Sampling & Interpolation
Spatial interpolation is the prediction of variables at unmeasured locations, and based on sampling of the same variables at known locations.
Spatial prediction also involves estimation of variables at unsampled locations but differs from interpolation in that estimates are based at least in part on other variables
Samples are spaced at uniformly at fixed intervals.
It is usually not the most statistically efficient sampling pattern because all areas receive the same sampling intensity, regardless of the level of variation on the landscape
Random sampling involves placing sample points at randomly generated XY locations. Random samples have an advantage over systematic samples in that they are unlikely to match any pattern in the landscape. However like systematic sampling random sampling does nothing to distribute samples in areas of high variation
Cluster sampling distributes sample points around centers that are generated by some random or systematic method. Reducing travel time is the primary advantage of cluster sampling
Adaptive or stratified sampling is characterized by frequent sampling in variable areas and sparse sampling in uniform areas
(IDW)
(IDW)
The top interpolation is using the 12 nearest points, with an exponent of two.
The lower interpolation is using the four nearest points, and an exponent of three.
Local influences are stronger as the exponent increases and the number of sample points decrease
Spatial prediction also involves estimation of variables at and between sampled locations, but differs from interpolation in that estimates are based at least in part on other variables
"everything is related to everything else, but closer things are more related."
Waldo Tobler's 1st Law of Geography
Spatial Autocorrelation is the tendency of nearby objects to vary in concert with one another. High values occur together, as do low values.
Includes three main components: