Continuous phenomena and field data
It is always important in the context of geospatial data to distinguish between the phenomenon itself and its representation in data. This is especially so with data that represent phenomena that are theoretically measurable at all locations, but in practice only measured at sample locations. The phenomenon (for example, atmospheric pressure) could, in principle, be measured at every location on Earth, but in practice, it cannot be. Instead, we work with a set of measurements at control points and interpolate these to generate a field that approximates the continuous phenomenon. How the control points and any subsequent interpolation results are stored can have important effects on subsequent analysis steps.
In this section, a variety of methods for storing field data are discussed, and some of their advantages and disadvantages considered. The schemes you are most likely to encounter in practice are point sampling schemes of one kind or another, particularly grid-based samples. Grid-based sampling schemes are particularly useful as the basis for surface analysis operations, because the equally-spaced measurements simplify many mathematical operations. Further, since any other surface data can be readily converted to this format by interpolation, in the remainder of this lesson, we assume that field data are available as a grid.
However, it is important to bear in mind that grids of field values often did not start out that way, but instead were produced by interpolation from a much sparser set of control points. As a consequence of this, you should check any associated metadata before assuming the accuracy of data values in a grid.