|
TABLE 1. Distribution of WUA (m2i km) for adult small-
<br />mouth bass at various depth-velocity combinations (data
<br />extracted from table I of Bovee 1978).
<br /> Velocity (m/s)
<br />Depth (m) 0.15.-0-30 0.30-0.45 0.45-0.60
<br />0.30-0.45 15 14
<br />0.45-0.60 .17 15 21
<br />0.60-0.75 17 15 6
<br />0.75-0.90 11 44 63
<br />0.90-1.05 15 15 9
<br />are biologically equal and Provide similar production rates
<br />unless each is an ex(r( t replica of the other unit. If each unit of
<br />WUA is biologically identical, a prediction of standing stock of
<br />fishes as a result of flow alteration may be ossi le /However, in
<br />the present form of calculations in PHABSIM, several combina-
<br />tions of depth, velocity. and substrate can give the same antrnmt
<br />of WUA, none of which may support a similar fish biomass We
<br />nave extracted data trom Bovee (1978) in Table I to show that at
<br />least live different combinations of depth and velocity provide
<br />the same amount of WUA (15m2lkni) for adult smallnu?uth
<br />bass in a stream. A lack of correlation between WUA and fish
<br />biomass may in part result from treating each habitat unit as
<br />F.fjects of Outliers on Relationship of LL'UA and Biomass
<br />The relationship between fish biomass (kilograms per hec
<br />tare) and WUA (expressed as percent of total area) is used in
<br />some investi ations. It is also weak and of uestiona
<br />relation. or example, of the 20 correlations presented by
<br />Orth and hlaughan (1982), 14 (70%) were not significant (P >
<br />0.05). Examination of these reported relationships shows that in
<br />most cases only two or three effective data points existed (a
<br />cluster of points near the origin and one or two extreme values).
<br />A cluster of points is equivalent to only one effective data point
<br />in the determination of the slope of the regression line. All data
<br />points should neigh equal) in determining a regression line
<br />(Draper and Smith 1979 Spurious correlations and therefore
<br />questionable conclusions can result when regressions are based
<br />The following examples will illustrate the influence of
<br />outliers on habitat relationships and subsequent interpretations.
<br />Removal of single outliers from the central stoneroller summer
<br />(X = 42.0, Y = 51.8) and fall (X = 42.9, Y = 17:7) data sets
<br />reduced the reported significant correlations (P < 0.01) from
<br />0.835 to 0:61 for the summer and from 0.737 to 0.416 for the
<br />fall, both correlations became nrnnignifir`ant (P > 0 05) A7
<br />therefore consider the original interpretation that physical
<br />habitat limits hoptilation of central stoneroller to h sus (`c't
<br />On the other hand, rcn?oval of oullicrs front five spring central
<br />stonerollev tlal7i sets ( t = 47.2. 40.0, t' 0.9,, 1. 1) incri-w,ctl
<br />the correlation from 0.159 (P > 0.05) to 0.892 (P < 0.01). Note
<br />that these are high values of WUA associated with low values of
<br />biomass (kilograms per hectare) and that their removal in no
<br />way biases the relationships that are assumed in IF1M. Using the
<br />same logic the original interpretation that the physical habitat
<br />did not limit the central stoneroller population becomes erron-
<br />eous. Similarly, the removal of a single outlier (X = 70.9. Y =
<br />0) from the adult smalbnouth bass summer data set increased
<br />the reported correlation from 0,423 (P > 0,05) to 0,857 (f
<br />changes may not result in corresponding changes in twal
<br />abundance of species life stage and should not he interpreted as
<br />such. The study also indicates, at the lead, that the a\ailability
<br />of usable area may not ha\e an immediate regulatory efletr ,it
<br />the population. In application of IFIAI, ho'.tever. it is ilalplicill;
<br />assumed that such chances \\ ill occur in filch o Iulations a, a
<br />result of flow alteraholls. Il Ollr Vlt't'., itltt,,llra[nrs Silwll i
<br />attempt to relate changes in physical habitat in terms rl thril
<br />effects on the standing styrck (abundance) in a stream as a k, hi-le
<br />rather than in temporary redistribution (if a scynu nt n1 the li II
<br />o ulation in a stretch ill a sit-Call),
<br />References
<br />nOVEE, K. D. 1978. The iecrcntental method of assessing habitat Pulent:al ;ur
<br />eoolwater species, with management implications, p. 3411-14o. Ill R. L.
<br />Kendall led. I Selected c+urln ater fishes of Nooh :America. Ain. I W) Soc
<br />Spec. Publ. No. 11
<br />1982. A guide to dream habitat analysis using the L'Is!rcmn Flu«
<br />Incremental D1ethodology. United State,. Fish and 'A ildlife Sen i:e
<br />Biological Smites I'lot•I I'll, Coulwralivc In,'rranl I L w Scl•ice GIvup.
<br />Inslreant Flom Infonn;ili ,r, I';qut Numlxr I?. I \'.;i !;nS S1.'.•!n 21?p
<br />11RAI'I.R, N. R., ANU 11. t'•tl l II 1'179. Applied ril'+?'?i?r::nr:rl'. i•. !u?! •',I Ir!'n
<br />Wiley R( Suns. I oc.. Pint 1 otl,. NY. 10911
<br />EMERY. A. R. 1973. Plchminan cumparisocs of day 7uld nigh! hal'its `r
<br />freshwater lisll in Onruio Likes. J. I Ish. Res. no;ud I. ul. ;11: 7(11-'7,1
<br />FRASER. J. C. 1972. Rrgulatcd discharge and the stream _ntin?nn:eat.
<br />p. 263-286. ht R. J Oglcshy. C. A. Carlson. and J. A NIct ann 1H.I
<br />River ecology and man. Academic Press, tie" Yvik. NY.
<br />FRASER, D. F., ANU E E. ENnioNs. 1984. Behasioral response of tltacknnse
<br />dace (RhiaicJult)a utrantlito R) %;wing densities of Predatory creek chute
<br />(Sernotilusatr(-t.,ut('ttlnnts). Can. J. Fish. Ayuat. Sci. 41:3b4-.370.
<br />GORE, J, A., ANn R. D- Jur)%-. JR. 1981. Predictise modus of benthic
<br />macroinvertebra,,e densuy For use in instream ll(,a studies and reguLnrd
<br />,
<br />flow manapanivnl. Clio J I'Irh, Aqt n, 5!'I )b' I W 070
<br />830 Can. J. Fob ,,law. St I_ IN. 42. 1'=?`
|