Study: Proposed offshore gulf LNG terminals will have minor effects on fish populations

Several new offshore LNG terminals that have been licensed or proposed for construction in the northern Gulf of Mexico will use open-rack vaporization (ORV) or the more general open-loop vaporization (OLV) technology.1 In accordance with the Deepwater Port Act,2 as part of the LNG facilities’ licensing requirements, environmental impact statements (EIS) were issued that predicted impacts to fisheries.

The central environmental issue regarding the use of OLV technology in the gulf is the potential impact on fishery stocks resulting from mortality of early-life stages of fishes (i.e., eggs and larvae, or ichthyoplankton) entrained in the seawater intakes of these facilities.3 Although the statements have concluded that overall impacts would be minimal, some comments on the impact statements have focused on the upper range of the fishery-impact predictions.

To address these concerns, the Center for LNG (CLNG) commissioned an independent study4 to answer the following questions:

Are the EIS assessments adequate?
Are there more scientifically valid approaches that could be used?
If so, do these approaches support the conclusions of the current statements that fishery impacts from the OLV systems will be minor?

The study found that, taken as a whole, data inputs and model approaches used in the impact statements use conservative assumptions and significantly overestimate the potential for adverse fisheries impacts. Although the statements conclude that predicted impacts are minimal, those effects are themselves overstated.

The study further found that the impact statements do not quantitatively account for all the factors that affect these estimates and so do not accurately describe the degree of overestimation associated with these values or their overall uncertainty. This is an important omission, given the potential for misinterpretation of the EIS findings.

Given the errors and oversimplified assumptions used, the potential upper-limit impacts to fisheries reported in the statements are greatly overstated. Moreover, the study found that the predicted reductions in the age-1 equivalents stated in the statements have been compared inappropriately with arbitrarily defined fishery weight statistics (e.g., landings for an individual state expressed in pounds of fish), thereby greatly inflating the apparent severity of the potential impact.

Such comparisons, whether with total Gulf of Mexico landings or state landings, are inappropriate because of the inherent uncertainties in the calculation of adult-equivalent weights based only on early-life-stage mortalities. This is highly uncertain and inappropriate because landing statistics depend on fisheries management decisions and because they do not account for population compensation effects.

Furthermore, the study found that the forward-projection (adult-equivalent) models used in the impact statements are inconsistent with fisheries stock-assessment methods, which are used to make other important fishery management decisions.

Despite these defects in the analyses conducted in the environmental impact statements, the underlying data do support and even reinforce the overall conclusions of the statements-that fisheries impacts from the use of OLVs would be minor. The conclusions of the statements are therefore appropriate for government decision making, although they are highly conservative.

Prevention, mitigation, and monitoring measures carried out as part of facility operations can provide additional information to further refine, and reduce, impact estimates. Preventive measures that could further minimize losses of ichthyoplankton from OLV systems may include the following:

Location of facilities away from spawning areas of commercially important species.
Use of fine-mesh screens and low intake velocities associated with intake structures.
Location of intakes at optimal depths.
Optimization of chlorination procedures during periods of ichthyoplankton occurrence to minimize mortalities while preventing fouling.

Monitoring programs implemented after facility start-up can provide data that will allow more specific and detailed assessments of potential impacts and further refinements of models, with results that are both more accurate and more precise than those in the present impact statements.

Modeling approaches using those data should be focused on an egg-equivalent endpoint that is compatible with stock-assessment models and that allows for the integration of potential OLV impacts with other factors that influence fish populations in the gulf.

Independent evaluation

Environmental impact statements prepared by contractors for the US Coast Guard (USCG) and the US Department of Transportation Maritime Administration (MARAD) have concluded that the environmental impacts from the OLV systems of the LNG facilities will be minor5-11 and will not likely be significant at the population level.12

In these impact statements, potential impacts to fisheries were assessed by comparing estimated fish egg and larval entrainment mortalities to fishery landings statistics, converting potential impacts (equivalent yield loss in pounds of fish per year) to a percentage of regional fish harvests, rather than as a percentage of impact on the fish population. Such estimates of fishery impacts are irrelevant from a fish population standpoint because fish landings are not population measures but, rather, represent fishing statistics that can change based on fisheries management practices.

The projected impacts to fisheries have resulted in a substantial number of comments on the statements from such agencies as the National Marine Fisheries Service (NMFS) and state departments of fish and wildlife, as well as such nongovernmental organizations as the Sierra Club and the Gulf Restoration Network. These comments interpret the EIS analyses as predictions of substantial impacts to fisheries and specifically to key species of recreational and commercial importance, such as red drum.

As written, the statements present a paradox in the overall conclusions concerning potential impacts on fishery resources.

For example, on one hand, the EIS for the Gulf Landing terminal concluded that the predicted red drum equivalent yield loss would represent “a long-term minor adverse impact on the taxa of concern” and that “the impact of the proposed ORV would be less than the impact of fishing on this resource.”6 (Taxa, plural for “taxon,” refers to a taxonomic group.) On the other hand, based on conservative assumptions, the assessment predicted an upper equivalent yield loss estimate that is 8.3% of the equivalent landings.

It appears that resource managers and others have misinterpreted this upper-bound estimate as a likely potential reduction in the fishery.13 To address these concerns, the CLNG’s independent evaluation of the EIS analyses4 was commissioned to assess the realism and accuracy of the predicted fisheries impacts.

The study reviewed the methods used in the statements for appropriateness and accuracy. It evaluated fisheries impact predictions for their degree of conservatism and relative uncertainty of scientific assumptions. It also evaluated the adequacy of the fisheries data used to predict impacts. In addition, the study reviewed relevant scientific reports, including analogous studies conducted for power plants, studies on the distribution of fish eggs and larvae (i.e., ichthyoplankton) in the gulf, and studies of the life history characteristics of key species.

The results of this evaluation indicate that the assumptions and uncertainties incorporated in the EIS estimates are conservative, on the whole, resulting in overestimation of likely impacts.

EIS predictions

The impact statements used ichthyoplankton sampling data, assumptions, and calculations to estimate the numbers of eggs and larvae that would likely be entrained, and they used models to project these to population-level impacts. Using ichthyoplankton data collected as part of the Southeast Area Monitoring and Assessment Program (SEAMAP), the impact statements estimated potential impacts to fish populations. They used assumptions to fill data gaps, resulting in uncertainties that may under or overestimate true environmental conditions.

The key issues associated with assessments of potential fisheries impacts are:

The adequacy of the fish population database: Do the data adequately characterize the seasonal, horizontal, and vertical distribution of ichthyoplankton relative to the locations of the LNG terminals?
Adequacy of life history data for key fish species: Are the life-stage durations and mortality rates supported by the scientific literature? What is the level of uncertainty associated with these data?
The appropriateness of the ichthyoplankton assessment modeling approach: How does the inherent variability of key model inputs propagate through and influence the validity of the model outputs? How appropriate are the endpoints used in predicting effects?

Population database

The SEAMAP database, which is administered cooperatively by the Gulf States Marine Fisheries Commission and the NMFS, contains measurements of fish egg and larval abundances from throughout the gulf, collected over a period of more than 2 decades. This database is generally a good starting point for calculating egg and larval abundances that may be affected by the proposed LNG facilities.

The best available data source, the SEAMAP database, however, has limitations for use in the environmental assessment of OLV systems. In particular, the SEAMAP data do not include depth-stratified samples, a proportion of the larvae collected could not be identified to the species level, there is no species-specific information for eggs, and there are limitations in the temporal and spatial data coverage in the vicinity of specific LNG facilities.

In all ichthyoplankton studies, some eggs and larvae cannot be identified to the species level and must be classified at a higher taxonomic level such as genus or family. In the environmental impacts statements, when fish larvae could not be identified to species, organisms that were identified only to the genus or family level were assigned to individual species in that genus or family, in proportion to the relative abundance of those species.

Click here to enlarge image

Although this approach is reasonable when describing the contents of each sample, it should not be applied to each data set as a whole, as it was in the impact statements, because abundances of different species vary seasonally (Fig. 1). For any species, its fraction of the ambiguously identified taxa therefore also varies seasonally.

Because the SEAMAP data set does not represent all periods of the year equally, the data should not be aggregated during correction for ambiguously identified taxa. This problem results in overestimates of abundances and entrainment mortalities of key species.

Uncertainty introduced

The impact predictions made in the statements relied on several important assumptions to fill data gaps. These assumptions are all useful to simplify a complex environmental evaluation, but each introduces uncertainty in the impact predictions, and some introduce bias:

Ichthyoplankton abundances near offshore LNG terminals are adequately represented by SEAMAP net samples that were taken up to a maximum of 50 miles away.

Ichthyoplankton are not distributed homogeneously throughout the northern gulf. There are distinct spatial gradients in the abundances of key species over the distances used to select SEAMAP sampling stations for use in the impact statements. Nevertheless, SEAMAP data were averaged across these gradients in the statements. Doing so, converted systematic variability into uncertainty.

For any given species and facility, oversimplification of spatial heterogeneity could lead to either an overestimate or an underestimate of the potential impact but, where there is systematic spatial variability, will always lead to an overestimate of the upper bound on the impact.

The net efficiency (i.e., the proportion of organisms retained on a plankton net) is the same for fish eggs and larvae.

Plankton nets used for sampling are not 100% efficient because some early-life-stage organisms may avoid or pass through the net. To account for this sampling problem and to allow for the estimation of actual ichthyoplankton abundances from the SEAMAP data, the impact statements assumed a net efficiency of 33% for both eggs and larvae.14 This value is within the likely range for the type of plankton net used by the SEAMAP program.

Because of their smaller size, however, the net efficiency for eggs is likely to be much lower than that for larvae. Consequently, the statements are likely to have underestimated egg abundances relative to larval abundances. The result is an underestimate of potential entrainment impacts.

All ichthyoplankton in the water column are equivalently susceptible to seawater intakes.

Fish larvae often display distinct depth preferences that vary diurnally and with age. Red drum larvae in the gulf, for example, are typically near the water surface (in the top meter or two) during the daytime but in deeper water (5 to 12 m) at night.15 Depth preferences such as these will at least partially isolate larvae from submerged seawater intakes.

The entrainment estimates used in the statements conservatively assume that all larvae occur at the depth of the seawater intake at all times. In fact, larvae may be more or less exposed to entrainment, depending on the depth preference of the species and the depth of the seawater intake. Actual entrainment could be half or less of that predicted from average abundance if seawater intakes are below the depths of peak abundance.

No entrained organisms survive (i.e., 100% entrainment mortality).

Entrainment mortality may also be reduced by active avoidance of the seawater intake by ichthyoplankton. One study has shown that the number of fish larvae entrained by both screened and open intakes at a flow velocity of 0.15 m/sec (which is similar to that planned for the LNG facilities)7 was as little as 10% of ambient abundance.16

In contrast, egg abundance was similar, indicating that fish larvae were actively avoiding the intake. Therefore, the assumption of 100% entrainment mortality used in the impact statements is likely to result in an overestimate of potential impacts.

No population-level compensation effects occur.

The statements extrapolated the estimated losses of ichthyoplankton directly to equivalent losses of adult fish, using life-history parameters. In reality, fish populations demonstrate an ability to compensate for mortality, for example, through excess fecundity. The use of a model in the statements that ignores compensation effects also leads to an overestimation of the population-level effects of entrainment.

Accuracy of data

The impact statements used mortality rates and life-stage durations for eggs, larvae, and juveniles that were estimated from field studies reported in the scientific literature. These values were estimated from observed age distributions and are not known with certainty, even for important species such as red drum and red snapper.

A recent synthesis of life-history studies provides alternative values for some of the life history parameters used in the statements.17 These values are plausibly the most accurate currently available and are recommended for use in future modeling of population-level impacts.

Model evaluation

The following systematic flaws were found in the quantitative analyses carried out in the environmental impact statements:

The variability and uncertainties associated with how the SEAMAP data are being used result in overestimates of potential impacts and a high degree of uncertainty, as described above.
The use of inappropriate calculation methods and highly uncertain life-history parameters in forward-projection (adult-equivalent) models produces biased and highly uncertain results and is inconsistent with fisheries stock-assessment methods, which are used to make other important fishery management decisions.
The direct comparison of adult-equivalent weights with fishery landing statistics is not a valid endpoint and may lead to inappropriate conclusions concerning the severity of potential impacts.

Seasonal spawning results in high variability in larval abundance month to month and even day to day throughout the spawning period. The impact statements lumped together all the abundance data to calculate a single annual average abundance and associated variability estimate. This approach converts systematic variability into uncertainty and so overestimates the upper bound on the average.

A better approach is to average successive abundance measurements over time periods that are short relative to the spawning season to achieve a more representative long-term average. Grouping of data over short time periods also allows the variability of each average abundance to be estimated.

For example, Fig. 1 shows the result of computing an 11-day moving average of red drum larval abundance over the spawning season. The result is that the measurements taken within each short time period are relatively similar (i.e., their variability is relatively low). The total variability of the larval abundance, when summed across all periods, is also relatively low-in particular, it is much lower than the variability of the entire data set taken as a whole.

The statements ignored seasonality when calculating confidence limits around estimated ichthyoplankton abundances, leading to a substantial inflation of the upper confidence limit. For red drum near the Gulf Landing LNG terminal, these inappropriate data analyses resulted in an estimate of 36 million larvae 1 year entrained, with 95% confidence limits ranging from 3 million to 69 million larvae.18

A recalculation that corrected for taxonomic ambiguity on a sample-specific basis and accounted for seasonality in an appropriate manner resulted in an estimate of 16 million larvae 1 year entrained, with 95% confidence limits ranging from 13 million to 19 million larvae.4

The forward-projection model used in the impact statements is one that has been endorsed by the US Environmental Protection Agency for evaluation of ichthyoplankton entrainment at power plants. “Forward-projection” refers to extrapolation of numbers of eggs and larvae to weights of adult fish (equivalent yield).

Evaluation of this model revealed that it uses an inaccurate approximation to represent the incremental mortality due to entrainment. The nature of this flaw results in an underestimate of the potential survival of entrained eggs and larvae. That is, more of the entrained eggs and larvae would have survived than the impact statements account for.

Thus, this error translates to an underestimate of impact in the EIS.

Use of a forward-projection model is inconsistent with fisheries stock-assessment methods, which use a “fecundity hindcasting” (or egg-equivalent) approach to make fishery management decisions. Thus the ORV impacts are not being judged by the same methods as those used to evaluate other stressors on fish populations such as recreational and commercial fishing.

Click here to enlarge image

To illustrate the effects of the uncertainty associated with the fisheries impact projections, Table 1 summarizes assessments of the accuracy and precision of mortality estimates for red drum at Gulf Landing. The net effect of these is a positive bias: that is, the EIS has substantially overestimated red drum mortality.

Recommended modeling approach

In the environmental impact statements, the effects of entrainment on the fish population were evaluated by estimating the number of equivalent adult fish that the entrained eggs and larvae would have grown into had they survived. The projected number of adult fish is less than the number of entrained eggs and larvae because natural mortality reduces the population size as the fish age.

Empirical relationships between fish age, length, and weight were then used to calculate the adult weight of these fish. The endpoint of the forward-projection model used in the statements is foregone production-specifically, the weight of fish of harvestable age that is not produced as a result of entrainment of eggs and larvae, or equivalent yield loss.

Predicted equivalent yield losses were then compared to fishery landings, treating them as if they were pounds of fish that would not be landed. The projection of egg and larval abundance to abundance and weights of fish of harvestable age requires extensive use of poorly known fish life-history parameters. In addition, comparison of the results to fishery harvest levels (recreational and commercial landings) provides no meaningful information about the potential impacts of LNG facilities on fish populations.

This benchmark is subjective because fishery landings can be changed by fisheries management decisions and other economic factors and is inappropriate because it does not account for population compensation effects. These weaknesses lead to a high degree of uncertainty and an overestimate of potential impacts.

The potential extent of fish mortality overestimation can be evaluated by contrasting the EIS results with the results of a revised modeling approach that includes:

A corrected abundance calculation, as discussed previously.
More accurate estimates of life-history parameters, such as those by Gallaway.17
A correction to the survival fraction calculation that accounts for the entrainment mortality rate, as well as the natural mortality rate.
A fecundity hindcasting, or egg-equivalent modeling approach.

Instead of the forward-projection approach used in the impact statements, the fishery impact assessments should use a fecundity hindcasting (or egg-equivalent) approach, in which entrainment losses of ichthyoplankton are converted to egg production and compared to total egg production at the population level. This approach provides a measure of the actual impacts to fish populations rather than to fishery landings.

Such assessments can be conducted with available data, do not require as many uncertain estimates of mortality rates, and provide more meaningful and interpretable endpoints. This approach is also compatible with stock-assessment methods that are used to evaluate fish populations as a whole.

Expressing the potential impact of OLV entrainment losses as an effective reduction in egg production would allow these effects to be incorporated into stock-assessment models that include the whole population and that can also account for natural mortality compensation effects.

To illustrate the overall effects of many of the issues identified in this review, the potential impacts on red drum at the Gulf Landing terminal were recalculated, with more accurate and precise alternative approaches. Table 1 summarizes the issues that affect the accuracy and precision of impact estimates and identifies those for which an alternative approach was used.

This recalculation resulted in a substantially lower estimate of impact for red drum. The EIS for the Gulf Landing facility predicts annual red drum mortality equivalent to about 28,000 age-1-equivalent fish.18

In contrast, the more accurate calculation predicts mortality of 5,600 age-1-equivalent fish, which is equivalent to 8 spawning females or 16 spawning fish of both sexes.

The primary cause of the difference between these two estimates is the use of uncertain life-history parameters for juvenile fish in the age 1-equivalent estimate. In practice, this uncertainty takes the form of a positive bias (overestimate) in the estimated impact. In this case, the Gulf Landing EIS estimate of impact is 1,750 times higher (i.e., 28,000 age-1 equivalents/16 spawning fish of both sexes) than the more accurate fecundity-based estimate.

References

Current/Planned Deepwater Ports. http://www.marad.dot.gov/dwp/deepwater_ports/index.asp. Deepwater Port Licensing, Maritime Administration (MARAD).
The Deepwater Port Act (DWPA) of 1974, as amended by the Maritime Transportation Security Act of 2002, 33 USC. 105 et seq.
Letter from Kathleen Babineaux Blanco, Governor of Louisiana, to John Jamian, Acting Maritime Administrator, US Department of Transportation, Maritime Administration, May 17, 2005. Available at www.gov.state.la.us/STUFF/ Press_Release_detail.asp?id=844.
Exponent. An evaluation of the approaches used to predict potential impacts of open loop LNG vaporization systems on fishery resources of the Gulf of Mexico. http://www.lngfacts.org/newsroom/Eval-OLV.pdf; 2005.
USCG and MARAD. Final environmental assessment of the El Paso Energy Bridge Gulf of Mexico LLC Deepwater Port License Application. US Coast Guard and Maritime Administration, 2003.
USCG and MARAD. Final environmental impact statement for the Gulf Landing LLC Deepwater Port license application. US Coast Guard and Maritime Administration, 2004.
USCG and MARAD. Draft environmental impact statement for the Compass Port LLC Deepwater Port license application. February 2005. Docket No. USCG-2004-17659. US Coast Guard and Maritime Administration.
USCG and MARAD. Draft environmental impact statement for the Pearl Crossing LNG Terminal LLC Deepwater Port license application. April 2005. Docket No. USCG-18474. US Coast Guard and Maritime Administration.
USCG and MARAD. Draft environmental impact statement for the Main Pass Energy Hub Deepwater Port license application. June 2005. Docket No. USCG-2004-17696. US Coast Guard and Maritime Administration.
USCG and MARAD. Final environmental impact statement for the Port Pelican. US Coast Guard and Maritime Administration 2005.
USCG and MARAD. Final environmental impact statement for the Port Pelican. US Coast Guard and Maritime Administration 2005.
USCG and MARAD. Beacon Port environmental report. US Coast Guard and Maritime Administration, 2004.
NMFS. Letter from S.A. Kennedy, Acting NEPA Coordinator, to M.A. Prescott, US Coast Guard, Jan. 3, 2005, regarding comments on the final environmental impact statement for the Gulf Landing LLC Deepwater Port license application. National Marine Fisheries Service.
NMFS. Potential impacts of liquid natural gas processing facilities on fishery organisms in the Gulf of Mexico. Memorandum from Nancy B. Thompson, Science Administrator, to Roy Crabtree, Regional Administrator, Feb. 18, 2004.
Lyczkowski-Shultz, J., and Steen Jr., J.P., Diel vertical distribution of red drum Sciaenops ocellatus larvae in the northcentral Gulf of Mexico. Fish. Bull. 89:631 641, 1991.
Zeitoun, I.H., Gulvas, J.A., and Roarabaugh, D.B., Effectiveness of fine mesh cylindrical wedge-wire screens in reducing entrainment of Lake Michigan ichthyoplankton. Can. J. Fish. Aquat. Sci. 38:120 125, 1981.
Gallaway, B.J., Proposed revisions for the early life history parameters being used for red drum Sciaenops ocellatus, red snapper Lutjanus campechanus and Penaeid shrimps in seawater-use assessments. Prepared for the Pearl Crossing LNG Terminal Project. Bryan, Tex.: LGL Ecological Research Associates, 2005, 48 pp.
e2M. Appendix G (Revised). Ichthyoplankton assessment model methodology and results for the Gulf Landing LLC deepwater port license application, environmental impact statement. Prepared for US Coast Guard; Washington: Engineering-Environmental Management Inc., 2005, 96 pp.

The authors

Click here to enlarge image

Dreas Nielsen (nielsend@ exponent.com) is a managing scientist in Exponent’s Environmental Sciences and EcoSciences practices who specializes in collecting, managing, and analyzing environmental data. He has been with Exponent and its predecessor firm, PTI Environmental Services, since 1987, and previously worked as an environmental scientist for Tetra Tech Inc. and Oregon State University. He holds a BS (1976) in biology from Union College and an MS (1982) in oceanography from Oregon State University.

Click here to enlarge image

Thomas Ginn (ginnt@ exponent.com) is a principal scientist and practice director of Exponent’s EcoSciences practice. He is one of the founders of what is today Exponent’s environmental division, formerly PTI Environmental Services. He is a biologist with 30 years of experience specializing in ecotoxicology, natural resource damage assessment, and ecological risk assessment. He holds a BS (1968) in fisheries and an MS (1971) in biological sciences, both from Oregon State University, and a PhD (1977) in biology from New York University. Ginn is a member of the Society of Environmental Toxicology and Chemistry, the American Chemical Society, the Water Environment Federation, and the American Institute of Fishery Research Biologists.

Click here to enlarge image

Linda Ziccardi (lziccardi@ exponent.com) is a senior ecologist in Exponent’s EcoSciences practice. She has 18 years of experience evaluating environmental impacts at industrial and development sites nationwide. Before joining Exponent, Ziccardi held consulting scientist positions at several other firms, including Blasland, Bouck, and Lee and Parsons Engineering Science. She holds a BS (1985) in natural resource management and applied ecology from Cook College of Rutgers University and has completed graduate courses in environmental science, aquatic toxicology, water law, and natural resource management at Rutgers University (1987-89). Ziccardi is a member of the Society of Environmental Toxicology and Chemistry, the Rocky Mountain Association of Environmental Professionals, and the Colorado Hazardous Waste Management Society.

Click here to enlarge image

Paul Boehm (pboehm@ exponent.com) is a group vice-president and principal scientist at Exponent-Health and Environmental. He has 30 years of experience specializing in environmental impact assessments of oil and gas facilities, environmental forensics, and natural resource damage assessments at oil spill and hazardous waste sites. He formerly held senior positions in the energy and environmental practices at Arthur D. Little Inc. and Battelle Memorial Institute. Boehm holds a BS (1970) in chemical engineering from the University of Rochester, an MS (1973) in oceanography from the University of Rhode Island, and a PhD (1977) in oceanography from the University of Rhode Island’s Graduate School of Oceanography. He is on the advisory board of the Institute for Energy Law; is an associate of the American Bar Association’s section on energy, environment, and resources; and is a member of the Society of Environmental Toxicology and Chemistry and the American Chemical Society.