Interpolation tips and best practices
1. Data interpolation, by its very nature, is fraught with "gotchas" (common mistakes based on invalid assumptions). The user must understand the limitations of the data interpolation methods and of the data used as the basis for interpolation. Be aware that data interpolation is merely an estimation based on probability. The results are only as reliable as the input data and the applicability of the interpolation method. If poor quality data is used, poor quality estimations will result. The user should "ground truth" results before using them to make decisions that may have social or economic consequences.
2. Many users try to generate DEMs (Digital Elevation Models) from contour data with the assumption that contours are a highly accurate and reliable data source. In fact, most contour layers are just estimations and generalizations, making their reliability as a data source quite low.
3. It is absolutely critical that the user choose the appropriate cell resolution for the interpolation. If data points are separated by 100 meters or more, do not expect to obtain sub-meter accuracy in the interpolated results.
4. The Spline command has a limitation that IDW and Kriging do not. The Spline algorithm generates a curved surface that may result in changes to surface volumes compared to other interpolation techniques.