Horvitz-Thompson estimators for double-platform line transect surveys.

David Louis Borchers, Stephen Terrence Buckland, PW Goedhart, ED Clarke, other 1

Research output: Contribution to journalArticlepeer-review

96 Citations (Scopus)

Abstract

Horvitz-Thompson estimators based on estimated inclusion probabilities are developed for line transect surveys in which some of the central assumptions of conventional line transect theory fail. The estimators are designed to accommodate uncertain detection of animals in the observers' path simultaneously with dependence of detection probabilities on explanatory variables other than perpendicular distance. They require that data be collected by two observer teams simultaneously searching the survey region, recording for each detection both positional data and details of any other variables that may affect detectability. One of the estimators is designed to correct for responsive animal movement. This method requires that the primary observer team searches independently of the secondary team, whereas the second estimator requires both observer teams to search independently of each other. The methods can be viewed as specific forms of mark-recapture experiments, in which individual capture probabilities are modelled as functions of observable explanatory variables and in which animals detected by both observer teams correspond to recaptures. They differ from existing mark-recapture Horvitz-Thompson estimators primarily in that they use an assumption about the distribution of perpendicular distances that is implicit in conventional line transect estimators. One of the estimators is used to estimate harbour porpoise abundance in the North Sea from a dual-platform line transect survey.

Original languageEnglish
Pages (from-to)1221-1237
Number of pages17
JournalBiometrics
Volume54
Publication statusPublished - Dec 1998

Keywords

  • SIZE BIAS
  • MODELS

Fingerprint

Dive into the research topics of 'Horvitz-Thompson estimators for double-platform line transect surveys.'. Together they form a unique fingerprint.

Cite this