Catchment, Landform, and River Network analysis - Sisyphe
Contents
1. Scilab is governed by the CeCILL license GPL compatible abiding by the rules of distri bution of free software since Scilab 5 family See the information delivered on the Free Software Foundation on CeCILL at the following URL http www fsf org licensing licenses You can use modify and or redistribute the software under the terms of the CeCILL license as circulated by CEA CNRS and Inria at the following URL http www cecill info Copyright c 2011 2013 Scilab Enterprises Copyright 1989 2012 INRIA Copyright c 1989 2007 ENPC Contents 1 Introduction 2 Getting started 2 1 Building and running CLaRiNet 211 Opete eme an asi a sde ds 212 Windows VETBEDEL o ie nonu Geek Du ncm dm aa ri Bl OEVER ce aa uana En eom LED edo ca ee 2 14 Demo AMAR ann ae aten bed A EUR ee 2 2 dale STATS ns aa We duel OR 221 Handling structures o rasas ae ee ou de de Rs 222 Gid IRE uos ae er ari B 24 9 FloatRaster structure cscs ho essa aa 224 FlowDirGrid structure 4 2 4 sa assen o 4 48 SER OY 2 25 MUCE unos D ne tia a Ri AUN a mb mas da cbe noo PCT Cc zal Online Help 2 4 doi a oem bacs om euge 2 02 OE WSS 22 2444 Bk B re ve 24 Beer Samples 22 2 2624 aaneen ae ee Bb ER EO a 241 Extracting a catchment oane koe2. 2 4 Basic examples 2 4 1 Extracting a catchment Demo script extract catchment sce n path libraryinfo toolbox clarinetlib DEMOPATH fullpath path demos fdrfile DEMOPATH fdr_eagleriver flt fdrcoding 32 64 128 16 9999 1 8 4 2 fdr CreateFlowDirFromBinary fdrfile f fdrcoding header GetFlowDirGridHeader fdr ncols header 1 nrows header 2 cellsize header 3 x11 header 4 yll header 5 LL header 7 outlet XY2rec header 929387 1901293 Create a new raster NODATA_value 9999 bas CreateFloatRaster ncols nrows cellsize x11 y11 NODATA_value LL SetRasterToNODATA bas Flag catchment pixels with value 1 in this new raster FlagCatchmentCells fdr outlet bas 1 Write raster to file WriteRasterToFile bas DEMOPATH bas_eagleriver flt Of course in this example we have flagged only one catchment in the raster so it would have made more sense to do this operation with a GIS software However it is possible to flag as many catchments as one wish in a single raster We will see that this functionality will be very usefull for generating topographic partitions e g for running a spatially distributed hydrological model 2 4 2 Computing a hypsometric curve Demo script hypsometric_curve sce 16 CHAPTER 2 GETTING STARTED m path libraryinfo toolbox_clarinetlib DEMOPATH fullpath path demos Run the c
3. ae age A 2 4 2 Computing hypsometric curve 24 3 Strahler ordering lt scoas crpio Rn area 3 Writing complex workflows Bibliography O0 O0 A AAN 10 10 10 12 12 14 14 15 15 15 15 16 19 21 CONTENTS Chapter 1 Introduction CLaRiNet is a library for geomorphological analysis using raster datasets It has been developed with the intent of providing low level functions for batch analyses as such it does not compete with more user friendly graphically interfaced packages that come with popular GIS softwares e g Hydrology tools within ArcGis Spatial Analyst In contrast the functions provided by CLaRiNet are designed to be interfaced within a more general programming environment in or der to perform advanced analyses of geomorphological variables sampled at numerous locations in the geographical space Currently CLaRiNet is distributed in the form of a main library written in C and Fortran and a Scilab gateway written in C even though the main library could be interfaced with any software The choice of Scilab as the host environment has been motivated by the fact that is a very general open source high level numerically oriented programming language It provides an interpreted programming environment with matrices as the main data type This allows users to rapidly construct models for a wide range of mathematical problems Hereafter 1t is assumed that the user has a basi
4. these 4 pixels are siblings which can be sorted by decreasing size 35 43 51 and 34 Pixel 43 is second and the only information we need to store is The next sibling of pixel 43 is pixel 51 The most convenient aspect of this approach is that each link in the tree is a fixed size structure containing only static elements an integer ID a first pointer to the first child structure and a second pointer to the next sibling structure If a link has no child then the first pointer is set to NULL If it has no subsequent sibling then the second pointer is set to NULL A CellTree structure is created by invoking the CreateCellTree function with the handle on flow direction grid and the outlet pixel as arguments m path libraryinfo toolbox_clarinetlib DEMOPATH fullpath path demos fdrfile DEMOPATH fdr_eagleriver flt fdrcoding 32 64 128 16 9999 1 8 4 2 fdr CreateFlowDirFromBinary fdrfile f fdrcoding header GetFlowDirGridHeader fdr outlet XY2rec header 929387 1901293 t CreateCellTree fdr outlet The function returns a handle on the newly created structure i e an integer ID of the tree Practically besides a pixel ID and two pointers each link of the linked list can be associated with a weight In the previous example each pixel has a unit weight so that the total weight of the tree rooted at a given outlet is the number of upstream sites on Figure 2 4 the t
5. CLaRi Net Catchment Landform and River Network analysis A geomorphological toolbox for Scilab User Manual draft Nicolas Le Moine UMR Metis Universit Pierre et Marie Curie 4 Place Jussieu 75252 Paris cedex 05 France CLaRiNet 0 1 0 December 2013 Copyright 2013 UPMC CLaRiNet is free software you can redistribute it and or modify it under the terms of the GNU General Public License as published by the Free Software Foundation either version 3 of the License or at your option any later version CLaRiNet is distributed in the hope that it will be useful but without any warranty without even the implied warranty of merchantabil ity or fitness for a particular purpose See the GNU General Public License for more details You should have received a copy of the GNU General Public License along with the program see the file COPYING txt if not write to the Free Software Foundation Inc 675 Mass Ave Cambridge MA 02139 USA ClaRiNet uses a subset of the LAPACK routines developed at Univ of Tennessee Univ of California Berkeley NAG Ltd Courant Institute Argonne National Lab and Rice University They are copyrighted by the University of Tennessee and covered by a BSD style license see LAPACK txt ClaRiNet also uses a subset of the BLAS routines developed by Jack Dongarra Argonne National Lab Jeremy Du Croz NAG Ltd lain Duff AERE Harwell Richard Hanson Sandia National Labs and Sven Hammarling NAG Ltd
6. FloatRaster structure The first data structure used in CLaRiNet is the FloatRaster structure it is used to store arrays of 32 bit floating point values The structure contains the raster data and its georeferencing information New FloatRaster structures can essentially be created in two ways 2 2 CLARINET DATA STRUCTURES 11 e the function CreateFloatRaster will create a new structure with specified dimensions and georeference information ras CreateFloatRaster ncols nrows cellsize x11 yll NODATA value LL where LL must be either center if x11 and 11 refer to the center of the lower left pixel or corner if these coordinates refer to the corner of the lower left pixel e the function CreateRasterFromBinary will load a raster dataset stored in a binary file on the disk ras CreateRasterFromBinary binfile datatype where binfile is the path relative or absolute to the binary file on the disk An ascii header file having the same and the extension hdr must be located in the same directory in order to provide the referencing information The possible values for the datatype argument are given in Table 2 1 Whatever the data type in the file it will be converted to 4 byte floating point in order to be stored in the FloatRaster structure Data type Record size byte datatype argument Unsigned char 1 uc Short integer 2 gh Unsigned short integer 2 Integer 4 Unsigned integer 4 ui Float
7. atchment flagging script exec DEMOPATH extract_catchment sce 1 Define the empirical frequencies of the desired quantiles P 0 1000 1000 elev CreateRasterFromBinary DEMOPATH elev_eagleriver flt f Compute elevation quantiles using catchment as a mask msk bas Qz RasterCDF elev msk P Plot the curve elevations are in cm in the raster plot 1 P Qz 100 xtitle Exceedance probability Elevation m The RasterCDF function can be used to compute any zonal CDF e g rainfall topographic index 2 4 3 Strahler ordering Strahler stream ordering in a catchment is a good example of a problem which is better addressed with the use of a tree structure The following code demo script strahler sce illustrates how to extract and order streams with a support area threshold for channelization of 1 km The resulting plot is shown in Figure 2 5 m path libraryinfo toolbox_clarinetlib DEMOPATH fullpath path demos fdrfile DEMOPATH fdr eagleriver flt fdrcoding 32 64 128 16 255 1 8 4 2 fdr CreateFlowDirFromBinary fdrfile f fdrcoding header GetFlowDirGridHeader fdr outlet XY2rec header 929387 1901293 t CreateCellTree fdr outlet cellsize header 3 pixelarea cellsize 2 supportarea 1e6 in square meters supportweight supportarea pixelarea STREAMS StreamOrder MakeStrahlerStreams t support
8. c knowledge of Scilab environment Docmenta tion and tutorials for Scilab can be found at the following URL http www scilab org en resources documentation In the second chapter of this short manual the basic data structures handled by CLaRiNet will be presented as well as some basic functions In the third chapter we will see how to combine these functions in order to write more complex workflows and to couple them with Scilab capabilities e g graphics optimization tools CHAPTER 1 INTRODUCTION Chapter 2 Getting started 2 1 Building and running CLaRiNet Currently the only documented way to use CLaRiNet library is through the Scilab interface though the library can be freely interfaced with any code This section describes how to compile the source and build the Scilab toolbox 2 1 1 Compatibility issues CLaRiNet interface has been developed with the Scilab 5 3 x API and there has been some substantial changes in this API for versions 5 4 x so that the current interface will not build with these latest versions It is recommended to use CLaRiNet with Scilab 5 3 3 either under Linux or Windows 2 1 2 Windows version The CLaRiNet library is written in C and Fortran and the Scilab interface is written in C so you will need a set of compilers to build the toolbox from the source The most secure way to do this is to 1 install the MinGW package Minimalist GNU for Windows which provides a set of compilers and t
9. erface and macros open the toolbox folder first run cleaner sce and the builder sce just drag and drop the files in the Scilab console At the end of the process a file loader sce should have been created run this file in order to load CLaRiNet 2 1 4 Demo dataset The demo dataset used in the following examples is a small tributary of the Colorado River in the Southern Rocky Mountains the Eagle River below Gypsum CO closed at USGS stream gage 09070000 see Figure 2 1 Geographic coordinates Latitude 39 38 58 Longitude 106 57 11 NAD 1983 Albers projected coordinates X 929387 m Y 1901293 m Catchment area 2445 km Elevation data comes from the hydro conditioned Digital Elevation Model DEM provided in the National Hydrography Dataset NHD Plus McKay et al 2012 2 derived from the 30 meter National Elevation Dataset NED USGS 2009 3 Data files are located in the demo directory of the toolbox 2 1 BUILDING AND RUNNING CLARINET 9 Figure 2 1 Overview of the Eagle River dataset Grid on map indicates false eastings and northings in meters in the NAD 1983 Albers projection Co lora do River Basin 870000 860000 1920000 1920000 o es 1870000 1880000 1890000 1890000 1870000 1860000 1860000 Z 940000 930000 920000 910000 900000 890000 880000 Elevation m 10 CHAPTER 2 GETTING STARTED 2 2 CLaRiNet data structures 2 2 1 Handling
10. hen 2 install the MinGW Compiler support for Scilab 5 3 written by Allan Cornet available through the ATOMS package manager in Scilab 1 You need to install MinGW package distributed by Equation Solution http www equation com servlet equation cmd fa programminglog ftp ftp equation com gcc On Windows 32 bits platform x86 ftp ftp equation com gcc gcc 4 5 2 32 exe On Windows 64 bits platform with Scilab 32 bits x86 ftp ftp equation com gcc gcc 4 5 2 32 exe with Scilab 64 bits x64 ftp ftp equation com gcc gcc 4 5 2 64 exe 2 Once you have installed MinGW run Scilab open the ATOMS package manager and browse for the MinGW toolbox in the category Windows Tools Install the package and restart Scilab The compiler support should be available you can check it by calling the function haveacompiler which should return True 7 8 CHAPTER 2 GETTING STARTED Now you can build the ClaRiNet library interface and macros open the toolbox folder first run cleaner sce and the builder sce just drag and drop the files in the Scilab console At the end of the process a file loader sce should have been created run this file in order to load CLaRiNet 2 1 3 Linux version In order to build CLaRiNet under Linux you will need to have Fortran and C C compilers check that you have installed gcc and gfortran or get them from the package manager of your distribution In order to the ClaRiNet library int
11. ing point 4 en Long integer 8 1 Unsigned long integer 8 ul Double 8 q Table 2 1 Data types supported by the CreateRasterFromBinary or CreateFlowDirFromBinary functions The following code shows how to load the elevation data for the Eagle River dataset m path libraryinfo toolbox_clarinetlib DEMOPATH fullpath path demos binfile DEMOPATH elev_eagleriver flt elev CreateRasterFromBinary binfile f Note the content of the ascii header file elev_eagleriver hdr associated with the binary file ncols 2772 nrows 2132 xllcorner 941985 888 yllcorner 1856628 955 cellsize 30 NODATA value 9999 byteorder LSBFIRST 12 CHAPTER 2 GETTING STARTED 2 2 4 FlowDirGrid structure In order to perform geomorphological analyses in river networks it is necessary to provide information about local flow directions CLaRiNet uses a classical D8 approach Jenson and Domingue 1988 1 in which each pixel can flow into one of its eight neighors You may read write flow direction grids from to binary rasters using any coding convention see Figure 2 3 for the ESRI ArcGis convention 1 64 16 1 128 EE Direction Code 4 4 K 64 gt E 1 Ta N BE 2 4 An I 1 4 5 4 v SW 8 8 B 16416 64 a W 16 46 16 16 2 4 X NW 32 32 16 2 4 4 T N 64 NE 128 Figure 2 3 ESRI flow direction coding elevation values left map and corresponding flow directions right map Because flow directions can
12. nto the same downstream pixel Figure 2 4 illustrates how the pixels of the catchment closed at pixel 49 can be transformed into a binary tree 2 2 CLARINET DATA STRUCTURES Y 12 13 18 19 le 22 z 34 35 36 37 38 13 41 eben 50 51 52 id pointer to next sibling 49 weight y v pointer to 42 5 41 50 first child v v v 35 43 51 34 y v T v 27 36 44 51 v T v 28 19 18 26 37 45 8 2 1 1 2 1 v M v v v v 29 20 11 38 4 3 1 1 v Y v v 21 22 30 12 S 13 v v v v v Figure 2 4 Transformation of the pixel topology into a binary tree in a catchment 14 CHAPTER 2 GETTING STARTED This structure stores the topology in a way which avoids redundancy Consider the tree rooted at pixel 43 e Pixel 43 has two children trees the one rooted at pixel 36 and the one rooted at pixel 44 both pixels 36 and 44 drain into pixel 43 The largest of these two trees is the one rooted at pixel 36 it has a total weight of 3 hence we only store the information The first larger child of pixel 43 is pixel 36 e Pixel 43 which has a total weight of 6 drains into pixel 42 its ancestor pixel so do pixels 34 size 1 35 size 14 and 51 size 2 Hence
13. only take 8 different values a 1 byte type is sufficient to store this information For this reason CLaRiNet uses a distinct data structure for storing flow direction grids instead of the 4 byte FloatRaster structure The following code shows how to load the flow direction grid for the Eagle River dataset which uses the ESRI convention by invoking the CreateFlowDirFromBinary function This function creates a FlowDirGrid data structure in memory and returns a handle on this structure in Scilab n path libraryinfo toolbox clarinetlib DEMOPATH fullpath path demos fdrfile DEMOPATH fdr eagleriver flt fdrcoding 32 64 128 16 255 1 8 4 2 fdr CreateFlowDirFromBinary fdrfile f fdrcoding 2 2 5 CellTree structure We have seen that the flow direction grid contains the exhaustive pixel topology and that it is sufficient to perform any search upstream or downsteam a given pixel However for some prob lems it can be usefull to translate this topological information into a different representation namely a binary tree The CellTree structure is designed for this purpose A cell tree is always rooted at some outlet pixel and spans upstream from this outlet It is a linked list in which each link contains the pixel ID and two pointers a pointer to its first child i e the outlet of largest catchment draining into the pixel and a second pointer to its next sibling i e the second smallest catchment draining i
14. ree rooted at pixel 35 has a total weight of 14 Instead of it the tree can be constructed with a different weight for each pixel this local value of the weight can be read in a raster For example we can build a topographic partition based not on equal areas but on equal inflow volumes by using a tree with the map of mean annual rainfall as a weighting raster see Figure XX the procedure will be explained in Chapter 3 2 3 Getting help 2 3 1 Online help You can access online help either by the toolbar or F1 shortcut or by the prompt by typing help followed by the name of a function gt help CreateFloatRaster CLaRiNet online help provides a quite exhaustive list of the functions in the library and a description of their syntax 2 4 BASIC EXAMPLES 15 2 3 2 Error messages Thanks to the Scilab API CLaRiNet provides minimal error management If you encounter an error it is often wiser to quit and restart Scilab In some cases CLaRiNet may crash without warning in particular recursive traversing in very large networks may cause stack overflow due to an excessive recursion depth Some important functions have dual implementations a recursive one and a slower but safer iterative one but not all of them In the cases where recursive calls fail repeatedly you may want to consider splitting the problem into smaller parts e g calling the function on several subcatchments More iterative implementations will be added in future
15. structures The basic purpose of the CLaRiNet main library is to handle large raster datasets which would otherwise not fit in the Scilab stack an important part of the code is dedicated to memory allocation and management In order to access data from Scilab CLaRiNet uses a system of handles a handle can be seen as an abstract reference to a data structure stored outside the Scilab stack In CLaRiNet every handle is simply an integer identifier which is used to grab an existing data structure 2 2 2 Grid indexation Since the main library is written in C CLaRiNet uses a row major order for internal storage of raster data Indices start at 0 so that the raster cells are numbered as shown in Figure 2 2 In many functions grid cells are referenced to by their memory offset 0 1 2 3 4 5 6 7 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 A Center of lower left pixel x yllcenter Corner of lower left pixel x yllcorner Figure 2 2 Illustration of row major cell indexation and georeferencing convention used in CLaR iNet with a small 8 x 8 grid Rasters are georeferenced using the x y coordinates of the lower left pixel Two conventions are possible can either indicate the position of the center of the lower left pixel or the position of the corner of this pixel 2 2 3
16. weight Plot streams COLOR jetcolormap max StreamOrder for i 1 length STREAMS XY rec2XY header STREAMS i order StreamOrder plot XY 1 XY 2 Color COLOR order LineWidth order end a gcal a isoview on 2 4 BASIC EXAMPLES 1 92e 06 a 1 91e 06 Y ses N nl j 1 90e 06 f dd 7 Adds 1 89e 06 X py 2 c i TA E 1 88 06 1 c d o E Y x 1 87 06 2 1 Pe order 2 order 3 7 order 4 order 5 1 85e 06 T T T T T T 9 4e 05 9 3e 05 9 2e 05 9 1 05 9 0e 05 8 9e 05 8 8e 05 8 7e 05 8 6e 05 X easting m Figure 2 5 Extraction of Strahler streams for the Eagle River catchment 18 CHAPTER 2 GETTING STARTED Chapter 3 Writing complex workflows 20 CHAPTER 3 WRITING COMPLEX WORKFLOWS Bibliography Jenson S K and J O Domingue 1988 Extracting topographic structure from digital elevation data for geographic information system analysis Photogrammetric Engineering and Remote Sensing 54 11 1593 1600 McKay L T Bondelid T Dewald et al 2012 NHDPlus Version 2 User Guide ftp ftp horizon systems com NHDPlus NHDPlusV21 USGS 2009 National Elevation Dataset edition 2 U S Geological Survey Sioux Fall SD 21
Download Pdf Manuals
Related Search