Steps to producing an interpolated surface

Page Shortcuts

Page Content

Steps to producing an interpolated surface

Steps to producing an interpolated surface

  1. Prepare a directory

    It is important that you have a project directory set up for your analysis which contains one directory for each interpolation which will be run.
    • AUNSPLIN produces 23 files and one grid for each interpolation.
    •  Because each of the files have the same name for each interpolation they must be kept separate.
  2.  Load the appropriate themes

    At least one point theme and a grid are required for an analysis.
  3. Assess the sampling density

    Determine what cell size the output surface should have. This will depend on the density of the point data (refer to the document; Integrated Toolset/spatial modeling training notes [IT_training.doc]).
  4. Analysis extent & cell size

    Check 'Analysis / properties' to determine the extent and cell size of the output grid and error surface. The extent can be set with the “Graphic on grid” tool, however be careful (especially in projections such as AMG) and ensure that the cell size will equally divide into the extent set.

  5. The ANUSPLIN interface

    Image of the ANUSPLIN interface

  6. Choosing an error surface

    Image of LAPGRD information

When the model fails

There are a few reasons why the model might fail, these are the most common:

  • Greater than three decimal places (for the Grid values)
    Image of error message when greater than three decimal places used
  • Duplicate point records (two records in the point theme have the same Latitude and Longitude)
    Image of error message when duplicate point records are used
  • Cell size is not equally divisible into the extent selected
    Image showing error message when cell size in not equally divisible into the extent selected

Important output files

In the directory specified in the ANUSPLIN interface ('Output Data File') 23 files will be produced (shown; right). Many of these files are generated for the model only.

image of 23 files produced

The important files are:

  • result.gcv: the data in this file can be graphed (X/Y) in a program such as Excel. This shows the smoothing parameter for the model.

image of the estimate of overall error

  • result.grd: This is the Binary Raster version of the final output grid from the model (which is automatically added to the view as a permanent grid.
  • result.log & resulterr.log: These text files contain all relevant information about the model surfaces, including input parameters and output statistics. Also, if the model fails whilst producing the surfaces the reason for the failure will be written at the end of these files.
  • resulterr.grd: This Binary Raster surface is the error surface produced by the model. It can be imported directly into ArcView.
  • result.dat: This file contains a copy of the data which was read from the original point theme.
  • result.lis: This file contains a list of all point data read by the model plus the predicted value for each record.
  • resultpnt.log: If the user withholds a subset of the point data (in the form of a text file). These values can be used by the model as another (fully independent) means to check for errors in the surface. The final output of this error analysis is written to this file.

The Integrated Toolset - Introduction | Data Requirements | Steps to Produce an interpolated surface | Integrated Toolset Manual | Integrated Toolset Training | Integrated Toolset Presentation | Download the Integrated Toolset



Last reviewed: 28 Mar 2007
Contact:
Email Bureau of Rural Sciences
About Bureau of Rural Sciences and Contact details
Media Inquiries