To estimate the impact of fishing activity on total
squid stocks, it is important to take into account variations
in fishing area formation by region and season. In
our study, the 7 classified areas show distinctive spatial
and temporal patterns in fishing areas. Examination of
the distribution and abundance of squid in each of the
7 areas classified herein could assist in the management
of stocks by providing additional or reduced
catch to each fishing area for stock management purposes.
For example, Class 5 corresponds to the spawning
areas reported previously by Murata (1990). It must
therefore be assumed that spent females are present in
this region, and therefore fishing activities should be
reduced and calculation of the total allowable catch
(TAC) should take this into account for squid stocks the
following year. We believe particular emphasis should
also be placed on evaluating catches in each of the 7
areas, particularly during main fishing periods. This
could be directly translated into specific policies for
squid catch management in the Sea of Japan.
To summarize, we examined the spatial and temporal
patterns of Todarodes pacificus fishing areas in the
Sea of Japan using a nighttime visible data set of
DMSP/OLS from 1994 to 1999. The methodology provided
information on the geographical location and
seasonal periodicity of T. pacificus fishing areas in the
Sea of Japan. Our results indicate that the DMSP/OLS
is capable of detecting fishing vessel lights at night in
the Sea of Japan. The lights move northward from the
Tsushima Strait from winter to summer before turning
southward again from autumn to winter.
An important aspect of this study has been its use of
image-classification to determine the temporal variability
of high DN values in DMSP/OLS nighttime visible
images. This approach also differs from that in
land-use studies that divides spectral data into classes
(Thomas et al. 2002), as well as other fishery research
that uses DMSP/OLS data (Waluda et al. 2002).
Although some problems with respect to mixed classes
still remain, they are unlikely to have markedly
affected our interpretations of the spatial and temporal
fishing area formation.
Classifying remotely-sensed images provides a new
way to derive detailed information on spatial and temporal
dynamics of fishing vessels in the Sea of Japan.
Remotely-sensing and classification methods were
used successfully to locate areas that had a high poten