[R-sig-Geo] How to model anisotropy and get final isotropic variogram for kriging using gstat package?

[Uzzal]

I have dataset contains hourly Particulate matter concentrations (PM10) of 1 march2012,1.00 am for 104 sites. Please download from HERE. My ultimate goal is to do ordinary Kriging (spatial kriging) analyses for this dataset. As far as I know, I need to plot a isotropic variogram for kriging analysis. For this purpose, I have plotted a omnidirectional variogram with following R code using gstat package:  seoul311
#######################################################library(sp)
library(gstat)
library(rgdal) seoul311 #assign a CRS
proj4string(seoul311) =  "+proj=longlat +datum=WGS84"  #Reprojecting data to utm by rgdal
library(rgdal)
seoul311 #plot Omnidirectional Variogram
seoul311.var #Model fit
model.311 After this, I wanted to check the anisotropy. so, I plotted directional variogram for every 10 degree with tolerance=5 degree by following code:
#Directional Variogram
seoul311.var1 Question: What should I do next to get a final isotropic variogram for Ordinary Kring analysis? How can I model anaisotropy using gstat package in R?  [I am kind of struck here. I have read many documents and example for 2 months but couldn't get a clear procedure to do this! Could anyone please explain in details that what code I should write and what aspect I should keep in mind before starting Ordinary Kriging for this data set? Overall. it will be very helpful for me if I got step by step procedure for doing variogram analysis  before starting kriging.]

Orpheus 




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://stat.ethz.ch/pipermail/r-sig-geo/attachments/20151011/20f3dd88/attachment.html>