Japanese Hatchery-based Stock Enhancement:
Lessons for the Chesapeake Bay Blue Crab

Three Case Studies

To clarify Japanese attempts at hatchery-based enhancement for P. trituberculatus, three case studies are presented here for systems visited during July 2001. Another limited study is also presented for the portunid crab, Scylla tranquebarica, which presented a novel case and method in which hatchery contribution was evaluated. The case studies represent a range of ecosystems, fisheries and methods used to evaluate hatchery-based enhancement (Table 1, Figure 1). The difference in sizes between systems offers an opportunity to improve our understanding of the importance of scale in assessing the effect of hatchery-based enhancement programs for P. trituberculatus. Recent research in small systems in Japan has supported the view that hatchery releases may be contributing to swimming crab fisheries that are depressed (Ariyama 2000, Okamoto unpubl ms, Kochi Fisheries Coop. Assoc.). Evidence for successful stock enhancement (i.e., increased population productivity) in all three systems is lacking, and in each of the systems fishing effort remained very high and habitats continued to decline.

Small Enclosed System: Hamana Lake

Hamana Lake is a c. 10,000 hectare coastal lagoon with a single inlet to ocean waters located in Shizuoka Prefecture (Figure 7). Principal nursery and feeding areas are associated with extensive eel grass (Zostera) beds within the lagoon; some spawning occurs in this region as well, but most spawning is believed to occur following an emigration out of the narrow inlet into nearshore coastal waters of the Pacific Ocean (Okamoto 2001). Yield of P. trituberculatus from the entire system (including adjacent shallow coastal ocean fisheries) is 14 metric tons (t) (Table 1). Harvests of swimming crab occur almost exclusively through "set-nets," a type of pound net. The wings of the set nets are positioned strategically to intercept fishes and crustaceans throughout the lagoon and inlet areas. Because continual harvest by nets in inlet areas could result in overexploitation, there is a regulation to lift nets out of the water for a 12-hr period (3 a.m. to 3 p.m.) each day. Set nets capture a diverse assemblage of species, most of which are brought to market (undersized fishes and crustaceans are released). The minimum size for legal P. trituberculatus is 120 mm CW.

Although P. trituberculatus is the most important crab harvested (~$170,000 per year), nearly a dozen other species of crabs are landed and sold. In 1984, a large decline in yield of P. trituberculatus occurred (Figure 8), thought to be the result of a shift in hydraulic conditions that no longer favored transport of coastal-spawned larvae and juveniles through the inlet (Okamoto 2001). In addition, increased shoreline resort and residential development, combined with low flushing rates in some parts of the lagoon, has resulted in increased persistence of hypoxic events that may have reduced system productivity. A program of releasing hatchery-produced crabs was initiated in an effort to compensate for the nearly 10-fold decline in recent harvest levels. Since 1990, approximately 1.4 x 106 C-4 crabs have been released annually into Hamana Lake. During the past decade, there has been no corresponding rise in landings with increased numbers of hatchery-released juveniles.

Hamana Lake is the only system for which fishery contribution rate has been examined through mark-recapture methods, one of the most robust approaches for evaluating such rates (see sidebar, Measurement of Hatchery Contribution). In June 1998 Dr. K. Okamoto, a Shizuoka Prefecture scientist, released 3,300 C-4 juveniles injected with micro-wire tags into the primary nursery area of Hamana Lake (Okamoto 2001; unpubl. ms). Over the next four months, he undertook a sampling routine of visiting three principal fish markets in regions adjacent to the site of release. Recaptures only occurred for the Washizu Fishing Cooperative, which was immediately adjacent to the site of release (Figure 7). With a sampling rate (number of crabs examined/ number of crabs harvested) of 29%, Okamoto observed 13 crabs that contained micro-wire tags. Based upon the number of crabs examined for tags and tag retention rate (90%), Okamoto calculated a harvest recapture rate (recapture in fishery/number released) of 1.2%.

A second method more commonly used to evaluate harvest recapture rates is the cohort method, whereby released juveniles occur as a discrete early juvenile cohort of similarly sized crabs. This cohort can sometimes be segregated from wild cohorts through length frequency analysis for 12 to 18 months after release. The cohort approach is supported by early spawn dates of hatchery produced crabs that tend to be 1-2 months earlier than most wild cohorts, and the relatively rapid and consistent growth rates exhibited by both wild and hatchery cohorts (in comparison to blue crabs) (Figure 9). The approach is particularly valuable when there is a single spring season of hatchery production and early summer release, rather than several cycles of production and release during the entire summer. The former situation exists for Hamana Lake, facilitating cohort identification. Okamoto used the lengths of recaptured crabs to verify his cohort interpretations (Figure 10). In general, lengths of recaptured crabs were contained within modes attributed to hatchery releases. Using the cohort analysis approach during the period 1992-1998, Okamoto predicted annual harvest recapture rates between 0.3% and 2.2%. Based upon a mean harvest recapture rate of 0.85%, then fishery contribution rate was estimated as 18% (see Table 1, harvest recapture rate, (0.0085) * release number (1.4 106) / harvest number (67,000)).

The small number of returned micro-wire tags in Okamoto's study conveys considerable uncertainty to the study. Micro-wire tags are sometimes difficult to detect (detection occurs through use of an electronic magnetic "wand"), and if only 6 additional tagged crabs were undetected, then the estimate of the contribution to the fishery would increase by ca. 50%. Still, the cohort and tagging analyses were in good agreement with each other, suggesting that contribution rates were moderate in Hamana Lake.

Heavily Impacted System: Osaka Bay

Osaka Bay (Figure 12) is a highly urbanized and industrialized system heavily impacted by continuing development of shorelines, most recently in association with the construction of the Kansai International Airport. Principal nursery grounds occur in shallow sub-tidal regions on the eastern and northern parts of the bay, including Osaka Harbor itself (Ariyama 2000). Summertime hypoxia (< 1 mg/L dissolved oxygen) results in seasonal absence of juvenile crabs in the Harbor that are otherwise relatively abundant during early summer and fall/ winter months (Ariyama 2000). Spawning occurs throughout Osaka Bay during summer months. The principal gear used is the Ishigeta dredge (Figure 13), a small 1.5 m wide trawl that contains a toothed dredge, similar to a Virginia winter dredge (Table 1). A small fishing vessel tows four of these simultaneously, and trawls bring up a diverse assemblage of shellfish, fishes and crustaceans. Other gears that land swimming crabs include set nets and other trawl types. The exploitation rates for the Ishigeta dredge is estimated to be extremely high (see below) and as opposed to other prefectures, no size limit is enforced, although a minimum size of 120 mm CW is encouraged. A visit to a retail market confirmed that crabs <100 mm CW are commonly harvested and marketed. The values for swimming crabs in Osaka region ($25/kg) seem to be slightly higher than other regions in Japan (Table 1), resulting in an important urban fishery (Ariyama 2000). During 1990-1998, average annual fishery yield and value were 57 t and ~$1.4 million, respectively. The fishery underwent two abrupt cycles of decline in the 1960s and 1980s (Figure 14), which are believed to be the result of loss of sandy nursery habitat through shoreline development (Figure 15), poor water quality, and perhaps increased exploitation (Ariyama 2000). Since 1992, Osaka Prefecture has released an average of c. 0.7 x 106 C-3 - C-4 juveniles each year to mitigate against reduced harvests (Ariyama 2000; Figure 14). Landings data give no indication that hatchery releases have resulted in improved fishery yields.

Figure 13. Ishigeta dredge diagram.The dredge is the principal gear used in Osaka Bay and Okayama Prefecture fisheries. In Hamana Lake, crabs are taken with a set net (not shown). Diagram from Ariyama (2000).

In the early 1990s, Dr. H. Ariyama, an Osaka Prefecture scientist, evaluated secondary rearing and rates of post-release survival, and conducted extensive research on fishery contribution. As discussed previously, this research showed that C-1 - C-3 stage crabs were highly vulnerable to predation and cannibalism in the field due to their pelagic behavior. He concluded that semi-natural secondary rearing systems should favor post-release survival. In 1990, Ariyama carefully reared C-1 - C-4 juveniles in net pens and then released 246,000 C4 crabs in a semi-contained region of sandy beach habitat. Over the following two months, he intensively sampled this and adjacent regions to monitor dispersal. From this release trial, Ariyama observed that juvenile crabs began dispersing into the adjacent regions one month after release (C-7 - C-9 crab instars), and by 6 weeks (C-9 - C-10 instars; c. 67 mm CW) most released crabs had dispersed from the region into which they were released. Ariyama measured a 66% decline in abundance in the release region and adjacent regions over the first ~4 weeks (Figure 16). Assuming that net dispersal into (by natural crabs) and out of (by released crabs) the study area was zero, Ariyama used 34% survival as a benchmark survival rate for released C-4 crabs to harvestable size (c. 80-90 mm CW). This study also confirmed that released crabs could grow rapidly and enter fisheries by late summer/early fall, during their first year of life.

Figure 14. Osaka Prefecture landings and C-1 and C-4 releases of P. trituberculatus during the past four decades. Note fishery crashes in 1959 and 1980. Data from Ariyana (2000).

Figure 15. Five-year average beach habitat perimeter (left panel) and 5-year landings of P. trituberculatus (right panel) in Osaka Prefecture during the past four decades. Note 4-5 fold decline in both beach habitat and landings over this period. Data from Ariyama (2000).

To estimate harvest recapture rate, Ariyama (2000) used a modeling approach (see sidebar, Measurement of Hatchery Contribution), rather than direct evidence of hatchery crabs taken in the catch. He coupled his post-release survival rate (34%) with a harvest model for P. trituberculatus in Osaka Bay. The post-release survival rate represented the fraction of released crabs that entered the fishery. The harvest model represented the fraction of hatchery crabs that were subsequently caught. The harvest model was based upon the standard yield equation

C = R[(F/(F + M)) (1 - e^{-(F + M)})]       (1)

where C is yield in the fishery, R is the abundance of the hatchery cohort at entry into the fishery, F is instantaneous mortality due to fishing, and M is instantaneous mortality due to natural causes (Gulland 1983). The time step over which mortality rates are tallied can vary according to what data are available just as long as instantaneous rates correspond to that time step. Ariyama chose a monthly time step for his model, which permitted him to track the hatchery cohort through seasonal changes in fishing effort. Natural mortality (M) was available from a previous tagging study (Kitada and Shiota 1990). To estimate F, Ariyama first determined the daily efficiency (q) with which an average fishing vessel extracted individuals from the population (Delury method; Hilborn and Walters 1992). The product of this coefficient and effort (number of fishing vessels x days) provided monthly estimates of F.

Using estimates of F and M, Ariyama calculated monthly yields for the 1990 cohort of 84,000 hatchery crabs expected to recruit into the fishery at the beginning of September. Monthly natural mortality was estimated at 4.6%, excluding months December to April, when mortality was assumed to be negligible. Monthly fishing mortality rates were estimated to range from 19% (January) to 46% (August), which suggests an exploitation rate considerably higher than that observed in other systems (e.g., on an annual basis, >90% of crabs in Osaka Bay are harvested compared with c. ~60% annual removal rates for Chesapeake Bay blue crabs). The harvest model predicted that 76,324 of the released crabs were harvested (91% harvest recapture rate) during their first year in the fishery; another 7014 crabs died of natural causes, and the remaining 662 crabs persisted into the next year of the fishery. Ariyama (2000) predicted that most released crabs (65%) were harvested within three months after entering the fishery. The high exploitation rates translated into harvest recapture rates of 31.3%. This in turn resulted in a relatively high fishery contribution of 59% for 1990 landings. See Table 1, harvest recapture rate, (0.313) * release number (0.7 x 106) / harvest number (372,000).

Osaka Prefecture releases C-3 - C-4 crabs throughout the summer, which precluded a size-based analysis of hatchery cohorts similar to those conducted for Hamana Lake. Currently, there is an active research program at Osaka Fisheries Experimental Station to develop tagging methods to support estimates of the hatchery contribution rate for P. trituberculatus and other crustaceans (Ariyama et al. 2001). Such studies are quite important because they will provide initial physical evidence of hatchery-produced crabs in the fishery, and thereby test Ariyama's post-release survival and harvest models.

Larger, Open System: Okayama Prefecture (Seto Inland Sea)

Coastal waters of Okayama Prefecture (Figure 17) occupy a relatively small segment of the Seto Inland Sea1 between Honshu and Shikoko Islands, but support a larger P. trituberculatus resource (145 t) than Hamana Lake or Osaka Bay (Table 1). Similar to Osaka Bay, the principal gear used is the Ishigeta trawl. Since 1990, Okayama Prefecture has released ~ 1 X 106 C-4 juveniles each year (Figure 18; data on releases after 1993 are not shown but have continued at a nearly constant level during most of the last decade [Fukuda, pers. comm.]). Higher releases of C-1 juveniles occurred prior to 1991 and were apparently coincident with higher landings, but prefectural scientists believed that any correlation was unrelated to hatchery releases because large releases of C-1 instars were unlikely to result in improvements of catch (Fukuda and Karakawa pers. comm; also, see above references to experimental work by Karakawa 1997 and Ariyama 2000). The correlation in patterns between release number and catch is mostly driven by three high release years (1981-1983), where more than 2 million C-1 crabs were released. Particularly curious is the cycle of hatchery releases. Typically, hatchery releases are relatively stable, determined by careful planning by the prefecture. Japanese scientists could offer no explanation for the apparent correlation between historical landings and releases of C-1 juveniles. During this past decade, landings have undergone an oscillation and are now at low levels (88 t in 1999). Release levels have been maintained at approximately 1 million C-4 juveniles during this period (Fukuda, pers. comm).

Secondary rearing in Okayama occurs in 0.8 ha coastal impoundments that are also used for secondary rearing of Japanese prawn and juvenile marine fishes. The impoundments are large excavations in coastal areas, contained in part by sea walls. The Yorishima Secondary Rearing Center was constructed at a cost of $2.4 million (Ariyama 2001). Sand bottoms are thought to allow juveniles to acclimate to natural conditions and reduce cannibalism. Harvesting C-4 - C-5 crabs after 15 days of rearing from the C-1 instar stage is accomplished by tidal flow out of the impoundment and a catchment basin. Three days of tidal out-flow are required to collect all crabs. Although expensive, Ariyama and other scientists interviewed believed that this method of secondary rearing was optimal because it substantially reduced cannibalism and favored post-release survival.

Dr. J. Karakawa (Okayama Prefectural Scientist) has used the size-based cohort method to estimate harvest recapture rates. This procedure is complicated and uncertain because two large releases occur each year that must be separated from each other and from wild cohorts. In an effort to verify the identification of hatchery cohorts, Karakawa developed an index of anomalous chela (claw) symmetry that was associated with loss of this limb during the hatchery collection of sC-4 crabs, and subsequent regeneration. The anomaly was observed in approximately 6% of landed crabs and this rate was attributed to hatchery individuals (Karakawa 2001). The anomaly was frequently observed to coincide with hatchery cohorts identified through length frequency analysis, although the anomaly appeared to occur at equal frequency for all sized crabs for any monthly sample (Figure 19). Further, the rate of natural chelae loss (autotomy) is not reported quantitatively but stated to be negligible (Karakawa pers. comm.). Autotomy rates observed for C. sapidus in the Chesapeake may be much higher (ca. 25%; Smith and Hines 1991). Based upon the 5.7% harvest recapture rate (measured using the cohort approach), the fishery contribution rate for 1998 was 9% (harvest recapture rate (1998: 0.057) * release number (1998: 1.1 106) / harvest number (1998: 716,000)).

To evaluate migration and habitat use patterns by hatchery-produced and wild P. trituberculatus, Okayama Prefectural scientists recently released marked juvenile crabs (~80 mm CW) at two sites and recaptured them over a two-year period (Karakawa 2001). Marked crabs initially occurred in shallow regions near the release site, then dispersed to deeper areas offshore to over-winter, and then returned to shallow nearshore areas to spawn the next year. These behaviors are similar to those of wild P. trituberculatus (Karakawa 1999). Two years after release, the scientific monitoring program recaptured 20-30% of released crabs (rates varied by release group). Importantly, all tagged crabs were recaptured within 20 km of the release sites; the majority were recaptured within 5 km of release sites.

Last modified October 07, 2002

Accessibility Statement

Maryland Sea Grant Publication Number UM-SG-TS-2002-02 (September 2002)
Cover photo from Japan Sea Farming Association

Blue Crabs in the Chesapeake
Blue Crab Stock Enhancement Lessons:  Contents • Acknowledgments
  Preface • Report Highlights • Executive Summary • Introduction
  World Fisheries for Swimming Crabs • Japanese Sea Ranching Program
  Hatchery Contributions • Fishing Cooperatives • Genetics & Survivorship
  Three Case Studies • Summary of Case Studies
  Implications for Chesapeake Bay Blue Crab Enhancements
  Conclusions • References • Download the PDF version

http://www.mdsg.umd.edu/crabs/stock_enhance/studies.html

For more information, report problems or provide comments, please contact webmaster@mdsg.umd.edu