| dc.description.abstract |
As management of commercial fisheries becomes more spatially oriented, collection
of
commercial fisheries data must adapt to accurately reflect location. An important
component of
accurate spatial data is scale. In an attempt to characterize patterns of scale in
fisheries data, I
tested the National Marine Fisheries Service Northeast bottom trawl survey data for
spatial
dependency using semivariogram analysis. Specifically, and more importantly for management,
I wanted to determine if the distance between sample locations is a good predictor
variable for
how much fish will be caught. Focusing on 1996-2002 catch data for Atlantic cod and
witch
flounder, I found that for current data collection techniques, the variance of catch
weight is
spatially independent from distance between observations. Thus, the scale and spatial
pattern of
the data can not be characterized based on distance for the range of space and time
analyzed.
This finding does not rule out the possibility that spatial dependence may be observed
in these
fisheries if we were to examine data sets with finer spatial distances and finer time
intervals.
Because ocean processes vary significantly across time, the effect of aggregating
the spatial data
across time may have acted to conceal some of the potential trends in the data set.
Determining the spatial patterns in the data is part of a sequential approach to understanding
ecological processes. Alternative hypotheses that may possibly explain the spatial
pattern of the
data need to be tested and include spatial patterns being dependent upon bottom habitat
complexity, water temperature, and/or prey availability. The goal is to find a variable
that
explains fish biomass patterns, allowing managers and scientists to begin to understand
what
proxy data they really need to collect and map, and at what scale, in order to predict
patterns of
fishes for effective and sustainable fisheries management.
|
|
