Remote sensing supported hydrological rainfallrunoff modeling Map Algebra
Remote sensing supported hydrological rainfall-runoff modeling Map Algebra – Hazard, Vulnerability and Risk Analysis Boud Verbeiren 1, Jef Dams 1 and Okke Batelaan 1, 2 1 Dept. of Hydrology and Hydraulic Engineering, Vrije Universiteit Brussel (VUB), Pleinlaan 2, 1050 Brussels, Belgium. 2 Geology Division, KU Leuven, Celestijnenlaan 200 e - bus 2410, 3001 Heverlee, Belgium. Luxembourg Earth Observation Day, Bourglinster, November 19, 2008. Boud Verbeiren 19 11 2008 | pag. 1 Department of Hydrology and Hydrauli engineering
Department of Hydrology and Hydraulic engineering Problem statement • Fully distributed hydrological models require spatially distributed input data! DEM RS SOIL ? ? ? METEO LANDUSE RS supported Hydrological Model 19 11 2008 | pag. 2 WETSPA model (Rainfall-Runoff)
Department of Hydrology and Hydraulic engineering MAMUD • MAMUD = Measuring and Modeling Urban Dynamics • Main objective = study the impact of urbanization on hydrology for Tolka River (Dublin) • Improve RAINFALL-RUNOFF modelling by INTEGRATION of Remote Sensing derived information • Adapt Wet. Spa model RS supported Hydrological Model 19 11 2008 | pag. 3
Department of Hydrology and Hydraulic engineering WETSPA DEM Soil Map Land use Meteo data RS supported Hydrological Model 19 11 2008 | pag. 4 Discharge Q = Qs + Qi + Qg Surface runoff Qs Interflow Qi Groundwater discharge Qg Time
WETSPA Department of Hydrology and Hydraulic engineering Digital Elevation Model Accumulation Direction Slope Land-use map Landuse Map Roughness coeff. Stream net Root depth Soil map Hydraulic conductivity Porosity Hydraulic radius Interception capacity Field capacity Flow velocity Evap. parameters Wilting point Residual soil moisture t 0 s RS supported Hydrological Model 19 11 2008 | pag. 5 c D Pore distribution index Wet. Spa Runoff coefficient & depression storage capacity
Department of Hydrology and Hydraulic engineering WETSPA • IDENTIFICATION: Which RS parameters? • IMPLEMENTATION: How implemented? • CODE ADAPTATION & TESTING Not considered: • DEM > SRTM (90 m) • Precipitation, Soil Moisture > RADAR RS supported Hydrological Model 19 11 2008 | pag. 6
![Department of Hydrology and Hydraulic engineering [1] Impervious surface • Runoff linked to imperviousness](http://slidetodoc.com/presentation_image/501143b6a5850ad3d20f64745251627c/image-7.jpg "Department of Hydrology and Hydraulic engineering [1] Impervious surface • Runoff linked to imperviousness")
Department of Hydrology and Hydraulic engineering [1] Impervious surface • Runoff linked to imperviousness Source: www. lakesuperiorstreams. org RS supported Hydrological Model 19 11 2008 | pag. 7
![[1] Adapting the Wet. Spa model: Runoff coefficient Department of Hydrology and Hydraulic engineering](http://slidetodoc.com/presentation_image/501143b6a5850ad3d20f64745251627c/image-8.jpg "[1] Adapting the Wet. Spa model: Runoff coefficient Department of Hydrology and Hydraulic engineering")
[1] Adapting the Wet. Spa model: Runoff coefficient Department of Hydrology and Hydraulic engineering Static Land-use classes Legend Forest 0% Grassland 0 % Agriculture 0% Sub-pixel impervious map Legend Percentage of impervious area per pixel 100% Urban 50% 0% Industry 65% % impervious constant for each class RS supported Hydrological Model 19 11 2008 | pag. 8 % impervious spatially distributed
![[1] Static landuse classes Department of Hydrology and Hydraulic engineering Discharge Virtual city Centre](http://slidetodoc.com/presentation_image/501143b6a5850ad3d20f64745251627c/image-9.jpg "[1] Static landuse classes Department of Hydrology and Hydraulic engineering Discharge Virtual city Centre")
[1] Static landuse classes Department of Hydrology and Hydraulic engineering Discharge Virtual city Centre 30% 30% URBAN 30% 30% Residential Suburb 30% IMPERVIOUS (%) PERVIOUS RS supported Hydrological Model 19 11 2008 | pag. 9 30% FOREST 30% 0% 30%
![Department of Hydrology and Hydraulic engineering [1] Spatially distributed sub -pixel imperviousness Discharge Sub-pixel](http://slidetodoc.com/presentation_image/501143b6a5850ad3d20f64745251627c/image-10.jpg "Department of Hydrology and Hydraulic engineering [1] Spatially distributed sub -pixel imperviousness Discharge Sub-pixel")
Department of Hydrology and Hydraulic engineering [1] Spatially distributed sub -pixel imperviousness Discharge Sub-pixel classification Dept. CCG (VUB) 90% 81% 63% 52% 18% 39% 48% 62% 15% IMPERVIOUS (%) PERVIOUS RS supported Hydrological Model 19 11 2008 | pag. 10 0% 9% 26% 23%
![Department of Hydrology and Hydraulic engineering [1] Results Chormanski et al. (2008) ≠ scenarios](http://slidetodoc.com/presentation_image/501143b6a5850ad3d20f64745251627c/image-11.jpg "Department of Hydrology and Hydraulic engineering [1] Results Chormanski et al. (2008) ≠ scenarios")
Department of Hydrology and Hydraulic engineering [1] Results Chormanski et al. (2008) ≠ scenarios • Spatially distributed land-cover information yields higher peak discharges RS supported Hydrological Model 19 11 2008 | pag. 11
![[1] Results Department of Hydrology and Hydraulic engineering One urban class I. S. =](http://slidetodoc.com/presentation_image/501143b6a5850ad3d20f64745251627c/image-12.jpg "[1] Results Department of Hydrology and Hydraulic engineering One urban class I. S. =")
[1] Results Department of Hydrology and Hydraulic engineering One urban class I. S. = 30% Runoff coefficient Spatially distributed Larger connectivity = Runoff coefficient Higher runoff coefficient Higher discharge • For hydrological modelling in urbanised areas: Impervious surfaces essential • Expert judgement reasonable, but RS estimates better RS supported Hydrological Model 19 11 2008 | pag. 12
![[2] Assimilation ET WETSPA model (Rainfall-Runoff) DEM RS SOIL Discharge Department of Hydrology and](http://slidetodoc.com/presentation_image/501143b6a5850ad3d20f64745251627c/image-13.jpg "[2] Assimilation ET WETSPA model (Rainfall-Runoff) DEM RS SOIL Discharge Department of Hydrology and")
[2] Assimilation ET WETSPA model (Rainfall-Runoff) DEM RS SOIL Discharge Department of Hydrology and Hydraulic engineering Time ? ? ? LANDUSE METEO RS ? ? ? • Can RS be used to obtain « measured » data for calibration purposes? Water balance > EVAPOTRANSPIRATION (ET) RS supported Hydrological Model 19 11 2008 | pag. 13
![Department of Hydrology and Hydraulic engineering [2] Assimilation ET • SEBAL: Estimate actual ET](http://slidetodoc.com/presentation_image/501143b6a5850ad3d20f64745251627c/image-14.jpg "Department of Hydrology and Hydraulic engineering [2] Assimilation ET • SEBAL: Estimate actual ET")
Department of Hydrology and Hydraulic engineering [2] Assimilation ET • SEBAL: Estimate actual ET (day) • Produce ET time series ET 1985 ET ET 2005 FUTURE • Derive ”measured” ET maps for Tolka to check and correct the model (assimilation) RS supported Hydrological Model 19 11 2008 | pag. 14 Page 14
Department of Hydrology and Hydraulic engineering Related research (1) • Ph. D Jef Dams (IWT): Integration of radar precipitation time series • Rainfall radars near Belgium RS supported Hydrological Model 19 11 2008 | pag. 15
Department of Hydrology and Hydraulic engineering RS supported Hydrological Model 19 11 2008 | pag. 16 Radar estimated rainfall
Department of Hydrology and Hydraulic engineering RS supported Hydrological Model 19 11 2008 | pag. 17 Data preparation
Department of Hydrology and Hydraulic engineering Distributed rainfall in hydrological modelling Radar Precipitation 1 map per model timestep RS supported Hydrological Model 19 11 2008 | pag. 18 Calibrated hydrological model
Department of Hydrology and Hydraulic engineering Related research (2) • Ph. D Wiesam Essa: Using thermal information to study urbanisation • Sharpening thermal bands using HR multispectral bands ASTER IMAGE = + TIR: 90 m RS supported Hydrological Model 19 11 2008 | pag. 19 VNIR: 15 m SHARPENED TIR: 15 m
Department of Hydrology and Hydraulic engineering Conclusion • Remote Sensing derived subpixel imperviousness was successfully integrated into the Wet. Spa model RESULTS: – Better estimation of impervious surface > runoff coefficients – Spatially distributed land-cover information yields higher peak discharges • This example shows the potential for Remote Sensing derived parameters in fully distributed hydrological models. • It is expected that use of RS derived information for data input and assimilation will yield more accurate estimates of the model outputs. RS supported Hydrological Model 19 11 2008 | pag. 20
Department of Hydrology and Hydraulic engineering THANK YOU! RS supported Hydrological Model 19 11 2008 | pag. 21
- Slides: 21
