[R-sig-Geo] Problem with semi variance estimates in space time variogram (gstat)

[Tom Gottfried]

Am 15.06.2012 00:06, schrieb Edzer Pebesma:
> I believe the sample variance of the data should equal the semivariance
> averaged over all distances (in space and time).

A helpful paper on that subject:

BARNES, RJ. „The variogram sill and the sample variance“. Mathematical 
Geology 23, Nr. 4 (1991): 673–678.

Tom

> You computed this over a very restricted (space time) range.
>
> A large scale trend in the data could lead to a strong difference
> between semivariances and sample variance; think of evaluating the
> variogram of z=x+y, where x and y both range from 0 to 1000, and the
> varioram is evaluated up to distance 10.
>
> On 06/14/2012 05:47 PM, Rawlins, Barry G. wrote:
>> Hello
>>
>> The variograms function in gstat is returning semi-variance estimates
>> which appear to be two orders of magnitude lower than what I expect from
>> calculating the variance of the data in the attached space time object
>> created with the STFDF function:
>>
>> library(gstat)
>> library(spacetime)
>> breaks<-seq(0.5,12.5,1)
>> var_all<- variogram(VMC~1,sp_time_var, bounds=breaks, cutoff = 12.5) #
>> VMC is the numeric data to calculate a sp-time varigram
>> plot(var_all) # note the low values ranges on the scale<0.15
>> var(sp_time_var at data$VMC) # variance is 11.5
>>
>> So I would expect the variance in the upper part of the graph to
>> approach a value around 11.
>>
>> Have I mis-specified the variogram function? I know from gstat that one
>> had to specify a gstat object, but here I thought specifying the
>> spacetime object was the right way to proceed and that VMC~1 was the
>> correct syntax?
>>
>> Any advice much appreciated.
>> Barry
>>
>>
>>
>> --
>> This message (and any attachments) is for the recipient only. NERC
>> is subject to the Freedom of Information Act 2000 and the contents
>> of this email and any reply you make may be disclosed by NERC unless
>> it is exempt from release under the Act. Any material supplied to
>> NERC may be stored in an electronic records management system.
>>
>>
>> _______________________________________________
>> R-sig-Geo mailing list
>> R-sig-Geo at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>