Module 2.1: Surfaces – Interpolation

Thiessen assumes the values of unsampled locations are equal to the value of the nearest sampled point. It creates a polygon around each data point and assigns that value – the advantage is that it is easy to understand, and values of the sample points remain. However, its disadvantage is it also easily generalize data that can create miscalculated values at unsampled locations. This can be use if you are okay with generalizing your model per area (district, zone, region) - sort of an overview.

Inverse Distance Weighting (IDW) uses an interpolation method that estimates values by weights inversely to its distance. The weight of a value decreases as it moves away from an unsampled location and vice versa (more influence or weight value). 

  

While spline creates a smooth curve across sample points and minimizes curvature as it pass through each data point. Regularized spline creates a smooth with gradually changing surface. Tension spline has control to the stiffness or surface weight which creates values closer to the sample data values.

Spline (Regularized)

 

Spline (Tension)

Overall for the Tambay BOD water quality analysis, I’d choose the Spline with tension as it minimizes curvature and passes through each data point taking consideration of the sampled min and max producing a better and smooth surface. In addition, our analysis is focused on water BOD concentration which does not change radically between close sample area points. With our test data being limited, then it is just right that we use Spline with tension to minimize inaccuracy of data.