Dear community,
Thanks to the answers we receive about our questions. But reading the
comments, I think I need to reformulate a bit...
I'm looking for reference about an objective way to discard a dataset
regarding the missing data distribution when kriging is used as
interpolation method.
We use kriging interpolation on satellite image.
So satellite images in the visible range can have some holes due to the
presence of clouds. Thus we use kriging to interpolate data.
However the cloud distributions can change drastically between days and these
distributions influence the interpolation results (large cloud patches vs small
could patches).
We are looking to quantify numerically this "patchiness" in a simple way.
We plot the number of point pairs in function of the distance between
the points (which is not a kernel neither the number of available points per
area). We define a treshold as an indication of the cutoff induced by presence
of very large cloud patches (for example if the number of available pairs
decrease after a distance d, we consider the image non recoverable by kriging).
We choose this function to do not add more calcul in the interpolation process
(the main objective is to give an introduction to kriging to people not very
familiar to this method). This number of pairs points in function of the
distance between the points is calculated during the variogram calculation.
I check in the bibliography the use of this function to quantify the patchiness
of a dataset and more generaly a way to check if a dataset can be reasonably
kriged, and I wasn't able to find any reference.
Does anyone have some information ?
Thanks in advance !
Laurent Dubroca
--
Laurent Dubroca
Stazione Zoologica Anton Dohrn
Villa Comunale 80121 Naples Italy
Phone/FAX: 0039 081 5833 247/224
___________________________________________________________________________
Découvrez une nouvelle façon d'obtenir des réponses à toutes vos questions !
Profitez des connaissances, des opinions et des expériences des internautes sur
Yahoo! Questions/Réponses
http://fr.answers.yahoo.com
+
+ To post a message to the list, send it to [email protected]
+ To unsubscribe, send email to majordomo@ jrc.it with no subject and
"unsubscribe ai-geostats" in the message body. DO NOT SEND
Subscribe/Unsubscribe requests to the list
+ As a general service to list users, please remember to post a summary of any
useful responses to your questions.
+ Support to the forum can be found at http://www.ai-geostats.org/
