[STATSGRASS] kriging results not within range of original data

[Edzer J. Pebesma]

Hi Carlos,

that is correct, and due to the fact that kriging weights are not 
constrained to be in [0,1]. In fact, they do become negative and may be 
larger than 1. Consider as alternatives either inverse distance weighted 
interpolation, or some form of constrained kriging (although I wouldn't 
know where to get the latter).
--
Edzer

Carlos "Guâno" Grohmann wrote:

> Hello,
>
> I'm dong some kriging of SRTM data, and noted that the kriged results
> don't lie within the range of the original data. It is not too far
> from the original, but there is a difference.
>
> original dem:
>
> Data attributes:
>     band1
> Min.   : 634.0
> 1st Qu.: 759.0
> Median : 814.0
> Mean   : 856.5
> 3rd Qu.: 923.0
> Max.   :1312.0
>
> resampled dem:
>   var1.pred
> Min.   : 633.2
> 1st Qu.: 759.2
> Median : 814.3
> Mean   : 856.4
> 3rd Qu.: 923.5
> Max.   :1307.4
>
>
> I konw thatt small deviations from the values range can occur, but I
> was looking for some deeper explanation on this. If someone could
> point me out any literature I can refer to, I will be thankful.
>
> Carlos
>