GOOMEX: Experimental Design and What the Data Mean
Presented by Dr. Paul Montagna
Presentation in PDF format
The Gulf of Mexico Offshore Operations Monitoring
Experiment (GOOMEX) was performed to develop and recommend sensitive and appropriate
techniques for monitoring activities of offshore oil and gas production. To accomplish
this goal, a broad range of biological, biochemical and chemical methodologies were tested
to detect and assess potential chronic, sublethal, and long-term effects of offshore oil
and gas production. GOOMEX study components included measurements of abiotic
characteristics to indicate environmental state (e.g., chemical patterns in sediments and
water, geological patterns, and physical patterns) and biotic responses (e.g., tissue body
burdens, detoxification response by fish and invertebrates to contaminant exposure,
sediment toxicity to invertebrates, meiofaunal, macrofaunal and megaepifaunal community
structure, harpacticoid reproduction and population genetic structure, megaepifaunal
reproduction). Program design started with two null hypotheses:
1. Ho: There are no differences in any sedimentological, chemical,
community, population, reproductive, toxic or detoxification measures between seasons,
among platforms (or platform groups), or with distance or direction away from platforms
(or platform groups).
2. Ho: There are no functional responses between chemical
contaminant gradients and community, population, reproductive, toxic or detoxification
measures.
The obvious, and usual, approach to test the first
hypothesis is to perform a series of univariate analysis of variance (ANOVA) computations
for all variables being measured at all stations in the study design. However, the
categorical variables in such a design, particularly distance from a platform, are just
surrogates for the expected contaminant gradient. The second hypothesis is designed to
test for changes with respect to any contaminant gradients that might exist among and
within platforms stations. Therefore, a multivariate approach is used to test the second
hypothesis.
The experimental design to test the two hypotheses
was based on a dose-response model. Briefly, there were four cruises(one in summer and
winter, in two consecutive years), three platforms (MAI-686, MU-A85, and HI-A389, each at
progressively deeper water depths), five distances away from a platform (50 m, 100 m, 200
m, 500 m, and 3000 m), where each distance was sample on a radial transect from the
platform (5). The five radii were symetrical around MAI-686 and MU-A85, but not at
HI-A389. At HI-A389, all radii were southeast of the platform deeper than the 100 m depth
contour, because there is a shallow (~ 50 m) reef (the Flower Gardens) northwest of the
platform. The experimental design is a complex partially hierarchical design, where the
radial transect is a random nested variable within platforms. A total of 25 boxcores were
taken at each platform during each cruise. Boxcores were subsampled for component
measurements. The design resulted in 300 sediment samples, which is sufficient to ensure a
robust ANOVA, but the real power lie in the fact that all measurements could be tied to a
specific sample ensuring a robust multivariate analysis. Megaepifauna and fish were
collected with trawls, not boxcores, so the design is simply reduced to two distances from
a platform Near 50 - 100 m, and Far 3000 m), but maintains robustness.
In general, results from the GOOMEX study indicate
that effects were limited to 100 m from platforms. Relative to background (i.e., 200 m),
the zone near platforms had sediments with higher levels of contaminants and toxicity;
reduced levels of abundance, species diversity, genetic diversity, and reproductive
success; and feeding guilds with more deposit feeders. The HI-A389A platform had much
higher levels of contaminants with concordant biotic responses that other platforms,
apparently because of near-bottom shunting to avoid dispersal to the nearby Flower
Gardens.
Mercury behaved like other divalent metals in
sediments. The average concentration of mercury at HI-A389 was twice as high as the other
two platforms. The highest average concentration (0.41 ug/g) was found within 50 m of the
platform, but decreased to 0.12 at 100 m. Although they are the highest found, they are
low relative to the probable effects level (0.7) thought cause biological effects. The
natural background level for sediments in the Gulf of Mexico appears to be around 0.04
ug/g).
Metal concentrations were measured in tissues for 37
species. Except for shrimp, there was no statistically significant evidence that metal
concentrations were higher near platforms that away from platforms, even though many of
these organisms might be preferentially using the platform as a reef-like habitat. There
was statistical evidence that tissue concentrations were higher at the HI-A389 platform
than other platforms. Fish tissue concentrations were generally low, for example the
average concentration was 0.45 ug/g for all flounder species, 0.39 ug/g all hake species,
0.24 ug/g and for all snapper species. Shrimp had statistically higher tissue
concentrations near (0.36 ug/g) platforms than far (0.19 ug/g) from platforms. These
values are well below the federal guidelines set by FDA to protect human health, which is1
ppm (i.e., ug/g). Most states also use this guideline, but Wisconsin uses 0.5 ppm to
conform with Canadian standards. Some regulating bodies develop site-specific criteria.
Proposed guidelines are based on actual consumption, not fish tissue levels. For flounder,
10 (5 near and 5 far) of 119 individuals were above 1 ppm. For hake, 3 of 42 (all from far
sites) were above the 1 ppm level. For shrimp, 2 of 57 (all from near sites) were above
the 1 ppm level. None of the 41 snapper were above the criteria.
For more information, contact James Cimato.
