Mapping the topsoil pH and humus quality of forest soils in the North Bohemian Jizerské hory Mts. region with ordinary, universal, and regression kriging: cross-validation comparison.

Vasat R., Pavlů L., Borůvka L., Drabek O., Nikodem A . (2013): Mapping the topsoil pH and humus quality of forest soils in the North Bohemian Jizerske hory Mts. region with ordinary, universal, and regression kriging: cross-validation comparison. Soil & Water Res., 8: 97–104. North Bohemia belongs to one of the most heavily industrialized and polluted regions in Europe. The enormous acid deposition which culminated in the 1970s has largely contributed to the accelerated acidification process in the soils and consequently to the wide forest decline in North Bohemian mountains. In this paper we map the active topsoil pH and humus quality with ordinary, universal, and regression kriging and compare the accuracy of resulting maps with cross-validation. For the regression kriging we use two types of spatially exhaustive auxiliary information, first the altitude derived from digital elevation model and second the land cover classes derived from satellite imagery. The leave-one-out (cross-validation) statistics, i.e. mean error, root mean squared error, and mean squared deviation ratio, are taken for comparison since they are widely accepted as measurements of the accuracy of digital soil maps. The results show that the regression kriging is superior over other kriging methods in this case. Out of 97 sampling sites the regression kriging with land cover classes is the best predictor at 32 sites for pH and at 30 sites for humus quality, the regression kriging with altitude at 31 and 25 sites, the universal kriging at 21 and 23 sites, and the ordinary kriging at 13 and 18 sites. The highest number of best predictions for regression kriging implies that the topsoil pH and humus quality are driven approximately equally by land cover and altitude and little less by pure geographic position. Furthermore, the universal kriging maps show a northeast to southwest spatial trend of topsoil pH and a northwest to southeast spatial trend for humus quality.

[1]  L. Borůvka,et al.  Distribution of aluminium among its mobilizable forms in soils of the Jizera Mountains region , 2018 .

[2]  Gerard B. M. Heuvelink,et al.  Sampling for validation of digital soil maps , 2011 .

[3]  R. Kodešová,et al.  A Numerical Study of the Impact of Precipitation Redistribution in a Beech Forest Canopy on Water and Aluminum Transport in a Podzol , 2010 .

[4]  L. Borůvka,et al.  Assessment of soil aluminium pools along three mountainous elevation gradients. , 2009, Journal of inorganic biochemistry.

[5]  G. Reinds,et al.  Modelling recovery from soil acidification in European forests under climate change. , 2009, The Science of the total environment.

[6]  Gerard B. M. Heuvelink,et al.  About regression-kriging: From equations to case studies , 2007, Comput. Geosci..

[7]  F. Oulehle,et al.  Modeling of the long-term effect of tree species (Norway spruce and European beech) on soil acidification in the Ore Mountains , 2007 .

[8]  Gerard B. M. Heuvelink,et al.  Accounting for change of support in spatial accuracy assessment of modelled soil mineral phosphorous concentration , 2006 .

[9]  M. Jurado-Expósito,et al.  Using geostatistical and remote sensing approaches for mapping soil properties , 2005 .

[10]  J. Hruška,et al.  Tree species (Picea abies and Fagus sylvatica) effects on soil water acidification and aluminium chemistry at sites subjected to long-term acidification in the Ore Mts., Czech Republic. , 2005, Journal of inorganic biochemistry.

[11]  L. Borůvka,et al.  Factors controlling spatial distribution of soil acidification and Al forms in forest soils. , 2005, Journal of inorganic biochemistry.

[12]  J. Ardö,et al.  Critical Loads of Acidity for Forest Soils and Relationship to Forest Decline in the Northern Czech Republic , 2004, Environmental monitoring and assessment.

[13]  Edzer J. Pebesma,et al.  Multivariable geostatistics in S: the gstat package , 2004, Comput. Geosci..

[14]  G. Heuvelink,et al.  A generic framework for spatial prediction of soil variables based on regression-kriging , 2004 .

[15]  Alex B. McBratney,et al.  A comparison of prediction methods for the creation of field-extent soil property maps , 2001 .

[16]  Hans Wackernagel,et al.  Multivariate Geostatistics: An Introduction with Applications , 1996 .

[17]  A. McBratney,et al.  Further results on prediction of soil properties from terrain attributes: heterotopic cokriging and regression-kriging , 1995 .

[18]  D. Brus,et al.  A comparison of kriging, co-kriging and kriging combined with regression for spatial interpolation of horizon depth with censored observations , 1995 .

[19]  R Core Team,et al.  R: A language and environment for statistical computing. , 2014 .

[20]  L. Borůvka,et al.  Profile distribution and temporal changes of sulphate and nitrate contents and related soil properties under beech and spruce forests. , 2013, The Science of the total environment.

[21]  Asi Building,et al.  Comparing Ordinary Kriging and Regression Kriging for Soil Properties in Contrasting Landscapes , 2010 .

[22]  Alfred Stein,et al.  Spatial prediction by linear kriging , 1999 .