I have to prepare a spatial map of various soil properties for that I am confused about the semi-variogram is compulsory or not?
Most often yes , unless grid size takes into account the area covered...
Anoop Kumar Srivastava for my analysis, sampling locations are very distant from each other. so whats in this situation
Where krigging is done , not covered within those grid sampling areas , and more importantly , we need to establish that such variograms could serve as decision support tool ..??
Calculate a semivariogram and compare the range of influence to the distance between sample locations. If the semivaariogram is flat i.e. pure nugget, your points are more than one range of influence apart and any interpolated values are essentially pure uncertainty.
Yes, you need to compute the semivariogram to run kriging. If you do not have measurements close enough to each other, then you miss information about the microvariability of your sample and you cannot distinguish very easily between this microvariability (variability in small scales) and actual measurements errors. If your measurements are way too far from each other then you will simply detect (through the semivariogram) a pure decorrelation regime between the measurements (sort of a flat line ) in which case I think kriging is highly problematic, you simply do not have data dense enough to perform it correctly.
Keep in mind that in general kriging is benefitted by measurements that are in variable distance from each other, as opposed to measurements that have more or less the same distance from each other. Variability in measurement distances informs the semivariogram on how the decorrelation functions for different distances. The more accurate the semivariagram the most positive the effect on the kriging estimation.
- Badges
- Science topic
- Similar topics
- Counseling Psychology
- Confusion