Grid-spacing and the quality of abundance maps for species that show spatial autocorrelation and zero-inflation

Olga Lyashevska, Dick J. Brus, Jaap van der Meer

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

The effect of grid-spacing on the quality of species abundance maps is explored for species that show zero-inflation and spatial autocorrelation. Using a zero-inflated Poisson mixture model multiple fields of the prevalence parameter π and the intensity parameter μ were simulated. A selected field was sampled by grid-sampling with 200, 400, 800, 1600, and 3200 m grid-spacing and used to predict at a fixed set of validation locations by simple kriging with an external drift. The external drift variables were silt, silt squared and altitude. The estimated sampling distribution of MSE against grid-spacing shows that beyond a spacing of 1600 m the mean of MSE increases at a much faster rate. Based on these findings the 1600 m grid which consists of 446 locations for our study area of 2400 km2 gives a compromise between sampling costs and prediction accuracy.

LanguageEnglish
Pages386-395
JournalSpatial Statistics
Volume18
DOIs
Publication statusPublished - 2016

Fingerprint

Zero-inflation
Spatial Autocorrelation
Autocorrelation
inflation
autocorrelation
Spacing
spacing
Silt
Sampling
Grid
silt
sampling
Poisson Mixture
kriging
Sampling Distribution
Kriging
Poisson Model
Mixture Model
Costs
prediction

Keywords

  • Autocorrelation
  • Count data
  • Generalized linear geostatistical modelling
  • Grid-spacing
  • Zero-inflation

Cite this

@article{9388dbfaeeb54110a9aaca477a5c1996,
title = "Grid-spacing and the quality of abundance maps for species that show spatial autocorrelation and zero-inflation",
abstract = "The effect of grid-spacing on the quality of species abundance maps is explored for species that show zero-inflation and spatial autocorrelation. Using a zero-inflated Poisson mixture model multiple fields of the prevalence parameter π and the intensity parameter μ were simulated. A selected field was sampled by grid-sampling with 200, 400, 800, 1600, and 3200 m grid-spacing and used to predict at a fixed set of validation locations by simple kriging with an external drift. The external drift variables were silt, silt squared and altitude. The estimated sampling distribution of MSE against grid-spacing shows that beyond a spacing of 1600 m the mean of MSE increases at a much faster rate. Based on these findings the 1600 m grid which consists of 446 locations for our study area of 2400 km2 gives a compromise between sampling costs and prediction accuracy.",
keywords = "Autocorrelation, Count data, Generalized linear geostatistical modelling, Grid-spacing, Zero-inflation",
author = "Olga Lyashevska and Brus, {Dick J.} and {van der Meer}, Jaap",
year = "2016",
doi = "10.1016/j.spasta.2016.08.001",
language = "English",
volume = "18",
pages = "386--395",
journal = "Spatial Statistics",
issn = "2211-6753",
publisher = "Elsevier",

}

Grid-spacing and the quality of abundance maps for species that show spatial autocorrelation and zero-inflation. / Lyashevska, Olga; Brus, Dick J.; van der Meer, Jaap.

In: Spatial Statistics, Vol. 18, 2016, p. 386-395.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Grid-spacing and the quality of abundance maps for species that show spatial autocorrelation and zero-inflation

AU - Lyashevska, Olga

AU - Brus, Dick J.

AU - van der Meer, Jaap

PY - 2016

Y1 - 2016

N2 - The effect of grid-spacing on the quality of species abundance maps is explored for species that show zero-inflation and spatial autocorrelation. Using a zero-inflated Poisson mixture model multiple fields of the prevalence parameter π and the intensity parameter μ were simulated. A selected field was sampled by grid-sampling with 200, 400, 800, 1600, and 3200 m grid-spacing and used to predict at a fixed set of validation locations by simple kriging with an external drift. The external drift variables were silt, silt squared and altitude. The estimated sampling distribution of MSE against grid-spacing shows that beyond a spacing of 1600 m the mean of MSE increases at a much faster rate. Based on these findings the 1600 m grid which consists of 446 locations for our study area of 2400 km2 gives a compromise between sampling costs and prediction accuracy.

AB - The effect of grid-spacing on the quality of species abundance maps is explored for species that show zero-inflation and spatial autocorrelation. Using a zero-inflated Poisson mixture model multiple fields of the prevalence parameter π and the intensity parameter μ were simulated. A selected field was sampled by grid-sampling with 200, 400, 800, 1600, and 3200 m grid-spacing and used to predict at a fixed set of validation locations by simple kriging with an external drift. The external drift variables were silt, silt squared and altitude. The estimated sampling distribution of MSE against grid-spacing shows that beyond a spacing of 1600 m the mean of MSE increases at a much faster rate. Based on these findings the 1600 m grid which consists of 446 locations for our study area of 2400 km2 gives a compromise between sampling costs and prediction accuracy.

KW - Autocorrelation

KW - Count data

KW - Generalized linear geostatistical modelling

KW - Grid-spacing

KW - Zero-inflation

U2 - 10.1016/j.spasta.2016.08.001

DO - 10.1016/j.spasta.2016.08.001

M3 - Article

VL - 18

SP - 386

EP - 395

JO - Spatial Statistics

T2 - Spatial Statistics

JF - Spatial Statistics

SN - 2211-6753

ER -