Hello,
I have a number of 10 in-situ soil loss and depositional rates in a small sloppy agricultural plot (ca. 500 km2), estimated by on site measurements. Now, I want to interpolate this data set to calculate a net soil loss through the entire experimental plot. Which is the best method for this interpolation in ArcGIS? Is the Inverse distance weighted (IDW) method more appropriate? Any justification? Is there any literature to be referred relevant to this type of interpolation indicating reasons for the preference to the method? Appreciate your answer.
Tilak
I would prefer the recent advancement of kriging method: Empirical Bayesian Kriging (EBK), which also come as an extension in ArcGIS-10.1, because EBK eases the most difficult aspects of producing a valid kriging model by creating several subsets of the data and performing multiple semivariogram simulations to derive a best fit. It also takes care of the non-stationary nature of geochemical data.
However, I must include here that the choice of proper model should be evaluated based on the prediction quality which depends on the numbers of data, spatial density of data distribution and nature of data.
Hello Tilak,
Good to see you here ! I guess your plot is 500 m² and not 500 km². What I would suggest you to do is to use some kind of model to make the extrapolation: either through correlating soil losses with terrain attributes such as slope gradient and curvature (and then use the variation of these parameters over your plot) or through the use of an erosion model that can handle complex terrain. WATEM/SEDEM is such a model (can be downloaded from http://ees.kuleuven.be/geography/modelling/watemsedem/download/.
You can find more inspiration in some papers like Govers et al., ESPL, 1996. There we interpolate 137Cs measurements to entire fields, based on simple model which was later developed into SPEROS (Van Oost et al., 2006 and other papers).
Kind regards, Gerard
Thanks to Gerard and all others for answering to this question. Tilak
Kriging method is more appropriate in this case compared to IDW, as IDW involves linear trend if you want to use ARCGIS. You can use some sediment models or sediment routing.
- Badges
- Science topic
- Similar topics
- Agriculture
- Soil Science