Dear All,
I tried to fit variogram model (spherical, Gaussian, exponential, and
linear) to each year data from the same production field. Exponential
function seemed to be the best function in each year. I want to assess the
stability of the spatial pattern by looking at the correlation of predicted
yield values among the year from ordinary kriging.
I tried to use option nmax=100 and it works perfectly. This implies I am
using local interpolation of 100 observations to predict a value at unknown
locations rather than global interpolation approach. I think this is better
considering large size of my observation.
Now, I find the range of projected spatial coordinates (easting and
northing) in meters for each year data from the same field. I intend to
probably consider the interpolation grid of 10 meters spacing for each
data. Does it make sense? I wish to look at the correlation of predicted
yield value among the years obtain from ordinary kriging in other to assess
the stability of the spatial pattern quantitatively.
Since the spatial location of each yield value varies from one year to the
other, I thought of using the interpolation value from block ordinary
kriging rather than point ordinary
kriging. Please advise me on this. If I am to use block ordinary kriging,
how do I modify my script below?
The table below shows how the spatial coordinates vary from one year to
others in meters within the same production field.
*Easting range (meter)*
*Northing range (meter)*
years
*Min*
*Max*
*Min*
*Max*
2008
299,678
301,298
5,737,285
5,738,128
2009
299,678
301,299
5,737,278
5,738,129
2010
299,678
301,298
5,737,279
5,738,128
2011
299,679
301,298
5,737,279
5,738,123
This is my script
### Create grid for the interpolation (prediction) through ordinary kriging
easting.range <- as.integer(range(canmod.sp@coords[,1]))
northing.range <-as.integer(range(canmod.sp@coords[,2]))
## now expand to a grid with 100 meter spacing:
grd <- expand.grid(x=seq(from=easting.range[1], to=easting.range[2],
by=100),
y=seq(from=northing.range[1], to=northing.range[2], by=100))
names(grd)<-c("easting","northing")
coordinates(grd)<-~easting+northing
proj4string(grd)<-CRS("+proj=utm +zone=12 +ellps=WGS84 +datum=WGS84
+units=m +no_defs +towgs84=0,0,0")
Pred_ok <- krige(id="yield",yield ~ 1, canmod.sp, newdata = grd,
model=exp.mod)
Thank you all for your assistance.
Moshood
[[alternative HTML version deleted]]
_______________________________________________
R-sig-Geo mailing list
R-sig-Geo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
