After the start of the fishery in Galapagos, catches rose and during the “peak period” 1999 to 2005 ]]> Fig. 1), but continuously decreased thereafter. During the fishing season of 2008 only about 0.86 million individuals were caught, although a total allowable catch (TAC) of 1.3 million had been agreed upon by the Participatory Management System (PMS), which is in place in Galapagos since 1998. The great decrease in catches after 2002 is generally explained by strong over fishing, driven by a Chinese market, ]]> et al. 2008).
]]> Isostichopus fuscus is gonochoric, but does not display sexual dimorphism. Fifty percent of individuals from Galapagos attain sexual maturity at 23.6cm, although individuals of 16cm with developing gonads have been recorded (Toral-Granda 1996). Reproductive activity occurs throughout the year and is thought not to be affected by temperature in the Galapagos Islands (Toral-Granda 1996); while in Mexico, increased reproductive activity was observed in the warm water season (Herrero-Perezrul et al. 1999). Larvae are planktotrophic (Toral-Granda 1996) and transparent, and display indirect development (Hamel et al. 2003). Under aboratory conditions, metamorphosis and settlement of 1mm juveniles can occur 28 days after fertilization (Hamel et al. 2003). Herrero-Perezrul et al. (1999) obtained preliminary population parameters of I. fuscus from Mexico, ]]>
The starting fishery was centered ]]> Fig. 2), where population densities were highest. A pilot population study off Fernandina in 1993 yielded a mean density of 6.24ind./m2 (Aguilar et al. 1993), and later assessments recorded between 0.8-6.2ind./m2 in the Canal Bolivar area (Aguilar et al. 1993, Richmond & Martinez ]]> et al. 2005) yielded initial densities of adults (>16cm TL), between 0.27-0.40ind./m2 from 1994 to 1997, similar to those obtained in 1999 (Toral-Granda & Martinez 2004). Currently, densities have decreased by about an order of magnitude to 0.03ind./m2 (Murillo et al. 2008, this study).
A strong recruitment event of the I. fuscus population in Galapagos was registered in 2000 in the Canal Bolivar area (Murillo et al. 2002, Hearn et al. 2005), which ended in 2002.
]]>
The steadily decreasing catches of the sea cucumber fishery over the past decade (following the peak of 2002), are mainly the results of a chronic over fishing that has not allowed the stock recuperation (Hearn et al. 2005). The main reasons are an overcapacity of the fleet, a reactive instead of proactive management, and a ‘race for fish’ situation, the lack of control as well as difficult enforcement of the fishery regulations in this large ]]> negotiated by the co-management system in place (except for the boom year of 2002). Only in 2009 a limit reference point (LRP) of 11ind./100m2 was implemented and the ]]> 2) observed during historical surveys and characteristic for a stock in recuperation. Below this density interval the stock is considered to be in a critical stage, above in a healthy stage (Castrejon et al. 2008). These measures caused two problems. The ]]>
Our research goals were to: (1) optimize stock evaluation procedures by considering macrozone differences in density and stock distribution, and (2) to provide a scientific basis for the estimation of an annual quota. Specific ]]>
Material and methods
Estimation of fishing areas (macrozones): The design of the new monitoring plan was derived from information of past monitoring surveys carried out annually from 1999 to 2008. Two types of monitoring had been carried out, pre-fishing and post-fishing. Since the prefishing data were generally more complete and consistent (Vadas 2008), we decided to focus on those for the analysis. The monitoring effort of all years was aimed to cover the principal fishing areas as well as no-fishing areas. This resulted in very large areas and for several of them in wide confidence limits around the density estimates.
The first 19 macrozones for the islands of Fernandina, Isabela, Espanola, Floreana, San Cristobal and Santa Cruz were defined considering the principal sea cucumber fishing zones as derived from fisheries statistics of the PNG and CDF of the period 2001-2008. Due to insufficient financial resources and because fishermen insisted that four of the previously defined macrozones (east Santa Cruz, East and South-East Isabela and East Floreana) are not currently used as fishing zones, the macrozones surveyed in this study were reduced to 15.
Each macrozone was delimited using georeferenced catch data from the participative monitoring programme conducted during this period by the following procedure: bathymetric data compiled by Chadwick (1994) for the depth range from 0-1 500m depth, as well as from the CDF Geographical Information System (coastline CDF/Clirsen-TNG/GNPS) data from 0-39m were subjected to an ArcGIS Triangulated Irregular Network (TIN) interpolation routine, which creates circumcircles around each selected sample point and their intersections are connected to a network of non overlapping triangles, being as compacted as possible. Based on this interpolation a -30m isobath, which is the maximum depth where fishing can be carried out, was created. With the isobath line as the outer and the coastline as the inner boundary, the area of each polygon (m2) was mapped. A histogram was created to describe the frequencies of points per depth strata and an analysis was done to describe mean distances between points by using nearest neighbor statistics (ArcGIS and Surfer). Only 50% of the estimated area was considered as suitable sea cucumber fishing area and habitat, given that approximately 20% has unsuitable sandy bottoms and another 30% is too shallow (<5m intertidal waters), or uninhabitable steep slopes (Banks et al. 2006).
Determination of transect number per macrozone: In order to estimate the required number of transects per macrozone, we proceeded as follows: first, means and standard deviations of stock densities obtained during past surveys were calculated for each zone. Then, 95% confidence intervals around the mean stock densities were estimated for each zone and 95% confidence intervals around the mean stock densities were calculated; second, the homogeneity/degree of patchiness of sea cucumber distribution per macrozone as reflected by those statistics was then used to calculate the minimum number of replicates necessary per macrozone to achieve a precisión of ±25% applying the following formula:
with n=minimum number of replicates required to achieve a precision of ±25% around our estímate of mean density; t=value of the t distribution (student’s t-test) for p<0.05; Av=annual average of the sea cucumber density per macrozone and SD = standard deviation.
Using the above formula, differences in spatial distribution and stock density per macrozone resulted in different numbers of replicates required for the different areas, with more replicates for areas of higher patchiness, and lower numbers where the population appears more evenly distributed. Based on this procedure population size estimates are considered substantially more precise than if the replicate number was just proportional to the size of the area.
Sampling: The type of sample unit used was a circular transect with a radius of 5.6m which covers an area of 100m2. This has been used in monitoring activities before and has proven to allow for the collection of available sea cucumbers with low error (Hearn et al. 2005). Moreover, using this sample unit unbiased comparisons with previous surveys became possible. The exact location of each transects in each macrozone was decided jointly between fishermen and scientists with the aim to adequately cover each area. The sampling sites were distributed in two depth strata: shallow (<15m) and deep (>15m), with the shallowest and deepest waters around five and 25m respectively.
The stock size per macrozone was estimated extrapolating the mean density value (ind./m2) to the corresponding total area. Then, the stock sizes of each macrozone were added up to estimate the stock for the whole fishing area.
The mean density value within the West Isabela macrozone was then compared with the sea cucumber reference points defined in the Fishery Chapter of the Management Plan for the Galapagos Marine Reserve (GMR) (Castrejon et al. 2008, Comision Tecnica Pesquera de la Junta de Manejo Participativo 2009). According to these reference points a healthy, abundant stock should have a West Isabela macrozone density >21ind./100m2, while densities of a stock in recuperation should range from 11-20.9 ind./100m2. If densities are <11ind./100m2, the stock is closed for fishing.
Precision of density estimates and density comparisons between depth strata: Since the density values of each macrozone proved to be not normally distributed, a bootstrap resampling routine (Efron 1981) was applied, which consisted in a random resampling (1 000x) of the data matrix for each macrozone. This yielded 1 000 normally distributed mean density values and allowed for the computation of the standard deviation and coefficient of variation (CV%) around the mean of these values.
In order to test for depth strata differences in sea cucumber densities, all density measurements taken in each depth stratum (shallow and deep), were considered for the calculation of overall means per depth strata. The resulting mean densities were bootstrapped for each depth stratum and compared using a t-test of means. For comparative purposes also a non-parametric Mann-Whitney U test for the medians was applied. We repeated this analysis by combining the information of sea cucumber density per depth strata for all those years where sufficient samples had been taken (1999-2002 and 2006-2009).
Calculating the catch quota. Annual stock production can be calculated by multiplying mean stock size by the rate of total production (P/B=Z) which is the sum of natural (M) and fisheries (F) mortalities. P/B tends to increase with fishing pressure while the stock biomass decreases. Gulland (1971) and Garcia & Le Reste (1981) proposed to approximate the maximum sustainable yield (MSY) of a stock from an estimate of its virgin biomass (B∞) and the rate of natural mortality (M). They propose: MSY = X* M*B∞ (With X=0.5 in the formula of Gulland and in the range of 0.32-0.44 in the formula of Garcia and Le Reste respectively). If a resource is heavily exploited and virgin stock levels can not be determined this formula should not be applied. Cadima (in Troadec 1977) proposed to use the following formula instead:
]]>
This formula is identical to MSY=M*B, if F=M at an exploitation rate of E=0.5.
Garcia et al. (1989) pointed out that Cadima’s formula only gives unbiased estimates of MSY if the stock is virgin and Z=M (when it is identical to Gullands formula mentioned above) or when the stock happens to be exploited at the level of MSY already (at B∞/2 and an exploitation rate of 0.5).
Considering this reasoning of the above authors and assuming that the current stock level of the Galapagos Sea Cucumber is far below B∞/2 it was decided to use a lower (precautionary) exploitation rate of E=0.3 when applying the Cadima formula. The M- value (0.17/year) of Hearn et al. (2005) was used, which was derived from a stock depletion model. This value was based on a rich data source and appears as a sound estimate in the right order of magnitude, when compared to other slow growing marine invertebrates reported in the literature (Brey 2001). A value double as high for this species in Mexico (M= 0.354/year) was proposed by Reyes-Bonilla & Herrero-Perézrul (2003), who, however used empirical formulas for its derivation, which were not established for holothurians.
Considering this M value of 0.17 the F value was calculated corresponding to an exploitation rate of E=0.3 by calculating:
By inserting the F and M values in the Cadima formula (above), we arrived at:
The annual quota here proposed is thus 12.2% of the standing stock.
Results
Estimates of macrozone areas and number of transects per macrozone: Figure 2 shows the distribution of sea cucumber fishing activities in the archipelago for the years 2001 and 2008 and the 15 macrozones using these data. Table 1 gives their extension and the numbers of replicates per macrozone required for a mean stock density estimates <±25% precision. It also provides the density estimates and quota per macrozone derived from the monitoring survey of 2009. The overall estimate of fishable area of all macrozones combined is 124.8km2.
The population monitoring was carried out in the last two weeks of May 2009 by fishermen, National Park and CDF staff coordinated and financed mainly by the National Park. San Cristobal is the island with the largest fishing area (52km2), followed by Isabela (33km2) and Santa Cruz (20.4km2). Floreana, Espanola and Fernandina combined only represent 20km2. While the coastline of Isabela where fishing takes place is much larger than the one of San Cristobal, the fishing area is smaller due to the very steep slope bringing the -30m isobath much closer to the coastline in Isabela than in San Cristobal. It is interesting to see that replicate numbers vary greatly between zones with Espanola requiring the highest number per area (49/5km2), while Santa Cruz requires only 31 transects in 20.5km2.
Densities and catch quota for the season 2009: Mean density for West-Isabela (3.72ind./100m2) is far below the critical value for the opening of the fishery (11ind./100m2) and was de lowest ever recorded in the last decade (Fig. 3). So, following the decision rule in place, it was recommended not to open the fishery for the year (2009). This recommendation was thereafter accepted by the co-management body (Interinstitutional Management Authority, IMA). The densities of all macrozones with the coefficient of variation around the mean (CV, %) are given in Table 1. If our rationale for a quota (TAC) estimate based on the stock size (4.9 million individuals) derived from the overall mean density (0.039ind./m2) and our combined area estimate (124.8km2) had been followed, the TAC would have been 59 889 individuals for the 2009 fishing season for the whole archipelago.
]]>
Figure 4 shows the mean densities (left) and the proportion of small (<20cm) sea cucumbers in the samples (right). Floreana, Isabela and Fernandina show a steady density decrease over the past years, while Santa Cruz has remained quite constant over the past six years. The density value of San Cristobal in 2009 was surprisingly the second highest since 1999.
The proportion of juveniles (recruits) in the stock has steadily decreased for the Western islands Isabela and Fernandina and is the lowest ever recorded in Espanola and Floreana. Santa Cruz and Cristobal revealed a slight increase over the last three years.
The 2009 monitoring data for the two depth strata showed higher densities in the shallow stratum compared to the deep stratum (4.5 compared to 3.2ind./100m2 respectively). Statistically, this difference was significant if the non-parametric Mann-Whitney U test of the median density values was used or if the mean values using the bootstrapping routine were compared (t-test, p<0.05). The density difference between depth strata was even more pronounced, when the data of eight years was combined (8.7 versus 17.9ind./100m2 for deep and shallow strata respectively).
Discussion
This study is the first in trying to estímate the fishing area of the sea cucumber stock of Galapagos, which seemed imperative if overall stock size was to be approximated for quota calculations. Since the topography and slope of the sea bottom varies substantially between macrozones, and available bathymetric data points were not covering all areas uniformly, the mapping of the polygons between the coastline and 30m isobath using ArcGIS required an interpolation procedure. The TIN interpolation used to define the -30m isobath for our 15 macrozones was supported by 8 615 geographical data points over the depth range from 0-1 500m, of which a large proportion (3 228 points, 37.5%) were situated within the macrozone depth limits. Based on a nearest neighbor analysis, a separation value for the mean distance between points was found equal to 0.0031589 with a standard deviation of 0.0054607 (degrees).
While we believe that our first fishing area estimate is a good approximation, it can definitely be improved through a more detailed bathymetric survey of the coastlines of the Galapagos Archipelago.
The assumption was then made that of the estimated area 20% is unsuitable Sandy bottom habitat and 30% too shallow (<5m) intertidal water or comprised by too deep waters ]]>
]]> Only 383 out of 547 planned transects (70%) could be carried out during the 2009 pre-fishery monitoring due to financial constraints. However, this transect number is the highest ever used in the Galapagos sea cucumber pre-fishery monitoring (the second highest was 303 transects taken in the year 2008). The precision around the mean densities differs between macrozones with a coefficient of variation (CV) ranging from 12.2% for Eastern Espanola (n=27/a=2.1km2) to 81.2% for Southern Fernandina ]]> 2). These differences can be attributed to the number of transects and the degree of patchiness of the sea cucumbers in each zone. When the transect data are integrated for each island, the respective density estimate is greatly improved with CVs always <25%. The two islands that had the highest transect numbers (Isabela with 125 and San Cristobal with 97) had the lowest CV around the mean density estimate (11.6% and 13.3% respectively). The density estímate for all transects combined (3.94ind./100m2,CV=6.8%) can be considered of a very good ]]>
Of all macrozones sampled, none ]]> 2 to open the fishery, which points to a very critical state of the stock in all macrozones. Several factors may explain this situation. During years 1999-2002 the seasonal fishery was opened without quota limits, and the stock was depleted to extreme low densities, not keeping a large enough residual biomass for the next season. In years, when catch quota were determined through the co-management process, these were often too high, representing too large a part of the available stock. This may be illustrated by the example of the year 2004, when a mean density of 7ind./100m2 ]]>
The density and recruit fraction ]]> I. fuscus was first published by Herrero-Perézrul et al. (1999). Santa Cruz is the only island, where the fraction of small specimens <20cm seem to have increased over the past years 2008 and 2009. However, since mean density has not increased in parallel, this relative increase in small specimens is rather a sign of heavy depletion of larger adults than of a successful recruitment. In Espanola a density increase during the year 2007 was paralleled by an increased fraction of small specimens, which may suggest a small ]]>
In this report, an empirical relationship between the pre-fishery (log) density obtained through the annual monitoring surveys and the catches obtained later in the same year has been assembled. While there is a great scatter around the regression line, possibly partly due to differences in monitoring precision between years (see discussion above), the figure shown suggests a significant relationship between both variables. The quota ]]>
]]> The often heard statement by the fishing sector that sea cucumber densities are higher in deeper waters giving the stock a strength in reserve if fishing pressure in shallower waters is high, could clearly be shown to be wrong for the year 2009, and also for the combined data of the eight year’s period. This is an important finding of our study since it removes the basis for the argument that a large portion of the stock is out there in deeper waters where it can not be caught.
]]>
We believe that the monitoring and management strategy of the Galapagos sea cucumber stock that has been presented in this paper is an important step towards a sustainable fishery of this species. It is the first time that the attempt was made to estimate the size of the whole fishing area and of the fishable stock. Moreover, the monitoring was adapted to specific conditions (size and sea cucumber patchiness) of each macrozone, and the quota was made a fraction of the stock size (12.2%), which makes it adaptive to ]]>
As a future step it may be considered to look for spatially-explicit tools for management of the Galapagos sea cucumber such as the implementation of territorial use rights (Defeo & Castilla 2005) or rotational closures of different areas from year to year. The first seems difficult in Galapagos because there are more fishermen with licenses than active fishermen and stock ]]>
Acknowledgments
]]> We want to thank all participants of the monitoring survey of 2009: H. Reyes, J. Paredes, G. Vasquez, M. Villalta, L. Garcia, W. Fuertes, O. Ricaurte, J. Garcia, Y. Llerena, Y. Mascarell, M. Ortega, J.L. Ballesteros, J. Pilamunga, P. Espinoza, W. Bran, J. Araujo, L. Lozano, F. Lucas, F. Parrales, Y. Araujo, A. Yamuca, F. Velez, J. Camacho, C. Lozano, I. Maffare, P. Tipan, L. Bonilla, G. Gil, J. Moreno, J. Delgado, S. Clarke, J.P. Tiernan, J.C. Murillo, S. Cruz, the National Park of Galapagos and the WWF. This publication is contribution number 2048 of the Charles Darwin ]]>