TOPOG: User Guide - Appendix A, Additional Information on Splin2h and Grdcon





	Appendix A
	Additional Information on Splin2h and Grdcon

A.1 More about Splin2h
An example of an .spl data file is shown below.

Line Data
01 ... 1 02 ... 1 03 ... 1.0 04 ... 0 05 ... 0.0 06 ... 5.0 10.0 40.0 07 ... 30 08 ... 600.0 1500.0 09 ... burra-all.res 10 ... 5.0 11 ... 6000.0 16000.0 12 ... 50000.0 68000.0 13 ... 20.0 14 ... 2 15 ... burra-all.cns 16 ... 5 17 ... (i6,f12.2,i2) 18 ... (8f10.2) 19 ... burra.bound 20 ... 4 21 ... (i6) 22 ... (8f10.2) 23 ... burra-all.grd 24 ... 0 25 ... (10f8.2) 26 ... burra-all.snk 27 ... burra-all.str

Line 01 This is the Drainage Option. It has 4 possible values; 0, 1, 2 and 3. The most useful options are 0 and 1. 0 means do not use drainage enforcement, and 1 means use drainage enforcement (recommended).

Line 02 This is the Contour Option. It has 2 possible values; 0 and 1. 0 means that the elevation data are mostly in the form of spot heights, and 1 means that the elevation data are primarily in the form of contours.

Line 03 This is the Residual Mean Square Error. Its value varies from 0.0 up to infinity effectively. It is set to 0.0 for data that are well chosen spot heights. It is set to 1% of input contour interval for digitally scanned contour data, and up to 12% of input contour interval for poor or hand digitised contour data. If the spot heights are not well chosen, the RMS may be increased.

Line 04 This is the Maximum Option. It has 2 possible values; 0 or 1. 0 means that maxima do not need to be on grid points, and 1 means that maxima must be on grid points. We recommend that you use option 0.

Line 05 This is the Derivative Trade-off. This sets the compromise between the sharp bends in contours along ridge and drainage lines and the minimum curvature required in planar areas. Where contours are the dominant input a value of 0.0 is recommended, i.e. follow what the contours say, and where well chosen spot heights are the dominant input a value of 0.5 is recommended, i.e. the spot heights are guides to the best lines, and use your judgement where the sharp curves are.

Line 06 This contains three tolerances for criteria to remove sinks. The first must be > 4*RMS, the second must be >= 2*first tolerance, and the third must be > 2*second tolerance. The supervisor program will check for these inequalities.

Line 07 This is the maximum number of iterations to achieve the desired RMS. Its value is between 20 and 30. Poor elevation data require more iterations.

Line 08 These are the lower and upper elevation limits of the input data. If input spot heights are outside this range, the program merely clips them out, but if an input contour is outside the range, Splin2h will produce an error message and stop.

Line 09-10 The first line contains a file name for large data errors. The file is called the residual file, hence the extension, and contains any grid point where the difference in input and calculated elevation exceeds a certain value. If this line is not blank, then the next line contains the minimum value written to the residual file. If this filename is a blank line, then no file is created and residuals are not checked.

Line 11 These are the x, or eastings, limits of the input data. They upper limit is usually rounded upwards to a power of 10, and the lower limit is usually rounded down to a power of 10. This makes grid spacing much simpler.

Line 12 These are the y, or northings, limits of the input data, and they are similarly rounded.

Line 13 This is the spacing between grid points of the final DEM. The smaller this number, the finer the resolution but the longer Splin2h will take to run. For catchment areas below approx 2 km², a grid of 200 to 400 each way is recommended. In our particular case, the area is 10 km by 18 km, so the grid resolution is (10000/20 + 1) by (18000/20 + 1) = 501 by 901. The grid spacing must be an exact divisor of both the x and y ranges.

Line 14 This is the number of datafiles of input data. It is not necessarily the case that more input means a better grid!

Line 15-18 These lines describe the first datafile. It is a contour line file and needs a filename, a type 5 indicator, a header format, and a data format. The supervisor programs can describe certain datafiles produced by topog_decode and apply default file names, types and formats. See the Splin2h documentation for more information of file type and formats.

Line 19-22 These lines describe the second datafile. In this case it is a polygon file that describes the outer extent of the data and that all areas outside should not be interpolated into.

Line 23 This is the result grid filename. This cannot be blank and will contain the DEM finally calculated.

Line 24 This is the Grid Type. It has 2 possible values; 0 and 1. 0 means that the grid is output row by row from the south most (bottom) to the north most (top), from west (left) to east (right). 1 means the file should be written out as (x,y,z) triplets. The supervisor program automatically assigns this to be 0.

Line 25 This is the output grid format. This must be a full FORTRAN format with parentheses, but may be blank and the file will be written in binary.

Line 26 This is the Sink Filename. It will contain a list of all the sinks in the final grid. This is a most useful file as it can give an indication of where data density is poor, or where input data are conflicting. If the filename is a blank line, no file will be produced.

Line 27 This is the Stream Filename. It will contain an ASCII output of the input and calculated streams. This is usually quite large and not normally used. If the filename is a blank line, no file will be produced.

Take me out of frames

Appendix A continued ....

last modified on 13 June 1997

Line 01	This is the Drainage Option. It has 4 possible values; 0, 1, 2 and 3. The most useful options are 0 and 1. 0 means do not use drainage enforcement, and 1 means use drainage enforcement (recommended).
Line 02	This is the Contour Option. It has 2 possible values; 0 and 1. 0 means that the elevation data are mostly in the form of spot heights, and 1 means that the elevation data are primarily in the form of contours.
Line 03	This is the Residual Mean Square Error. Its value varies from 0.0 up to infinity effectively. It is set to 0.0 for data that are well chosen spot heights. It is set to 1% of input contour interval for digitally scanned contour data, and up to 12% of input contour interval for poor or hand digitised contour data. If the spot heights are not well chosen, the RMS may be increased.
Line 04	This is the Maximum Option. It has 2 possible values; 0 or 1. 0 means that maxima do not need to be on grid points, and 1 means that maxima must be on grid points. We recommend that you use option 0.
Line 05	This is the Derivative Trade-off. This sets the compromise between the sharp bends in contours along ridge and drainage lines and the minimum curvature required in planar areas. Where contours are the dominant input a value of 0.0 is recommended, i.e. follow what the contours say, and where well chosen spot heights are the dominant input a value of 0.5 is recommended, i.e. the spot heights are guides to the best lines, and use your judgement where the sharp curves are.
Line 06	This contains three tolerances for criteria to remove sinks. The first must be > 4RMS, the second must be >= 2first tolerance, and the third must be > 2*second tolerance. The supervisor program will check for these inequalities.
Line 07	This is the maximum number of iterations to achieve the desired RMS. Its value is between 20 and 30. Poor elevation data require more iterations.
Line 08	These are the lower and upper elevation limits of the input data. If input spot heights are outside this range, the program merely clips them out, but if an input contour is outside the range, Splin2h will produce an error message and stop.
Line 09-10	The first line contains a file name for large data errors. The file is called the residual file, hence the extension, and contains any grid point where the difference in input and calculated elevation exceeds a certain value. If this line is not blank, then the next line contains the minimum value written to the residual file. If this filename is a blank line, then no file is created and residuals are not checked.
Line 11	These are the x, or eastings, limits of the input data. They upper limit is usually rounded upwards to a power of 10, and the lower limit is usually rounded down to a power of 10. This makes grid spacing much simpler.
Line 12	These are the y, or northings, limits of the input data, and they are similarly rounded.
Line 13	This is the spacing between grid points of the final DEM. The smaller this number, the finer the resolution but the longer Splin2h will take to run. For catchment areas below approx 2 km², a grid of 200 to 400 each way is recommended. In our particular case, the area is 10 km by 18 km, so the grid resolution is (10000/20 + 1) by (18000/20 + 1) = 501 by 901. The grid spacing must be an exact divisor of both the x and y ranges.
Line 14	This is the number of datafiles of input data. It is not necessarily the case that more input means a better grid!
Line 15-18	These lines describe the first datafile. It is a contour line file and needs a filename, a type 5 indicator, a header format, and a data format. The supervisor programs can describe certain datafiles produced by topog_decode and apply default file names, types and formats. See the Splin2h documentation for more information of file type and formats.
Line 19-22	These lines describe the second datafile. In this case it is a polygon file that describes the outer extent of the data and that all areas outside should not be interpolated into.
Line 23	This is the result grid filename. This cannot be blank and will contain the DEM finally calculated.
Line 24	This is the Grid Type. It has 2 possible values; 0 and 1. 0 means that the grid is output row by row from the south most (bottom) to the north most (top), from west (left) to east (right). 1 means the file should be written out as (x,y,z) triplets. The supervisor program automatically assigns this to be 0.
Line 25	This is the output grid format. This must be a full FORTRAN format with parentheses, but may be blank and the file will be written in binary.
Line 26	This is the Sink Filename. It will contain a list of all the sinks in the final grid. This is a most useful file as it can give an indication of where data density is poor, or where input data are conflicting. If the filename is a blank line, no file will be produced.
Line 27	This is the Stream Filename. It will contain an ASCII output of the input and calculated streams. This is usually quite large and not normally used. If the filename is a blank line, no file will be produced.