Abstract
<p>The red snapper stock in the Gulf of Mexico has been classified as severely overfished, and overfishing continues, therefore several management actions have been imposed on the Gulf red snapper fishery, including closed seasons, quotas and minimum sizes. For this reason the red snapper is often referred to as one of the most intensively managed stocks in the United States. A recovery plan for red snapper, which includes a reduction of the shrimp bycatch mortality to 50% of the 2001-2003 level, had been established so that a 20% SPR would be attained by the year 2032. Due to the very long history of high exploitation levels in the red snapper fishery (>130 yrs) there was a need to assess whether the minimum size regulations were serving their intended purpose of protecting the spawning stock, reducing the probability of recruitment failure and increasing the yield from the stock. To meet this objective a good descriptor of red snapper growth was required in order to simulate the existent exploitation pattern of the gears involved in the fisheries. An extensive analysis of red snapper growth was conducted using data supplied by the National Marine Fisheries Service (NMFS) Panama City, Florida Laboratory. The data set (database 1 (DB1)) contained information on otolith ageing and the annuli measurements made along the dorsal side of the sulcus acousticus as well as total length measurements collected from 370 fish between age 2 and 17. Data for the previous growth studies on red snapper, which were collected from the directed fisheries, indicated an underrepresentation of several age categories of fish. Due to the selectivity of the gears a major portion of the age two fish and even possibly some segment of the age three fish (smaller, slowgrowing individuals) were not included in the data collection, therefore it is necessary to ensure that proportionality exists between the otolith and somatic growth. The results of the analysis here indicated that the back-calculated otolith radius-total length measurements were uncoupled for several age classes, meaning that the growth in total length was not proportional to the otolith growth in all age classes. This, therefore, indicated that the data was not suitable for providing a good descriptor of the growth of the red snapper when used in backcalculations of size-at-age. To resolve this issue it was necessary to evaluate the use of the annulus measurements made along the dorsal side of the sulcus acousticus (best for ageing procedures) for growth estimation through backcalculation procedures. It was apparent that otoliths grow asymmetrically and the sulcus acousticus radius is a partial measure along the proximal-distal axis. Following this observation, since otolith growth is at a maximum along the dorsoventral axis, it was proposed that radius measurements along the dorso-ventral axis and of the area of annuli along the dorso-ventral axis of the otolith would likely be more representative of the total otolith’s growth, which was expected to be directly proportional to somatic growth in length. To investigate these propositions additional data sets and otoliths were requested from the NMFS Panama City Laboratory. Data and otoliths from a total of 519 individual fish between 0 and 8 years old were received (database 2 (DB2)). This database, after adding the measurements made along the dorso-ventral axis showed that there was a stronger relationship between the total length and the dorso-ventral axes measurements than for the proximal-distal axes (R2 = 0.9596 vs R2 = 0.8884). More importantly, the analysis indicated that there was coupling of the total length-otolith radius data in all the represented age classes, which indicated that there was a proportional relationship between the total length and the otolith measurements. Both the dorso-ventral axis and area measurements proved to be more directly proportional to somatic growth than the sulcus acousticus measurements. The area measurements, however, were deemed to be too impractical for widespread use due to the length of time required to determine and accurately measure the area of each annulus within the transverse section of the otolith. The measurements made along the dorso-ventral axis were used for all subsequent analyses. Several back-calculation models were selected to generate growth function parameters for comparison purposes. These models included several multiple regression models, the age-effects model (Morita & Matsuishi, 2001), a log-power model (Ehrhardt, 1992) and the Modified Fry model (Vigliola & Meekan, 2009). The result of each model was tested by comparing the mean of the observed lengths-atage to the mean of the back-calculated lengths-at-age. It was found that all of the models used approximated the observed data as there was no significant difference between the observed mean lengths-at-age and the back-calculated mean-lengths-at-age at the 0.05% level. The von Bertalanffy growth function parameters for a 2-way multiple regression model using the data from DB2 gave values of L∞ = 1232.913 mm, k = 0.114 yr-1, and t0 = -0.353 yr. The 3-way multiple regression model gave results of L∞ = 1081.697 mm, k = 0.128 yr-1, and t0 = -0.158 yr. The parameters determined from the age-effects model were L∞ = 1109.627 mm, k = 0.145 yr-1, and t0 = -0.204 yr. Growth parameter values of L∞ = 927.558 mm, k = 0.179 yr-1, and t0 = 0.156 yr were generated from the power model while for the Modified Fry model the results were L∞ = 924.870 mm, k = 0.184 yr-1, and t0 = -0.087 yr. Next, how well the age structure of the samples represented the stock was investigated. It was found that the age two fish were underrepresented due to the selectivity of the fishing gears. Also, the larger and older fish in the size-at-age distributions were mostly absent, perhaps also due to fishing gear selectivity and the resulting exploitation pattern (Walter & Ingram, 2009). This problem was approached by treating the age two and the older age classes as containing missing data, which would suggest a slower growth rate for smaller fish and perhaps a larger maximum size for older fish. A Monte Carlo approach was used to generate experimental length-at-age data from the Modified-Fry model which would then be used to simulate an unbiased virgin red snapper length-at-age population structure based on the results of the Modified Fry model. The growth parameters that resulted for the unbiased population structure were L∞ = 1201.40 mm, k = 0.124 yr-1, and t0 = -0.177 yr. Finally, to evaluate the appropriateness of the use of minimum size in the management framework to recuperate the red snapper stock it was necessary to consider the life history, the fishing mortalities due to the shrimp fishery, to the discards below the minimum size and to the closed season discards, and the temporal distribution of the fish stock. The minimum size limits are responsible for the majority of the discards from the directed fisheries. The release mortality rates for these discards are estimated to range from 71-82% for the commercial fishery and 15-40% for the recreational fishery (NMFS, 2006). Although the commercial fishery has the higher discard mortality rate, the recreational fishery releases a substantially higher number of red snapper each year due to the very large number of participating individuals in that fishery. This is perhaps the primary reason the minimum size limits have not been effective in meeting the desired management goals. A new deterministic size-age equilibrium SPR model was developed that utilizes a transitional mortality schedule matrix to generate a yield per recruit, for the purpose of estimating the effectiveness of the minimum size to control exploitation in the directed fisheries. Several management scenarios, different minimum sizes under different levels of shrimp bycatch, were investigated that estimated the likelihood of red snapper recuperation to the management goal of 20% SPR by 2032. The results indicated that only in the complete absence of any shrimp fishery bycatch would the SPR approach 18.5%, still short of the goal, regardless of any minimum size selected. At the target 50% shrimp bycatch reduction level the SPR is only projected to increase to about 7.2%, regardless of the minimum size implemented. These results show the minimum size regulations to be ineffective in meeting their intended goal of protecting the spawning stock for red snapper in the Gulf of Mexico.</p>