community assemblages present within many other streams surveyed in the region
(Department of Conservation 1999-2001), although overall fish density was low. The low
observed fish density is likely due to low observed invertebrate densities and inhospitable
habitat within the gorge.

Galaxiid community composition was similar to that observed in other similar streams within
the region, with inanga/banded kokopu habitat in the lowest gradient section of the river
(Studholme 1999), grading to banded/shortjaw kokopu habitat in the intermediate gradient
section of the river immediately below the gorge outlet, finally changing to koaro and longfin
eel habitat only, within the higher gradient gorge section.

The density of koaro located in favourable run/riffle habitat above the dam and gorge, is over
3 times higher than the density of koaro located in favourable run/riffle habitat within the
gorge itself. Density of juvenile longfin eels is slightly lower in favourable run/riflle habitat
within the gorge than in favourable run/riflle habitat above the gorge.

The average density of invertebrates in the lower gradient river section above the dam and
proposed intake site, was over 4 times higher than density of invertebrates in the top of the
gorge below the dam, in July 2004. Invertebrate samples collected in May 2004 from the
lower gorge, still give an indication of a similar low-density invertebrate trend throughout the
gorge. This trend fits with the observed differences in koaro density above and within the
Parapara gorge. Both fish survey and invertebrate collection results, give evidence to the
observation that the Parapara Gorge is a low-productivity, inhospitable freshwater
environment, but does provide some habitat for freshwater organisms.

The density of shortjaw kokopu in the Parapara River immediately below the gorge outlet
(0.0025 fish m2) represented a low to intermediate density of fish located when compared with
other population areas surveyed 1999-2001, possibly due to the relatively unproductive nature
of the river (Studholme 1999; Jack 2000; Jack 2001). These densities are calculated through
estimating the proportion of the river width able to be effectively spotlighted, and multiplying
this by the length of the survey reach. The number of fish located is then divided by this
figure. Usefulness of spotlighting for total population estimates of this fish species is limited
although semi-quantitative results can be obtained where a large proportion of the waterway is
represented by slow runs/pools. Where a large percentage of disturbed water such as riffles
and white water of cascades exist, quantitative work using spotlighting is very difficult.
Nevertheless, within these variables, results from this work can be compared with work from
previous 3 years surveys, many waterways of which were similar in nature, to give an index of
comparison, in terms of fish numbers observed per area surveyed.

The density of shortjaw located within the area surveyed appeared low-moderate when
compared with results of survey work undertaken 1999-2001 (R. Barrier pers. obs.). When
the density of fish located in the Parapara River is compared with the average density of 34

Fish & Game New Zealand Nelson Marlborough Region Sports Fish and Game Bird Management ...p20