The economic value of freshwater inputs to an estuarine fishery

Author’s Accepted Manuscript The economic value of freshwater inputs to an estuarine fishery Chris J. Kennedy, Edward B. Barbier

www.elsevier.com/locate/wre

PII: DOI: Reference:

S2212-4284(15)30016-5 http://dx.doi.org/10.1016/j.wre.2015.11.003 WRE65

To appear in: Water Resources and Economics Received date: 15 December 2014 Revised date: 25 November 2015 Accepted date: 27 November 2015 Cite this article as: Chris J. Kennedy and Edward B. Barbier, The economic value of freshwater inputs to an estuarine fishery, Water Resources and Economics, http://dx.doi.org/10.1016/j.wre.2015.11.003 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Title: The economic value of freshwater inputs to an estuarine fishery Corresponding author: Chris J Kennedy, Department of Environmental Science and Policy, George Mason University, VA, USA. Email: [email protected]. Phone: 703.993.5471. Address: George Mason University, MSN 5F2 4400 University Drive Fairfax, VA 22030 Second author: Edward B Barbier, Department of Economics and Finance, University of Wyoming, WY, USA. Email: [email protected] Address: University of Wyoming 1000 E University Ave. Dept. 3985, BU 225 East Laramie, WY 82071

The economic value of freshwater inputs to an estuarine fishery Draft date: December 4, 2015 Abstract The health of many estuarine and coastal ecosystems depends on upstream hydrology; however, there is limited information on the economic effects of restricting freshwater flow to the coast. We investigate the role of freshwater inputs on the blue crab fishery in Georgia, USA. Blue crabs are known to respond to salinity changes in the estuaries in which they reside, and declining freshwater flow to the coast over the last 50 years is correlated with increases in average estuarine salinity and falling commercial harvests. A structural, bioeconomic model reveals freshwater inputs have significant effects on fishery outcomes. Simulations of a counterfactual minimum-flow standard for three different rivers, set at 25% of 50-year averages, suggest such a policy would result in measurable benefits for the fishery, improving profits by 35% ($1.7 million) during the period 2002-2012 in the three fishing areas for which data were available. Aggregating

1

the necessary additional water releases to achieve this standard yields an approximate average value of water of between $1 and $7 per acre-foot.

1.

Introduction Estuarine and coastal ecosystems depend on freshwater inputs to maintain

moderate salinity (Alber, 2002), and reduced flows can substantially alter coastal ecology and the productivity of fishery resources (Das et al., 2012; Gillson, 2011; Mallin, Paerl, Rudek, & Bates, 1993). In the southeast United States, rapid population growth and changing precipitation patterns have put significant pressure on surface water resources (Olsen, Padma, & Richter, 2006; Vörösmarty et al., 2010), in most cases reducing the average volume of freshwater reaching coastal areas and increasing estuarine salinity (Alber, 2002). Blue crab fisheries along the coasts of the mid- and south-Atlantic and Gulf of Mexico have been identified as particularly vulnerable to elevated estuarine salinities, and reductions in catch have been correlated with reduced river flow to the coast since the 1990s (Posey, Alphin, Harwell, & Allen, 2005; Powell, Matsumoto, & Brock, 2002; Rogers, Arredondo, & Latham, 1990; Wilber, 1994). However, existing studies on this topic have tended to either (a) use an iterative process to identify an ad hoc structural model relating harvest to river flow (Powell et al., 2002; Wilber, 1994); or, (b) examine one or multiple specific mechanistic hypotheses without extending the analysis to fishery-level questions (Parmenter, 2012; Wrona, Wiegert, & Bishop, 1995). We attempt to both improve upon these previous efforts, as well as fill a gap in the ecosystem service valuation literature. There is a large body of work on the value of coastal ecosystem goods and services (Barbier et al., 2011), as well as the value of water (Young & Loomis, 2014). However, to our knowledge, there has been no attempt to quantify the economic implications of freshwater interactions with coastal fishery performance1, and water managers have almost universally ignored these impacts when 1

One related study was found during an exhaustive literature review. Fisher, Hanemann, and Keeler (1991) examine the impact of freshwater flow on California’s Central Valley salmon fishery using a simple simulation model. Chinook salmon migrate upriver to spawn, but are harvested in the ocean. They find likely benefits associated with increased

2

setting minimum flow standards (MFSs) and issuing withdrawal permits for rivers entering the sea.2 While freshwater inputs could be considered a driver of habitat size in estuarine environments, this study most closely fits within the research on the economic value of water quality improvements for marine fisheries (Finnoff & Tschirhart, 2011; Huang & Smith, 2011; Knowler, Barbier, & Strand, 2002; Massey, Newbold, & Gentner, 2006). This is because of the transient nature of salinity changes (the primary mechanism by which freshwater impacts the blue crab fishery), as well as the specific biological pathways by which salinity affects crabs (e.g., metabolic stress, disease). In this paper, we develop a structural, bioeconomic model consisting of four interdependent equations describing the evolution of adult crab biomass, juvenile crab biomass, harvests, and effort, with the goal of more thoroughly investigating the biological mechanisms driving the relationships between crab populations and salinity levels, and the implications for fishery performance. The model is estimated using highresolution, spatially-explicit data from the Georgia blue crab fishery. Specific, mechanistic hypotheses related to the impacts of salinity on unique crab life stages are embedded in this model. The parameterized model—along with a coupled model linking salinity to river flow in six estuarine systems (sounds)—is then used to simulate the effects of a hypothetical minimum flow standard (MFS) on fishery outcomes in three sounds fed by rivers. We find elevated salinity in these three sounds to be strongly correlated with nearby water flow readings, and model results reveal three crab life stages experience statistically-significant, negative impacts as a result. The MFS simulation predicts sound-specific profits would have been 28-83% ($300,000-$900,000) greater had the standard been binding during the period 2002-2012. This final step allows us to generate estimates for the average value of additional water released in accordance with the MFS, which is found to be between $1 and $7 per acrefoot, on the same order of magnitude as estimated values for other uses of water in this region. These figures should aid policymakers in their efforts to allocate water to freshwater release; however, economic benefits are not monetized, and the relationship between natural and hatchery salmon complicates the analysis. 2 The few instances in which US jurisdictions enforce minimum flow standards to the sea are not the result of a thoughtful weighing of benefits, but instead almost always driven by litigation under the Endangered Species Act (Alber, 2002; Powell et al., 2002).

3

alternative uses, and contributes to recent calls for embedding ecosystem service measures in coastal planning efforts (Arkema et al., 2015; C. J. Kennedy & Cheong, 2013). The next section provides a description of the commercial fishery in Georgia and its historical performance, a review of the biological motivations for specific hypotheses regarding salinity-crab hypotheses, and the relevance of this study in light of current stressors affecting surface water resources in the state. The bioeconomic model and description of the datasets used in the estimation are introduced in Section 3, followed by estimation results and a discussion of alternative model assumptions and structural specifications that were tested but ultimately rejected. Section 4 describes the counterfactual MFS simulation and discussion of results. Section 5 concludes.

2.

Background

2.1

The Georgia blue crab fishery The common blue crab (Callinectes sapidus) is known for its sweet, tender meat

that—in lump form—is the basis of Chesapeake Bay Crab Cakes or, alternatively (and with a significant amount of effort) is picked apart at crab bakes every summer throughout the Atlantic and Gulf of Mexico coastal regions of the United States. The commercial fishery in Georgia is the second most valuable in the state (Evans, 1998), with revenues at nearly $4 million (year-2010 dollars) on landings of 3.2 million pounds in 2013.3 This is despite a fall in landings since a modern high of 9.1 million pounds in 1995 (Figure 1). By 2003, landings were under two million pounds, and a disaster was declared by the National Marine Fisheries Service (GDNR, 2004).

< INSERT FIGURE 1 > Caption: Total landings of blue crab from 1950 to 2012. 1950-1988 from NMFS; 19892012 from GDNR.

3

Personal communication, NOAA National Marine Fisheries Service, Fisheries Statistics Division.

4

The fishery is prosecuted exclusively by small vessels, often operated alone by captains. Traps with floats and juvenile escape rings are the only allowed gear. Crabbers are territorial in response to congestion externalities associated with gear entanglement and heterogeneous fishing ground quality (Frizzelle, 1993), and will generally retrieve, harvest, bait, and replace traps every day or two, depending on the season and other conditions. Regulations designed to control effort and participation in this fishery are effectively non-existent: a license limit set equal to the number of active crabbers was introduced in 1995 (154), and has never been binding (approximately 80 licenses are actively fished now); a restriction on the number of traps allowed by a single license—set at 200—is avoided by employing inactive licenses, which are held mostly by active crabbers, or reducing the harvest interval.4 The Georgia coast is comprised of an extensive network of marshes that tolerate salinity and provide the structural foundation for a series of barrier islands. As a result, the fishery is effectively divided into a series of sub-fisheries occurring in hydrologicallyindependent sounds (Figure 2).

< INSERT FIGURE 2 > Caption: Map of the Georgia coast. Courtesy of V.J. Henry and The New Georgia Encyclopedia.

This unique structure is useful for estimation, as it allows relationships between salinity and fishery performance to be investigated at the sound level.

2.2

Blue crabs and freshwater Inhabiting a network of salt marshes and estuaries, crabs are impacted by salinity

through disruptions to mating and reproduction, reduced recruitment of juveniles to the adult stock, and increases in disease incidence, metabolic stress, and predation (Powell et al., 2002; Rogers et al., 1990; Wilber, 1994).

4

Personal communication with crabber informants and representatives of GDNR.

5

Elevated salinity impedes effective reproduction in blue crabs by inducing eggbearing females to spawn too far upstream, from where the eggs and larvae are not able to reach the low-turbidity, high-salinity marine environments necessary for development (GDNR, 2008). Nine months to a year post-spawning, juveniles return to estuaries, where elevated salinity is associated with metabolic stress, increased predation by marine predators, and disease, resulting in lower subsequent recruitment to the adult population (Aguilar et al., 2008; Hines, 2007). Finally, adult crabs have been observed to contract a suffocating blood parasite, Hematodinium perezi, in high-salinity, high-temperature areas, with infection rates reaching as high as 50% in some Georgia estuaries during periods of drought (Lee & Frischer, 2004). While drought conditions are a significant factor behind acute instances of elevated estuarine salinity, increasing terrestrial demands have exacerbated the situation. Since 1960, water consumption in Georgia has risen by 300%, driven in part by population growth (USGS, 2006). These effects are reflected in river flow patterns: flow readings from the eastern-most stations of the two largest rivers in Georgia (Altamaha and the Savannah) have declined about 35% since 1960 (Figure 3).

< INSERT FIGURE 3 > Caption: Yearly river flow at eastern stations of the Altamaha and Savannah Rivers. Data from USGS.

Like most US states, Georgia does not consider coastal ecosystem resource needs when setting minimum flow rates and evaluating withdrawal applications (GWC, 2008), and while efforts have been made to include coastal resources within management objectives, subsequent regulatory initiatives have not materialized.5 Thus, during periods of crisis, the ecological health of coastal areas (and any the value of any associated economic activities) are given low priority.

3. 5

Material and Methods

Personal communication, Georgia Environmental Protection Division (EPD) staff.

6

3.1

Empirical approach Blue crabs are have a relatively short life span in actively fished areas, and are

particularly sensitive to environmental variables, including salinity and temperature. 6 These factors suggest crab abundance may fluctuate significantly in response to shocks to even a single cohort, a hypothesis supported by the observed variability in historical harvests. Our empirical approach is designed to capture environmental impacts to multiple life stages occurring on a relatively high temporal resolution (quarterly). While fishery-level stock-recruitment relationships have been identified for several blue crab fisheries (e.g., Chesapeake Bay (Lipcius & Van Engel, 1990) and North Carolina (Eggleston, Johnson, & Hightower, 2004)), this approach is not appropriate for Georgia due to the hydrological independence of sounds. Additionally, examining the harvest data reveals individual crabbers rarely shift operations between sounds. For these reasons, we assume stock-recruitment relationships and effort adjustment occur at the sound level. Stock assessment data are not available for the Georgia fishery. As a result, we rely on raw abundance measures from fishery-independent survey trawls. The data and associated transformations are described in more detail in Section 3.2. Figure 4 provides a visual timeline of our approach.

< INSERT FIGURE 4 > Caption: Conceptual timeline of the bioeconomic model linking salinity to the blue crab fishery. The three impacts of salinity are represented by A (spawning), B (disease), and C (juvenile mortality).

Adult crab abundance is reduced by fishing, and increased by recruitment. Abundance is assumed to evolve according to Equation 1

6

Crabs reach maturity in 12-15 months, and rarely survive more than 2-3 years in active fisheries (Archambault, Wenner, & Whitaker, 1990; Hill, Fowler, & Moran, 1989).

7

௓೟షభǡೞ

ᇩᇭᇭᇭᇭᇪᇭᇭᇭᇭᇫ ܺ௧ǡ௦ ൌ ݂൫ܵ௧ǡ௦ ǡ ܶ௧ǡ௦ ൯ ൅ ሺͳ െ ܰሻ ൫ܺ ௧ିଵǡ௦ െ ‫ܪ‬௧ିଵǡ௦ ൯

(1)

൅ ܴ൫ܺ௧ିହǡ௦ ǡ ܵ௧ିଵǡ௦ ǡ ܵ௧ିହǡ௦ ǡ ܶ௧ିଵǡ௦ ǡ ܶ௧ିହǡ௦ ൯ where subscripts t and s index for quarter and sound, respectively. ܺ௧ǡ௦ is the abundance of adult crabs; ݂൫ܵ௧ǡ௦ ǡ ܶ௧ǡ௦ ൯ represents the impact of physical variables (salinity and temperature7) on current adult abundance; ܼ௧ǡ௦ is a measure of escapement abundance, ܺ௧ǡ௦ net of harvests, ‫ܪ‬௧ǡ௦ , adjusted to abundance units; ܰ is natural mortality; and, ܴሺǤ ሻ is juvenile abundance (recruitment), assumed to be impacted by physical variables during spawning (‫ ݐ‬െ ͷ) and settlement (‫ ݐ‬െ ͳ). One year passes between spawning and the juvenile settlement, and juveniles recruit to the adult stock one quarter after that (Wrona, 2004). The complex life history of blue crabs requires specific attention to be paid to juvenile crabs, and we do so by considering impacts to juveniles separately from adults. Since both ܺ and ܴ are predetermined, separate estimating equations are used to describe the evolution of each abundance measure. Without specific guidance from the biological literature, the basic structural specification assumes a linear relationship between salinity and adult abundance.8 The estimating equation for adult abundance is

7

While this research is concerned primarily with salinity, we include estuarine temperature—which can be influenced by the same factors affecting salinity—as a control variable. Elevated temperatures have been implicated in the proliferation of Hematodinium (Lee & Frischer, 2004), and can cause coastal eutrophication (Conley et al., 2009), which comes with a host of negative impacts for crabs. 8 A number of structural specifications for salinity and temperature are examined as robustness checks. The model selection process has indicated this simple, additive and linear structural form offers the strongest explanatory power. Equations 1 and 2 suggest the stock could regenerate spontaneously under certain physical conditions, even if no adult or juvenile crabs were present in the previous period. However, as mentioned, direct measures of stock are not available, and we rely on abundance estimates from fisheryindependent survey trawls. While these measures act as rough proxies for stock, extreme values (i.e., measures at or near zero) should be interpreted with caution. This, combined with the weak connectivity between adjacent fishing areas that is not observable, implies that the relationships described by the set of equations should be viewed as approximations, and not interpreted as violating global constraints on underlying biological processes.

8

ܺ௧ǡ௦ ൌ ߠ଴ ൅ ߠଵ ܵ௧ǡ௦ ൅ ߠଶ ܶ௧ǡ௦ ൅ ߠଷ ܼ௧ିଵǡ௦ ൅ ߠସ ܴ௧ିଵǡ௦ ൅ ܻ௤ ൅ ‫ܦ‬௤ ൅ ‫ܦ‬ௌ ൅ ߳

(2)

ܻ௤ is a year trend variable, ‫ܦ‬௤ and ‫ܦ‬௦ are quarter and sound-specific dummy variables, respectively, and ߳ is an i.i.d., normally distributed error term. Recruitment is modeled via a variation of the Ricker stock-recruitment specification (Ricker, 1975), used to model blue crab fisheries in North Carolina (Eggleston et al., 2004) and the Chesapeake Bay (Lipcius & Van Engel, 1990). Adjusting for our quarterly approach, the structure is given as ܴ௧ ൌ ߟ଴ ܺ௧ିସ ݁ ఎభ ௑೟షర , where ߟ଴ ൐ Ͳ and ߟଵ ൏ Ͳ. Following work by Tang (1985), salinity and temperature are assumed to affect the leading coefficient, ߟ଴ , multiplicatively. Dividing through by ܺ௧ିସ and logtransforming yields Equation 39 ܴ௧ǡ௦ ݈݊ ቆ ቇ ൌ ߟ଴ ൅ ߟଵ ܺ௧ିସǡ௦ ൅ ߟଶ ܵ௧ǡ௦ ൅ ߟଷ ܶ௧ǡ௦ ൅ ߟସ ܵ௧ିସǡ௦ ൅ ߟହ ܶ௧ିସǡ௦ ൅ ܻ௤ ൅ ‫ܦ‬௤ ܺ௧ିସǡ௦

(3)

൅ ‫ܦ‬ௌ ൅ ߳ Salinity is assumed to affect recruitment in two ways: contemporaneously, due to the sensitivity of juveniles to high salinity, and four quarters prior through the interruption of spawning. Both effects are referred to as “indirect” impacts. Harvests are assumed to occur according to Cobb-Douglas production technology Ž൫‫ܪ‬௧ǡ௦ ൯ ൌ ߙ଴ ൅ ߙଵ ݈݊൫‫ܧ‬௧ǡ௦ ൯ ൅ ߙଶ ݈݊൫ܺ௧ǡ௦ ൯ ൅ ܻ௤ ൅ ‫ܦ‬௤ ൅ ‫ܦ‬ௌ ൅ ߳

(4)

where ‫ܧ‬௧ǡ௦ indicates effort (trips taken). As discussed above, due to the non-binding license limitation, open access management of the fishery is assumed. Consequently, effort is assumed to respond recursively to profits ‫ܧ‬௧ǡ௦ െ ‫ܧ‬௧ିସǡ௦ ൌ ܾ଴ ൅ ܾଵ ൫ߨ௧ିଵǡ௦ െ ߨ௧ିହǡ௦ ൯ ൅ ܻ௤ ൅ ‫ܦ‬ௌ ൅ ߳

(5)

where profit ߨ௧ǡ௦ ൌ ‫݌‬௧ǡ௦ ‫ܪ כ‬௧ǡ௦ െ ܿ௧ǡ௦ ‫ܧ כ‬௧ǡ௦ . ‫݌‬௧ǡ௦ is the average price-per-pound in year2010 dollars and ܿ௧ǡ௦ is variable cost per trip. Both ‫݌‬௧ǡ௦ and ܿ௧ǡ௦ are adjusted to year-2010 dollars. ܾଵ is the speed of the effort-adjustment process, occurring on a quarterly basis. There is strong seasonality in effort and the number of participating crabbers in this fishery and product markets. Thus, rather than focus on absolute rents, we instead assume

9

For ease of presentation, coefficients are not represented in a log-transformed manner.

9

that if rents in t – 1 (one quarter ago) were higher relative to rents in t – 5 (the same quarter in the previous year), effort in t should increase relative to t – 4. A summary of model variables is given in Table 1.

< INSERT TABLE 1 > Caption: Variables used in the modeling approach. * Subscripts refer to quarter/sound. † GDNRa refers to scientific survey trawls. Records are from (a) inshore and (b) near-shore surveys. GDNRb refers to commercial harvest records. Both datasets from GDNR. Variable cost is the average of estimates from Virginia and Maryland. ‡ Juvenile implies individuals about to recruit to the adult stock (carapace width > 50mm).

3.2

Data description and sources Abundance and harvest data obtained from the Georgia Department of Natural

Resources (GDNR) Coastal Resources Division. Abundance data are the result of fisheryindependent survey trawls, conducted monthly since 1976. Six trawls per month assess crab populations in six different sounds (Ossabaw, Sapelo, St. Andrews, St. Marys, St. Simons, and Wassaw, see Figure 2) in creeks, the lower estuary, and continental shelf, recording weight, number of adults, mature females, and juveniles 10 , and physical variables, including salinity and temperature. We censor observations from the shelf, where oceanic salinity levels lead to few observed crabs and fishing operations are nonexistent. To generate abundance indices, we sum the number of adults and juveniles in each quarter, standardized where necessary to account for missing observations. The second dataset contains all commercial trip records from 1989-2012 (~200,000 records). These records include pounds caught, price received, license number, and general geographic information. Prior to 2001, trip records were completed by dealers. In 2001, crabbers were designated as dealers and that responsibility shifted to them. These changes are the likely cause of some irregularities in the data. Specifically, a significant portion of trips prior 2001 do not include location or license entries, and show a significant upward bias in catch per trip, suggesting wholesalers may have been 10

The juvenile count includes only those juveniles that are about to enter the adult stock (minimum carapace width of approximately 50mm).

10

aggregating multiple trips and/or licenses to cut down on paperwork. Therefore, we focus the analysis on 2001-2012, providing 48 quarterly observations per sound. Summary statistics for crab abundance, salinity, harvests, and trips are given in Table 2.

< INSERT TABLE 2 > Caption: Sound-specific summary statistics. Adult and juvenile measures are standardized to the sum of 12 quarterly surveys. Salinity is the average recorded surface salinity, measured in practical salinity units (PSU). Pounds and trips are totals.

Estimates of the costs of fishing are not available, and so we rely on a survey of crabbers from the Chesapeake Bay conducted in 1999 (Rhodes, Liption, & Shabman, 2001). The survey found average daily variable costs—which include fuel, bait, gear, and the opportunity cost of labor—to be approximately $188 per day (trip). Further, as GDNR does not provide stock estimates based on surveys, it is necessary to develop an estimate of the harvest rate, ‫ ܨ‬, to reconcile aggregate harvest data with abundance indices. As an independent stock assessment is beyond the scope of this research, we instead suggest that for each quarter/sound combination, the maximum possible fishing mortality is 50%, occurring when data reveal the highest ratio of harvests-to-abundance, ൫‫ܪ‬௧ǡ௦ Ȁܺ௧ǡ௦ ൯

௠௔௫

period, ‫ݐܨ‬ǡ‫ ݏ‬ൌ

. This baseline ratio is then used to calculate the harvest rate in every

ͲǤͷൈ൫‫ݐܪ‬ǡ‫ ݏ‬Ȁܺ‫ݐ‬ǡ‫ ݏ‬൯ ݉ܽ‫ݔ‬

൫‫ݐܪ‬ǡ‫ ݏ‬Ȁܺ‫ݐ‬ǡ‫ ݏ‬൯

. While unconventional, the approach yields an average harvest

rate of 11.6% that, when annualized, falls in the estimated range for the North Carolina fishery (43-56%) (Eggleston et al., 2004). 11 Escapement is then computed as ܼ௧ǡ௦ ൌ ൫ͳ െ ‫ܨ‬௧ǡ௦ ൯ܺ௧ǡ௦ .

3.3

Estimation and results12 The regressors for each equation are either pre-determined or exogenous to the

system, and therefore may be estimated independently. Thus, model estimates can be 11

Numerous alternatives were tested (e.g., constant harvest rate, 25% and 75% max rates). Sensitivity analysis confirms estimation results are only minimally sensitive. 12 Estimation results from robustness checks and alternative structural specifications.

11

generated via equation-by-equation OLS. However, it is likely there are mechanisms linking the different equations that are not explicitly accounted for. This is particularly true for the biological equations, which may be impacted by unobserved physical or environmental variables. In this case, allowing the error structures of the various equations to be related may offer efficiency gains. To test this hypothesis, we perform a Breusch-Pagan test of independence of equations on the various structural specifications of the model. Cross-equation error independence is strongly rejected; therefore, we employ a seemingly unrelated regressions (SUR) procedure (Zellner, 1962), well suited for systems of independent equations in which the error structures are correlated (Greene, 2003). A Hausman specification test for the SUR specification indicates an improvement over OLS. Estimation is performed using STATA 14 software, with results in Table 3.

< INSERT TABLE 3 > Estimation results for the primary specification. t – statistics are given in parentheses. The interpretation of Equation 3 is more complicated due to the structure of the estimating equation. This model restricts parameters to be identical across sounds. There is not a strong argument for marginal effects differing systematically across sounds; however, to test this assumption, we perform a Chow test on the parameter estimates from independently estimated equations from each sound. The null is that there is no difference, and cannot be rejected. Durbin-Watson (DW) statistics for the four, sound-specific equations were computed, and Breusch-Pagan tests for heteroskedasticity were performed. While heteroskedasticity is indicated in three of the four equations, standard errors are still very small after correcting for non-constant errors. As suggested by the biological literature, current adult abundance is positively impacted by both escapement and juvenile abundance. Salinity, which was hypothesized to negatively affect adult abundance through the promotion of Hematodnium is negative, as expected, but less significant. Temperature has a positive influence on adult abundance, in line with other evidence suggesting growth is positively impacted by moderately increased temperatures (Wrona, 2004).

12

Table 3 similarly confirms assumptions regarding the evolution of juvenile abundance in the fishery. The coefficient on adult abundance is significant and negative, indicating density dependence and consistent with previous studies. Salinity—concurrent and lagged—has significant and negative impacts on abundance, confirming a priori hypotheses regarding juvenile settlement and spawning processes. Temperature also has a negative impact on concurrent juvenile settlement, but not on spawning. Harvests are strongly dependent on both adult abundance and the number of trips taken. While production appears to be locally increasing in scale with respect to the number of trips taken, this is possibly due to the composite nature of the effort variable. Due to gaps in the data, we did not include the number of traps fished or soak time as explanatory variables. Additionally, the standard practice is to check traps once every one or two days. This choice results, in part, from common knowledge about the way in which tides and crab movement and foraging interact, and therefore, arbitrarily increasing the number of trips taken (at least during good conditions) may not always be feasible. Finally, effort responds positively to quarter-specific, relative changes in profits. However, the explanatory power is low. A number of additional variables were evaluated for inclusion in Equation 5, including fuel prices and state-level employment and wage indices. None were found to have a measureable influence on effort. A number of robustness checks were performed. To test whether it is necessary to explicitly consider multiple life stages in the specification, a sparse, three-equation model was estimated, in which the right hand side variables of Equation 3 are embedded in Equation 2. The four-equation model offers a better fit, and the coefficients on the salinity variables are consistent. The model is also estimated using the full data series (1989-2012), including dummy variables for observations occurring after 1995 (new commercial regulations), and 2001 (dealer designation). Interaction effects between these rule changes and harvest and effort are also tested. We also examine alternative structural specifications for the impact of salinity. Our primary specification is linear; ܵ௧ǡ௦ . We also test four transformations, as well as interaction effects between salinity and abundance indices. The transformations are performed to test specific hypotheses regarding the way

13

in which salinity affects crabs.13 Results indicate salinity has consistent, negative, direct (contemporaneous) and indirect impacts on juvenile abundance. The coefficients are robust across all specifications. Additionally, the linear specification for the impacts of salinity is retained, as alternatives provide no obvious improvement.

4.

Policy Simulation

4.1

Counterfactual flows and simulation To understand the economic implications of the above findings, we develop a

counterfactual policy simulation of a minimum flow standard (MFS), a common regulatory approach for river and stream flow management (Alber, 2002). To do this, it is necessary to understand how salinity and river flow interact in the three riverine sounds in our sample—Ossabaw Sound (Ogeechee River); St. Andrews Sound (Satilla River); and, St. Marys Sound (St. Marys River). Flow data are recorded from the eastern-most reporting stations, and obtained from the USGS National Water Information Service. When modeling the relationship between salinity and flow, our strategy is to minimize mean squared error, constrain errors to be homoscedastic, and limit complexity. A log-linear relationship between flow and salinity achieves these objectives (Figure 5).

< INSERT FIGURE 5 > Caption: ln(flow) vs. salinity for the three sounds with measurable river flow entering the estuaries.

We evaluate the potential impact of a 25% MFS. Such a standard would require flow in each river to be maintained at or above 25% of the quarterly average from 1960 to 2012 (53 years). The MFS is binding in 12 quarters of the simulation period for Ossabaw Sound, 18 for St. Andrews, and 16 for St. Marys (see Table 5).

< INSERT TABLE 4 > 13

Specifically: (a) salinity levels above 25 PSU (practical salinity units) impede growth, and (b) existence of an “optimal” level of salinity of 15 or 20 PSU (Wrona, 2004).

14

Caption: Quarterly river flow and salinity over the simulation period.

Predicted flows under the minimum standards are then used to generate counterfactual salinity profiles. These are used to simulate the impact of the MFS on the fishery using a dynamic, recursive adjustment process, parameterized with Table 3. In each period, new values for harvest, trips taken, variable profits, adult abundance, and juvenile abundance are calculated and used to adjust the original series.

4.2

Simulation results Our model assumes crabbers adjust the number of trips they take in response to

variable profits. Table 5 reports total profits, or revenues net of variable and fixed costs, which includes boat and engine maintenance, insurance, and depreciation, estimated to be $254 per day in 1999 (Rhodes, Liption, and Shabman 2001). Table 5 reveals fishery performance in all three sounds responds positively to the counterfactual MFS, with estimated improvements to sound-level profit of 29-83% over the period 2002-2012.14

< INSERT TABLE 5 > Caption: Counterfactual simulation results.

The most significant biological response is observed in juvenile abundance, with improvements of 93-153%. Adult abundance experiences a more modest effect. Figures 6, 7, and 8 display counterfactual paths for harvests, effort, and adult and juvenile abundance, revealing our model’s inability to predict the large spikes in adult abundance observed in the survey data. It is not clear why these spikes occur; however, juveniles are subject to significant cannibalism (Moksnes, Lipcius, Pihl, & van Montfrans, 1997), which could perhaps be behind cycles of adult population booms and busts following a collapse.

< INSERT FIGURE 6 > 14

The simulation period starts one year after the estimation period due to the lags in the juvenile abundance and effort estimating equations.

15

Caption: Counterfactual simulation for Ossabaw Sound.

< INSERT FIGURE 7 > Caption: Counterfactual simulation for St. Andrews Sound.

< INSERT FIGURE 8 > Caption: Counterfactual simulation for St. Marys Sound.

4.3

Value of water An important policy question is concerned with the value of freshwater to this

fishery, relative to other economically-important uses. As a rough estimate, the additional profit per acre-foot associated with the 25% MFS can be calculated, yielding approximately $0.63 per acre-foot for St. Andrew’s Sound, $1.94 per acre-foot for St. Mary’s Sound, and $6.97 per acre-foot for Ossabaw Sound (see Table 5). These estimates are average values: the MFS binds in less than half of the total quarters in the sample, and it would be difficult to interpret the effect of a one-time “pulse” of freshwater through the system of equations—which presumably would yield a marginal value—as estimates would be dependent on the current state of the resource and previous river flow. However, it is still useful to compare this value to other numbers from the literature. Frederick, Vandenberg, and Hanson (1996), reviewing water value estimates for the United States, report mean and median values for all uses in the South-Atlantic/Gulf of Mexico region to be $17.65 and $7.35 per acre-foot, respectively (2010 dollars). The mean value associated with estimates for terrestrial recreational/fish and wildlife habitat are approximately $4.41. More recently, Petrie and Taylor (2007) used a hedonic approach to calculate an annuity value of $35.38 per acre-foot of water per year in Georgia, based on price differentials associated with water rights. Despite the small scope of the value embedded in our estimate, it represents a non-trivial addition to the estimated value of surface water in Georgia, suggesting it

16

makes economic sense to at the very least consider coastal resources and the economic activity they generate when designating surface water management plans.15

5.

Discussion and Conclusions This study is novel in that it appears to be the first to examine the value of

freshwater inputs to a coastal fishery. The empirical model is estimated with fisheryindependent abundance data for adult and juvenile blue crabs, allowing a decoupled approach, 16 and identification of the key biophysical mechanisms driving fishery performance. The multiple ways in which salinity affects blue crabs—across multiple life stages occurring more than a year apart—suggests reduced-form models may misrepresent the impacts of an environmental disturbance on fishery dynamics and lead to incorrect predictions of economic and biological consequences. A simulation suggests a modest MFS for rivers, set at 25% of 50-year averages, would have significant impacts on fishery profits. It is important to note that, despite improvements to the data-collection process, the accuracy of commercial harvest data is still an issue. Given that each sound is fished by between five and fifteen crabbers, one individual who consistently combines multiple trips into single trip record (or underreports harvests for some reason) could bias the entire harvest series for a sound. Additionally, the Georgia coast is a complex, interconnected web of rivers, creeks, estuaries, and marshes, many of which are known by multiple, sometimes informal names, possibly leading to misidentification in the data series. Third, we have assumed a fixed price for crab. However, a significant fraction of total output is exported directly to other mid-Atlantic markets (notably Washington and Baltimore), in which case local supply and demand forces are less relevant. While various robustness checks were performed to investigate the possibility of threshold effects associated with salinity, the selected model assumes a linear functional form. However, historical abundance measures are far lower in the simulation period than 15

It is likely that the marginal value of water is higher when flow is very low, and therefore correlated with the opportunity cost of releasing water to the coast. 16 Decoupled models estimate biological and economic coefficients separately, difficult without fishery-independent abundance data (Smith, 2008).

17

the period 1976-2000, suggesting the stock may have undergone a permanent or semipermanent shift in productivity. It is possible the aforementioned drought, which has been implicated in a subsequent collapse in the fishery, led to a period of fragility that prevented full recovery subsequent to the next serious drought in 2010. If true, a minimum flow standard may also act as insurance against collapse—if the fishery ever recovers to pre-drought productivity. Alternatively, the large-scale dieback of coastal marshes that occurred concurrent with drought in the early 2000s may have reduced the carrying capacity of the fishery. This analysis does not consider physical habitat interactions, and is unable to effectively investigate fishery-level changes prior to 2001. There are a number of factors that suggest the estimated values from our policy simulations are a lower bound on the overall benefit of a MFS. As discussed above, problems with the data prevented us from estimating the model for the entire time series. However, cursory experimentation with the simulation model shows the quarterly differences in abundance and harvests under the counterfactual policy increase as the starting time period is pushed backward from 2002. This is due in part to amelioration of effects associated with the historic drought that hit Georgia and the greater southeast United States in the early 2000s. Had a flow maintenance scheme been in place before and during that time, baseline abundance levels would have likely been higher. Additionally, the scope of this analysis is small, and true benefits of a statewide MFS are likely far greater. The policy simulation is applied to only the 35% of state-wide harvests that could be definitively assigned to the three riverine sounds. An additional 10% of harvests could justifiably be assumed to have occurred in the same sounds due to ambiguous geographical classifications, and another 10% occurred in areas influenced by the rivers in our dataset but, lacking more advanced understanding of sound-specific hydrology, were censored. More significantly, due to a lack of abundance data, the two largest rivers entering the coast in Georgia (Altamaha and Savannah) were not included. These rivers have produced more than 30% of commercial harvests since 2001. Further, we have ignored the recreational component of the fishery, which is likely significant in terms of added welfare (Evans, 1998). Finally, a MFS may likely result in benefits apart from the blue crab fishery. For instance, there is evidence that shrimp are sensitive to salinity in their larval phase

18

(McKenny & Neff, 1979), and poor performance of the shrimp fishery in Georgia and other southeastern states in the early 2000s was, according to some scientists, at least in part caused by drought conditions (Breed, 2001). Drought has also been implicated as a contributing factor in the marsh dieback events that were observed during the same period (Ogburn & Alber, 2006; Silliman & Bertness, 2002). Salt marshes are some of the most productive ecosystems in the world and provide a number of ecosystem services beyond fishery habitat, including storm surge protection and shoreline stabilization, habitat for birds and mammals, carbon sequestration, and water filtration (Barbier et al., 2011). If a MFS improves the integrity of marsh habitat, subsequent benefits could be significant. Future work should include an evaluation of the relative benefits of alternative river management plans, emphasizing the total economic value of freshwater inputs to coastal areas. An additional question relates to growth-overfishing and resilience. Whereas in 1950, it would be common to see a large proportion of three- and four-yearold crabs in a given day’s catch, almost every crab caught in modern fisheries is less than two years old. In theory, the reduction in the number of cohorts would increase the susceptibility of a stock to a deleterious environmental shock. The degree to which this type of relationship is relevant to the Georgia fishery deserves investigation, as well as its applicability to other fisheries that are susceptible to environment-driven shocks. Acknowledgements Funding from a NOAA/NMFS - Sea Grant Fellowship in Marine Resource Economics (NOAA Grant# FLUNV47991) is greatly appreciated, along with an associated mentorship with Larry Perruso at the NMFS Southeast Fisheries Science Center. Assistance from Dorset Hurley and Pat Geer of the Georgia Department of Natural Resources was vital to this research. This paper has benefited from comments by attendees of IWREC 2014 in Washington, DC, and the EAERE 2013 meetings in Toulouse, as well as seminar participants at the University of Iceland, Clarkson University, and the University of Maryland. Any remaining errors are ours.

References Aguilar, R., Johnson, E., Hines, A., Kramer, M., & Goodison, M. (2008). Importance of

19

Blue Crab Life History for Stock Enhancement and Spatial Management of the Fishery in Chesapeake Bay. Reviews in Fisheries Science, 16(1), 117–124. Alber, M. (2002). A conceptual model of estuarine freshwater inflow management. Estuaries, 25(6), 1246–1261. Archambault, J. A., Wenner, E. L., & Whitaker, J. D. (1990). Life history and abundance of blue crab, callinectes sapidus rathbun, at Charleston Harbor, South Carolina. Bulletin of Marine Science, 46(1), 145–158. Arkema, K. K., Verutes, G. M., Wood, S. A., Clarke-Samuels, C., Rosado, S., Canto, M., … Guerry, A. D. (2015). Embedding ecosystem services in coastal planning leads to better outcomes for people and nature. Proceedings of the National Academy of Sciences, 112(24), 201406483. Barbier, E. B., Hacker, S. D., Kennedy, C., Koch, E. W., Stier, A. C., & Silliman, B. R. (2011). The value of estuarine and coastal ecosystem services. Ecological Modelling, 81(2), 169–193. Breed, A. G. (2001, May 6). South Parched by Historic Drought. Los Angeles Times. Retrieved from http://articles.latimes.com/2001/may/06/news/mn-59954 Conley, D. J., Paerl, H. W., Howarth, R. W., Boesch, D. F., Seitzinger, S. P., Havens, K. E., … Likens, G. E. (2009). Ecology. Controlling eutrophication: nitrogen and phosphorus. Science, 323(5917), 1014–5. Das, A., Justic, D., Inoue, M., Hoda, A., Huang, H., & Park, D. (2012). Impacts of Mississippi River diversions on salinity gradients in a deltaic Louisiana estuary: Ecological and management implications. Estuarine, Coastal and Shelf Science, 111, 17–26. Eggleston, D. B., Johnson, E. G., & Hightower, J. E. (2004). Population dynamics and stock assessment of the blue crab in North Carolina. Chapel Hill, NC: Final Report for Contracts 99-FEG-10 and 00-FEG-11 to the North Carolina Fishery Resource Grant Program, North Carolina Sea Grant, and the North Carolina Department of Environmental Health and Natural Resources, Division of Marine Fisheries. Evans, C. (1998). Conservation and management of the blue crab fishery in Georgia. Journal of Shellfish Research, 17(2), 451–458. Finnoff, D., & Tschirhart, J. (2011). Inserting Ecological Detail into Economic Analysis: Agricultural Nutrient Loading of an Estuary Fishery. Sustainability, 3(12), 1688– 1722. Fisher, A. C., Hanemann, W. M., & Keeler, A. G. (1991). Integrating Fishery and Water

20

Resource Management: A Biological Model of a California Salmon Fishery. Journal of Environmental Economics and Management, 20, 234–261. Frederick, K. d., Vandenberg, T., & Hanson, J. (1996). Economic values of freshwater in the United States (No. 97-03). Washington, D.C. Frizzelle, P. T. (1993). The Crabbers of Savannah: A Study of Defensive Behavior and its Role in Conflict Resolution and Resource Regulation. MS Thesis. Athens, GA: The University of Georgia. GDNR. (2004). Georgia Blue Crab Fishery Crisis, January 2004. Brunswick, GA: Georgia Department of Natural Resources, Coastal Resources Division. GDNR. (2008). Management Plan: Blue Crab, June 2008. Brunswick, GA: Georgia Department of Natural Resources, Coastal Resources Division. Gillson, J. (2011). Freshwater Flow and Fisheries Production in Estuarine and Coastal Systems: Where a Drop of Rain Is Not Lost. Reviews in Fisheries Science, 19(3), 168–186. Greene, W. H. (2003). Econometric Analysis (5th ed.). Upper Saddle River, NJ: Pearson Education. GWC. (2008). Georgia Comprehensive State-wide Water Management Plan. Atlanta, GA: Georgia Department of Natural Resources, Environmental Protection Division: Georgia Water Council. Hill, J., Fowler, D. L., & Moran, D. (1989). Species Profiles : Life Histories and Environmental Requirements of Coastal Fishes Fish and Wildlife Service: Blue Crab. US (Vol. 82). Washington, D.C.: US Fish and Wildlife Service. Hines, A. H. (2007). Ecology of juvenile and adult blue crabs. In V. S. Kennedy & L. E. Cronin (Eds.), the Blue Crab: Callinectes sapidus (p. 774). College Park, MD: Maryland Sea Grant. Huang, L., & Smith, M. D. (2011). Management of an annual fishery in the presence of ecological stress: The case of shrimp and hypoxia. Ecological Economics, 70(4), 688–697. Kennedy, C. J., & Cheong, S.-M. (2013). Lost ecosystem services as a measure of oil spill damages: a conceptual analysis of the importance of baselines. Journal of Environmental Management, 128, 43–51. Knowler, D., Barbier, E. B., & Strand, I. (2002). An open-access model of fisheries and nutrient enrichment in the Black Sea. Marine Resource Economics, 16(3), 195–218.

21

Lee, R., & Frischer, M. (2004). The decline of the blue crab - Changing weather patterns and a suffocating parasite may have reduced the numbers of this species along the Eastern seaboard. American Scientist, 92(6), 548–553. Lipcius, R. N., & Van Engel, W. A. (1990). Blue Crab Population Dynamics in Chesapeake Bay: Variation in Abundance (York River, 19721988) and StockRecruit Functions. Bulletin of Marine Science, 46(1), 180–194. Mallin, M. A., Paerl, H. W., Rudek, J., & Bates, P. W. (1993). Regulation of Estuarine Primary Production by Watershed Rainfall and River Flow. Marine Ecology Progress Series, 93(1-2), 199–203. Massey, D. M., Newbold, S. C., & Gentner, B. (2006). Valuing water quality changes using a bioeconomic model of a coastal recreational fishery. Journal of Environmental Economics and Management, 52(1), 482–500. McKenny, C. L., & Neff, J. M. (1979). Individual Effects and Interactions of Salinity, Temperature, and Zinc on Larval Development of the Grass Shrimp Palaemonetes pugio. I. Survival and Developmental Duration Through Metamorphosis. Marine Biology, 52, 177–188. Moksnes, P.-O., Lipcius, R. N., Pihl, L., & van Montfrans, J. (1997). Cannibal–prey dynamics in young juveniles and postlarvae of the blue crab. Journal of Experimental Marine Biology and Ecology, 215(2), 157–187. Ogburn, M. B., & Alber, M. (2006). An investigation of salt marsh dieback in Georgia using field transplants. Estuaries and Coasts, 29(1), 54–62. Olsen, S. B., Padma, T. V., & Richter, B. D. (2006). Managing Freshwater Inflows to Estuaries: A Methods Guide. Washington, D.C.: USAID, The Nature Conservancy, The Coastal Resource Center - University of Rhode Island. Parmenter, K. (2012). The effects of drought on the abundance of the blue crab, Callinectes sapidus, in the ACE Basin Nerr in South Carolina. PhD Thesis. Clemson, SC: Clemson University. Petrie, R. A., & Taylor, L. O. (2007). Estimating the Value of Water Use Permits: A Hedonic Approach Applied to Farmland in the Southeastern United States. Land Economics, 83(3), 302–318. Posey, M., Alphin, T., Harwell, H., & Allen, B. (2005). Importance of low salinity areas for juvenile blue crabs, Callinectes sapidus Rathbun, in river-dominated estuaries of southeastern United States. Journal of Experimental Marine Biology and Ecology, 319(1-2), 81–100.

22

Powell, G. L., Matsumoto, J., & Brock, D. A. (2002). Methods for Determining Minimum Freshwater Inflow Needs of Texas Bays and Estuaries. Estuaries, 25(6B), 1262–1274. Rhodes, A., Liption, D., & Shabman, L. (2001). A SocioEconomic Profile Of The Chesapeake Bay Commercial Blue Crab Fishery. Baltimore, MD: Bi-State Blue Crab Advisory Committee. Ricker, W. E. (1975). Computation and interpretation of biological statistics of fish populations. Bulletin of the Fisheries Research Board of Canada, 91, 1–382. Rogers, S. G., Arredondo, J. D., & Latham, S. N. (1990). Assessment of the effects of the environment on the Georgia blue crab stock. Brunswick, GA: Georgia Department of Natural Resources, Marine Fisheries Section. Silliman, B. R., & Bertness, M. D. (2002). A trophic cascade regulates salt marsh primary production. Proceedings of the National Academy of Sciences, 99(16), 10500. Smith, M. D. (2008). Bioeconometrics: empirical modeling of bioeconomic systems. Marine Resource Economics, 23(1), 1–23. Tang, Q. (1985). Modification of the ricker stock recruitment model to account for environmentally induced variation in recruitment with particular reference to the blue crab fishery in Chesapeake Bay. Fisheries Research, 3, 13–21. USGS. (2006). USGS Georgia Water Science Center. Retrieved from http://ga.water.usgs.gov/projects/projectwateruse.html Vörösmarty, C. J., McIntyre, P. B., Gessner, M. O., Dudgeon, D., Prusevich, A., Green, P., … Davies, P. M. (2010). Global threats to human water security and river biodiversity. Nature, 467(7315), 555–61. Wilber, D. (1994). The influence of Apalachicola River flows on blue crab, Callinectes sapidus, in north Florida. Fishery Bulletin, 92(1), 180–188. Wrona, A. (2004). Determining movement patterns and habitat use of blue crabs (Callinectes sapidus rathbun) in a Georgia saltmarsh estuary with the use of ultrasonic telemetry and a geographic information system (GIS). PhD Thesis. Athens, GA: The University of Georgia.

Wrona, A., Wiegert, R., & Bishop, T. (1995). Initial report of settlement patterns of brachyuran megalopae at Sapelo Island. Bulletin of Marine Science, 57(3), 807–820.

23

Young, R. A., & Loomis, J. B. (2014). Determining the Economic Value of Water: Concepts and Methods. London: Routledge. Zellner, A. (1962). An Efficient Method of Estimating Seemingly Unrelated Regressions and Tests for Aggregation Bias. Journal of the American Statistical Association, 57(298), 348–368.

24

Pounds (thousands)

Figure1

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

1950

1956

1962

1968

1974

Landings Value

Year

1980

1986

1992

1998

2004

2010

0

1000

2000

3000

4000

5000

6000

7000

8000

Value (thousands of year 2000 dollars)

)LJXUH

Figure3

Flow (ft3/s)

0

5000

10000

15000

20000

25000

30000

1960

1965

1970

1975

1980

Year

1985

1990

Altamaha Savannah

1995

2000

2005

2010

Figure4

Trips

Time t+1

Time t

Time t-4

Trips

Profits

Relative harvest

Juvenile abundance

C

Adult abundance

Escapement abundance

Adult abundance

B

Salinity

Adult abundance

A

Salinity

Empirical Approach

Salinity (practical salinity units)

)LJXUH

35

Ossabaw Sound

30 25 20

y = -4.5x + 54.8 R² = 0.79

15 10 4

5

6

7

8

9

10

Salinity (practical salinity units)

ln(flow) (ft3/s)

40

St. Andrews Sound

35 30 25

y = -3.1x + 47.5 R² = 0.69

20 15 10 3

4

5

6

7

8

9

10

7

8

9

Salinity (practical salinity units)

ln(flow) (ft3/s)

40

St. Marys Sound

35 30 25

y = -2.2x + 40.7 R² = 0.41

20 15 10 2

3

4

5

6

ln(flow) (ft3/s)

Figure6

Pounds

Adult abundance

0 2001

20

40

60

80

100

120

140

160

180

0 2001

50000

100000

150000

200000

250000

300000

350000

400000

2002

2003

2004

2003

2004

25% minimum flow

Observed

2002

25% minimum flow

Observed

2005

2005

2006

2006

Year

2007

Year

2007

2008

2008

2009

2009

2010

2010

2011

2011

2012

2012

2013

2013

Trips Juvenile abundance 0 2001

10

20

30

40

50

60

70

80

0 2001

100

200

300

400

500

600

700

2002

2002

2004

2003

2004

25% minimum flow

Observed

2003

25% minimum flow

Observed

2005

2005

2006

2006

2007

Year

2007

Year

2008

2008

2009

2009

2010

2010

2011

2011

2012

2012

2013

2013

Figure7

Pounds

Adult abundance

0 2001

10

20

30

40

50

60

70

80

90

2004

2003

2004

25% minimum flow

Observed

2003

25% minimum flow

Observed

2002

2002

0 2001

50000

100000

150000

200000

250000

2005

2005

2006

2006

Year

2007

Year

2007

2008

2008

2009

2009

2010

2010

2011

2011

2012

2012

2013

2013

Trips Juvenile abundance 0 2001

10

20

30

40

50

60

70

0 2001

100

200

300

400

500

600

700

800

2002

2002

2004

2003

2004

25% minimum flow

Observed

2003

25% minimum flow

Observed

2005

2005

2006

2006

2007

Year

2007

Year

2008

2008

2009

2009

2010

2010

2011

2011

2012

2012

2013

2013

Figure8

Pounds

Adult abundance

0 2001

10

20

30

40

50

60

70

80

90

2004

2003

2004

25% minimum flow

Observed

2003

25% minimum flow

Observed

2002

2002

0 2001

20000

40000

60000

80000

100000

120000

140000

160000

2005

2005

2006

2006

Year

2007

Year

2007

2008

2008

2009

2009

2010

2010

2011

2011

2012

2012

2013

2013

Trips Juvenile abundance 0 2001

10

20

30

40

50

60

70

0 2001

50

100

150

200

250

300

350

400

450

2002

2002

2004

2003

2004

25% minimum flow

Observed

2003

25% minimum flow

Observed

2005

2005

2006

2006

2007

Year

2007

Year

2008

2008

2009

2009

2010

2010

2011

2011

2012

2012

2013

2013

7DEOH

Variable*













































Description Average # adult crabs Average # juvenile crabs‡ Average salinity Average temperature Total harvest Total effort (# trips) Ex-vessel price per pound Variable cost per trip Profit Estimated fishing mortality Escapement

Source† GDNRa GDNRa GDNRa GDNRa GDNRb GDNRb GDNRb Rhodes Calculated Calculated Calculated

7DEOH

Variable

Obs

Ossabaw Sound (2001 - 2012) Mean Std. Dev. Min

Max

Obs

St. Mary's Sound (2001 - 2012) Mean Std. Dev. Min Max

Xt,s

48

18.76

30.11

0

166

48

11.18

17.14

0

83

Rt,s

48

10.82

13.26

0

52.5

48

7.48

10.24

0

48.5

St,s

48

24.58

4.72

13.78

32.94

48

29.67

3.20

23.38

37.55

Ht,s

48

129,129

75,744

4,042

279,944

48

57,213

25,483

6,493

142,997

Et,s

48

313.79

126.76

52

691

48

230.69

79.00

58

422

Max

Obs

Obs

Sapelo Sound (2001 - 2012) Mean Std. Dev. Min

St. Simon's Sound (2001 - 2012) Mean Std. Dev. Min Max

Xt,s

48

17.35

18.77

0

71

48

51.86

61.94

0

313

Rt,s

48

20.66

28.89

0

124.5

48

40.63

52.31

0

230.5

St,s

48

28.46

2.79

21.9

34.45

48

24.90

3.91

17.65

33.72

Ht,s

48

146,711

88,312

13,177

361,388

48

46,509

20,568

2,502

118,835

Et,s

48

410.65

147.98

58

774

48

193.60

80.56

10

444

St. Andrew's Sound (2001 - 2012) Mean Std. Dev. Min Max

Obs

Obs

Wassaw Sound (2001 - 2012) Mean Std. Dev. Min

Max

Xt,s

48

13.91

20.89

0

77

48

14.79

19.56

0

88.5

Rt,s

48

7.45

10.23

0

44.5

48

11.63

20.89

0

132.5

St,s

48

26.53

3.91

17.13

35.17

48

29.09

2.65

22.81

33.50

Ht,s

48

115,009

52,768

12,924

205,151

48

50,168

31,553

8,117

130,365

Et,s

48

460.90

151.76

131

710

48

108.23

46.05

16

208

7DEOH

Variables

2

Equation 3 4

5

Zt-1,s

0.1821*** (3.132)

-

-

-

Rt-1,s

0.4871*** (6.943)

-

-

-

Xt-4,s

-

-0.0297*** (-10.529)

-

-

ln(Xt,s)

-

-

0.1396*** (12.289)

-

-

-

1.2061*** (27.921)

-

-

-

-

0.0007*** (4.383)

St,s

-0.8025* (-1.821)

-0.1293*** (-4.662)

-

-

Tt,s

1.9520* (1.736)

-0.0680** (-2.553)

-

-

-

-0.0861*** (-2.962)

-

-

-

-0.0512 (-0.766)

-

-

Sapelo

-2.6819 (-0.485)

1.0448*** (2.903)

-0.2220*** (-3.580)

9.1354 (0.457)

St. Andrews

-1.4162 (-0.269)

0.4659 (1.389)

-0.4664*** (-7.295)

10.8005 (0.540)

St. Marys

-2.0291 (-0.356)

1.2813*** (3.265)

-0.2258*** (-3.645)

9.7331 (0.487)

St. Simons

12.2213** (2.229)

1.1393*** (3.384)

-0.4631*** (-6.923)

0.7870 (0.039)

Wassaw

1.0161 (0.181)

0.9155** (2.441)

0.4117*** (5.439)

-10.4186 (-0.521)

q2

-15.2086 (-1.178)

-0.1948 (-0.195)

-0.1637*** (-2.842)

-

q3

-32.0527* (-1.769)

-0.1321 (-0.092)

0.2971*** (5.060)

-

q4

2.7310 (0.324)

0.1373 (0.215)

0.3173*** (5.993)

-

0.1029 (0.237)

-0.0180 (-0.659)

0.0205*** (3.989)

-2.9901* (-1.789)

-207.9018 (-0.239)

42.2926 (0.770)

-36.7352*** (-3.576)

6005.3543* (1.790)

0.47 240.53

0.36 181.00

ln(Et,s)

(πt-1,s - πt-5,s)

St-4,s

Tt-4,s

year

constant

288

n r2 chi2

0.88 2225.96

0.07 23.48

7DEOH

Sound (river)

Flow (ft3/s) Flow (ft3/s) (2001-2012) (1960-2012)

25% min 3

flow (ft /s)

Salinity 25% min # binding (PSU) salinity (PSU) quarters

Mean SD Ossabaw (Ogeechee) Min Max

1,496 1,566 72 6,611

2,242 2,048 72 13,003

1,514 1,531 277 6,611

24.30 4.77 13.78 32.94

23.14 4.12 13.78 29.16

12 of 44

Mean St. Andrews SD (Satilla) Min Max

1,641 2,035 43 9,213

2,280 2,456 43 12,513

1,748 1,938 301 9,213

26.25 3.94 17.13 35.17

24.92 2.94 17.13 29.45

18 of 44

Mean SD St. Marys (St. Marys) Min Max

538 664 24 2,693

622 613 24 3,507

560 638 97 2,693

29.53 3.30 23.38 37.55

27.74 2.24 23.38 30.98

16 of 44

7DEOH

13,471,293 15,273,622 13.4%

44,430 45,605 2.6%

St. Marys

Combined 25% Min

% Dif

Actual

2,464,449 2,840,456 15.3%

10,254 10,472 2.1%

Actual 25% Min % Dif

St. Andrews

5,147,541 5,642,917 9.6%

20,385 20,718 1.6%

Actual 25% Min % Dif

5,859,303 6,790,249 15.9%

13,791 14,416 4.5%

Pounds

Actual 25% Min % Dif

Trips

Ossabaw

Sound

$12,181,820 $14,083,601 15.6%

$2,056,656 $2,401,202 16.8%

$4,681,558 $5,228,024 11.7%

$5,443,606 $6,454,375 18.6%

Revenue

Total

$4,890,814 $6,599,764 34.9%

$373,965 $682,783 82.6%

$1,336,359 $1,828,245 36.8%

$3,180,490 $4,088,736 28.6%

$1,708,950

$308,818

$491,886

$908,246

Profit

14.45 15.56 7.7%

11.49 12.46 8.4%

13.28 14.81 11.5%

18.59 19.42 4.5%

8.05 18.12 125.2%

7.03 17.79 152.9%

7.07 17.22 143.7%

10.03 19.34 92.8%

Adults Juveniles

$1.61

$1.94

$0.63

$6.97

Value of water

Average