentwasalsohighlighted. 2.1 Method1 TheGIS/HEC1Model ArcGISDesktopsoftwarewasusedforthepurposeofthemodeldevelopment.Thefirst stepwastocreateaDEMandforthatahardcopyof10mcontourmapprocuredfrom the local authority, Ministry of Housing and Lands was used. The contours were first digitised using AutoCAD and exported to Arc GIS. Using the Kriging Interpolation method, the DEM was generated and rasterised. From the predefined raster DEM, the 490

INTERNATIONALJOURNALOF GEOMATICSANDGEOSCIENCES Volume1,No 3,2010

Copyright2010AllrightsreservedIntegratedPublishingservices

Researcharticle

ISSN0976 4380

flow definition model was developed. In this study, the DEM developed have the followingproperties: NumberofColumns(478), NumberofRows(552),and Cellsize(109.11x109.11m). 2.2FlowRouting(Definingflowdirections) Theflowdirectionsarecomputedbycalculatingthesteepestdescentfromneighbouring cells,usingthesurfaceelevationvalues.Whenflowdirectionshavebeendefined,aflow accumulationgridisobtained.Thisstepinvolvestheassigningofaunitweighttoeachof thecells,andtheaccumulatednumberofcellsupstreamforanoutletpointcanbeknown, bythesummationofalldownwards flowaccordingtotheflowdirection.Areassmaller thanthethresholdvaluewereconsideredtobedominatedbysheetoroverlandflow.For thisstudy,thethresholdusedwas10cells,whichreferstoanareaofabout0.11km.In essence,sheetflow isconsideredtooccuroveramaximumradiusofabout100m ifthe contributingareaisconsideredasasemicircularregion. Whenthestreamnetworkhasbeendefined(Figure1),thecatchmentforagivenstream is delineated (Figure 2). This hence has defined the catchment grid according to the elevation of the surface. Hence, the natural barriers considered are only the slope and aspect of the land. The subcatchment delineation is derived primarily for a stream. A streamisconsideredandthedrainagearearelativetoitisdelimited.Whenthisstephas beencompleted,thebatchcatchmentdelineationisdone.

Figure1: Catchmentoutline(red),flowlines(blue)andsurfacestorage(green)

491

INTERNATIONALJOURNALOF GEOMATICSANDGEOSCIENCES Volume1,No 3,2010

Copyright2010AllrightsreservedIntegratedPublishingservices

Researcharticle

ISSN0976 4380

Figure2: Catchmentdelineation Incatchmentsdelineationmap(Figure2)itisseenthatthereareseveralareasnotlinked with any basins. Practically, they form part of a drainage area or catchment. They are howevercharacterisedbyarelativelyflatprofilewheredrainageislesslikelytooccur,or have boundaries that have higher elevations than the central parts of that area or the adjacentcatchments.Therefore,flowoccurringovertheseregionsismostlytowardsthe central parts of the area and thus making these regions prone to accumulation. These regionscanhenceberegardedasfloodsensitivezones. Afterthedrainagelinesandpointshavebeenobtained,thelongestflowpathwithamajor catchment can be computed. This links up the outlet for a major basin to the furthest pointupstreamofthedrainagelineforeachsubcatchment. 2.3 Method2:HydrographSimulationsusingtheSCSModel Thepurposeofsecondstagewastosimulateafloodhydrographfortwoselectedbasins. AunithydrographforthesetwobasinswasderivedusingtheoutputoftheGISmodelto work out the SCS curve number. The soil types and land use were analysed for each basinandtheSCScurvenumberwasderivedusingdatafromtables1and2.

492

INTERNATIONALJOURNALOF GEOMATICSANDGEOSCIENCES Volume1,No 3,2010

Copyright2010AllrightsreservedIntegratedPublishingservices

Researcharticle

ISSN0976 4380

Table1:LandUseTypes LandUseType UrbanandRoads SugarcaneandTea CorrespondingSCSLandUse Residential Sugarcane, planted on contours, variable hydrologicconditions Woods,FairHydrologicConditions [Source:NCRS,2004] From the Hydrologic Soil Cover Complexes tables (NEH 2004), the corresponding valuesfortheSCSrunoffcurvenumberswereidentifiedandareincludedinTable2. Table2:SCSLandUseTypesandCurveNumbers SCSLandUseCategory Residential&Paved Sugarcane WoodsNaturalForests HydrologicSoilGroup A B C 57 72 81 25 59 75 36 60 73

ForestandMountains

D 86 83 79

For the areas under consideration, the paved road areas have been combined with the residential use type, given the relatively low area occupied by these, and the sugarcane data is based upon the conditions prevailing in Hawaii and have been adapted for MauritiusIsland. 3.Studyarea The study areas chosen are the catchments of Rivire du Poste and Rivire Des Anguilles.Bothcatchmentsarelocatedinthesouthernregionandarealmostadjacentto one another. The characteristics of each catchment (WRU, 2003) are as described in Table3. 3.1 Basin1:RivireDuPoste(CatchmentCodeJ) This basin is denoted as a major basin bearing the catchment code J. There is an important stream network within this basin, almost all of which contribute to the river commonlyknownasRivireDuPoste.

493

INTERNATIONALJOURNALOF GEOMATICSANDGEOSCIENCES Volume1,No 3,2010

Copyright2010AllrightsreservedIntegratedPublishingservices

Researcharticle

ISSN0976 4380

3.2 Basin2:RivireDesAnguilles(CatchmentCodeM) This basin is a major basin, bearing catchment code M. It also consists of a significant stream network, which contribute to the river, commonly known as Rivire des Anguilles(Table3).

CatchmentJRiverDuPoste

CatchmentMRiverDesAnguilles

Figure3: Imagesshowingcatchmentregion

494

INTERNATIONALJOURNALOF GEOMATICSANDGEOSCIENCES Volume1,No 3,2010

Copyright2010AllrightsreservedIntegratedPublishingservices

Researcharticle

ISSN0976 4380

3.3 Basin1:CatchmentJ For this catchment up to the point of consideration, the land use is mostly made up of naturalforests,andtheremainingpartoftheareaisundersugarcaneplantation.Also,the forestareasarefoundonrockytoveryrockygrounds,requiringaweightedaveragefor thecurvenumber. 3.4 Basin2:CatchmentM The land use type forthis catchment consistsof a mix of sugarcane and tea plantations and some residential areas. Again a weighted average for the curve number was calculated.ValuesobtainedforeachregionaregiveninTable4: Table4:AreaDistributionoverCatchment LandUse% Forest SugarcaneandTea Residential WeightedCurveNumber 3,5TimeofConcentration Thetimeofconcentrationwascalculatedforeachofthecatchmentsabove.TheSCSLag equationwasused.Datarequiredsuchasthelongestflowpathandtheaveragewatershed slopewereextractedfromtheGISwhereverpossible.TheresultsareshowninTable5. Table5:TimeofConcentration Parameters LongestFlowPath(m) Elevationatstart(m) Elevationatstation(m) Averagewatershedslope(%) Hydrauliclengthofwatershed(ft) Lagtime,L(hrs) Timeofconcentration,Tc (hrs) Basin1 6,750 587 351 4.5 22,146 2.07 3.46 Basin2 16,150 605 85 3.6 52,990 4.66 7.76 Basin1 70 30 0 75 Basin2 15 65 20 75

From the relationship L= 0.6 Tc, the time of concentration is hence obtained. These valuesofTc requirefurtherrefining,whichdependsonotherfactorssuchastypeofflow andlanduse.Thiscanhencebeunderstoodtobearoughorrawvalue.

495

INTERNATIONALJOURNALOF GEOMATICSANDGEOSCIENCES Volume1,No 3,2010

Copyright2010AllrightsreservedIntegratedPublishingservices

Researcharticle

ISSN0976 4380

3.6UnitHydrographDerivation Based upon the physical characteristics of the catchments, SCS curve numbers, time of concentration, duration of excess rainfall, the time to peak and peak discharge were obtained for each sub catchment, following a unit rainfall (Table 6). This unit hydrographwasthenusedtopredictpeakdischargelikelytobeobtainedfromfloodtype rainfall,whichtendstooccurduringcyclonicperiods. Table6: TimetoPeakandPeakdischargeforbasins1and2 UnitHydrographcharacteristics Durationofexcessrainfall,D(hr) TimetoPeak,Tp(hrs) PeakDischargem/s Basin1 0.45 2.3 39.20 Basin2 1 1 28.2

Figure4:SubbasinwisedelineationofBasin1 3.7SimulatedFloodHydrograph ThefloodhydrographswereobtainedfromsimulationusingtheHECHMSsoftwareand thefollowingdata:basemapofthedrainageareaandsubbasincharacteristics:subbasin areas,lagtime,SCScurvenumber,rainfalldata,andInitialAbstraction. Basin 1 was then divided into 8 sub basins (Figure 3) to study the contribution to the floodhydrographfromeachsubbasin. 3.8CatchmentCharacteristics For the derivation of the flood hydrographs, several catchment properties were used to derivecharacteristicsandrequiredinputsforthesimulation.Thesearelistedinthetable below:

496

INTERNATIONALJOURNALOF GEOMATICSANDGEOSCIENCES Volume1,No 3,2010

Copyright2010AllrightsreservedIntegratedPublishingservices

Researcharticle

ISSN0976 4380

Table7:CatchmentCharacteristicsandLagTimes
Sub Basin 1 2 3 4 5 6 7 8 Area (km) 5.7 2.5 1 1 1 1 1.5 3 Longest flow Average catchment Curve path(m) slope(%) Number 6,423 2,535 1,883 2,544 2,057 2,038 2,261 3,803 2.65 3.04 6.16 5.03 6.17 6.38 4.87 4.44 75 75 75 75 75 75 75 75 Maximum potential retention(mm) 85 85 85 85 85 85 85 85 Lagtime (mins) 155.88 69.17 38.29 53.89 41.05 40.08 49.87 79.10

Fromthistable,thevariationsbetweenthedifferentsubcatchmentscanbeseeninterms of area, length of longest flow path and slope.These factors incur variations onthe lag times. For this analysis, a single weighted curve number was used. Given the uniform landuse patterns within each single subcatchment, it is preferable to use a single weighted curve number, as compared to the complex derivation of a singular curve numbersforeachofthesubbasins,evenifthelatterwouldhaveprovidedbetterresults. ThelagtimewascomputedusingtheSCSlagequation.Thiswascarriedouttosimulate floodhydrographsfromeachsubbasinfollowingafloodrainfallevent. 3.9StormeventCycloneGerry(11/02/2003 13/02/2003) This storm event is defined by a total precipitation of 228.4mm of rainfall over the 3 days, with an isolated period of intense rainfall the rainfall intensity varying between 10mm/hrto30mm/hrwithinaperiodof7hours(Figure4).Thefloodhydrographwas simulatedusingthesoftwareHEC1.

Figure5: HydrographresultingfromStormevent: cycloneGerry

497

INTERNATIONALJOURNALOF GEOMATICSANDGEOSCIENCES Volume1,No 3,2010

Copyright2010AllrightsreservedIntegratedPublishingservices

Researcharticle

ISSN0976 4380

The outflow at the junction of two sub basins, 1 and 2, was also simulated. Figure 5, illustratesthecontributiontothefloodhydrographfromsubbasins1and2.

Figure6: Floodhydrographscontributionfrom2subbasins 4.DiscussionandConclusion This study aimed to provide an insight on the combined possibilities of the use of geographical information systems and hydrological models in the assessment of a particularnaturalhazard,morepreciselyfloods.Thisproblemwasanalysedwithaview to provide a decisionmaking tool for these events, and thus the applicability of these techniques to such use was analysed. While the GIS study helped in provided more detailedinformationaboutthenatureofthesubbasinsmakingupamainbasin,theSCS 498

INTERNATIONALJOURNALOF GEOMATICSANDGEOSCIENCES Volume1,No 3,2010

Copyright2010AllrightsreservedIntegratedPublishingservices

Researcharticle

ISSN0976 4380

Hydrological Model, was used to simulate the flood hydrograph which was generated from each subbasin. This sub basin wise information helps in identifying which sub basinismorevulnerableandwherecorrectivemeasurescouldbefocussedtoalleviatethe consequencesoffloods. This approach can therefore help in terms of decision making for floodprone regions, where the focus for remedial actions that can be implemented, is shifted from a larger catchmenttothe smaller subcatchment.Thus, further consideration can be provided to theaspectsofplanninganduseofcountermeasuressuchasfloodroutingandprovision ofwaterretainingstructures. Acknowledgements The authors would like to express their sincere thanks to UGC TEC for providing financial support. The authors also wanttothank Dean, Science Faculty, University of Mauritius and Head, Department of Geology, University of Pune for providing the facilities.Atlast,authorswanttothankanonymousreviewersfortheirusefulcomments onthemanuscript. 5.References 1. Abebe,A.J.&Price,R.K.,2001.Decisionsupportsystemforfloodwarningin urban areas. Proceedings of 29th IAHR Congress,Beijing, China, 1621 September2001,C,TsinghuaUniversityPress,Beijing,pp7884. 2. AlSabhan W., Mulligan M., Blackburn G. A., 2003. A realtime hydrological model for flood prediction using GIS and the WWW. Computers, Environment andUrbanSystems27(1),pp932. 3. Catelli,C.,Pani,G.&Todini,E.,1998.FLOODSS:FLOodOperationalDecision Support System. Proceedings of Hydroinformatics 98. Balkema, Rotterdam, 1, pp255262. 4. Chow V.T., Maidment D. R., Mays L.W., 1988. Applied Hydrology, McGraw Hill.MinistryofPublicInfrastructuresWaterResourcesUnit,HydrologyData Book(19951999),2003,GovernmentPrinter. 5. Eastman,J.R.,Jin,W.,Kyem,P.A.K.andToledano,J.,1995.Rasterprocedures formulticriteria/multiobjectivedecisions.PhotogrammetryandRemoteSensing. 61(5)539547. 6. Haimes,Y.Y.,Li,D.&Stakhiv,E.Z.,1990.Selectionofoptimalfloodwarning threshold.InRiskBasedDecisionMakinginWaterResources(ed.Y.Y.Haimes &E.Z.Stakhiv),pp7490.ASCE,NewYork.

499

INTERNATIONALJOURNALOF GEOMATICSANDGEOSCIENCES Volume1,No 3,2010

Copyright2010AllrightsreservedIntegratedPublishingservices

Researcharticle

ISSN0976 4380

7. Malnes,R.,2008.ClimateScienceandtheway weoughttothinkaboutDanger EnvironmentalPolitics17(4):pp660672. 8. Water Resources Unit, 2003. Hydrological data yearbook, Ministry of Public Utilities. 9. Crossley, D., (2008). History: The worst disasters ever. Available at: http://www.livescience.com/history/080506naturaldisasters.html accessed on February,2011. 10. European Commission, 2005. Natural Hazards. Available at: http://ec.europa.eu/research/environment/index_en.cfm?pg=hazards accessed on February,2011. 11. NEH, 2004. National Engineering Handbook:, Section 4 HydrologyAvailable at: http://www.hydrocad.net/neh4.htm accessedonFebruary,2011.

500