Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
Can. J. Fish. Aquat. Sci. 57: 19401956 (2000) 2000 NRC Canada Is the sampling strategy interfering with the study
 

Summary: Can. J. Fish. Aquat. Sci. 57: 1940­1956 (2000) © 2000 NRC Canada
1940
Is the sampling strategy interfering with the study
of spatial variability of zooplankton communities?
Carol Avois, Pierre Legendre, Stéphane Masson, and Bernadette Pinel-Alloul
Abstract: Surveys at the whole-lake scale take some time to carry out: several hours or several days. For logistic rea-
sons, the sites are not sampled simultaneously or in a random sequence. Traditional limnological sampling methods re-
quire an appreciable amount of time at each site. Any sampling strategy that is not random or simultaneous introduces
dependencies among the observations, which must be taken into account during the analysis and interpretation of the
data. What is the real nature of the variation measured using a given sampling design? This question is approached us-
ing sites sampled by two boat teams during two consecutive days. Statistical modelling was used to partition the varia-
tion of zooplankton size-class data into environmental and spatial components. The conclusions reached after an
analysis that did not control for the sampling design are erroneous and quite different from those reached when the ef-
fect of the sampling design (factors Day, Boat, and Hour) was taken into account. Clearly, when a significant effect of
the sampling design is found, one must control for it during the analysis and interpretation of ecological variation.
Résumé : Réaliser un relevé à l'échelle panlacustre peut prendre du temps : plusieurs heures ou même plusieurs jours.
Pour des raisons de logistique, les stations ne sont pas échantillonnées simultanément ni en ordre aléatoire Les métho-
des classiques d'échantillonnage limnologique imposent de consacrer un temps appréciable à chaque station. Une stra-
tégie d'échantillonnage qui n'est ni simultanée ni aléatoire introduit des dépendances entre les observations, ce qui doit
être pris en compte dans l'analyse et l'interprétation des données. Quelle est la vraie nature de la variation mesurée

  

Source: Avois-Jacquet, Carol - Station d'Hydrobiologie Lacustre, INRA

 

Collections: Environmental Sciences and Ecology