Module 2.2: Interpolation

Module 2.2: Interpolation

This weeks module was about using different interpolation techniques to represent different data points. The different interpolation techniques we used were Thiessen, Inverse Distance Weighted (IDW), and Spline (Regularized and Tension). Each one of the techniques shows a different way to represent data and it all depends on what the person wants to represent changes which technique you are going to use. The techniques were used to represent water quality in Tampa Bay, Florida.

The first technique was the Thiessen or Nearest Neighbor interpolation. One advantage of using Thiessen interpolation is that it assigns a value for any unsampled location that is equal to the value found at the nearest sample location. This is helpful because it makes each point its own polygon. Another advantage is that it provides an exact interpolator which means that the interpolated surface equals the sampled values at each sample point. It means that each location is preserved and there is not a difference between the true and interpolated values at the points. A major disadvantage is the sharp edges between the different points. It causes the image to look harsh with the edges. A final disadvantage is that the areas there are no points causes bigger polygons than what might actually be in the area. An example would be if they were on a terrain that changes quickly and there is no point in which the polygon would not represent that change.

The second technique is IDW. When looking at the data distribution of the IDW it is based on the assumption that the points that are close to one another are more alike than the points farther away. This could mess up the areas that have significantly different elevations next to each other. The IDW does not provide prediction standardized error which makes the use of the DEM a little problematic. Also, the IDW is oversensitive to outliers which could cause “bullseye” areas that have one point as a high or low point.

The third technique is Spline (Regularized and Tension). Spline, it is helpful because it is smoother and looks like it could be more accurate than the IDW. It is important to note that the DEM only works if the data is very accurate. The use of Spline is more adaptable because they may be used for lines or surfaces, and they may be estimated. Also, Spline is points that is used are considered “guides” that helps make the lines smoother. Lastly, a major problem to consider is that it can create unpredictable surfaces (overestimation or underestimation) in areas with low data density or high topographic change.

What I learned about surface interpolation is that there is not one method that is better than another. When choosing the technique all depends on different factors such as cost of sampling, available resources, and accuracy. Also, it was interesting to see that the sampling techniques can change how you try to represent your data. The results that surprised me were how spline interpolation can cause a big change just because one extra point was in the area. All I did was change the spline regularized interpolation by taking away one point. The way I would decide on a different technique is if I knew the sampling techniques that were used. I feel like the systematic technique would help make all the interpolations better whereas the cluster and random sampling would need to be more with spline interpolation. Lastly, when doing the adaptive sampling pattern would work best with the spline tension interpolation because it seems to be like the original data points.

The image below is Spline Tension interpolation. This image to me represented the data points the best. The reason is because it seemed to be the most like the non-spatial data that was represented. Also, the spline interpolation is smoother when it is representing the data points. Lastly, I chose this technique because they can be changed to represent the data points, and they cab ne used for lines and surfaces unlike the other techniques.