Manual - SASWE Research Group
Contents
1. DESIGN SPECIFICATION FOR OPEN BOOK WATERSHED MODEL BY NITIN KATIYAR May 2007 Design Specification for Ka npur 1 0 Open book Watershed Model PREFACE 3 DISCLAIMER 3 INPUT FILES 4 NAMING CONVENTION 5 DEFINING SOME GLOBAL VARIABLES 6 TECHNICAL SPECIFICATIONS 7 Algorithm Formulation of the Model and Computer Implementation 7 Main Routine 9 Fn get rainfall data Routine 10 Fn populate slope file Routine 10 Fn populate cos file Routine 11 Fn linear storage discharge 11 Fn overland flow processing 11 Fn sheet flow processing 12 Fn total inflow 2 pixel 13 Fn total inflow 2 pixel channel 13 Fn laminar overland flow processing 13 Fn turbulent overland flow processing 14 Fn channel flow processing 14 Fn river flow processing 14 Nitin Katiyar 6 4 2007 Last saved by Nitin Page 2 Design Specification for Ka npur 1 0 Open book Watershed Model Preface Before reading this manual reader is advised to read the thesis Development of an open book watershed model for rapid prototyping of satellite based flood forecasting in international river basins submitted to Dept of Civil Engineering Tennessee Tech University Cookeville TN and Paper Katiyar N and Hossain F 2007 An Open book Watershed Model for Prototyping Space borne Flood Monitoring Systems in International River Basins Environmental Modeling and Software In press doi 10 1016 j envsoft 2006 12 005 to understand the concept illustrated and applied in m2. Ka npur 1 0 Open book Watershed Model If the time remaining in the time step calculated in overland flow processing is less than half of the time required no movement of water takes place otherwise water moves to next pixel If it moves to next pixel this same routine is called recursively for movement of water in the remaining time until the time remaining is less than half of the time required for the next movement When it comes out of this routine the location of water packet will be the destination pixel value And with these values destination pixel matrix will be updated later On the other hand if flow is channel flow fn channel flow processing is called using l v channel flow flow value Note the unit of l v channel flow is m s not m s Fn total inflow 2 pixel This routine calculates total inflow to pixel Flow from the pixel will be inflow to pixel under consideration and rainfall taking place of the pixel during the given time step q inflow to pixel rainfall Rainfall is calculated using this formula rainfall i L COS t Fn total inflow 2 pixel channel This routine calculates total inflow to pixel for a channel pixel Processing of this routine is same as routine fun total inflow 2 pixel except the unit for flow in channel pixel is not unit width Fn laminar overland flow processing This routine calculates the velocity of the sheet flow if the flow regime is laminar Resistant coefficient is calcula
3. anual and implemented code Disclaimer This manual is in no way comprehensive Users with no background in or understanding of distributed hydrology are strongly advised against using this code in any mode particularly in operational mode Besides knowledge of basic hydrology experience with typical numerical techniques used in physically based hydrodynamic models is recommended as it will help the user grasp capabilities and limitations of this model Users are encouraged to experiment with the model and venture in hydrology textbooks and journal papers to learn more about the topics touched upon in this manual No claims are made regarding the suitability of the model for any particular purpose The model was written for research and educational purposes with a particular focus on rapid prototyping of satellite based flood forecasting systems in International River Basins Nitin Katiyar 6 4 2007 Last saved by Nitin Page 3 Kanpur 1 0 Design Specification for Open book Watershed Model Input Files File Name Format dem txt No of pixels in x direction TAB No of pixels in y direction TAB Pixel Length of the square pixel this program assumes square grid NEW LINE Elevation values in grid format gFTTPPPTA Dem txt channel width txt This File is in grid format with the x and y number of grid according to dem txt file It contains 0 if the land belonging to this pixel is NOT a channel flow othe
4. ed in v velocity Since we know the velocity and the length of the movement Again if the remaining time is less than the half of the v travel time no more movement takes place Water packet has reached its destination for the current time step Otherwise again rountine fn channel flow processing is called recursively Fn river flow processing This routine is used for velocity calculation in channel flow To start with new depth l v new depth is taken as 1 00 and the old depth 1 v old depth is assumed to Nitin Katiyar 6 4 2007 Last saved by Nitin Page 14 Design Specification for Ka npur 1 0 Open book Watershed Model be very small number 9999 0 Until we get the negligible difference between new calculated depth and depth in the previous step keep doing the following processing Newton s Method is used for faster convergence Q I n So B Y BY B 2Y and Yn 1 Ys 1 Q Qy SB 6Y J GY BA 2Y After these iterations the final depth will be the stable depth for Manning s equation Velocity is calculated by flow depth formula Nitin Katiyar 6 4 2007 Last saved by Nitin Page 15
5. gn Specification for Open book Watershed Model Defining some global variables a slope n river width 250 150 3 This 3 D array contains Elevation River width Slope a resultant direction 250 150 2 This 3 D array contains resultant flow direction of the pixel a cosine 250 150 Cosine of the slope angle a rainfall 250 150 Rainfall for a particular rainfall period a inflow 2 pixel 250 150 Inflow to pixels for all the pixels from other pixels a destination pixel 2 Location of the destination pixel a temp inflow 2 pixel 250 150 Temporary storage of inflow to pixels a soil storage Soil storage a grid Response time 250 150 Grid response time a saturation excess flow 250 150 Saturation excess flow a porosity Porosity a depth Depth of bed rock Nitin Katiyar 6 4 2007 Last saved by Nitin Page 6 Design Specification for Ka npur 1 0 Open book Watershed Model Technical Specifications Algorithm Formulation of the Model and Computer Implementation Computer processing of the open book model involves several routines Details about how the code works and input output files are documented in manual prepared for the model These routines have been written in computer language C to implement equations mentioned in previous sections Figure illustrates the flow diagram of the model The model basically requires seve
6. h values in a slope n river width i j 1 Call subroutine fn populate slope file Call subroutine fn populate cos file Read viscosity manning s roughness threshold reynolds number gravitational acceleration time step saturated hydraulic conductivity and field capacity from parameter file into v viscosity v mannings coefjcient v thre reynids no v_gravitation_accln v_time_step v sat hydraulic conduc and v field capacity Read porosity values into global array a porosity and depth values into global array a depth Read percentage soil storage into global array a soil storage after multiplying this value by porosity and depth of the corresponding pixel Convert unit of saturated hydraulic conductivity from cm hr to m s Call subroutine fn populate grid response time array Nitin Katiyar 6 4 2007 Last saved by Nitin Page 9 Design Specification for Ka npur 1 0 Open book Watershed Model Rainfall file may contain several sets of dataset time period and data in grid format Store it in v time period and Call fn get rainfall data Every period of rainfall storm period may contain several time steps So for each time step Call fn linear storage discharge and fn overland flow processing subroutines Print final output at the end of this time step in the output file and reinitialize array a inflow 2 pixel Currently subroutine which check if a inflow 2 pixel to all pixels is zero and rainfall is also zero or not for termination condi
7. ion for Ka npur 1 0 Open book Watershed Model Fn populate cos file Routine Store the cosine of the slope angle into a cosine array This array is calculated using cos arc tan slope formula Fn linear storage discharge Excess rainfall is calculated from input rainfall using linear discharge conceptualization Time step and v time period are in min so for further calculation it needs to be multiplied with 60 for it to be in sec For each pixel on the watershed following processing is performed If the soil moisture is greater than field capacity a soil storage i j gt v field capacity subsurface flow is calculated using qs S t Sy t and stored in v subsurface flow otherwise it is zero If soil moisture storage is greater than storage capacity a soil storage i j gt a porosity i j a depth i j overland flow is calculated using qse S t Sy A t and stored in a saturation excess flow i j otherwise it is zero qse is also known as excess rainfall Using qs and qs values update the soil moisture storage S t A t S t Rainfall dss qse time step Fn_overland_flow_processing Time step is taken as input for this routine Basic purpose of this routine is to rout the flow as far as it can reach in this given time step In this beginning l v time remaining v time period For each pixel following processing takes place Check if the pixel is a river pixel If river width value for thi
8. me k Time k 1 Overland Flow Turbulent Flow Routing Update Flow Distribution Is all inflow routed Channel Flow Generate Hydrograph Terminate Processing Flow diagram of computer data processing of the model Nitin Katiyar 6 4 2007 Last saved by Nitin Page 8 Design Specification for Ka npur 1 0 Open book Watershed Model step it is checked 1f the flow is still If flow is turbulent based on Reynolds number criteria Manning s Equation is used to rout the flow until the flow meets the channel Overland sheet flow meets the channel flow and becomes part of channel For channel flow Manning s Equation is used and for faster convergence of this equation Newton method is used This routed flow is reported for the outlet pixel of the open book watershed as the flow for a given time of the flood hydrograph For the benefit of future users of the developed computer code a user s manual has also been prepared that details how the code is to be applied on a computer system Main Routine Open files dem txt channel width txt rainfall txt parameter txt soil storage txt porosity txt depth txt in read mode and output txt in write mode If error in opening file output file will contain the error Read x and y direction pixels pixel length and dem values in v x pixels v y pixels v pixel length and a slope n river width i j 0 a slope n river width is three dimensional array Read river widt
9. n different types of ASCII input files with strict formatting namely DEM file Rainfall Data Channel Width data Depth or soil column Porosity Initial Soil Moisture Storage and Parameter files Model input is converted into SI units Preprocessing of the data is required by the model The slope values for each pixel are generated which is the steepest gradient among the surrounding eight pixels in its neighborhood Time varying rainfall is read from the input file and then converted to overland flow For each pixel excess rainfall is calculated For this purpose a linear storage discharge reservoir scheme is followed Subsurface flow is nonzero if soil moisture storage in the soil column for that pixel is greater than field capacity Overland saturation excess flow is calculated if soil moisture storage is greater than soil storage capacity otherwise it is zero Using Overland flow and subsurface flow soil moisture is updated for every time step using the mass balance equation 1 The overland flow is assumed to be sheet flow over the surface Using a Reynolds number threshold value provided in the parameter input file identification of the overland flow regime as laminar or turbulent is conducted If the overland flow in laminar flow is routed For every time Nitin Katiyar 6 4 2007 Last saved by Nitin Page 7 Design Specification for Ka npur 1 0 Open book Watershed Model Rainfall Data Flow Distribution Pixel i j Ti
10. processing where channel flow processing takes place la temp inflow 2 pixel is a temporary array which stores the inflow to pixel values for all the destination pixels v out flow quantity which is nothing but the v flow channel value This value is added to the mentioned temporary array Assign zero to v out flow as it has already been added to l a temp inflow 2 pixel array Add these temporary inflow to pixel values to the a inflow 2 pixel array Fn sheet flow processing This routine is called for each pixel in routine fn overland flow processing If the pixel under consideration is a river pixel follow the following instructions Calculate the Reynolds number of the flow and store in l v reynolds no Reynolds Number 4q Now the Reynolds number criteria given in the parameter input file and stored in l v thres reynlds no decides about the type of regime laminar turbulent If the flow is laminar velocity of the flow is calculated using fn laminar overland flow processing function otherwise it is calculated using fn turbulent overland flow processing Length of movement will be pixel under consideration to the destination pixel from the vector matrix Knowing the length and velocity time taken for movement of water packet from pixel under consideration to destination pixel is calculated This calculated time is stored in v travel time variable Nitin Katiyar 6 4 2007 Last saved by Nitin Page 12 Design Specification for
11. rwise it will have channel width Note wherever we have a channel pixel the width of the pixel will be taken from this file not from dem file so dem file will have extra pixel gU channel width txt rainfall txt Rainfall Period in min NEW LINE Rainfall Values in grid format unit mm hr l Rainfall txt parameter txt Viscosity in MKS TAB Manning s Roughness TAB Threshold Reynold s Number to decide about the laminar turbulent TAB Gravitational Acceleration TAB Time Step for the calculation in min TAB Saturated Hydraulic conductivity in cm hr TAB Field Capacity Nitin Katiyar Last saved by Nitin 6 4 2007 Page 4 Kanpur 1 0 Design Specification for Open book Watershed Model parameter txt soil storage txt Percentage Storage in grid format Soil storage will be calculated using depth of bed rock and porosity Soil storage txt porosity txt Porosity in grid format Porosity txt depth txt Depth of bed rock in grid format in m Depth txt output txt Flow at outlet point of watershed in m s Dm Output txt Naming convention Function Subroutine fn Input File Name fl in Output File Name fl out File Pointer Ip Array a_ Variable v Local Variable v andl v Counters DS ii s US Nitin Katiyar 6 4 2007 Last saved by Nitin Page 5 Kanpur 1 0 Desi
12. s pixel is nonzero i e it has a nonzero entry in channel width file for this pixel It can be confirmed by checking that a slope n river width i j L if very small For this code small value considered is 0 001 some one may choose a very small value as the double precision is considered If the flow is not river flow then calculate flow by calling routine fn total inflow 2 pixel and store it in v flow Initialize inflow to pixel array for this pixel a inflow 2 pixel i j Nitin Katiyar 6 4 2007 Last saved by Nitin Page 11 Design Specification for Ka npur 1 0 Open book Watershed Model 0 0 Callfn sheet flow processing for overland sheet flow processing la temp inflow 2 pixel is a temporary array which stores the inflow to pixel values for all the destination pixels v out flow quantity which is nothing but the v flow value This value is added to the mentioned temporary array Destination pixel is calculated and discussed in fn sheet flow processing routine Assign zero to v out flow as it has already been added to a temp inflow 2 pixel array If the flow is channel flow following processing takes place An important thing that should be noticed is that unit here is m s not m s as in sheet flow Channel flow is calculated by calling fn total inflow 2 pixel channel routine and stored in v flow channel Initialize inflow to pixel array for this pixel a inflow 2 pixel i j 0 0 Call fn channel flow
13. ted using formula Resistance Coefficient Cr 96 108 rainfall i and stored in local variable l v resis coeff 1 v friction factor is calculated by dividing resistant coefficient by Reynolds number Water depth is calculated using Nitin Katiyar 6 4 2007 Last saved by Nitin Page 13 Design Specification for Ka npur 1 0 Open book Watershed Model Darcy Weisbach Equation Depth y Friction Factor f Flow qo 2 8 Gravitation Acceleration g Slope So 1 3 Hence velocity of flow is Velocity of flow V Flow qo Depth y Fn turbulent overland flow processing This routine calculates the velocity of the sheet flow if the flow regime is turbulent Depth of water is calculated using Manning s Equation Water Depth y Manning s Equation n Flow qo sqrt Slope So 3 5 and stored in v water depth Using this value v velocity is calculated by dividing v flow by v water depth Fn channel flow processing Since the outlet point of the watershed is at the end of channel flow As soon as water flows to the outlet point no more movement takes place for this water packet If during processing pixel under consideration is outlet pixel provided in the parameter file Assign the whole flow to this pixel for the current time step and terminate the processing for this time step for the pixel under consideration Velocity of flow is calculated by calling fn river flow processing and velocity is stor
14. tion has been removed It can be again added at the same place Currently the program runs for 200 more timesteps after rainfall ends to stabilize itself In every run it calls Call fn linear storage discharge and fn overland flow processing subroutines Close all the open files at the end Any error that occurred while closing is written in output file Fn get rainfall data Routine Read rainfall data from the file and convert the unit from mm hr to m s by multiplying it with 2 777777X10 Declaration of the variable as double is used for storing this Fn populate slope file Routine For every pixel in the grid get the 8 neighboring pixels for boundary pixels neighboring pixels will be less than 8 If the neighboring pixel s x coordinate or y coordinate is same as that of the pixel under consideration then the slope to that pixel will be elevation difference upon pixel length because both pixels are side by side On the other hand if the pixel is neighboring pixel without satisfying the above criteria then the slope will be elevation difference divided by pixel length multiplied with square root of two because both pixels are diagonally situated For every pixel under consideration find out the steepest slope and store it in a slope n river width i j 2 and resultant pixel direction in a resultant direction i j 0 and a resultant direction i j 1 Nitin Katiyar 6 4 2007 Last saved by Nitin Page 10 Design Specificat
Download Pdf Manuals
Related Search