A review of spatio-temporal modelling of quadrat count data with application to striga occurrence in a pearl millet field

D. Hess, M.C. van Lieshout, W. Payne, A. Stein

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

This paper describes how spatial statistical techniques may be used to analyse weed occurrence in tropical fields. Quadrat counts of weed numbers are available over a series of years, as well as data on explanatory variables, and the aim is to smooth the data and assess spatial and temporal trends. We review a range of models for correlated count data. As an illustration, we consider data on striga infestation of a 60 x 24 m2 millet field in Niger collected from 1985 until 1991, modelled by independent Poisson counts and a prior auto regression term enforcing spatial coherence. The smoothed fields show the presence of a seed bank, the estimated model parameters indicate a decay in the striga numbers over time, as well as a clear correlation with the amount of rainfall in 15 consecutive days following the sowing date. Such results could contribute to precision agriculture as a guide to more cost-effective striga control strategies.
Original languageEnglish
Pages (from-to)133-138
JournalInternational Journal of applied Earth Observation and Geoinformation
Volume3
DOIs
Publication statusPublished - 2001

Fingerprint Dive into the research topics of 'A review of spatio-temporal modelling of quadrat count data with application to striga occurrence in a pearl millet field'. Together they form a unique fingerprint.

Cite this