Home

Representative Elementary Watershed Model User Manual

1. 64 12 Summary of I O files of the postprocessor 65 13 Summary of TARDEM system filesi 86 14 Summary of rewanalysis system files 87 15 Summary of preprocessor system files 87 16 Summary of solver system files 88 bk dt OOD ea ew ee 89 90
2. 20 21 22 23 23 25 25 5 The terrain analysis package TARDEM 5 1 The TARDEM package structure 5 1 1 flood exe 00 4 51 2 dBeze wwa 5 1 3 dinf eze 5 14 aread8B eze 5 1 5 areadinf exe 5 1 6 gridnet exe 9 1 7 Net SUPER IIIA 5 1 8 arclinks eze 5 1 9 arcstreams exe 5 1 10 subasinsetup eke 6 The Rewanalysis package 6 1 REWANALYSIS package structure 6 1 1 rewanalysis eke 6 1 2 recanalysis exe 6 1 3 zbins exe 7 The Preprocessor package 7 1 The preprocessor structure l l preprocessor exe 7 1 2 The preprocessor parameter file 7 1 3 Meteorological forcing information 27 27 30 30 31 32 32 33 33 35 35 36 37 37 38 38 39 41 7 1 4 Reservoirs 7 2 The preprocessor I O filesystem 7 2 1 The boundary and initial condition files 2 7 2 2 The Matlab I O filesl 8 The Solver package 8 1 The solver structure amp Li solyer exel aeris a poak AA Mae a a a E e Als 2 8 1 2 The solver parameter fill ooa aaa 8 2 The solver I O filesystem Ss ee ee eee TUUU ee ee ere ee eae ooh eee eae A aeeueeoenie es 10 The river
3. yp y channel subsurface inflow eP es no no y exfiltration gi yes no no channel in and outflow erin erout no es es y Table 1 Summary of internal external and inter REW fluxes 22 3 Model Installation and setup To perform REW model simulations a specific file and directory structure is rec ommended The software is however flexible to allow different directory structures if one wishes to In Table 2 the directory organization is described The directory D data lt name gt contains the specific model project files for a particular applica tion while D exec contains all executable and library files A listing and a short description of all executables and library files is given in Tables 13 17 in Appendices A E directory name directory scope D data lt name gt Project root directory This directory carries the name of the project D exec Executable file repository This direc tory contains all executable files dy namic link libraries and static libraries Table 2 Working directories setup 3 1 Creating the work environment for the model sim ulations The REW model working environment is set up following the steps described next These steps will also be rehashed briefly in the Mosel case Tutorial to be found at the end of this document e Firstly create two basic working directories preferably on the D drive 23 Set the system search path to D exec directory to en
4. Arguments lt name gt Input lt name gt p drainage directions file from D8 method Function Calculates various path lengths and the Horton Strahler stream orders on a per pixel basis Operation By executing gridnet exe three raster maps are generated Output lt name gt plen ArcView exchange format file containing the longest path length to each grid point along D8 directions lt name gt tlen ArcView exchange format file containing the total path length to each grid point along D8 directions lt name gt gord ArcView exchange format file containing the Strahler order for grid network defined from D8 flow directions 5 1 7 net_setup exe Group TARDEM Command net_setup lt name gt Arguments lt name gt m p1 p2 p3 p4 tresh xy coord_x coord_y pl p2 p3 p4 are parameters for the particular algorithm chosen see Table 5 thresh is the stream threshold area expressed in number of pixels coord_x coord_y are the outlet coordinates m and xy are argument separators 33 Input lt name gt fel filled DEM lt name gt sd8 slope matrix from D8 method lt name gt ad8 accumulated area matrix from D8 method Function Extracts stream network from a digital elevation model Operation For the extraction of the channel network the location of the channel outlet has to be indicated first The selection of the outlet coordinates is performed most easily by loading the raster file which contains the Strahler orders n
5. By executing recanalysis ere the unsaturated zone in the REW is sepa rated into columns based on the respective combination of land use soil and eleva tion information Output lt name gt rec dat an ASCII file showing the REW connectivities and ge ometries The Matlab file lt name gt rec mat containing the REW and the REC information The file lt name gt w an ArcView exchange format file containing a grid with REC masks 6 1 3 zbins exe Group REWANALYSIS CommanD zbins exe lt name gt dz Argument lt name gt dz Input lt name gt fel Function Generates a raster map where the DEM is separated into different eleva tion categories with a distance dz im metres A possible value for dz is for example 100m meaning that RECs are formed on the basis of zones with 100 m surface elevation intervals Operation By executing zbins ere the elevation class file lt name gt z is generated zbins exe is only used when a subdivision of the REWs on the basis of elevation zones is required e g for simulations in presence of snow Output an ArcView exchange format file containing a grid with REC masks based on elevation categories called lt name gt z see also input files for recanalysis exe 39 The following Table 6 summarizes all I O files of Rewanalysis Rewanalysis files input file name file information directory lt name gt rew mat Matlab file containing summary D data lt name gt tardem informati
6. flow zone as shown in Figure The length scales over which piezometric head differences are dissipated between adjacent REWs is an unknown quantity that is re calculated at chosen time steps by respecting the Kirchhoff laws For this scope the Hardy Cross 1936 network balancing method is used see Figure 6 Given a piezometric head distribution calculated from the mass balance for the saturated zone of each REW at a given point in time and given known groundwater losses across the watershed boundaries the dissipation length scales are calculated by successive approximation The procedure is parsimonious and is based on a non linear system of equations that preserve continuity at each network node and energy expressed as head losses along a closed triangular loop Kirchhoff laws as shown in Figure 6 The aquifer flow field velocity is subsequently calculated by resolving the momentum balance equation for the REW elements The inertial term is neglected under assumption of 15 Loop 124 yA Figure 6 Implementation of the Kirchhoff laws for the groundwater system conceptualized as a resistor network slow flow thus effectively reducing the momentum equation to a system of algebraic equations An example of a vector of aquifer flow velocities for the river Geer groundwater system Belgium is shown in Figure The REW average groundwater levels are interpolated at selected time steps through bicubic spline functions providing a sm
7. for drainage network extraction Operation By executing flood exe on a raw DEM anew de pitted DEM is generated Output lt name gt fel ArcView exchange format file containing the de pitted filled and processed DEM 5 1 2 d8 exe Group TARDEM Command d amp lt name gt Arguments lt name gt Input lt name gt fel filled DEM file Function Processing a digital digital elevation model and deriving drainage di rections with the D8 algorithm Eight possible drainage directions towards the 8 neighbouring pixels differing by II 4 radians angles are considered 30 Operation By executing d8 exe on a filled DEM a drainage directions map is gen erated Output lt name gt p ArcView exchange format file containing a raster map with the drainage directions as an integer between 1 and 8 lt name gt sd8 amp ArcView exchange format file containing a raster map with the local surface slopes in radians between 0 and II 2 radians 5 1 3 dinf exe Group TARDEM Command dinf lt name gt Arguments lt name gt Input lt name gt fel filled DEM file Function Processing a digital elevation model and deriving drainage directions with the Doo algorithm Infinite possible drainage directions between 0 and 211 radians angles towards the 8 neighbouring pixels are considered Operation By executing d8 ere on a filled DEM a drainage directions map is gen erated Output lt name gt ang ArcView exchange format file containi
8. from D data mosel tardem into the directory D data mosel matlab e Change directory to D data mosel e Execute the dos command preprocessor exe addParams skipParams mosel The flag addParams causes all parameters to be overwritten with the values read from the paramter file mosel prm while the flag skipParams will run the pre processor without overwriting the parameters also if the values in the param eter file have changed e Edit the boundary conditions file D data mosel ascii mosel be and verify that the flags indicating the flux noflux boundary conditions are set to 0 for REWs within the watershed already set to zero by default to 1 for REWs that have a no flux boundary in common with the external watershed bound ary and to 2 for REWs that have a flux boundary in common with the external watershed boundary e Re run the preprocessor executing the dos command preprocessor exe addParams mosel This step makes sure that all boundary conditions are set as specified 75 in the mosel bc file e After the execution the binary files forcing mat moselrec1 mat and mosellinks1 mat are to be found in the Matlab file directory D data mosel matlab 10 7 Running the Solver The actual simulations and solution of differential equations is performed by the solver kernel The parameters for the solver are set in the parameter file mosell prm The solver is executed by going through the following steps e Execute the dos co
9. gt logs inoue mat Matlab file containing the inter polated water table at sucessive time steps The matrices can be plotted in Matlab with the surf command DA data lt name gt matlab Rew000r out ascii files containing the solver simulation results For each REW one file is written The file name carries the REW number x D data lt name gt results 60 REsolve000x out ascii files containing the Richards D data lt name gt results equation results For each REW one file is written The file name carries the REW number x Table 10 Summary of I O files of the solver 9 The Postprocessor package The postprocessor is an application which transforms the output of the REW model in formats of choice The native output of the model is given is ASCII tables These can be transformed into either XML files for import into FEWS or to binary files in Matlab file format Transformation into of the output into NetCDF format is also possible Additional file formats will be added if needed in future applications 9 1 The postprocessor structure The postprocessor package consists of a single application which is described here under 9 1 1 Postprocessor exe Group POSTPROCESSOR Command postprocessor exe 61 Arguments lt name gt Input lt name gt 2 prm REW000z out Function transforms output of solver exe into different formats Operation By running postprocessor
10. if the XML switch is selected in the lt name gt 2 prm file The file names carry the respective REW number x D data lt name gt results lt name gt results mat Matlab file containing the solver simulation results One single files is written for all REWs if the Matlab switch is selected The file can be loaded directly into Matlab D data lt name gt matlab lt name gt results nc NetCDF file containing the solver simulation results One single files is written for all REWs if the netcdf switch is selected The file can be visualized with a NetCDF visu alization tool such as ncBrowse D data lt name gt results Table 12 Summary of I O files of the postprocessor 65 10 The river Mosel case tutorial The present chapter presents an application of the REW model to a study case the river Mosel a 60 000 km river basin in Germany For the tutorial the following data are provided e The 75x75 m digital elevation model DEM for the Mosel basin file mosel asc e A NetCDF file containing the P T Etp Rh Dt time series and called mosel stations nc e A boundary conditions file mosel bc 10 1 Creating the working directories To begin with we need to set up all necessary working directories to operate the program Firstly we need to set up all the program executables to install the model then we need to set up a project directory for the specific watershed we
11. next The parameter file is organized in a series of data blocks separated by hashed lines Block 1 contains hydraulic information such as Manning coefficients and the Leopold and Maddock at a station and downstream channel geometry Moreover the hy draulic conductivity of the channel bed and the thickness of the aquifer channel bed transition layer is assigned Block 2 contains the data to be attributed uniformly to the subsurface zone These include hydraulic conductivity porosity soil parameters for the unsaturated zone porosity conductivity and thickness of the shallow subsurface flow layer and quanti ties to specify the initial water content in the unsaturated zone as well as the initial 59 position of the water table Block 3 contains all the information inherent to kriging and block kriging and the time series The possible options include the choice of variogram type the sill nugget and other variogram parameters the recording frequency of the P T Etp Rh Dt data The no data missing value flag used in the data series need to be indicated as well Block 4 contains the data inherent to the finite element method used for fitting a bedrock surface form irregular data points or a smooth water table across the REW average groundwater elevation points see also solver exe The fitting method is based on the Inoue 1986 algorithm The values that need to be given as input are the Finite Element mesh bins along x and y t
12. number of surface pixels forming the threshold area for the beginning of a stream channel The coordinates outlet_x outlet_y are the watershed outlet coordinates selected inter actively as described earlier For the present tutorial we choose a parameters combination with pre set values Type in the following command net_setup exe mosel m 4 4 1 05 200 xy 399106 5579658 The application is executed generating the output files mosel src and mosel ord 70 The first one contains a network mask based on channel source rules the second one a grid with Strahler order for mapped stream network Moreover two data files moselcoord dat and moseltree dat are written These are both ASCII files describing the stream pixel coordinates and the networks structure It is possible to experiment with other network extraction methods 10 3 3 Extracting the subbasins The subbasins are extracted with the application subbasinsetup exe e Type the command subbasinsetup ere mosel order_tresh e For the present example select order_tresh equal to 2 A raster map moselw containing the masks for the sub basin areas is generated The sub basins are identified on the basis of an accumulation analysis for the individual network links For the visualization in GIS a number of shape shp shx and data base files dbf are generated Two shape file groups are thus written mosel shp mosel shz mosel dbf containing the polygon lines delineating the n
13. pp 80 Luce C H D G Tarboton and K R Cooley 1998 The influence of the spatial distribution of snow on basin averaged snowmelt Hydrol Proc 12 10 11 1671 1683 McCarthy G T 1940 Flood Routing Chap V Flood Control The Engineer School Fort Belvoir Virginia pp 127 147 Naden P P Broadhurst N Tauveron and A Walker 1999 River routing at the continental scale use of globally available data and an a priori method of parameter estimation Hydrol Earth Syst Sci 3 1 109 124 Nash J E 1953 The form of the instantaneous unit hydrograph IUGG General Assembly of Toronto Vol III IAHS Publ 114 121 Peuker T K and D H Douglas 1975 Detection of surface specific points by local parallel processing of discrete terrain elevation data Comput Graphics Im age Process 4 375 387 Reggiani P M Sivapalan and S M Hassanizadeh 1998 A unifying framework of watershed thermodynamics 1 Balance equations for mass momentum energy and entropy and the second law of thermodynamics Adv Water Resour 22 4 367 398 81 Reggiani P M Sivapalan M and S M Hassanizadeh 2000 Conservation equa tions governing hillslope responses Water Resour Res 38 7 1845 1863 Reggiani P M Sivapalan S M Hassanizadeh and W G Gray 2001 Coupled equations for mass and momentum balance in a stream network Theoretical deriva tion and computational experiment
14. the preprocessor are generated The entire analysis starts from a raw DEM file in ESRI ArcGrid ASCII exchange format lt name gt asc of the study site where lt name gt is the project name prototype An example for a typical header of an Esri exchange format file is shown in Table 4 A Digital elevation model in the ESRI ArcGrid ASCII exchange format can be exported from most third party GIS software applications ncols 4868 nrows 6012 xllcenter 100930 yllcenter 5220311 cellsize 30 nodatavalue 9999 0 6 0 7 0 9 1 1 Ves 1 6 1 7 1 8 2 0 QA si Table 4 Example of ESRI ArcGrid ASCII exchange format file IMPORTANT e It is essential that the DEM is prepared such that it amply covers the entire area of the watershed If a part of the watershed is clipped and thus portions 28 of the watershed lie outside the boundaries of the digital elevation model TARDEM is not able to complete the analysis To run the applications the extension asc if present needs to be removed from the DEM file lt mosel gt asc to retain a file name without extension As good practice it is suggested to name the DEM with a name prototype lt name gt which is in line with the denomination of the project e g lt mosel gt indicating that an analysis of the Mosel river basin is performed The project name forms the file name prototype for all output files generated in the downstream analysis Table in Appendix A summarizes the command l
15. the watershed name as project name E g in case of a project for the river Nile the model directory will be called D data Nile In the parameter files the paths chosen for the various directories to from which the model reads writes various I O files need to be set Table 3 shows how the paths are set in the parameter files given the directory structure suggested in Table If the user decides to change the directory structure i e chooses other directory names or deeper subdirectory layers for the I O of the files the path indications need to be updated correspondingly in the parameter file 4 Organization of the software application The entire REW model consists of a series of applications or modules stand alone executable files that can be bundled into four principal software components or packages e The TARDEM watershed analysis package e The REW REC analysis package Rewanalysis e The Preprocessor e The Solver e The Postprocessor 25 The working of the four components will be explained in separate sections The applications are compiled to run as console applications and are not embedded in any graphical user interfaces GUI The chosen I O output formats of the REW model allow the use of third party software or open source products such as common text editors ArcView ArcGIS ILWIS GRASS open source Matlab Ocatve open source R open source or similar for reading and manipulating I O files 4 1 Source code compi
16. to be set manually to 2 otherwise it is left equal to 1 i e the preprocessor sets by default all boundaries as no flux boundaries 50 IMPORTANT e When changing the lt name gt bc manually it is always necessary to re run the preprocessor exe in order to update the Matlab files If this is not done the changes will not take effect e If a parameter value e g porosity initial water table condition initial soil moisture content in the parameter file lt name gt prm is changed the old lt name gt bc file needs to be renamed or cancelled because preprocessor exe will read in the existing file Thus in absence of a lt name gt bc file a new one will be written in which the flux no flux boundary conditions need to be spec ified by re setting the appropriate flags correctly Afterwards preprocessor exe needs to be re run again in order to update the Matlab files e The groundwater module makes use of the Hardy Cross 1936 algorithm to balance discharges in respect of the Kirchoff laws In this context it is necessary that at least one REW has a flux boundary In case of doubt it is best practice to set the outlet REW with a flux boundary If all boundary REWs are set as no flux i e the flags are left equal to 1 the Hardy Cross algorithm for the determination of the horizontal groundwater exchange parameters will not be able to converge The following Table 8 provides an overview of all I O files read and written b
17. want to model For this purpose we perform the following steps on a virgin hard drive of a Windows works station e Create a directory D exec on the D drive whereby any other drive besides D can be used if needed It is recommended however not to install the model and project directories on the drive where the operating system resides e Create the project root directory D data mosel e Create the subdirectories D data mosel stations D data mosel matlab 66 D data mosel logs D data mosel tardem D data mosel ArcView D data mosel ascii D data mosel results e Add the path D exec to the system path by accessing the Windows settings under Control Panel System Advanced Environment Variables 10 2 Installing executable and project files e Copy all executable files and libraries into the D exec directory on the D drive e Copy the DEM file mosel asc into D data mosel tardem e Rename mosel asc the Digital Elevation Model file to mosel without exten sion e Copy the parameter file mosel prm mosell prm and mosel2 prm into the project root directory D data mosel 10 3 Running the TARDEM terrain analysis package 10 3 1 DEM preprocessing Once the system and file are set up the terrain analysis package TARDEM needs to be executed In principle the first analysis steps can be included in a prepare bat batch execution file and run all at once However in order to make the individual step mor
18. 3 p4 tresh lt name gt ord order and drainage network coordinates xy coord_x coord_y lt name gt tree dat lt name gt coord dat pi p4 are parameters for the particular algorithm chosen and thresh is the stream threshold area expressed in pixels co ord_x coord_y are the outlet coordinates m and xy are argument separators arclinks exe lt name gt lt name gt li e00 Determination of network of the links Generates an output file in the GIS ex change format e00 85 Determination of the streams Generates an output file in the GIS exchange format st00 arcstreams exe lt name gt st e00 lt name gt subbasinsetup exe lt name gt vw lt name gt tresh lt name gt w sha lt name gt w shp lt name gt w dbf lt name gt shz lt name gt shz lt name gt dbf Determinates the polygon shapes and drainage areas of the various sub basins contributing to the network links The output files are ArcView shape files con taining the info for the visualization of the REWs and the channel network Table 13 Summary of TARDEM system files B REWANALYSIS system files executable files and argu ments output file names scope revanalysis exe lt name gt lt name gt links mat lt name gt rew mat lt name gt links dat lt name gt rew dat Performs geometric anay sis of REWs recanalysis exe lt name gt lt name g
19. 8 Figure 10 Pixels in the moselgord asc file coloured according to the Strahler order e Rename the file moselgord into moselgord asc Import the file into ArcView and visualize it on the screen e Change the colours such that the cell corresponding to higher order pixels stand out clearly as shown in Figure 10 e Identify the outlet which is positioned at the UTM coordinates x 399106 and y 5579658 e Remember these two coordinates 69 10 3 2 Extracting the river network The following step is the extraction of the stream network with the aid of the application net_setup eze TARDEM offers a series of possibilities on how to extract the channel network from the DEM It is possible for example to use the threshold area method or other methods which look at the curvature of the land surface For more in depth explanation the reader is referred to the description of the TARDEM package in Section If we decide for example to employ the accumulation area of upward curved grid cells method method 4 in Table 5 for the identification of the channel network the following command line arguments are passed to the application net_setup ere lt name gt m pl p2 p2 p4 tresh xy outlet_x outlet_y whereby the arguments pi p4 indicate specific parameter values to be used for TARDEM tresh is the threshold accumulated area beyond which it is assumed that stream channels begin The threshold area is expressed in terms of the
20. I O file system 7 2 1 The boundary and initial condition files The directory D data lt name gt ascii contains the ASCII file lt name gt bc and lt name gt init the first one being the file containing the model initial and boundary conditions and the second one a log file of the watershed geometric quantities such as link lengths surface areas contributing areas and link areas etc The boundary condition file lt name gt bc is read by default by the preprocessor if already present If no lt name gt bc file is found a new one is created by re running the preprocessor The file is written based on the parameter values set in the lt name gt prm file By default constant and spatially uniform values for the various hydro dynamic parameters are assigned in the watershed model If spatially distributed values for quantities such as parameter values or various initial conditions e g initial soil moisture content of the soil column initial water table positions etc are required for a particular watershed the lt name gt bc file needs to be edited manually and the values changed An important aspect is to set the flags for the flux no flux boundary conditions for the external boundary in the watershed By default the flag is set to 0 if a REW is situated within the watershed i e has no external boundaries or equal to 1 if the REW has an part of an external boundary If that boundary is to be modelled as a flux boundary the flag needs
21. Mosel case tutorial 10 1 Creating the working directories 10 2 Installing executable and project files 10 3 Running the TARDEM terrain analysis package 2 10 3 1 DEM preprocessing 10 3 2 Extracting the river network 10 3 3 Extracting the subbasins 10 4 Extracting Representative Elementary Watersheds REWs 10 5 Extracting Representative Elementary Columns RECs 73 be ee Gee a Bee 75 ee ne ee ee eee TUUU 76 ee ne ee se E 77 78 78 85 86 87 88 88 1 Introduction The REW model is an integrated hydrological simulation tool which has been developed to simulate the entire hydrological cycle including the saturated and un saturated zone channel and overland flow The model is suited for water yield and water balance studies as well as for rainfall runoff simulations The REW model is based on the integration of point scale conservation equations for mass and momentum to the scale of characteristic control volumes called Representative Elementary Watersheds REWs The underlying concepts are described in the papers by Reggiani et al 1998 1999 2000 2001 and Reggiani and Rientjes 2005 2010 The REW is a spatially distributed model which preserves the description of mass exchange and flow based on physical principles such as piezometric head differences and gravity in contrast to lumped con
22. Qout h 1 where I is the inflow from upstream areas or REWs and Qout is a tabulated con trolled outflow It is not possible to introduce more than one reservoir per REW The information for the reservoirs is stored in the reservoirs mat file which is if not already existent created as a template by the preprocessor This requires that in Block 4 in the file lt name gt prm at least one REW is indicated as containing a reservoir For each REW reservoirs mat contains a table with 5 rows pre set default row number and 4 columns All rows and columns contain zeroes The file can be edited in Matlab by loading it Now the first column needs to be filled with absolute water level data the second column with corresponding surface area values The third column is filled with absolute water level data and the fourth column with discharge values at the reservoir outflow corresponding to the different water levels in column 3 Column 1 and 2 constitute a tabulated characteristic A h reservoir curve and column 3 and 4 a characteristic Qout h reservoir outflow curve After the values have been inserted the file is saved as reservoirs mat by overwriting the template file initially created by the preprocessor Successive runs of the preproces sor will leave the file unaltered unless additional reservoirs have been added in the meantime in the parameter file The reservoirs mat file will subsequently be used by the solver 49 7 2 The preprocessor
23. Representative Elementary Watershed Model User Manual P Reggiani WaterIntellect October 2 2012 atmosphere REW boundary unsaturated zone u mantle concentrated surface A overland flow c water table saturation overland flow 0 channel reach r Abstract The present document is a technical and user manual for the use of the Representa tive Elementary Watershed REW model The document provides a brief outline of the modelling approach a detailed description of the usage of the software pack ages To introduce new users to the model a tutorial for the river Mosel basin as a test case is provided Contents 1 Introduction ee re eee 2 3 Sub REW variability tee 2 4 1 Unsaturated zone U zone 2 4 2 Saturated zone S zone 2 4 3 Saturation excess flow Dunne type flow O zone 2 4 4 Subsurface Stormflow P zone 2 4 5 Infiltration excess flow Horton type flow C zone 2 4 6 Snow layer F zone 2 4 7 Channel flow R zone 2 4 8 Summary of exchange fluxes in the REW model 3 Model Installation and setup 3 1 Creating the work environment for the model simulations 3 2 Model directory path settings 4 Organization of the software application 4 1 Source code compilation 4 2 Compatibility with third party software 10 12 13 15 17 19 19
24. Rientjes 2010 The solution of a system of non linear Ordinary Differential Equations and Richards equation including the groundwater flow module are performed by optimized C code The preparation of input data can be facilitated by using appropriate GIS software like ArcView The following sections describe the basic hydrological pro cesses simulated by the REW model 2 Modelled processes 2 1 Model capabilities The REW model is a complex hydrological simulation tool which is designed and developed for the simulation of the entire hydrological cycle of a watershed system underlain by a regional aquifer which extends beyond the topographic boundaries of the watershed The modelling tool can be used for a series of hydrological studies which look at different components of the hydrological cycle and at processes that play a role at different time scales For example it can be used for event based studies such as the response of a watershed to an extreme hydrological event or the behaviour of the hydrological system under forcing conditions that are changing over longer time periods Examples of possible applications and hydrological studies are 1 hydrological water balance 2 rainfall runoff studies 3 groundwater recharge and development studies 4 impact of climate change on the hydrological cycle 2 2 Spatial discretization of the landscape In the REW model a watershed is partitioned into a series of discrete spatial un
25. The binary file moselresults nc can be inspected with the NetCDF browsing tool and also imported into Delft FEWS like the XML files 11 Conclusions The present document has given a comprehensive introduction to the use of the REW model If you have additional questions please do not hesitate to contact Paolo Reggiani at paolo reggiani deltares nl Have fun 12 References Abbott M B Bathurst J C Cunge J A O Connell P E and Rasmussen J 1986a An Introduction to the European Hydrologic System Systeme Hydrologique Europ en SHE 1 history and philosophy of a physically based distributed model ing system J Hydrol 87 45 59 for reference see http www epic noaa gov java ncBrowse 78 Abbott M B Bathurst J C Cunge J A and O Connell P E 1986b An In troduction to the European Hydrologic System Systeme Hydrologique Europ en SHE 2 structure of a physically based distributed modelling system J Hydrol 87 61 77 Band L E 1986 Topographic partition of watersheds with digital elevation models Water Resour Res 22 1 15 24 Bergstr m S 1995 The HBV model In Singh V P Ed Computer Models of Watershed Hydrology Water Resources Publications Highlands Ranch CO pp 443 476 Cross H 1936 Analysis of Flow in Networks of Conduits or Conductors Bull 286 Univ of Ill Urbana Chow V T 1968 Handbook of Hydrology MCGraw Hill New York N
26. The runoff on the saturated areas is calculated by analytical solution of the mass and momentum balance equations for overland flow i e kinematic wave The overland flow zone discharges laterally into the river channel yielding a lateral channel inflow flux e The saturation excess zone is fed directly by precipitation and is exposed to potential evaporation during dry periods periods If infiltration excess flow is generated on the unsaturated part of the REW surface the infiltration excess flow is discharged into the saturated 18 overland flow zone through a flux e 2 4 4 Subsurface Stormflow P zone Subsurface stormflow is generated in a shallow layer beneath the surface with high conductivity For some watershed the use of this zone is essential to capture certain rapid runoff phenomena This zone can also be used to represent a perched aquifer system thus the denomination P zone which constitutes a shallow suspended reser voir of groundwater The subsurface stormflow or the perched system is fed by direct infiltration of precipitation and discharges towards the channel through the flux term e In case of saturation of the subsurface layer the excess flow is dis charged directly into the saturated overland flow zone as flux e The governing equations for the subsurface storm flow are the mass and momentum balance equa tions for subsurface flow which are combined into a kinematic wave equations and solved anal
27. The values that need to be given as input are the Finite Element mesh bins along x and y the number of initial subdivision of the length and width into a grid cells that are interpolated via a cubic spline the maximum number of Gauss Seidel iterations the roughness and tension of the sur face to be fitted Some more specific parameters are generally left at default values For a general explanation of the interpolation method the reader is referred to the cited paper For general applications it is recommended to leave the parameter and the mesh sub division information at their pre selected default values Block 7 contains the paths setting referring to the I O directories of the particular model following the structure D data lt name gt matlab D data lt name gt stations D data lt name gt ascii D data lt name gt logs D data lt name gt tardem D data lt name gt results 44 The directory D data lt name gt matlab contains the Matlab files lt name gt rew mat and lt name gt rec mat lt name gt links mat written by rewanalysis exe and recanaly sis exe These are most conveniently copied or moved after the execution of the two programs from the directory D data lt name gt tardem into the directory D data lt name gt matlab This can be done either manually or with the aid of a batch file The directories D data lt name gt logs and D data lt name gt tardem point to a log file repository and to the TARDEM I O fil
28. Y Crawford N H and R K Linsley 1966 Digital simulation in hydrology Stanford Watershed Model IV Technical Report No 39 Stanford University Palo Alto CA Cunge J A 1963 On the subject of a flood propagation computation method Muskingum Method Delft The Netherlands J Hydr Res 7 2 205 230 79 Dooge J C I 1974 Linear Theory of Hydrologic Systems USDA Tech Bull 1468 U S Department of Agriculture Washington DC Hyami S 1951 On the propagation of flood waves Bulletin no 1 Disaster Pre vention Research Institute Kyoto University Japan Inoue H 1986 A least squares smooth fitting for irregularly spaced data Finite element approach using the cubic B spline basis Geophysics 51 11 2051 2066 Liu Z and E Todini 2002 Towards a comprehensive physically based rainfall runoff method Hydrol Earth Syst Sci 6 5 859 881 Liu Z and E Todini 2004 Assessing the TOPKAPI nonlinear reservoir cascade approximation by means of a characteristic lines solution Hydrol Processes 19 10 1983 2006 Liu Z M L V Martina and E Todini 2005 Flood forecasting using a fully distributed model application of the TOPKAPI model to the Upper Xixian Catch ment Hydrol Earth Syst Sci 9 4 347 364 Leopold L B and Maddock T Jr 1953 The Hydraulic Geometry of Stream Channels and Some Physiographic Implications U S Geological Survey Professional Paper 252 56
29. able R v R m R w R p Table 9 Solver parameter file 8 2 The solver I O file system The following Table 10 summarizes all input and output files of the solver input file name file information directory lt name gt recl1 mat Matlab file containing the up dated information on REWs and RECs D data lt name gt matlab lt name gt links1 mat Matlab file containing the up dated stream links structure D data lt name gt matlab timeinfo mat Matlab file containing informa tion such as recording time step start and end time of the forcing series D data lt name gt matlab forcing_5x mat Matlab file containing the meteo forcing at the REW centroid D data lt name gt matlab 59 reservoirs mat Matlab file containing the A h and Q h tables for the reser voirs The tables need to be filled in with the data of the reservoir characteristics D data lt name gt matlab lt name gt 1 prm parameter file for solver D data lt name gt output file name file information directory globals log general log file for the solver D data lt name gt logs loops log log file for the groundwater net D data lt name gt logs work resolution inoue log log file for the surface interpola D data lt name gt logs tion masserror log log file containing mass balance errors D data lt name
30. able 8 provides an overview of all I O files in Matlab format used and written by the preprocessor 8 The Solver package For the actual simulation the the console application solver exe is executed The application reads the parameter file lt name gt 1 prm and a series of Matlab files listed in Table 8 1 The solver structure 8 1 1 solver exe Group SOLVER Command solver exe lt name gt Arguments lt name gt Input lt name gt 1 prm lt name gt recl mat lt name gt link1s mat lt forcing gt mat Function Solves equations and performs hydrological simulation Operation By running solver exe with appropriate input and parameter files the 54 REW model is activated Output The timeinfo mat time info inoue mat interpolated water table surface dump _state0000 mat files internal model states the output files REW000z out REW scale variables and REsolve000x out Richards equation solution 8 1 2 The solver parameter file The parameter file lt name gt 1 prm allows to set the time information to run the model such as start and end time time step size output step size and others Moreover there are parameter for the Richards Equation solver Ross 2003 and for the fitting of the water table surface between REWs which have already been explained in the context of the lt name gt prm file of the preprocessor An expla nation of the various blocks making up the lt name gt 1 prm shown below are given
31. able program execu tion from anywhere within the system The Windows path in NT ME XP can be set as an environment variable Under Windows go to Control Panel System Advanced Environment Variables SystemVariables Path and add D exec to the path environment variable Create the basin specific root directory d data lt name gt where lt name gt refers to the river basin to be modelled e g mosel if the Mosel basin is modelled in a specific project Under the project root directory D data lt name gt it is necessary to set up a range of subdirectories which allow to organize the necessary working and output files The subdirectory structure shown in Table 3 must be created A full description of files in each subdirectory is given in subsequent sections search path settings in parameter files meteo files D data lt name gt stations Matlab files D data lt name gt matlab log files D data lt name gt logs ASCII files D data lt name gt ascii TARDEM files D data lt name gt tardem simulation output files D data lt name gt results Table 3 I O file path settings for model setup 24 3 2 Model directory path settings The module preprocessor exe reads the parameter file lt name gt prm while solver exe reads lt name gt 1 prm and preprocessor eze reads lt name gt 2 prm The parameter files should be conveniently located in the project root directory D data lt name gt It is recommended to use
32. ative and water storage consistent variable param eter Muskingum Cunge approach Hydrol Earth Syst Sci 11 1645 1659 2007 doi 10 5194 hess 11 1645 2007 Todini E and Ciarapica 2001 The TOPKAPI model Chapter 12 in Mathemati cal models lof large watershed hydrology V P Singh et al Edts Water Resources Publications Littleton CO Van der Kwaak J E and K Loague 2001 Hydrologic response simulations for the R 5 catchment with a comprehensive physics based model Water Resour Res 83 37 5 999 1013 Zhao R J 1992 The Xinanjiang model applied in China J Hydrol 135 371 381 84 A TARDEM system files executable files and arguments output file names scope flood exe lt name gt lt name gt fel Pit removal from DEM d8 exe lt name gt lt name gt p lt name gt sd8 Drainage directions along eight 45 di rections D8 dinf exe lt name gt lt name gt ang Drainage directions along infinite possible lt name gt slp angles between 0 and 360 Doo aread8 exe lt name gt lt name gt ad8 Contributing area determination based on D8 method areadinf exe lt name gt lt name gt sca Contributing area determination based on Doo method gridnet ere lt name gt lt name gt plen Flow path length calculation lt name gt len lt name gt gord net_setup exe lt name gt lt name gt scr Setup of the drainage network stream m pl p2 p
33. ceptual models which are based on a sys tem of inter linked reservoirs for which mass exchange terms are parameterized in terms of simple power laws or other type of ad hoc relationships Typical exam ples of lumped conceptual models are the Swedish HBV Bergstr m 1995 or the Sacramento model In principle the REW model is not fully distributed such as the SHE Abbot et al 1986a and 19886b or the INHM model Van der Kwaak and Loague 2001 which solve systems of Partial Differential Equations PDEs in high spatial detail In the REW approach the integration of the conservation equations over charac teristic control volumes yields a system of Ordinary Differential Equations ODEs that can be solved either numerically or analytically Moreover the REW model contains also a Richards equation solver based on the procedure proposed by Ross 2003 for the simulation of vertical moisture diffusion in the unsaturated zone and an estimation of infiltration and water table recharge fluxes The groundwa ter module is based on mass conservation equations for irregular elements of the subsurface zone whereby the horizontal distribution or groundwater and respective groundwater fluxes is based on piezometric head distributions that are dissipated over characteristic length scales The characteristic length scales are estimated by solving flow distributions in a flow resistor network following the Kirchoff The procedure is described in Reggiani and
34. data lt name gt matlab weights mat Matlab file containing Kriging weights DA data lt name gt matlab lt name gt bc boundary and initial conditions file It is written only if not al ready present D data lt name gt ascii lt name gt init log file summarizing a list of geo metric properties for the various REWS D data lt name gt ascii preprocessor log general log file of the preproces sor D data lt name gt logs 7 2 2 The Matlab I O files Table 8 Summary of I O files of the preprocessor The I O of the preprocessor is read written from to a series of Matlab files in the directory D data lt name gt matlab The preprocessor exe module reads the files lt name gt rec mat and lt name gt links mat and writes the files lt name gt links1 mat lt name gt recl1 mat restored mat forcing mat and timeinfo mat The first two files 53 contain the geometric information of the various REWs and RECs including the parameter values and initial conditions restored mat contains the reconstructed P T Etp Rh Dt time series at the stations reconstruction of missing values effectuated through Kriging from neighbouring stations and forcing mat the block kriged time series with P T Etp Rh Dt values at the centroids of the REWs timeinfo mat contains all time related information including the data recording time step and the simulation start and end time T
35. e directory Block 8 contains some switches on the I O file formats and is additional debug information needs to be print to screen and in the log files during run time Table 7 shows an example of a parameter file for the preprocessor Lines that begin with a are ignored at input HHH preprocessor parameter file KAFA Block 1 basic hydraulic information steady state base flow event mm h 0 01 overland flow Manning roughness parameter 0 120 channel flow Manning roughness parameter 0 050 at a station depth scaling exponent 0 40 discharge area scaling exponent 0 8 hydraulic conductivity for channel bed m s 0 0000001 river bed transition zone thickness m 1 5 Block 2 subsurface water table depth m 5 0 bedrock depth m 300 45 soil porosity 0 5 saturated hydraulic conductivity Szone m s 0 0005 depth of saturated subsurface flow layer m 0 25 exponent for surface precipitation partitioning 0 20 depth of top soil layer for saturation averaging m 0 30 exponent in power relationship p 1 linear 0 35 Block 3 kriging variogram circle exponential gaussian polynom spline circle sweep reduced maximum distance station and maximum distance reduced variogram type event climatological event calculate variance no scaling parameter alphad4 1 scaling parameter beta 1 Block 4 reservoirs reservoir node in REW 80 r
36. e visible we execute them one by one e Change directory to D data mosel tardem 67 e Run the flooding algorithm by typing the dos command flood ere mosel The application is executed generating the filled elevation data file moselfel in ArcView ASCII exchange format e Run the the D8 algorithm by typing the dos command d8 ezxe mosel The application is executed generating the ArcView ASCII exchange file moselp and moselsd e Run the the Doo algorithm by typing the dos command dinf ere mosel The application is executed generating the file moselang and moselslp e Run the accumulation algorithm for the D8 method by typing the dos com mand aread amp exe mosel The application is executed generating the file mose lads e Run the accumulation algorithm for the Doo method by typing the dos com mand areadinf exe mosel The application is executed generating the file mose lasca e Run the gridnet application to extract the Horton Strahler orders of the net work from the D8 method by typing the dos command gridnet exe mosel The application is executed generating the file moselgord containng the Strahler orders for each pixel moselplen and moseltlen containing the partial and total path lengths respectively All files generated are in the ArcView ASCII exchange format and can be imported into any GIS visualization package To do so it is however necessary to add the extension asc to the respective file names 6
37. ecific infrastructure categories i e urban non urban surfaces can be extracted Make sure the infrastructure category file is present in the D data mosel tardem directory and is named moselis without file extension If only one of the files moselz mosellu mosels or moselis is present the RECs will be determined only by taking the particular file into account If two three or all four files are present all maps will be taken into account for the REC analysis If one ore more map files need to be excluded from the REC analysis it suffices to rename the file so that it is not named after one of the four categories i e moselz mosellu mosels moselis listed above If none of the four map file types are present a REC analysis will be performed by skipping the breakdown of a REW into RECs altogether Execute the REC analysis by typing the command recanalysis exe mosel The execution generates the files moselrec dat and moselrec mat and mosellinks mat in the directory D data mosel matlab 74 10 6 Running the Preprocessor To run the preprocessor a series of preparatory steps are needed before the appli cation can be executed These steps are required to make sure that all necessary input files are in the respective directories The parameters for the preprocessor are set in the parameter file mosel prm The preprocessor is executed by going through the following steps e Copy the files moselrec mat and mosellinks mat
38. es are present a corresponding subdivision of the REC is performed zbins exe is a utility to create a subdivision of REWs into elevation zones which generates the lt name gt z file Subsequently recanalysis exe needs to be run with lt name gt z as input 37 6 1 1 rewanalysis exe Group REWANALYSIS Command subbasinsetup lt name gt order_thresh Arguments lt name gt order_thresh Input lt name gt w lt name gt fel Function Calculates the REW geometries connectivities and properties Operation By executing rewanalysis exe three output files are generated The threshold order order_thresh should be equal or larger than the one used in the operation of subasinsetup Output lt name gt rew dat and lt name gt links dat two ASCII files showing the REW and the stream network connectivities and geometries Two Matlab files are also written lt name gt rew mat containing REW information and lt name gt links mat containing link information 6 1 2 recanalysis exe Group REWANALYSIS Command recanalysis lt name gt Arguments lt name gt Input lt name gt w lt name gt fel lt name gt rew mat lt name gt links mat lt name gt s soil information optional lt name gt lu land use information optional lt name gt is infrastructure information optional lt name gt z elevation classes information op tional Function Calculates the REC geometries connectivities and properties 38 Operation
39. eservoir node in REW 81 Block 5 snow model parametetrs latitude of catchment centroid deg 49 00 capillary water retention fraction 0 05 solid precipitation threshold temperature C 1 0 Block 6 surface fitting parameters roughness rou gt 0 1 0 tension 1 gt tau gt 0 0 5 46 number of sucessive grid sub divisions 1 0 margin of mesh grid 100 0 Block relative directory paths input data files Stations datools run info file DaTools Block 8 miscellaneous parameters time series files netcdf xml netcdf save forcing in XML format daTools no debug mode no Table 7 Preprocessor parameter file 7 1 3 Meteorological forcing information The meteorological forcing information consists of equally long series of precipi tation P mm recording unit Temperature T C Potential Evaporation Etp mm recording unit Relative air humidity Rh dimensionless and daily temper ature excursion Dt C These data can be supplied either as Delft FEWS XML PIPI format or in the standard NetCDF file format for mete orological data The meteo data will be supplied as each input series relative to the centroid of each REW These series can be generated by external preprocess For reference visit the URL https publicwiki deltares nl display FEWSDOC Home 47 ing using the shape file containing REW polygons to calculate REW centroids Alternativel
40. essing of terrain information and the extraction of drainage network features as well as subbasins in the REW model is performed with the terrain analysis software TARDEM Band 1986 Tarboton et al 1991 1992 Tarboton 1997 However also different types of software could in principle be used e g Rivertools TOPAZ The choice to use TARDEM for digital terrain preprocessing has been made because of the availability of open source code 5 1 The TARDEM package structure TARDEM consists of a series of applications or modules that carry out analysis steps The steps follow the standard analysis methods for the extraction of a channel network and drainage surfaces from digital elevation data For a closer description of the software or the employed algorithms the reader is referred to the manuals and respective scientific literature published on TARDEM This information can be retrieved from the web site of the Utah State University where TARDEM has been TARDEM is a set of watershed analysis tools to extract a o the subbasins and the drainage network Software and manuals on TARDEM are available http hydrology usu edu taudem 27 developed and is maintained The names of the TARDEM executable file suite are summarized in Table 13 in Appendix A Through the sequential execution of the applications listed in Table 13 the topogra phy described by the digital elevation model DEM is analyzed and specific model input files for
41. essor exe addParams skipParams lt name gt Arguments addParams skipParams lt name gt Input lt name gt prm lt name gt rec mat lt name gt links mat lt name gt bc optional Function Prepares and pre processes all information required by the solver kernel Operation By running preprocessor exe with appropriate input and parameter files all initial and boundary conditions as well as meteo forcing files required for the operation of the REW model are prepared By using the option addParams the parameter values in the lt name gt rec mat and lt name gt links mat files are updated By using the option skipParams the parameter values are not updated Output The Matlab files timeinfo mat model time info forcing_5z mat spatially interpolated meteo forcing series restored_5z mat reconstructed original input se 42 ries series weights mat Kriging weights and reservoirs mat reservoir template files The following Matlab files are also written lt name gt recl mat containing updated REW information and lt name gt links1 mat containing update link infor mation The following two ASCII files are also written lt name gt bc containing the boundary conditions and lt name gt init containing the initial conditions 7 1 2 The preprocessor parameter file The parameter file is organized in a series of data blocks separated by hashed lines Block 1 contains hydraulic information such as Manning coefficients and the Le
42. et_setup lt name gt gord asc into ArcView In this file it is possible to distinguish the principal drainage network based on the largest Strahler order when the stream pixels are coloured in ArcView according to Strahler stream order Output lt name gt ord ArcView exchange format file containing a grid with Strahler order for mapped stream network lt name gt src containing a network mask based on channel source rules lt name gt tree dat with a connectivity matrix of the tree structure lt name gt coord containing the pixel coordinates of the stream network method explanation of parameters pl 1 drainage area threshold A gt p2 pl 2 area slope threshold A gt p2 pl 3 length area threshold A gt p2 L 3 L is the maximum drainage length to each cell in the lt name gt plen file pl 4 Accumulation area of upward curved grid cells The DEM is first smoothed by a kernel with value p2 at its centre p3 on its edges and p4 on diagonals The Peuker and Douglas 1975 method is then used to identify upwards curved grid cells and contributing area computed using only these cells A threshold Auc gt thresh on these cells is used to map the channel network 34 pl 5 grid order threshold method The desired threshold order is selected as O gt p 2 pl 6 Use existing channel networks specified in a fdrn file The fdrn file is created from fdr file by flood exe The fdr file is created from shape
43. etwork and moselw shp moselw shz moselw dbf for the visualization of the sub basin shapes These files can be loaded into ArcView and visualized as shown in Figure The number of subbasins and the density of the network is governed by the pa rameters i e threshold area selected in the routine net_setup exe and the selected Strahler order threshold value order_tresh passed as an argument For order_tresh equal to 1 sub basins are attributed to all links including first order streams For order_tresh equal to numbers large than 1 links are attributed only to streams 71 BILZEN HERMALLE E ANGLEUR E LA FIZE FONTAINE Figure 11 A river basin separated into 73 REWs with an order equal or larger than order_tresh Some criteria to set the appropriate Strahler order are described in the TARDEM manual The subbasinsetup command can also be executed through a suitable batch file 104 Extracting Representative Elementary Watersheds REWs The REWs are extracted by executing the following command e Type the command rewanalysis mosel order_tresh 72 e For the present example select order_tresh equal to 2 The files moselrew dat and moselrew mat and mosellinks mat are generated The threshold order should be greater or equal the value used for the execution of sub basinsetup exe 10 5 Extracting Representative Elementary Columns RECs REC are a subdivision of the unsaturated zone into sub REW sca
44. exe with appropriate input and parameter file the REW model output files are transformed into a format of choice Output The output files REW000z x2ml containing output variables for each REW respectively in case the XML output option is selected the binary file lt name gt results mat containing output variables for all REWs in case the Matlab output option is selected or the binary file lt name gt results nc containing output variables for all REWs in NetCDF format in case the NetCDF output option is selected 9 1 2 The postprocessor parameter file Block 1 contains generic information such as the switch which allows to indicate if the output should be generated as a Matlab file as NetCDF file or as XML file for import into FEWS In Block 2 the user provides information to be included in the XML file such as the region of origin the geodatum time units time zone etc For reference on this information the reader is referred to the description of the Published Interface PI format header information Block 3 contains the directory paths for I O operations of the preprocessor These are unless specified otherwise the same as for the preprocessor or the solver Block 4 allows to specify which variables one wants to export By using this op tions the size of the output files can be significantly reduced by only exporting the 62 variables that are specifically needed Table shows an example of a parameter file for the po
45. f output files i a series of XML files one for each REW for potential use in the Delft FEWS forecasting platform ii a single binary Matlab file called moselresults mat which contains all the results in a single file or iii a single binary NetCDF file called moselresults nc which contains all the results in a single file The parameters for the postprocessor are set in the parameter file mosel2 prm Further extensions of the postprocessor are planned These will allow to transform the output also in other formats such as the Delft standard his data storage format for exchange with Delft software e Execute the dos command postprocessor ere mosel e If the Matlab file output flag is set in the parameter file the binary file mosel results mat is written to the directory D data mosel matlab e If the NetCDF file output flag is set in the parameter file the binary file moselresults nc is written to the directory D data mosel netcdf 77 e If the XML output file type is selected in the parameter file an tml file for each REW named after the respective REW number is written to the directory D data mosel results In this case no Matlab file moselresults mat will be generated The resulting binary file moselresults mat can be simply loaded into Matlab by using the Matlab command load moselresults mat There are a series of Matlab scripts available that can be used to visualize different computed variables and internal fluxes
46. file by streamtogrid ere This method is als referred to as burning an pre existing network e g form a shape file in Table 5 Parameters for net_setup 5 1 8 arclinks exe Group TARDEM Command arclinks lt name gt Arguments lt name gt Function Determination of network of the links Generates an output file in the GIS exchange format e00 Operation arclinks exe computes link properties associated with the channel net work Input are the ASCII files lt name gt coord dat and lt name gt tree dat files Output The generic GIS interchange files lt name gt li e00 5 1 9 arcstreams exe Group TARDEM 35 Command arcstreams lt name gt Arguments lt name gt Function Determination of the network of links Generates an output file in the GIS exchange format e00 Operation arcstreams exe computes stream properties associated with the channel network Input are the ASCII files lt name gt coord dat and lt name gt tree dat files Output The generic GIS interchange files lt name gt st e00 5 1 10 subasinsetup exe Group TARDEM Command subbasinsetup lt name gt order_thresh Arguments order_thresh Function Defines the shapes and areas of the subbasins and creates shape files for easy import into ArcView Operation subbasinsetup exe computes the subbasins after indicating the threshold order passed as a program argument by the user All basins with a Horton Strahler order larger than t
47. form solver output into XML or Matlab files Table 17 Summary of postprocessor system files List of Figures 1 Organization of REWs around the structure of the network 10 2 A Representative Elementary Watershed as a 3 D spatial entity 11 3 Subdivision of the unsaturated zone of a REW into RECs 12 4 A view of hydrological processes represented in the REW model 13 Or Conceptualization of the REW as a zero dimensional element with ive Sie PG h ERS Oe ee Fhe 8S eo 14 6 Implementation of the Kirchhoff laws for the groundwater system eke AA aver ea 16 7 The groundwater flow field River Geer basin Belgium 17 TEETE 18 9 Cross section of the REW showing the river channel 21 10 Pixels in the moselgord asc file coloured according to the Strahler order 69 11 A river basin separated into 73 REWs 72 List of Tables 1 Summary of internal external and inter REW fluxes 22 89 2 Working directoriessetup 23 3 I O file path settings for model setup 24 4 Example of ESRI ArcGrid ASCII exchange format file 28 5 Parameters for netsetup 35 6 Summary of rewanalysis and recanalysis I O files 40 AA ee ee ee eee 47 bg Wa n ea eee FY 53 Cota ee Aha eo ee eee ae oe 59 10 Summary of I O files of the solver 61 11 Postprocessor parameter file
48. ge head differences between the REW average groundwater level and the river The water between the two zones is exchanged through the river bed for which an own hydraulic conductivity and a thickness can be specified For situations in which the average water level in the channel reach is higher that the water level in the surrounding aquifer the river groundwater exchange flux causes the groundwater to be fed from the channel If on the other hand the average water 21 level in the aquifer increases with respect to the channel the groundwater feeds the channel This principle is explained schematically in Figure p which features the REW average water level the actual water level the water table interpolated via the optimized bicubic spline surface and the average water level in the channel 2 4 8 Summary of exchange fluxes in the REW model The most relevant model internal and internal fluxes are shown in Table The table specifies which fluxes are within zones in a REW and which ones between a REW and either neighbouring REWs or the outside environment i e across watershed boundaries Flux description symbol REW internal inter REW External river saturated zone es es no no y water table recharge ets es no no yi surface infiltration eu es no no y Horton type runoff flux e es no no yp y channel surface inflow e es no no yi regional groundwater flux e no es es y channel Dunne type inflow eP es no no
49. he number of initial subdivision of the length and width into a grid cells that are interpolated via a cubic spline the maximum number of Gauss Seidel iterations the roughness and tension of the sur face to be fitted Some more specific parameters are generally left at default values For a general explanation of the interpolation method the reader is referred to the cited paper For general applications it is recommended to leave the parameter and the mesh sub division information at their selected default values Block 5 contains the paths setting referring to the I O directories of the particular model following the structure D data lt name gt matlab D data lt name gt stations D data lt name gt ascii D data lt name gt logs 56 D data lt name gt tardem D data lt name gt results The directory D data lt name gt matlab contains the Matlab files lt name gt rew mat lt name gt rec mat and lt name gt links mat written by rewanalysis ere and recanaly sis exe These are most conveniently copied or moved after the execution of the two programs from the directory D data lt name gt tardem into the directory D data lt name gt matlab This can be done either manually or with the aid of a batch file The directories D data lt name gt logs and D data lt name gt tardem point to a log file repository and to the TARDEM I O file directory Block 6 contains some switches on the I O file formats and is additional deb
50. he threshold order order _thresh will be identified as basins order _thresh must be at least equal to 1 Output lt name gt w ArcView exchange format file containing a grid with REW masks lt name gt shp ArcView shape file containing the shapes of the channel net work lt name gt w shp ArcView shape file containing the shapes of the REWs 36 6 The Rewanalysis package 6 1 REWANALYSIS package structure The REWANALYSIS group of modules is made up by three applications rewanal ysis exe recanalysis ere and zbins ezxe rewanalysis exe identifies the REWs as 3 D spatial regions establishes inter connectivity between REWs and calculates REW specific geometric quantities such as surface areas average surface elevation and others recanalysis exe performs a subdivisions of the unsaturated zone of the REC into vertical columns so called RECs The subdivisions can be carried out on the basis of a superimposition of different types of GIS layers that need to be prepared in the standard ArcView ASCII exchange format These files have the following extension according to their content lt name gt s is a raster map indicating different soil types lt name gt z is a map containing masks for different elevation classes lt name gt lu contains different land use classes and lt name gt is contains different infrastructure classes If these are not present a single REC is identified within each REW default situation If the specific fil
51. ine instructions and the argument list to be passed to the executables the names of the output files and the purpose of the individual routines By executing the sequence of the executable files 1 6 in Table raster output files are generated which contain the local drainage direction and drainage area maps The raster output can be viewed in ArcView by renaming the file and adding the extension asc For example the raster output files lt name gt sca lt name gt d lt name gt fel are renamed to lt name gt fel asc lt name gt fel asc and lt name gt fel asc From arclinks exe and arcstreams exe two extension files are created that allow easy import into ArcView The first six analysis steps can be conveniently triggered by a batch file prepare bat supplied with the installation By editing the batch file setting the right project file name prototype and launching the batch process the applications are called automatically in the right sequence and the directory D data lt name gt tardem is populated with the raster output files The following subsections address each 29 module of TARDEM describe I O and provide instructions on the usage of each module 5 1 1 flood exe Group TARDEM Command flood lt name gt Arguments lt name gt Input lt name gt fel filled DEM file Function Processing a digital digital elevation model de pitting of holes and prepa ration and smoothing of the surface in order to make it ready
52. its called Representative Elementary Watersheds REWs REWs are identified by performing an analysis of the watershed topography and constitute a set of the interconnected elements that are organized around the tree like structure of the stream channel network as shown in Figure I REWs constitute 3D regions with a vertical prismatic mantle surface defined by the REW boundaries The REW boundaries coincide with the topographic divides They delineate a well defined area of the land surface that captures the precipitation The contour of a REW mantle surface coincides with the shape of the ridges defining a subbasin A schematic representation of a REW element is depicted in Figure A REW is delimited by the atmosphere at the top and by an impermeable layer at the bottom The impermeable layer can be either defined by a horizontal surface or can be given by interpolation of bedrock depth for a series of irregular points REW 1 REW 2 M REW 6 REW 10 REW11 REW 5 N REW 7 REW 3 REW 4 REW 8 REW 12 REW 9 REW 13 watershed outlet Figure 1 Organization of REWs around the structure of the network 2 3 Sub REW variability In order to be able to account for hydrological variability within a REW with characteristics smaller than the REW e g attributable to factors such as land use or soil properties the unsaturated zone can be subdivided into smaller units or columns labelled Representative Elementary Columns RECs These REC
53. lation The software REW model software project can be compiled either under Windows or under LINUX For compilation under windows the source code has been organized into respective projects under MS Visual Studio 2008 For compilation it is necessary to install the Visual C and the Visual Fortran Compiler as there is source code written in C C and Fortran languages For compilation under LINUX makefiles can be supplied which allow compilation with wither the Intel Fortran and C C compilers or the GNU GCC compiler We note that the TARDEM source code can at the moment only be compiled under Windows using MS Visual Studio 2008 4 2 Compatibility with third party software It has been deliberately decided to develop the software without a GUI as pure DOS commandline applications to ensure as much as possible a stand alone character of the application and to support the integration of the REW model into particular type of open architecture products such as the flood forecasting operational shell Delft FEWS 26 The stand alone structure of the REW model application facilitates the develop ment of model specific adapters for transfer of I O information between third party software and the model An extension for making the REW model application compatible with the Open Modelling Interface OpenMI http www openmi org standards is foreseen for the near future 5 The terrain analysis package TARDEM The part related to the pre proc
54. le columns based on a series of criteria The subdivision can be based on elevation zones or on the basis of three categories of spatial information map files in ESRI ArcGrid ASCII ex change format containing elevation zones landuse soil category or infrastructure categories respectively Each pixel belonging to a particular elevation soil lan duse or infrastructure category must be assigned an integer number in the spatial information map file e To create RECs based on elevation height zones of e g 100m height bins exe cute the application zbins exe by typing the command zbins ere mosel 10 The execution can be carried out directly in the directory D data mosel tardem if the path to the executable file is set correctly e The execution generates the ESRI ArcGrid ASCII exchange format file moselz using the elevation file moselfel without extension already present in the di rectory e Incase of presence of a land use file RECs based on a specific land use category is can be extracted 73 Make sure the landuse category file is present in the D data mosel tardem directory and is named mosellu without file extension In case of presence of a soil type file RECs based on specific soil categories can be extracted Make sure the soil category file is present in the D data mosel tardem direc tory and is named mosels without file extension In case of presence of an infrastructure type file RECs based on a sp
55. mmand solver exe mosel e The output time step and the maximum computational step can be set in the parameter file mosell prm Note that the time intervals for the time steps are in seconds The start and end time of the computation are read automatically from the timeinfo mat file This file is written by the preprocessor which saves the start and end time extracted form the forcing time series to the timeinfo mat file e All model results are written to the directory D data mosel results in a se ries of ASCII files named after the respective REW numbers i e rew0001 out etc The results files contain all fluxes and dynamic state variables which are printed at every output step It is possible to select the fluxes and variables which need to be printed for each REW and the specific REWs for which results need to be printed by appropriate selections in the parameter file e It is recommended not use a maximum time step larger than 3 hours in simu 76 lation mode as the solution of the differential equations becomes numerically inaccurate Longer time steps than 3 hours can be used in warm up mode e g to generate a model state as in such a context accuracy of the simulation may be of lesser importance The content of the ASCII output files can be visualized in Matlab by executing the respective m Matlab script files 10 8 Running the Postprocessor The postprocessor is used to transform the solver output into three types o
56. ng a raster map with the drainage directions between 0 and 2II radians lt name gt sip ArcView exchange format file containing a raster map with the local surface slopes in radians between 0 and II 2 31 5 1 4 aread8 exe Group TARDEM Command aread amp lt name gt Arguments lt name gt Input lt name gt p drainage directions file from D8 method Function Calculates contributing areas from the D8 derived flow directions The units are specific drainage area i e number of grid cells times cell size Operation By executing aread8 ere on a digital elevation model a drainage area accumulation matrix is generated Output lt name gt ang ArcView exchange format file containing a raster map with the accumulated area expressed in number of pixels times cell size 5 1 5 areadinf exe Group TARDEM Command dinf lt name gt Arguments lt name gt Input lt name gt ang drainage directions file form Doo method Function Function Calculates contributing areas from the Doo derived flow di rections The units are specific drainage area i e number of grid cells times cell size Operation By executing areadinf exe on a digital elevation model a drainage area accumulation matrix is generated Output lt name gt sca ArcView exchange format file containing a raster map with 32 the accumulated area expressed in number of pixels times cell size 5 1 6 gridnet exe Group TARDEM Command gridnet lt name gt
57. on on REWs lt name gt links mat Matlab file containing summary D data lt name gt tardem information on the stream links structure lt name gt rew dat Ascii file summarizing the D data lt name gt tardem REW geometries en inter connectivities lt name gt links dat Ascii file summarizing the D data lt name gt tardem Link geometries en inter connectivities Recanalysis files lt name gt rec mat Matlab file containing summary D data lt name gt tardem information on REWs lt name gt rec dat Ascii file containing the informa D data lt name gt tardem tion on REWs and RECs Table 6 Summary of rewanalysis and recanalysis I O files IMPORTANT e After running rewanalysis exe and recanalysis exe it is necessary to copy the 40 files lt name gt rew mat and lt name gt rec mat manually to D data lt name gt matlab where they will be read during following analysis steps running preproces SOT EXE 7 The Preprocessor package The preprocessor is an application that performs a series of operations which are preliminary to the actual simulations These operations include e Assigning the model parameters and material properties to the various model entities such as REWs and RECs e Preprocessing precipitation data by reconstructing missing data over a selected period via Kriging optional e Perform block Kriging of the precipitation for each REW op
58. ooth groundwater surface between REW average groundwater points The fitting of the smooth surface is based on the Finite Element Method FEM which calculates the surface by minimizing the elastic ten sion energy in the surface The same procedure can be applied for the definition of 16 Geer basin aquifer flow field 200 400 600 800 number of pixels 30x30m 1000 1200 200 400 600 800 1000 1200 1400 number of pixels 30x30m Figure 7 The groundwater flow field River Geer basin Belgium the impermeable lower boundary of the watershed if sparse measurement points of the bedrock depth are available Figure 8 shows an example of a fitted water table surface 2 4 3 Saturation excess flow Dunne type flow O zone The saturation excess flow or also referred to as Dunne type flow is caused by direct precipitation onto saturated areas In the REW model the growth of the sat urated areas is linked directly to the raise and fall of the REW average groundwater 17 Figure 8 Water table interpolated with the bi cubic spline surface elevation y in Figure 5 By default it is assumed that the relation between the ground water levels and the growth of the saturated areas is linear The saturated areas are fed through exfiltration from the saturated zone on conceptual seepage faces which coincide with the saturated areas The model calculates the saturated REW area fraction w as a dynamic variable
59. opold and Maddock at a station and downstream channel geometry Moreover the hy draulic conductivity of the channel bed and the thickness of the aquifer channel bed transition layer is assigned Block 2 contains the data to be attributed uniformly to the subsurface zone These include hydraulic conductivity porosity soil parameters for the unsaturated zone porosity conductivity and thickness of the shallow subsurface flow layer and quanti ties to specify the initial water content in the unsaturated zone as well as the initial position of the water table Block 3 contains all the information inherent to kriging and block kriging and the time series The possible options include the choice of variogram type the sill nugget and other variogram parameters the recording frequency of the P T Etp Rh Dt data The no data missing value flag used in the data series need to be indicated as well 43 Block 4 contains the assignment of reservoirs to specific REWs One reservoir per REW is allowed Block 5 contains the parameters for the snow energy balance model including snow density thermal properties and limit temperatures for solid and liquid precipitation Block 6 contains the data inherent to the finite element method used for fitting a bedrock surface form irregular data points or a smooth water table across the REW average groundwater elevation points see also solver exe The fitting method is based on the Inoue 1986 algorithm
60. porous media flow subsurface zones as well as overland and channel flow land surface zones The modelling of the various flow processes is described 12 separately in the following paragraphs Figure 4 gives an overview of the processes listed above unsaturated zone u overland flow zone 0 concentrated overl flow c A water table p T channel reach r in Figure 4 A view of hydrological processes represented in the REW model 2 4 1 Unsaturated zone U zone The unsaturated zone is modelled by means of a dual system First the vertical water moverment is solved by means of a Richards equation solver Ross 2003 The chosen solver for the Partial Differential Equation PDE governing flow in unsaturated soil has the property of linearising the mass flux between cells and allows a very fast solution of the equation avoiding to search for iterative solutions With respect to full non linear solvers the accuracy of the numerical solution is somewhat lower But given the high uncertainty in the choice of the soil parameters the errors approximations made in the approximation choices of the numerical method are to 13 be considered of second order and thus negligible Next to a complete numerical solution the mass balance for the unsaturated zone is solved at the REW scale via an analytical solution whereby the infiltration fluxes and the water table recharge fluxes are given by those calculated via the solu
61. s Proc Royal Soc A 457 157 189 Reggiani P and T H M Rientjes 2005 Internal Flux Parameterisation in the Representative Elementary Watershed REW approach application to a natural basin Water Resour Res 41 W04013 doi 10 1029 2004WR003693 Reggiani P and T H M Rientjes 2010 Closing horizontal groundwater fluxes with pipe network analysis An application of the REW approach to an aquifer Environmental Modelling and Software DOI 10 1016 j envsoft 2010 04 019 Rodriguez Iturbe I and Vald z J B 1979 The Geomorphologic structure of hy drologic response Water Resour Res 15 1409 1420 Ross P J 2003 Modeling Soil Water and Solute Transport Fast Simplified Nu merical Solutions Agron J 95 1852 1361 Snell J and M Sivapalan 1995 Application of the meta channel concept Con 82 struction of the meta channel hydraulic geometry for a natural channel Hydrol Proc 9 485 495 Tarboton D G R L Bras and I Rodriguez Iturbe 1991 On the Extraction of Channel Networks from Digital Elevation Data Hydrol Processes 5 1 81 100 Tarboton D G R L Bras and I Rodriguez Iturbe 1992 A Physical Basis for Drainage Density Geomorphology 5 1 2 59 76 Tarboton D G 1997 A New Method for the Determination of Flow Directions and contributing Areas in Grid Digital Elevation Models Water Resour Res 33 2 309 319 Todini E 2007 A mass conserv
62. s are defined by overlapping a series of GIS maps such as land use and soil type The procedure of subdividing the unsaturated zone allows to attribute different soil properties and evapo transpitation rate depending on plant species to each smaller 10 unsaturated land surface saturated areas impermeable mantle surface reference system Figure 2 A Representative Elementary Watershed as a 3 D spatial entity unit The units can have an irregular shape which is dictated by the particular combination of the REW shape combined with the local soil and land use pattern Figure 3 shows an example of a watershed separated into RECs bottom right map by overlying a map with REWs top left map derived purely from topographic information with a landuse map top right map and a soil map bottom left map 11 Figure 3 Subdivision of the unsaturated zone of a REW into RECs 2 4 Modelled hydrological processes The volume occupied by a REW contains typical flow zones encountered in a wa tershed The following zones can be modelled explicitly and for every REW 1 the unsaturated zone 2 the saturated zone 3 the subsurface stormflow zone 4 the saturated overland flow area 5 the infiltration excess overland flow 6 the channel reach and 7 the snow pack The flow within the various domains extends over very different temporal scales and encompasses flow phenomena such as unsaturated and saturated
63. stprocessor Lines that begin with a are ignored at input HHH HHH postprocessor parameter file HHH AHA Block 1 basic hydraulic information convert to matlab xml netcdf matlab file name prototype mosel debug mode no Block 2 XML info version 1 2 region Vlaanderen geodatum Belgian National Reference source organisation WaterIntellect source system IBM pc unit second timezone with respect to GMT 1 0 multiplier 3600 missing value 999 0 Blocks relative directory paths matlab files Matlab log files Logs tardem files Tardem ascii files Ascii NetCDF files Netcdf output files Results 63 HHH Block4 print info number of REWs variable variable variable variable Table 11 Postprocessor parameter file 9 2 The postprocessor I O file system The following Table 12 summarizes all input and output files of the postprocessor input file name file information directory Rew000z out ascii files containing the solver D data lt name gt simulation results lt name gt 2 prm parameter file for postprocessor D data lt name gt output file name file information directory postprocessor log general log file for the solver D data lt name gt logs 64 Rew000x xml XML files containing the solver simulation results For each REW one file is written
64. t rec mat lt name gt rec dat Performs geometric analy sis of RECs 86 zbins exe lt name gt zbin lt name gt z Utility to subdivide REWS into elevation zones for REC definition Table 14 Summary of rewanalysis system files C PREPROCESSOR system files executable files and argu output file names scope ments preprocessor Le forcing mat Performs data preprocessing lt name gt addParams weights mat reser skipParams voirs mat time info mat lt name gt bc lt name gt init lt name gt recl mat lt name gt links1 mat preprocessor log Table 15 Summary of preprocessor system files 87 D SOLVER system files executable files and argu ments output file names scope solver exe lt name gt REW000z out REsolve000z out lt name gt results mat lt name gt inoue mat globals log inoue log loops log masser ror log Performs core simulations XML 2forcing exe lt name gt lt name gt forcing_5x mat Adds perturbed forcing to lt name gt forcing_5x mat file Used for Data assimi lation in the DA tools en vironment Table 16 Summary of solver system files E POSTPROCESSOR system files executable files and argu ments output file name scope postprocessor exe lt name gt REW000z xml lt name gt results mat lt name gt results nc 88 Trans
65. tion at the snow surface the sensible and latent heat fluxes the energy advected by the precipitation in liquid format and the energy removed by meltwater Al these energy fluxes are balanced by the internal energy of the snowpack The snow pack is represented as a single layer and a surface zone of infinitesimal thickness as described by Luce et al 1998 The surface temperature is calculated iteratively by balancing the different surface flux components The snow surface temperature is also dependent on the average snow core temperature which is calculated diagnostically form the internal energy of the snow pack The meltwater flux is indicated as e and is discharged directly to the P zone 20 mantle surface average water table interpolated water table unsaturated zone real water table i A NWN ar mantle flux sm recharge flux saturated zone average bottom boundary datum elevation Figure 9 Cross section of the REW showing the river channel 2 4 7 Channel flow R zone The channel flow zone is recharged by fluxes from upstream links the outflow to the down stream reach and lateral inflow fluxes from the overland flow zone O zone the aquifer S zone and the subsurface stormflow zone P zone The lateral inflows due to overland flow and the shallow subsurface storm flow zone are calculated form the governing equations for these zones The interaction with the groundwater is controlled by the avera
66. tion of Richards equation The solution of the Ordinary Differential Equation ODE yields an average saturation and depth to the water table value for each REW The soil moisture profiles and the average values for saturation and depth to the water table variable y in Figure are printed to the respective output files at chosen time steps average water table y Y unsaturated zone saturated zone TX saturated area eee s projection X za ai a tS unsaturated area ae KA projection X ee TS REW surface area projection Figure 5 Conceptualization of the REW as a zero dimensional element with sub volumes 14 2 4 2 Saturated zone S zone The saturated zone is modelled as a 2 D aquifer with only horizontal flow The groundwater is recharged vertically through percolation from the unsaturated zone The groundwater is then distributed laterally between REWs through piezometric head differences between REWs The piezometric head is expressed in terms of the average water table elevation calculated for a REW via the mass balance equation The mass balance equation is an Ordinary Differential Equation ODE solved an alytically given the recharge flux from 1 the unsaturated zone e 2 the lateral groundwater distribution fluxes e between adiacent REWs 3 the seepage flux e and 4 the exchange flux of groundwater with the river channel across the bed area e The seepage flux e feeds the overland
67. tional e Define the presence of a reservoir for a particular REW optional e Assign initial and boundary conditions to the model for each REW e Add pumping flows and water abstractions to the basin optional 7 1 The preprocessor structure The preprocessor consists of a single executable preprocessor exe This executable is run several times during the model setup phase The preprocesor prepares all necessary input files for the operation of the downstream package solver The pre processor assigns all necessary material properties such as hydraulic conductivities soil texture and structure data and river channel geometry to the various REWs 41 and zones The assignment of these properties is performed in an spatially uniform manner If values need to be changed in space the file lt name gt rec1 mat needs to be edited and values changed manually or automatically with an appropriate Matlab script The preprocessor also assigns all initial and boundary conditions These are stored in the lt name gt bc and lt name gt ic ASCII files One of the most important tasks of the preprocessor is to prepare the meteorological forcing infor mation For this purpose the meteorological forcing which is either supplied in the lt name gt stations cml file in XML format or in the lt name gt stations nc file in NetCDF format are read and interpolated towards the REW centroids 7 1 1 preprocessor exe Group PREPROCESSOR CommanD preproc
68. ug information needs to be printed to screen and in the log files during run time Table 9 shows an example of a parameter file for the solver Lines that begin with a are ignored at input HHH solver parameter file 44 TT Block 1 infos for computations guess integration time step secs 100 0 min integration time step secs 0 max integration time step secs 3600 0 print step secs 3600 0 mesh update step secs 604800 0 number of iterations 1 Block 2 Richards Equation amp miscellaneous maximum delta s 0 01 unsaturated zone cell thickness m 0 1 57 calibration factor for water table recharge 1 0 Block 3 Richards Equation amp miscellaneous channel routing Runge Kutta Muskingum kinematicwave kinematicwave Block 4 surface fitting parameters roughness rou gt 0 1 0 tension 1 gt tau gt 0 0 5 number of successive grid sub divisions 1 0 margin of mesh grid 100 0 Block 5 relative directory paths input data files Stations datools run info file DaTools Block 6 additional info retain model states no debug mode no Block 7 output info print output for REW no 1 print output for REW no 2 print output for REW no 50 Block 8 print variables variable P m variable E m variable T m variable R q 58 variable variable variable vari
69. y a Kriging facility is available in the model which allows to perform reconstruction of missing values through serial Kriging and mapping of series to wards the REW centroid through spatial interpolation block Kriging This facil ity is however only available if the meteorological data are supplied in the XML PI file format The meteorological data are commonly sorted in the model directory D data lt name gt stations but can also be stored in a different folder in which case the path needs to be changed in the parameter file The naming convention for the meteorological forcing files is as follows Pi XML files two files need to be prepared i lt name gt timeseries cml con taining the actual time series and ii lt name gt locations xml containing the latitude longitude coordinates of each location The locations can be either arbitrary station positions or REW centroids NetCDF files one single file lt name gt timeseries nc is supplied Following the NCAR Climate and Forecast CF Metadata Conventions the file contains the time series with respect to a reference base date 01 01 1970 the data values and the latitude longitude coordinates of the observing location of each series 3For reference see http www unidata ucar edu software netcdf 48 7 1 4 Reservoirs It is possible to specify one reservoir for each REW The model solves a common reservoir equation with the explicit Euler method d A h dt I t
70. y the preprocessor input file name file information directory 5l lt name gt rec mat Matlab file containing the infor mation on REWs and RECs D data lt name gt matlab lt name gt links mat Matlab file containing the stream links structure DA data lt name gt matlab lt name gt bc boundary and initial conditions file If not present it is written by the preprocessor D data lt name gt ascii lt name gt prm parameter file for preprocessor D data lt name gt output file name file information directory lt name gt reci mat Matlab file containing the up dated information on REWs and RECs D data lt name gt matlab lt name gt links1 mat Matlab file containing the up dated stream links structure D data lt name gt matlab timeinfo mat Matlab file containing informa tion such as recording time step start and end time of the forcing series D data lt name gt matlab forcing_5x mat Matlab file containing the meteo forcing at the REW centroid D data lt name gt matlab restored_5x mat Matlab file containing the origi nal reconstructed meteo forcing series D data lt name gt matlab 52 reservoirs mat Matlab file containing the A h and Q h tables for the reser voirs The tables are initially empty and need to be filled by editing the file in Matlab D
71. ytically 2 4 5 Infiltration excess flow Horton type flow C zone The infiltration excess flow or also called Horton type flow in the literature is caused by precipitation which exceeds the infiltration capacity of the soil As a result water builds up on the surface and flows off subsequently In the REW model the infiltration excess flow is modelled through analytical solution of the mass and momentum balance Ordinary Differential Equations ODE The overland runoff is discharged directly into the saturated overland flow zone as flux e The infiltration excess flow is fed by the precipitation rate during precipitation events 19 2 4 6 Snow layer F zone The snow pack is represented in the REW model as a separate zone the frozen zone F zone The F zone is a layer of water at the solid state snow which deposits on the land surface The snow layer is build up when the 2m air temperature is just above zero e g 1 deg C From there onwards the precipitation is starting to fall as solid The mix of snow and water changes linearly with all the snow water mix becoming water at the upper temperature limit and snow only at the lower limit e g 1 deg C The upper and lower temperature limits can be set explicitly in the parameter file The dynamics of the snowpack i e snow build up and ablation are simulated using an energy balance model which balances various the energy fluxes of the snow pack These include the solar radia

Representative Elementary Watershed Model User Manual

Contents

Download Pdf Manuals

Related Search

Related Contents