For Project 5, you will explore the world of spatial interpolation, specifically

For Project 5, you will explore the world of spatial interpolation, specifically with respect to
bathymetric data in the Gulf of Mexico. The techniques you will learn are not specific to the Gulf of
Mexico. You will learn about various interpolation methods, when they are appropriate, how to
compare them, and their levels of accuracy when compared to continuous bathymetric data.
In a perfect world (for analysis), we would have spatial data covering every phenomenon, at every
scale, and we could represent and operationalize it. In the real world, we have data collection points
that are geographically dispersed, sometimes systematically, and sometimes based on a combination
of factors including natural barriers, cost, and convenience.
Interpolation is the process of estimating what is happening between data collection points by
analyzing the location and attributes of those known points to fill in the unknowns. This involves
taking a series of points and making a surface. Because this is spatial data, the concept is guided by
Tobler’s First Law of Geography: points that are closer together will generally share more similarities
than points further apart. While intuitive, the concept raises many questions. What criteria constitute
similarity? What spatial factors interrupt it? What are the consequences of aggregating characteristics
across space? As with most spatial analysis techniques, interpolation ends up being intuitive in
theory but more complex in practice.
Interpolation can be used to understand any spatially continuous phenomenon, such as elevation,
precipitation, or concentrations of pollutants in the air. This project employs point bathymetry data
– the depth under water of sea/lake floor – to interpolate a bathymetric surface. It is also possible to
estimate bathymetry values using remotely sensed imagery. These methods are not mutually
exclusive; some of the readings associated with this assignment will show the way in which these
methodologies can complement each other.
There are a variety of interpolation techniques, and the decision of which one to use will have an
effect on your conclusions. These techniques include inverse distance weighting, kriging, nearest
(natural) neighbor, thin-plate spline, and trend analysis. Part of this exercise will include learning
how these different techniques differ from one another.