appendix b final 2008 biological surveys of los angeles and long ...

4.0 Ichthyoplankton
The numbers were adjusted to the number of each taxon per 100 m 2 (referred to in this report
as weighted mean abundance) by multiplying the number of each taxa per 100 m 3 by the depth
of water that the net sampled. It was estimated that the manta net sampled the upper 0.16
meters of the water column, the epibenthic net sampled the lower 0.70 meters of the water
column, and the stepped oblique net sampled the rest of the water column (station water depth
– [0.16 + 0.70]). For example, if there were 2.5 individuals of Taxon A per 100 m 3 in a stepped
oblique sample that was collected in 12.2 meters water depth, the total density would equal
28.35 per 100 m 2 [2.5 times (12.2 minus 0.86)]. If this same density of fish was found in the
neuston net there would be 0.4 per 100 m 2 (2.5 times 0.16). This process of weighting the
strata allowed the three sample types from each station to be added together to generate an
estimate of the number of eggs and larvae in the entire water. Diversity at each station was
calculated as the number of taxa, Shannon-Weiner index, Margalef index, and dominance (see
Section 3.2.4 for a description of these measures of diversity).
The data on larval fishes collected by the three sampling methods during the three surveys were
combined and compared among the nineteen stations using multivariate non-metric
multidimensional scaling (MDS). The analysis was completed by first transforming the data to
the 4 th root to correct for the large-scale differences among densities, and then computing Bray-
Curtis distances among each set of data. The Bray-Curtis distance is a commonly used
measure of the degree of similarity between sets of data based on differences in species and
their abundances (Clarke and Warwick 2001). The Bray-Curtis distances were then analyzed
using MDS that optimizes the spatial differences into two-dimensions. The MDS analysis was
completed using the PRIMER analysis package (Clarke and Gorley 2001).
The relationship between larval fish composition and sediment grain size was investigated using
the RELATE analysis procedure in the PRIMER statistical package (Clarke and Gorley 2001).
The RELATE procedure computes a rank correlation between the rank order of Bray-Curtis
distances among stations based on the biological data and the rank order based on the
Euclidean distances among the stations based on the grain size data.
4.2.4 Method Comparison Study
A comparison of larval species composition and
abundance collected during the current study
(referred to below as “three nets”) and the
modified CalCOFI-type net deployment method
was undertaken during Surveys 2 (April) and 3
(July). The objective of this special study was to
determine if more recent and widely accepted
sampling methods could produce comparable
data and decrease the overall field and laboratory
effort for future baseline surveys. During this
special study the bongo nets were deployed at
ten of nineteen ichthyoplankton stations (LA1,
LA2, LA4, LA5, LA6, LB1, LB2, LB3, LB5, and
LB6; Figure 4.2-1) to collect a sample throughout
the entire water column following similar methods to those employed by CalCOFI. In general,
the bongo nets were fished from the surface to the bottom and then back to the surface, with the
nets collecting a similar volume of water at all depth strata. The deeper stations (depths
between 40-50’) were selected for the comparison study because (1) these stations are
relatively easy to sample (fewer deployment and retrieval cycles to filter a target volume) and (2)
if there is stratification of the larvae, such as between the surface and bottom, differences are
2008 Biological Surveys of Los Angeles and Long Beach Harbors 4–3
April 2010