Transcript How many fish are discarded at sea? Estimating discards

How many fish are discarded at sea?
Estimating discards using fisheries-independent data
Discarding fish at sea is one of the most controversial
aspects of fisheries management. More recently a
high profile public campaign has influenced a reform
of the EU’s Common Fisheries Policy, such that
discarding will become outlawed under a land-all
obligation.
This PhD research, currently in its early stages, aims to estimate the
number of fish discarded at sea, by using fisheries-independent data. It
will focus initially on haddock and whiting in ICES Area VIa (West of
Scotland), with a view to extending the methods developed to other areas
and other species.
Although its effects are deemed to be undesirable,
discarding is a normal function of fishing activity,
particularly in mixed fisheries where it is impossible to
be entirely selective. The sorting and discarding
process is influenced by market forces and regulation
alike. While knowledge of discard mortality is far from
complete, it is generally held that most demersal
species do not survive.
Estimating the number of fish discarded at sea is an
important part of fish stock assessment. However,
there is a danger that a ban on discarding may
compromise the availability of data obtained by direct
observation.
Jackie McLean
University of Strathclyde
Department of Mathematics & Statistics
[email protected]
Key approaches that will inform the development of the research
include Chai (1991), Casey (1996), Harley (2000), Dingsør (2001), Palsson
(2003), McBride and Fotland (1996) and Piet et al (2009).
The initial steps include:
 Analysing the length composition of landings and the relationship
between known discarding and other variables, using data published
in the ICES Working Group reports on the Celtic Sea Ecosystem.
 Using trawl survey data to estimate commercial catch composition.
For this, it will be necessary to take into account differences in
selectivity between survey and commercial gears. The survey gear,
designed to assess the abundance of species, is necessarily designed
to capture smaller fish than are allowed in commercial fishing.
It is recognised that commercial fishing grounds are not chosen
randomly, therefore it can be expected that the catch composition in the
commercial catch will differ inherently from survey vessel catch
particularly in relation to species size. To address this to some extent,
VMS data will be consulted, and statistical rectangles that appear to
have been fished will be used as the basis of selecting research vessel
data.
There currently exists a body of work on estimating unregulated, unreported or illegal fishing mortality
around the world, which provides insight into alternative methods of obtaining discard estimates. Broadly,
there are two approaches to estimating discards where direct data is limited:
Once obtained, the length composition of the estimated catch will be
compared with that in known landings to assess discarding levels, taking
into account length-based and quantity-based discarding. Known data
about discarding can be used to assess the accuracy of the estimates
derived in this way.
 Statistical modelling, using known relationships between discarding and other variables to estimate
discarding where there is missing data but where data on the other variables is available
 Estimating catch directly and then applying a statistical model of discarding behaviour. This approach
would be taken where there is no data available on discarding at all, and is found less frequently in the
literature.
The graph below is an example of preliminary research; proportions of fish
reaching MLS (from survey data) are compared with proportions discarded
(from ICES data).
Caught haddock 1981-2012 Area VIa
100.00
The survey data may then be applied to measures of fishing effort in
order to obtain estimates of commercial catch.
80.00
60.00
% ≤ MLS
40.00
% Discarded
20.00
0.00
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012