Spatial prediction of two soil properties using topographic information

dc.contributor.advisor	Kalivas, Dionissios P.	en
dc.contributor.advisor	Καλυβάς, Διονύσιος Π.	el
dc.contributor.author	Triantakonstantis, Dimitrios P.	en
dc.contributor.author	Τριαντακωνσταντής, Δημήτριος Π.	el
dc.contributor.author	Kollias, Vasiliki J.	en
dc.contributor.author	Κόλλια, Βασιλική Ι.	el
dc.date.accessioned	2015-06-22T11:36:20Z
dc.date.available	2015-06-22T11:36:20Z
dc.date.issued	2002-09-30
dc.identifier.uri	http://62.217.125.90/xmlui/handle/123456789/6407
dc.identifier.uri	http://journal.gnest.org/sites/default/files/Journal%20Papers/kalivas.pdf
dc.rights	CC0 1.0 Παγκόσμια	*
dc.rights.uri	http://creativecommons.org/publicdomain/zero/1.0/	*
dc.subject	Co-kriging	en
dc.subject	Clay	en
dc.subject	Sand	en
dc.subject	Distance to river	en
dc.subject	Kriging	en
dc.subject	Regression kriging	en
dc.title	Spatial prediction of two soil properties using topographic information	en
heal.recordProvider	Γεωπονικό Πανεπιστήμιο Αθηνών/Αξιοποίηση Φυσικών Πόρων	el
heal.publicationDate	2002-03
heal.bibliographicCitation	Kalivas, D. Spatial prediction of two soil properties using topographic information, Global Nest Journal, vol. 4 (1), pp. 41-49, Global NEST, 2002	en
heal.abstract	The objective of this study was to determine whether the use of the co-regionalization of the distance-to-river topographic variable with the soil properties topsoil clay and sand can improve their mapping. The interpolation techniques: ordinary kriging, kriging combined with regression (two models) and heterotopic co-kriging were applied to data from 153 observation points. The two models of kriging combined with regression involve: (a) linear regression of the two soil variables with the distance-to-river variable on the 153 observation points followed by kriging and (b) summation of the kriged regression values and kriged regression residuals. For co-kriging 350 additional observations for the distance-to-river-variable were employed. The distance-to-river data were easily obtained from the map of the area which was stored in a Geographical Information System (GIS). The performances of the methods were evaluated and compared using the cross-validation method. The mean error of prediction indicates reasonably small bias of prediction for the two soil variables by almost all the methods. The mean square error showed that heterotopic co-kriging produced better estimates of the soil variables than kriging but there was a clear advantage in using the first model of kriging combined with linear regression technique. The second model of kriging combined with regression does not show any particular advantage over the other methods	en
heal.publisher	Global NEST	en
heal.journalName	Global Nest Journal	en
heal.journalType	Open Access
dc.identifier.doi	http://journal.gnest.org/sites/default/files/Journal%20Papers/kalivas.pdf