Assessing the Influence of Climate Datasets for Quantification of Water Balance Components in Black Sea Catchment: Case Study for Melen Watershed, Turkey Gökhan Cüceloğlu, İzzet Öztürk Istanbul Technical University Department of Environmental Engineering 34469 Maslak/IstanbulTurkey
Outline Introduction Aim and Scope Study Area Model Setup Results Conclusion
Introduction Melen Watershed is the integral part of the Istanbul Water Resources System Quantification the water quantity and quality in Melen Watershed is important. Watershed Model River Water Quality Model Reservoir Water Quality/Ecology Model - Discharge - Water Balance - Diffuse Pollution - Point Sources - Sediment Trans. - Water Quality Parameters - Bathmety, Hydrodynamics - Water Quality Parameters
Aim and Scope Developing a hydrological model for Melen Watershed, Turkey Comparing of local data set with reanalysis data (CFSR) Precipitation is one of the predominant input uncertainity sources Data scarce region Spatial distribution, measurement quality, missing data Representative weather data for watershed Ungauged mountainous region
Study Area (Istanbul, Turkey) Melen Watershed
Study Area (Melen Watershed) BLACK SEA 188 km SEA OF MARMARA Water Treatment Plant Area = 2444 km 2 Q ave = 45.7 m 3 sec -1 Precip = 823mm??? MELEN WATERSHED
Model Setup (Data Sources) Digital Elevation Map SRTM 9m Land Use CORINE Soil FAO Climate Data Set CFSR (1979-214) Local Climate Datasets (196-213) Discharge Data DSI (6 stations)
Model Setup Simulation period (1995) 2 212 5 years warmup 2 Subbasins Hargreaves Method SWAT-CUP using SUFI-2 algorithm 6% of streamflow used for calibration NSE as an objective function Performance criteria NSE, R2, PBIAS as well as P-factor and R-factor
Methodology Change Climate Data Land Use Climate Data Soil SWAT Model Setup Preliminary Analysis Evaluation Elevation Band Adj. Evaluation Parameter Adjustment Evaluation Topography Observed Streamflow Re-Setup Model Calibration Complete ArcSWAT SWAT-CUP
Climate Data Rainfall (mm) 16 14 12 1 8 6 4 2 Station Number Rainfall (mm) 14 12 1 8 6 4 2 Arithmetic Avg. 1715 177 1772 Station Number Weighted Avg.
Discharge Discharge (m3s-1) Discharge (m3s-1) (m3s-1) 25 15 15 2 2 2 2 15 1 15 15 1 15 1 1 5 1 5 5 5 5 D13A33 D13A32 E132 D13A38 E13A4 D13A59 1 4 7 1 13 16 19 22 25 28 31 34 37 4 43 46 49 52 55 58 61 64 1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 13 19 115 121 127 133 Observed CFSR MGM Observed CFSR MGM Observed CFSR MGM Observed CFSR MGM Observed CFSR MGM 1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 13 19 115 121 127 133 139 1 4 7 1 13 16 19 22 25 28 31 34 37 4 43 46 49 52 55 58 61 64 67 7 73 76 79 82 1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 11 16 111 116 1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 13 19 115 121 127 133 139
Elevation Band Adjustment To consider the orographic effects on precipitation and temperature in watershed, Five elevation bands have been applied to model developed by local data (MGM) Tlapse Rate is chosen as -6.5 C / km Plapse Rate is chosen as 6mm/km
Discharge Discharge (m3s-1) Discharge (m3s-1) (m3s-1) 15 1 2 15 2 5 1 15 5 1 5 D13A32 D13A38 D13A59 1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 13 19 115 121 127 133 Observed CFSR MGM 1 4 7 1 13 16 19 22 25 28 31 34 37 4 43 46 49 52 55 58 61 64 67 7 73 76 79 82 Observed CFSR MGM 1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 13 19 115 Discharge Discharge (m3s-1) (m3s-1) 121 127 133 139 Observed CFSR MGM 15 2 12 15 15 1 5 1 5 5 Discharge (m3s-1) D13A32 D13A38 D13A59 1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 13 19 115 121 127 133 Observed CFSR MGM (Elev Adj) 1 4 7 1 13 16 19 22 25 28 31 34 37 4 43 46 49 52 55 58 61 64 67 7 73 76 79 82 Observed CFSR MGM (Elev Adj) 1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 13 19 115 121 127 133 139 Observed CFSR MGM (Elev Adj)
SWAT Model Parameters CFSR MGM (Elev Adj.) SWAT Parameters Initial Range Final Range r CN2.mgt -.5 to.5 -.92 to.2 -.23 to.34 r GWQMN.gw -.5 to.5 -.44 to.18 -.93 to.2 r GW_REVAP.gw -.5 to.5 -.98 to.4 -.83 to.5 r SOL_AWC().sol -.5 to.5 -.86 to.4 -.71 to.9 r REVAPMN.gw -.5 to.5 -.9 to.72 -.82 to.5 r ESCO.hru -.2 to.2 -.5 to.22 -.33 to.2 r ALPHA_BF.gw -.5 to.5 -.33 to.22 -.7 to.78 r SOL_K().sol -.5 to.5 -.3 to.9 -.23 to.28 r SOL_BD().sol -.5 to.5 -.9 to.7 -.68 to.1 48 simulations has been done
Calibration Results Data Set CFSR Station ID Elevation (m) Calibration Validation p-factor r-factor R2 NS PBIAS p-factor r-factor R2 NS PBIAS D13A59 1.76 1.8.82.8 13.1.88 1.16.82.8 1.6 E13A4 23.77.93.8.76 14.7.86 1.1.79.78 9.4 E13A2 115.87 1.21.81.79-6.89 1.19.8.79-2.5 D13A38 276.85 1.74.78.41-27.5.82 1.59.76.44-36 D13A33 276.64.73.7.52 31.8.78.84.7.67 13.6 D13A32 873.47.55.33.1 47.1.6.78.45.31 37 MGM E13A4 23.88.99.85.85 2.8.86 1.1.84.83-3.9 E13A2 115.88 1.27.85.76-19.5.84 1.2.84.79-18.8 D13A38 276.77 1.48.74.59 22.6.89 1.5.71.69 9.4 D13A33 276.54.58.6.4 64.6.74.8.46.25 43.2 D13A32 873.64.99.63.48-32.3.5 1.37.53 -.17-74.1 (Adj. Elev)D13A59 1.92 1.13.89.88-4.2.88 1.14.85.83-1.2 Moriasi et al. 27 Krause et al. 25 Abbaspour et al. 215
Water Budget Components Water Height (mm) 14 12 1 8 6 4 2 *for all basin Water Budget Components* PREP PET ET WYLD SW CFSR MGM U95ppu L95ppu
Conclusion Effects of climate data input are investigated using local data (MGM) and CFSR data in Melen Watershed, Turkey. SWAT model has been succesfully applied to quantify water budget components of Melen Watershed. Global freely available gridded reanalysis data (CFSR) represents watershed better than local data (directly used) in Melen Watershed. After using elevation band features of SWAT model for local data, findings of model results has become quite satisfactory at outlet of watershed Model results for both data set in ungauged (climate stations) and mountainous region of watershed still poorly represented. Using different sources of model input (if possible) is very important to define and quantify of unceartinity sources
Thanks for your attention...