New perspectives in the analysis of fish distributions: a case study

Color profile: Disabled
Composite Default screen
52
New perspectives in the analysis of fish
distributions: a case study on the spatial
distribution of largemouth bass (M icropterus
salmoides)
Timothy E. Essington and James F. Kitchell
Abstract: Analyses of fish distributions rarely account for spatial arrangement of habitat types, are typically conducted
at a single scale, and use a null model of random distributions without considering other null models. This study
demonstrates a procedure to circumvent these difficulties by analyzing telemetry data collected on largemouth bass
(Micropterus salmoides) in Long Lake, Michigan. Bass were highly aggregated within the littoral region, showing
peaks of aggregation at small (<60 m) and large (>160 m) spatial scales. A neutral movement model (no taxis or
kinesis within habitat types) could explain some of the observed aggregation, yet substantial aggregation remains
unexplained. Much of the large-scale aggregation could be generated by including a taxis towards the eastern half of
the basin, but taxes towards shallower cells or cells containing woody emergent macrophytes were unable to generate
the observed degree of small-scale aggregation. Our results highlight the utility of analyzing spatial distributions at
multiple scales and the importance of the spatial arrangements of habitat types and suggest that nonrandom
distributions at one scale may be due to processes occurring at different scales.
Résumé : Les analyses des distributions des poissons rendent rarement compte de l’organisation spatiale des types
d’habitat, sont généralement effectuées à une seule échelle, et ont recours à une hypothèse nulle de distribution
aléatoire sans considérer d’autres hypothèses nulles. Nous présentons ici une méthode permettant de contourner ces
difficultés en analysant des données télémétriques recueillies sur l’achigan à grande bouche (Micropterus salmoides) du
lac Long (Michigan). Les achigans étaient fortement regroupés dans la zone littorale, et montraient des pics de
regroupement à des échelles spatiales petite (<60 m) et grande (>160 m). Un modèle de déplacement neutre (ni taxie
ni cinèse au sein des types d’habitat) a pu expliquer une partie du regroupement observé, mais non l’importance de
cette agrégation. Une bonne portion de ce phénomène pourrait être engendrée par une taxie orientée vers la moitié est
du bassin, mais l’intégration de taxies orientées vers des cellules peu profondes ou des cellules contenant des
macrophytes ligneux émergés n’a pas pu générer le degré observé d’agrégation à petite échelle. Nos résultats
soulignent l’utilité d’une analyse des distributions spatiales à des échelles multiples et l’importance de la configuration
spatiale des types d’habitats, et font ressortir que les distributions non aléatoires observées à une seule échelle peuvent
être dues à des processus survenus à des échelles diverses.
[Traduit par la Rédaction]
Essington and Kitchell
Introduction
Despite the applied and ecological implications of the
spatial distribution of organisms (Hassell and Pacala 1990;
Ives 1992; Murdoch et al. 1992; McCauley et al. 1996), we
have only modest understanding of the processes underlying
causes of aggregation, avoidance, or preference for different
habitats for many fish species. To some extent, this may be a
consequence of methodological problems in the analysis of
spatial distribution data; accounts of fish distribution and
habitat tend to be descriptive, and habitat preferences are
Received October 14, 1997. Accepted September 8, 1998.
J14253
T.E. Essington1 and J.F. Kitchell. University of WisconsinMadison, Center for Limnology, 680 N. Park St., Madison,
WI 53706, U.S.A.
1
Author to whom all correspondence should be addressed.
e-mail: [email protected]
Can. J. Fish. Aquat. Sci. 56(Suppl. 1): 52–60 (1999)
J:\cjfas\cjfas56\Fish Sup\F99-213.vp
Tuesday, November 30, 1999 7:37:01 AM
60
often ascertained from correlations between local abundance
and physical or biological attributes of that locality. This
approach does not account for the spatial arrangement of
habitat types present and thereby ignores the spatial autocorrelation in the data. Moreover, nonrandom distributions
of fishes are routinely attributed to habitat preferences, yet
alternative hypotheses are rarely addressed. Finally, most
habitat analyses are conducted at a single spatial scale, and
these scales may not be the same between studies. Collectively, these shortcomings make descriptive accounts of fish
distributions limited in their ability to provide general conclusions about where fish are likely to be located and the underlying processes responsible for observed distributions.
Excellent examples of these difficulties come from studies
on the role of structural habitat complexity in the spatial distribution of largemouth bass (Micropterus salmoides). Conventional wisdom holds that bass most commonly associate
with areas of structural complexity. Indeed, some telemetry
studies have indicated that bass associate with areas containing macrophytes (Mesing and Wicker 1986), areas with
© 1999 NRC Canada