TASK 1 - Field sampling
1.1 Implementing Alternate Sampling Strategies
You will generate ten random points for Gippsland.
- The polygon shape file you should use is called region.shp. This ensures that it has the same spatial extent as the other sampling strategies.
- Select Create Random Points in the Data Management Toolbox (under Feature Class).
- Select a name for your new point feature class, the constraining feature class, and the number of points to be created.
You will need to generate four sampling strategies: Rand10 where ten points are located within the shapefile called region, Rand55 where 55 points are randomly distributed within the Region shapefile,oneperquad where one point is located in each 20km by 20km cell, and twoperquad where two points are located in each cell.
You will need to spatially join Rand10.shp and Rand55.shp to quadra20.shp to get the associated Quadrat Id for those randomly generated points. This will be helpful for comparing statistical values.
NOTE: Because we are dealing with random numbers every student should have a different answer for this section!
1.2 Sampling elevation using ArcGIS and GRID data
- Ensure that the Spatial Analyst extension is on.
- Import the DEM file using ArcToolbox. Download the DEM called dem1000.asc to your working directory. This is currently a zipped file and needs to be unzipped before use.
This is how you will import your DEM file which is in ASCII format:
ArcToolbox> Conversion Tools > To Raster > ASCII to Raster
You should now see the DEM for Gippsland that we will sample. We will use the Extract Values to Points tool in ArcToolboxto extract z-values from the DEM.
ArcToolbox>Spatial Analyst Tools>Extraction>Extract Values to Points
NB: The Spatial Analyst extension must be turned on for this tool to work.
- Add all random point files to your view.
- For each sampling theme, select the layer and use the Extract Values to Points tool (in Spatial Analyst Tools -> Extraction).
- View the attribute table for the sampling theme. You should see the new elevation model values added to this table.
- Export for use in Excel. The statistical analysis you are required to perform is described below.
1.3 Obtaining statistics for the entire DEM
There are a number of ways of obtaining the statistics for the entire DEM.
One way is to approximate it by generating a very large random sample across the Region.shp file and calculate the statistics of this - perhaps 500 random points. Another way is to convert your DEM1000 to a Shapefile by utilizing ArcToolbox: Conversion Tools > From Raster > To Point or To Polygon.
You can then view and export the attribute table of this shape to Excel as it will contain a value for every grid cell in your data.
1.4 Statistical analysis of sampling strategies
The analysis you are expected to complete will utilise the skills developed in earlier assignments, namely the creation of histograms of population distributions and consequent statistical analysis using Excel. Hence these steps will not be repeated here. Compare the results from each sampling strategy.
Q1. What is the difference between the region.shp and Quadrat20.shp files? What does Quadrat20.shp allow us to do?
Q2. What is the difference between stratified and random sampling?
Q3. What were the similarities and differences between oneperquad and twoperquad? Be sure to explain pattern, dispersion, and density.
Q4. What is variance-mean-ratio? Explain.
Q5. Address the consequences of various sampling strategies in the context of real applications using your Gippsland data as a case study.