requirements. The commercial fishery captures returning fish with gill-nets (effort in
“fathom-days”) in Bristol Bay, but fishing is forbidden in the inlets to the rivers. The
yield of the gill-net fishery is measured by counting salmon. Once they are swimming
up the rivers, people at the fish ladders and weirs watch and count the fish that
escaped the fishery. By including results from multiple years, the correlation of Y/X to
B = (Y + river counts) can be examined (Fig. 17.6a).
Fig. 17.6 (a) CPUE for Bristol Bay sockeye salmon versus total run size from catch
plus river counts.
(After Tanaka 1962.)
(b) Comparison of the CPUE for the June south Peninsula sockeye purse seine fishery
and the size of the western Alaska sockeye run, 1975–2008.
(After Martin 2009.)
The CPUE estimates appear to be excellent. CPUE estimates are also used for much
less exact comparisons of catch rates to stocks. For example, Martin (2009) has
compared the June sockeye catch (mixed gear with effort in boat-days) just near the
south tip of the Alaska Peninsula with estimates of the total annual sockeye runs for
western Alaska (Fig. 17.6b), predictably finding a much less exact fit.
(^) Another test can be applied to fish (or clams) which move infrequently (or very
slowly) so that they can be totally fished out of some modest area without influence
from migrants moving into the vacated space. Tropical snappers (Lutjanidae) are
territorial around reefs, moving only short distances from their base to grab prey (or
baits). Over 10 weeks, King (2007) directed fishers in Western Samoa as they fished
down the Etelis coruscans stock with hook and line on a 12 km^2 patch of reef not
previously fished. A time-series plot (Fig. 17.7a) of CPUE vs. total catch, Ysum, has q
as its slope. Leslie and Davis originally used this technique to estimate rat