High-resolution spatial (1-min logging interval) VMS (vessel
monitoring system) data were obtained via TOREDAS from a
pole-and-line fishing vessel for the period 2007–2009. These
data consist of latitude and longitude positions logged by the
vessel’s GPS and a date–time stamp. The distance between adjacent
positions was calculated in ArcGIS 9.2, assuming that the
vessel moved between each logged position along a straight line.
Vessel speed (knots) was calculated using the distance travelled
between adjacent polling points and the time taken (usually
1 min). A histogram of estimated vessel speeds was plotted and
used to categorize vessel activity. A spatial filter excluded all
VMS data received from within 10 km of the Japanese coast, so
eliminating movement into and out of ports, which trigger the
false fishing activity because of slow vessel speed. Only data transmitted
between 06:00 and 18:00, when skipjack tuna fishing is normally
conducted, were retained. Finally, a speed filter retained all
data associated with speeds of 0.1–3 knots, indicative of fishing
activity as inferred from the VMS speed histogram.
SST and chlorophyll a (Chl a) from monthly Aqua MODIS
standard mapped images (resolution 4 km), and the delayed
time and merged SSHA products from AVISO (http://
www.aviso.oceanobs.com/en/data/products/sea-surface-heightproducts/global/msla/index.html)
were downloaded for the
period 2007–2009 and mapped in ArcGIS 9.2. The data were
re-gridded to uniform resolution, matched with the
VMS-derived fishing positions, and sampled in ArcGIS 9.2.