[STATSGRASS] gstat probability maps

[Wolf Bergenheim]

Pierluigi,

That would be indicator kriging:

I use gstat, so I can't give you any specific R commands, but the
principle is the same: You have to create a new dataset which is 1 if
the threshold is exceeded, and 0 if not. Now your data is 0 or 1. Then
do normal ordinary kriging on that new data column. The resulting map
shows the probability of exceeding the threshold. In gstat you have to
add  extra checks to see that your result is strictly [0.0 - 1.0]

--Wolf

On 27.03.2007 18:18, Pierluigi De Rosa wrote:
> Dear All,
> 
> I need to prepare a probability map of exceeding a value...
> I can make the simulation, for example I tried to make it with data meuse:
> 
> v <- variogram(log(zinc)~1,loc= meuse) 
> m <- vgm(.59, "Sph", 874, .04)
> sim <- krige(formula = log(zinc)~1, locations = meuse, model = m, newdata = 
> meuse.grid, nmax = 15, nsim = 5)
> names(sim)
> [1] "sim1" "sim2" "sim3" "sim4" "sim5"
> 
> Now How can I create a probabilility map of exceding a value (for example 5)??
> 
> Thanks
> Pierluigi
> 

-- 

<:3 )---- Wolf Bergenheim ----( 8:>