IMPACT OF LAND USE/COVER CHANGES ON STREAMFLOW: THE CASE OF HARE RIVER WATERSHED, ETHIOPIA Kassa Tadele and Gerd Foerch University of Siegen July 06, 2007
Presentation outline 1. Introduction Study area Objective 2. Methodology Model setup Existing watershed practices 3. Results Land use dynamics Model evaluation Seasonal Streamflow variability Downstream impacts 4. Conclusions
1. Introduction 1.1 Study area Abaya-Chamo Basin Hare Watershed Introduction-Materials and Methods-Results-Conclusions
1.2 Objectives of the study Examine the extent of past and present land use/cover dynamics and analyse their implications on streamflow at a watershed and sub-watershed levels Analyse the seasonal streamflow variability and understand the upstream-downstream linkages with respect to irrigation water use Introduction-Materials and Methods-Results-Conclusions
2. Methodology 2.1) Model setup I) DEM and stream network A DEM was derived from digital contour lines A stream network was digitized from top map 15 sub-watersheds and 92 HRUs were created
II) Meteorological Data To establish elevationrainfall relation, 15 weather stations, Elevation bands were developed in SWAT to account for orographic effect of PCP Mean annual (m m ) 1500 1400 1300 1200 1100 1000 900 800 700 y = -0.0002x 2 + 1.1022x - 201.74 R 2 = 0.9573 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 P Poly. (P) Elevation (m) a.s.l.
III) Soil data Sample locations identified (random sampling) Sample were taken to determine physical & chemical parameters Soil polygons were developed for the point location samples
Soil sampling and analysis Texture distribution Sample site preparation Hygrometer Permeability (HC)
IV) Land use/cover mapping Spatial databases were developed using aerial photographs (1967 &1975), satellite image (2004) and intensive on field land use mapping (2005) Hybrid of automated classification (supervised classification based on maximum livelihood approach) and visual interpretation (based on tone, texture, proximity) was adopted post-classification comparison method
V) Streamflow data Observed daily streamflow (1980 2005) at the outlet of the watershed Mean daily discharge 6.0 5.0 Q (m3/s) 4.0 3.0 2.0 1.0 0.0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
2.2 ) Existing watershed practices a) Downstream practices 3 diversions to irrigate 2224 ha (depend on daily streamflow)
b) Upstream practices
3. Results 3.1 Land use dynamics Land use/cover -1967 N Land use/cover 1975 12 12 11 10 15 13 8 11 10 15 13 8 9 9 1 2 3 4 1 2 3 4 14 5 6 14 5 6 7 7 4 0 4 8 12 16 Kilometers Land use/cover -2004 4 0 4 8 12 16 Kilometers 12 15 11 10 13 8 Legend 1 14 2 5 3 9 6 4 Subbasins Land use/cover Farmlands and settlement Grazing/pasture land Wood/Bush land Forest land Riverine tree/bamboo 7 4 0 4 8 12 16 Kilometers
Cont d Farmlands increased mostly associated with a decrease in forest cover 2004 1975 1967 52.0 16.2 11.8 13.8 6.2 28.3 28.4 13.8 17.0 12.5 16.5 34.4 12.6 19.6 16.9 Farmland Forest Woodlands Grasslands Riverine trees 0% 20% 40% 60% 80% 100% Sub-watersheds adjacent to villages 60 50 40 30 20 10 0-10 -20-30 more affected -40 change (%) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Subwatersheds Farmland Forest Woodlands Grass Riverine
I) Sensitivity Analysis (SA) 8 most crucial parameters Curve number (CN), Soil Available Water Capacity (SOL_AWC), Soil depth (SOL_Z), Soil Evaporation Compensation factor (ESCO), Saturated hydraulic conductivity (SOL_K), Slope (SLOPE), Groundwater revap coefficient (GW_REVAP) and Groundwater recession factor (ALPHA_BF) Controls quick flow generation CN Controls overland & lateral flow SOIL_Z SOL_AWC SOL_K ESCO SLOPE Controls Water mov t through soil profile GW_REVAP Controls Slow flow generation ALPHA_BF
II) Calibration and validation 1975 land use/cover map 2004 land use/cover map Index Calibration (1980-85) Validation (1986-91) Calibration (1992-97) Validation (1998-04) Daily Mon. Daily Mon. Daily Mon. Daily Mon. Coef. dete(r 2 ) 0.63 0.72 0.52 0. 55 0.71 0.85 0.62 0.71 N-S coeff. (E) 0.52 0.63 0.43 0.45 0.62 0.82 0.58 0.67 Q (m3/s) 8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0 a) Calibration (1980-1985) Jan-80 Jan-81 Jan-82 Jan-83 Jan-84 Jan-85 Measured Simulated Q (m3/s) b) Validation (1986-1991) 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0 Jan-86 Jan-87 Jan-88 Jan-89 Jan-90 Jan-91 Measured Simulated
3.3 Seasonal streamflow variability (1992-2004) sub-watersheds Farmland & settlement class change (%) Mean monthly flow change (%) Wet season (Mar.-May) Dry season (Nov.-Feb.) 11 + 5.1 + 7.1-13.8 13 + 12.8 + 8.1-26.9 4 + 18.2 + 11.6-31.8 2 + 18.8 + 13.3-39.6 15 + 18.9 + 11,7-43,3 Entire WS + 10.4 + 12.5-30.5
3.4 Downstream impacts on irrigation project Irrigation demand VS water availability Irrigation water demand at the intakes was computed Q (m3/s) 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 Jan-10 Feb-11 Mar-11 Apr-10 May-10 Jun-09 Jul-09 Aug-08 Sep-07 Oct-07 Nov-06 Dec-06 water requirement 75% diversion 100% diversion Water deficit & surplus periods for irrigation were determined Irrrigation water deficit and surplus periods 3.5 3 2.5 2 1.5 1 0.5-0.5 0-1 -1.5-2 -2.5 534.5 ha 75% diversion 100% diversion 671.8 ha Jan-10 Jan-30 Feb-21 Mar-11 Mar-31 Apr-20 May-10 May-30 Jun-19 Jul-09 Jul-29 Aug-18 Sep-07 Sep-27 Oct-17 Nov-06 Nov-26 Dec-16
4. Conclusions Hare watershed had experienced land use/cover dynamics during the past four decades Model performance assessment verified that the model simulation results are dependable and SWAT can be utilized in similar watersheds Simulation results illustrated that land use/cover dynamics has had significant impacts on streamflow at present Hare River only satisfies 15.75% of downstream irrigation water demand