[ai-geostats] var. regularisation and proportional effect
- Dear all,
I d`appreciate your help in two points. Firstly, the data I deal with are a series of circle shaped sample areas, each of them consisting of 25-35 basic units. The data subject to geostatistical analysis are average values per each sample area assigned to the center of the area. The first question relates to variogram regularisation. Since the number of basic units (in fact trees in the sample area) is unequal, how the regularisation should be accomplished in this case? As I read in a literature, the lenght of core is often mentioned in this regard to produce regularized variogram, but I am not just clear how is it in my case.
The second point is something like proportional effect. The mean values per each area increases along with variability (st.dev vs mean is R=0.76). How this affects the variography and kriging? I might mathmematicaly stabilize the variability, but I am afraid of spoiling the data. Besides, this phenomenon has certain interpretation in forest science. Is there some specific geostatistical treatment of this?
Thank you very much