NONLINEAR SCALE DEPENDENCE AND SPATIOTEMPORAL VARIABILITY IN PLANKTONIC FOOD WEBS

Zooplankton food webs of Lake Okeechobee, Florida, were analyzed in or der to examine spatiotemporal variability and test for scale dependenc e of major food web properties. 46 webs were constructed from data col lected during several years (1988-92) at six locations. For all these webs, the following food web properties were calculated: number of spe cies, total links, links per species, connectance, percentage of top a nd intermediate species, and food chain length. We did not find any st atistically significant spatiotemporal variation in these properties. Still, there were consistent differences seasonally (summer vs winter) and spatially (littoral vs pelagic habitats). There was also a very c lear nonlinear scale dependence in most food web properties: links per species, food chain length, and proportions of top and intermediate s pecies vs number of species. The scale dependence was strong for small webs, but became weaker for larger webs. The relatively simple food w ebs and consistently collected data used in our study provide some of the dearest and most statistically significant results to date. They h elp reconcile the debate about scale invariance vs dependence in major food web properties.