Mapping water table depths in wetlands and polder areas by probability sampling

Martin Knotters*, Dennis Walvoort, Paul Gerritsen

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Information on water table depth (WTD) in polder areas and wetlands is important in, for example, estimating emissions of greenhouse gases, assessing the agricultural and ecological potential, and flood risk management. The seasonal variation of WTDs is summarized with averages of the yearly highest (shallowest) and lowest (deepest) water tables (MHW and MLW). These characteristics show short-distance variations within the fields in polder areas, which cannot be mapped using geostatistical interpolation techniques or physical modelling against reasonable costs or with acceptable accuracy. The within-field variations depend on soil type and water management. MHW and MLW were determined from auger hole measurements of WTDs at locations being selected following stratified simple random sampling in subareas classified by soil type and water management. Within these subareas, a further classification was made on the basis of distance to ditches. For each subarea spatial distribution functions of MHW and MLW were made, taking censored observations into account. Uncertainty was quantified by 10,000 bootstrap realisations of the spatial distribution functions. From these realisations maps depicting summary statistics for the spatial distribution of WTD-characteristics within the subareas were derived, as well as a map with probabilities of exceedance of a critical level that can serve as input for risk analysis.

Original languageEnglish
Article number115928
JournalGeoderma
Volume422
DOIs
Publication statusPublished - 15 Sep 2022

Keywords

  • Censored observations
  • Choropleth map
  • Design-based approach
  • Nonparametric regression
  • Raster map
  • Resolution

Fingerprint

Dive into the research topics of 'Mapping water table depths in wetlands and polder areas by probability sampling'. Together they form a unique fingerprint.

Cite this