Prediction of soil pH patterns in nature areas on a national scale

Research output: Contribution to journalArticleAcademicpeer-review

1 Citation (Scopus)

Abstract

Question: To assess the acidification process, nationwide information about soil pH on a site level is called for. Measurements of soil pH may be used, however there are not sufficient measurements available to map soil pH nationwide on site level. Instead we developed a soil pH map based on vegetation data. Location: Natural terrestrial areas in The Netherlands. Methods: 271,693 vegetation plots were used to estimate average soil pH per plot with indicator values, based on field measurements, of plant species. By spatial interpolation average pH values between the plots, with the soil type, groundwater table and vegetation management type as ancillary explanatory variables we created a soil pH map. The map covers all terrestrial nature areas (all areas that are not built up areas, agricultural areas and infrastructural areas) in the Netherlands with a map resolution of 25 × 25 m2 raster cells. Results: The predicted pH of the map varied between 3.0 and 8.6 with standard errors between 0.13 and 0.93. Most of the standard errors range from 0.4 to 0.55, with an average just below 0.5 pH unit. Cross-validation shows that for 33% the difference between observed and predicted is between −0.1 and 0.1 pH-unit and for 83% the difference is between −0.5 and 0.5 pH-unit. Validation shows that the pH map is unbiased (mean error is almost zero), accurate (root mean squared error is 0.64) and nicely captures spatial patterns (r = 0.77). We applied the pH map to assess the impact of acidification on the abiotic quality of nature areas in the Netherlands. Conclusions: The model fit in the predicted soil pH is in good resulting in a low standard error and a high correlation. The measures taken to prevent acidic deposition causing further acidifying of nature areas can be considered as successful.

Original languageEnglish
Pages (from-to)189-199
JournalApplied Vegetation Science
Volume22
Issue number2
Early online date7 Jan 2019
DOIs
Publication statusPublished - Apr 2019

Fingerprint

prediction
soil
acidification
vegetation
raster
interpolation
soil type
agricultural land
groundwater

Keywords

  • acidity
  • kriging
  • pH
  • relevé
  • soil type
  • vegetation

Cite this

@article{b6acab5e94334daaa325810d63afa9e4,
title = "Prediction of soil pH patterns in nature areas on a national scale",
abstract = "Question: To assess the acidification process, nationwide information about soil pH on a site level is called for. Measurements of soil pH may be used, however there are not sufficient measurements available to map soil pH nationwide on site level. Instead we developed a soil pH map based on vegetation data. Location: Natural terrestrial areas in The Netherlands. Methods: 271,693 vegetation plots were used to estimate average soil pH per plot with indicator values, based on field measurements, of plant species. By spatial interpolation average pH values between the plots, with the soil type, groundwater table and vegetation management type as ancillary explanatory variables we created a soil pH map. The map covers all terrestrial nature areas (all areas that are not built up areas, agricultural areas and infrastructural areas) in the Netherlands with a map resolution of 25 × 25 m2 raster cells. Results: The predicted pH of the map varied between 3.0 and 8.6 with standard errors between 0.13 and 0.93. Most of the standard errors range from 0.4 to 0.55, with an average just below 0.5 pH unit. Cross-validation shows that for 33{\%} the difference between observed and predicted is between −0.1 and 0.1 pH-unit and for 83{\%} the difference is between −0.5 and 0.5 pH-unit. Validation shows that the pH map is unbiased (mean error is almost zero), accurate (root mean squared error is 0.64) and nicely captures spatial patterns (r = 0.77). We applied the pH map to assess the impact of acidification on the abiotic quality of nature areas in the Netherlands. Conclusions: The model fit in the predicted soil pH is in good resulting in a low standard error and a high correlation. The measures taken to prevent acidic deposition causing further acidifying of nature areas can be considered as successful.",
keywords = "acidity, kriging, pH, relev{\'e}, soil type, vegetation",
author = "G.W.W. Wamelink and Walvoort, {Dennis J.J.} and Sanders, {Marlies E.} and Meeuwsen, {Henk A.M.} and Wegman, {Ruut M.A.} and Rogier Pouwels and Martin Knotters",
year = "2019",
month = "4",
doi = "10.1111/avsc.12423",
language = "English",
volume = "22",
pages = "189--199",
journal = "Applied Vegetation Science",
issn = "1402-2001",
publisher = "Wiley",
number = "2",

}

TY - JOUR

T1 - Prediction of soil pH patterns in nature areas on a national scale

AU - Wamelink, G.W.W.

AU - Walvoort, Dennis J.J.

AU - Sanders, Marlies E.

AU - Meeuwsen, Henk A.M.

AU - Wegman, Ruut M.A.

AU - Pouwels, Rogier

AU - Knotters, Martin

PY - 2019/4

Y1 - 2019/4

N2 - Question: To assess the acidification process, nationwide information about soil pH on a site level is called for. Measurements of soil pH may be used, however there are not sufficient measurements available to map soil pH nationwide on site level. Instead we developed a soil pH map based on vegetation data. Location: Natural terrestrial areas in The Netherlands. Methods: 271,693 vegetation plots were used to estimate average soil pH per plot with indicator values, based on field measurements, of plant species. By spatial interpolation average pH values between the plots, with the soil type, groundwater table and vegetation management type as ancillary explanatory variables we created a soil pH map. The map covers all terrestrial nature areas (all areas that are not built up areas, agricultural areas and infrastructural areas) in the Netherlands with a map resolution of 25 × 25 m2 raster cells. Results: The predicted pH of the map varied between 3.0 and 8.6 with standard errors between 0.13 and 0.93. Most of the standard errors range from 0.4 to 0.55, with an average just below 0.5 pH unit. Cross-validation shows that for 33% the difference between observed and predicted is between −0.1 and 0.1 pH-unit and for 83% the difference is between −0.5 and 0.5 pH-unit. Validation shows that the pH map is unbiased (mean error is almost zero), accurate (root mean squared error is 0.64) and nicely captures spatial patterns (r = 0.77). We applied the pH map to assess the impact of acidification on the abiotic quality of nature areas in the Netherlands. Conclusions: The model fit in the predicted soil pH is in good resulting in a low standard error and a high correlation. The measures taken to prevent acidic deposition causing further acidifying of nature areas can be considered as successful.

AB - Question: To assess the acidification process, nationwide information about soil pH on a site level is called for. Measurements of soil pH may be used, however there are not sufficient measurements available to map soil pH nationwide on site level. Instead we developed a soil pH map based on vegetation data. Location: Natural terrestrial areas in The Netherlands. Methods: 271,693 vegetation plots were used to estimate average soil pH per plot with indicator values, based on field measurements, of plant species. By spatial interpolation average pH values between the plots, with the soil type, groundwater table and vegetation management type as ancillary explanatory variables we created a soil pH map. The map covers all terrestrial nature areas (all areas that are not built up areas, agricultural areas and infrastructural areas) in the Netherlands with a map resolution of 25 × 25 m2 raster cells. Results: The predicted pH of the map varied between 3.0 and 8.6 with standard errors between 0.13 and 0.93. Most of the standard errors range from 0.4 to 0.55, with an average just below 0.5 pH unit. Cross-validation shows that for 33% the difference between observed and predicted is between −0.1 and 0.1 pH-unit and for 83% the difference is between −0.5 and 0.5 pH-unit. Validation shows that the pH map is unbiased (mean error is almost zero), accurate (root mean squared error is 0.64) and nicely captures spatial patterns (r = 0.77). We applied the pH map to assess the impact of acidification on the abiotic quality of nature areas in the Netherlands. Conclusions: The model fit in the predicted soil pH is in good resulting in a low standard error and a high correlation. The measures taken to prevent acidic deposition causing further acidifying of nature areas can be considered as successful.

KW - acidity

KW - kriging

KW - pH

KW - relevé

KW - soil type

KW - vegetation

U2 - 10.1111/avsc.12423

DO - 10.1111/avsc.12423

M3 - Article

VL - 22

SP - 189

EP - 199

JO - Applied Vegetation Science

JF - Applied Vegetation Science

SN - 1402-2001

IS - 2

ER -