Nondetection sampling bias in marked presence-only data

Trevor J. Hefley, Andrew J. Tyre, David M. Baasch, Erin E. Blankenship

Research output: Contribution to journalArticle

14 Citations (Scopus)

Abstract

Species distribution models (SDM) are tools used to determine environmental features that influence the geographic distribution of species' abundance and have been used to analyze presence-only records. Analysis of presence-only records may require correction for nondetection sampling bias to yield reliable conclusions. In addition, individuals of some species of animals may be highly aggregated and standard SDMs ignore environmental features that may influence aggregation behavior. We contend that nondetection sampling bias can be treated as missing data. Statistical theory and corrective methods are well developed for missing data, but have been ignored in the literature on SDMs. We developed a marked inhomogeneous Poisson point process model that accounted for nondetection and aggregation behavior in animals and tested our methods on simulated data. Correcting for nondetection sampling bias requires estimates of the probability of detection which must be obtained from auxiliary data, as presence-only data do not contain information about the detection mechanism. Weighted likelihood methods can be used to correct for nondetection if estimates of the probability of detection are available. We used an inhomogeneous Poisson point process model to model group abundance, a zero-truncated generalized linear model to model group size, and combined these two models to describe the distribution of abundance. Our methods performed well on simulated data when nondetection was accounted for and poorly when detection was ignored. We recommend researchers consider the effects of nondetection sampling bias when modeling species distributions using presence-only data. If information about the detection process is available, we recommend researchers explore the effects of nondetection and, when warranted, correct the bias using our methods. We developed our methods to analyze opportunistic presence-only records of whooping cranes (Grus americana), but expect that our methods will be useful to ecologists analyzing opportunistic presence-only records of other species of animals.

Original languageEnglish (US)
Pages (from-to)5225-5236
Number of pages12
JournalEcology and Evolution
Volume3
Issue number16
DOIs
StatePublished - Dec 1 2013

Fingerprint

sampling bias
aggregation behavior
sampling
methodology
animal
biogeography
researchers
animals
crane
group size
method
ecologists
geographical distribution
linear models
detection
distribution
modeling

Keywords

  • Grus americana
  • Inhomogeneous Poisson point process
  • Missing data
  • Nondetection
  • Sampling bias
  • Species distribution model
  • Whooping crane

ASJC Scopus subject areas

  • Ecology, Evolution, Behavior and Systematics
  • Ecology
  • Nature and Landscape Conservation

Cite this

Hefley, T. J., Tyre, A. J., Baasch, D. M., & Blankenship, E. E. (2013). Nondetection sampling bias in marked presence-only data. Ecology and Evolution, 3(16), 5225-5236. https://doi.org/10.1002/ece3.887

Nondetection sampling bias in marked presence-only data. / Hefley, Trevor J.; Tyre, Andrew J.; Baasch, David M.; Blankenship, Erin E.

In: Ecology and Evolution, Vol. 3, No. 16, 01.12.2013, p. 5225-5236.

Research output: Contribution to journalArticle

Hefley, TJ, Tyre, AJ, Baasch, DM & Blankenship, EE 2013, 'Nondetection sampling bias in marked presence-only data', Ecology and Evolution, vol. 3, no. 16, pp. 5225-5236. https://doi.org/10.1002/ece3.887
Hefley, Trevor J. ; Tyre, Andrew J. ; Baasch, David M. ; Blankenship, Erin E. / Nondetection sampling bias in marked presence-only data. In: Ecology and Evolution. 2013 ; Vol. 3, No. 16. pp. 5225-5236.
@article{1a26011c1bf84b3d8da307e5691de8ec,
title = "Nondetection sampling bias in marked presence-only data",
abstract = "Species distribution models (SDM) are tools used to determine environmental features that influence the geographic distribution of species' abundance and have been used to analyze presence-only records. Analysis of presence-only records may require correction for nondetection sampling bias to yield reliable conclusions. In addition, individuals of some species of animals may be highly aggregated and standard SDMs ignore environmental features that may influence aggregation behavior. We contend that nondetection sampling bias can be treated as missing data. Statistical theory and corrective methods are well developed for missing data, but have been ignored in the literature on SDMs. We developed a marked inhomogeneous Poisson point process model that accounted for nondetection and aggregation behavior in animals and tested our methods on simulated data. Correcting for nondetection sampling bias requires estimates of the probability of detection which must be obtained from auxiliary data, as presence-only data do not contain information about the detection mechanism. Weighted likelihood methods can be used to correct for nondetection if estimates of the probability of detection are available. We used an inhomogeneous Poisson point process model to model group abundance, a zero-truncated generalized linear model to model group size, and combined these two models to describe the distribution of abundance. Our methods performed well on simulated data when nondetection was accounted for and poorly when detection was ignored. We recommend researchers consider the effects of nondetection sampling bias when modeling species distributions using presence-only data. If information about the detection process is available, we recommend researchers explore the effects of nondetection and, when warranted, correct the bias using our methods. We developed our methods to analyze opportunistic presence-only records of whooping cranes (Grus americana), but expect that our methods will be useful to ecologists analyzing opportunistic presence-only records of other species of animals.",
keywords = "Grus americana, Inhomogeneous Poisson point process, Missing data, Nondetection, Sampling bias, Species distribution model, Whooping crane",
author = "Hefley, {Trevor J.} and Tyre, {Andrew J.} and Baasch, {David M.} and Blankenship, {Erin E.}",
year = "2013",
month = "12",
day = "1",
doi = "10.1002/ece3.887",
language = "English (US)",
volume = "3",
pages = "5225--5236",
journal = "Ecology and Evolution",
issn = "2045-7758",
publisher = "John Wiley and Sons Ltd",
number = "16",

}

TY - JOUR

T1 - Nondetection sampling bias in marked presence-only data

AU - Hefley, Trevor J.

AU - Tyre, Andrew J.

AU - Baasch, David M.

AU - Blankenship, Erin E.

PY - 2013/12/1

Y1 - 2013/12/1

N2 - Species distribution models (SDM) are tools used to determine environmental features that influence the geographic distribution of species' abundance and have been used to analyze presence-only records. Analysis of presence-only records may require correction for nondetection sampling bias to yield reliable conclusions. In addition, individuals of some species of animals may be highly aggregated and standard SDMs ignore environmental features that may influence aggregation behavior. We contend that nondetection sampling bias can be treated as missing data. Statistical theory and corrective methods are well developed for missing data, but have been ignored in the literature on SDMs. We developed a marked inhomogeneous Poisson point process model that accounted for nondetection and aggregation behavior in animals and tested our methods on simulated data. Correcting for nondetection sampling bias requires estimates of the probability of detection which must be obtained from auxiliary data, as presence-only data do not contain information about the detection mechanism. Weighted likelihood methods can be used to correct for nondetection if estimates of the probability of detection are available. We used an inhomogeneous Poisson point process model to model group abundance, a zero-truncated generalized linear model to model group size, and combined these two models to describe the distribution of abundance. Our methods performed well on simulated data when nondetection was accounted for and poorly when detection was ignored. We recommend researchers consider the effects of nondetection sampling bias when modeling species distributions using presence-only data. If information about the detection process is available, we recommend researchers explore the effects of nondetection and, when warranted, correct the bias using our methods. We developed our methods to analyze opportunistic presence-only records of whooping cranes (Grus americana), but expect that our methods will be useful to ecologists analyzing opportunistic presence-only records of other species of animals.

AB - Species distribution models (SDM) are tools used to determine environmental features that influence the geographic distribution of species' abundance and have been used to analyze presence-only records. Analysis of presence-only records may require correction for nondetection sampling bias to yield reliable conclusions. In addition, individuals of some species of animals may be highly aggregated and standard SDMs ignore environmental features that may influence aggregation behavior. We contend that nondetection sampling bias can be treated as missing data. Statistical theory and corrective methods are well developed for missing data, but have been ignored in the literature on SDMs. We developed a marked inhomogeneous Poisson point process model that accounted for nondetection and aggregation behavior in animals and tested our methods on simulated data. Correcting for nondetection sampling bias requires estimates of the probability of detection which must be obtained from auxiliary data, as presence-only data do not contain information about the detection mechanism. Weighted likelihood methods can be used to correct for nondetection if estimates of the probability of detection are available. We used an inhomogeneous Poisson point process model to model group abundance, a zero-truncated generalized linear model to model group size, and combined these two models to describe the distribution of abundance. Our methods performed well on simulated data when nondetection was accounted for and poorly when detection was ignored. We recommend researchers consider the effects of nondetection sampling bias when modeling species distributions using presence-only data. If information about the detection process is available, we recommend researchers explore the effects of nondetection and, when warranted, correct the bias using our methods. We developed our methods to analyze opportunistic presence-only records of whooping cranes (Grus americana), but expect that our methods will be useful to ecologists analyzing opportunistic presence-only records of other species of animals.

KW - Grus americana

KW - Inhomogeneous Poisson point process

KW - Missing data

KW - Nondetection

KW - Sampling bias

KW - Species distribution model

KW - Whooping crane

UR - http://www.scopus.com/inward/record.url?scp=84890882740&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84890882740&partnerID=8YFLogxK

U2 - 10.1002/ece3.887

DO - 10.1002/ece3.887

M3 - Article

C2 - 24455151

AN - SCOPUS:84890882740

VL - 3

SP - 5225

EP - 5236

JO - Ecology and Evolution

JF - Ecology and Evolution

SN - 2045-7758

IS - 16

ER -