Laserfiche WebLink
<br />,. <br /> <br />.0 <br /> <br />Perry et al. 1986, among many others). However, very few studies actually demonstrate a <br />statistically valid relationship between biomass or some other abundance measure and flow <br />variables that apply to different streams or even different stream segments. Indeed, Morhardt <br />(1986) reviewed and annotated 72 studies that attempted to derive a general instream flow model <br />that would accurately predict productivity to flow variables in different streams. Only one (Binns <br />and Eiserman 1979) produced a statistically valid result, and Morhardt (1986) concluded that was <br />because the streams were in the same region and were biophysically very similar. Armitage (1989) <br /> <br />was able to predict the occurrence and biomass of macroinvertebrates from a suite of environmental <br /> <br />. <br /> <br />and Statzner et al. (1988). <br /> <br />In rivers that are large and complex most studies are site specific by design because it is <br />widely recognized that unbiased replication of sites across streams is difficult, if not impossible, <br /> <br />owing to the stochastic nature of large rivers. In fact, it is very difficult to replicate within a stream <br /> <br />segment because flow mechanics produce so many different microhabitats that it is almost <br />impossible to take enough samples to describe biotic distributions. Pseudoreplication is a problem <br />in many studies. All streams are ecologically different and therefore mechanistic models must <br />compromise reality to gain generality. The alternative is essentially a trial and error approach. In <br />other words, multivariate analyses may show that certain flow variables influence biotic productivity <br />in a regulated stream; therefore, a particular flow pattern should optimize productivity. The only <br />way to verify that prediction is to implement the flow regime and monitor productivity. <br /> <br />47 <br />