Chinese Journal of Oceanology and Limnology   2017, Vol. 35 issue(4): 894-901     PDF       
http://dx.doi.org/10.1007/s00343-017-6045-1
Institute of Oceanology, Chinese Academy of Sciences
0

Article Information

WANG Haijun(王海军), LIANG Xiaomin(梁小民), WANG Hongzhu(王洪铸)
Sustainable fisheries in shallow lakes: an independent empirical test of the Chinese mitten crab yield model
Chinese Journal of Oceanology and Limnology, 35(4): 894-901
http://dx.doi.org/10.1007/s00343-017-6045-1

Article History

Received Mar. 2, 2016
accepted in principle May. 11, 2016
accepted for publication Jun. 6, 2016
Sustainable fisheries in shallow lakes: an independent empirical test of the Chinese mitten crab yield model
WANG Haijun(王海军), LIANG Xiaomin(梁小民), WANG Hongzhu(王洪铸)        
State Key Laboratory of Freshwater Ecology and Biotechnology, Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan 430072, China
ABSTRACT: Next to excessive nutrient loading, intensive aquaculture is one of the major anthropogenic impacts threatening lake ecosystems. In China, particularly in the shallow lakes of mid-lower Changjiang (Yangtze) River, continuous overstocking of the Chinese mitten crab (Eriocheir sinensis) could deteriorate water quality and exhaust natural resources. A series of crab yield models and a general optimum-stocking rate model have been established, which seek to benefit both crab culture and the environment. In this research, independent investigations were carried out to evaluate the crab yield models and modify the optimum-stocking model. Low percentage errors (average 47%, median 36%) between observed and calculated crab yields were obtained. Specific values were defined for adult crab body mass (135 g/ind.) and recapture rate (18% and 30% in lakes with submerged macrophyte biomass above and below 1 000 g/m2) to modify the optimum-stocking model. Analysis based on the modified optimum-stocking model indicated that the actual stocking rates in most lakes were much higher than the calculated optimum-stocking rates. This implies that, for most lakes, the current stocking rates should be greatly reduced to maintain healthy lake ecosystems.
Key words: Chinese mitten crab     sustainable fishery     yield model     optimum-stocking model     independent test     Changjiang lakes    
1 INTRODUCTION

Intensive aquaculture is becoming an important stressor threatening lake ecosystem health. In China, industrial Chinese mitten crab (Eriocheir sinensis) culture is an intensive fishery widely practiced in shallow lakes. Although this crab is an invasive species in Europe and America, causing great ecological and economic loss (Rudnick et al., 2003; Dittel and Epifanio, 2009) through damaging dykes and other installations, it has long been a delicacy in China. Crab culturing has a long history in China. In the 1990s, farming of this crab developed rapidly, particularly in the mid-lower Changjiang Basin. This species is now cultured in all of the Chinese inland provinces, cities, and autonomous regions, except for Tibet. Chinese mitten crab is becoming an important component of the freshwater aquaculture industry in China. The total yield in 2011 was 64.9×104 tons, corresponding to a value of 45.2×108US$ (FAO, 2013), accounting for approximately 2% of total production and 10% of the value of Chinese inland aquaculture in that year.

However, rapid crab culture development comes at a cost, bringing ecosystem health deterioration in shallow lakes. For example, high crab densities negatively affect macrophytes as demonstrated by both mesocosm experiments (Jin et al., 2001) and a presence-absence study in a lake (Xu et al., 2003). The negative effects tend to increase with increasing crab stocking density, as suggested by a study in rice and crab culturing systems (Li et al., 2007). In our recent large-scale analysis with a combination of long-term (12 years) monitoring in a lake and multilake comparisons (20 sub-areas in four lakes), high crab densities led to a decrease in submerged macrophytes and transparency, and, hence, weakened macrophyte-transparency positive feedback, the basic mechanism that maintains the clear-water state in shallow lakes (Wang et al., 2016). In the mid-lower Changjiang Basin, where crab culture has been practiced for decades, many shallow lakes have experienced a regime shift from macrophyte-to phytoplankton-dominated states (Wang et al., 2014). Besides excessive nutrient loading, continuous intensive crab stocking may play an important role. Loss of macrophytes may in turn negatively influence crab culture development because macrophytes are important for crab growth.

A series of maximal crab yield (CYMax) and general optimum-stocking rate (SROpt) models with a combination of adult crab body mass (BW) and recapture rates (RR, calculated as a ratio of the amount of stocked crabs to the amount of harvested crabs) have been established based on a 1-year investigation of 20 sub-areas in four lakes. The aim was to determine the reasonable natural resource consumption levels for sustainable crab culture (Wang et al., 2006; see also Section 1.1). A test of the yield models based on an independent investigation is needed to evaluate the performance of these models because BW and RR used in the optimum-stocking model were roughly estimated in Wang et al. (2006). A modification of the optimum-stocking model with a more reasonable estimation of BW and RR is also needed.

The purpose of the present research is twofold: 1) to test the predictive ability of the crab yield models developed by Wang et al. (2006) using an independent dataset; and 2) to modify the optimum-stocking model by exploring the key factors affecting adult crab body mass and recapture rates.

1.1 Background of the crab yield models tested

The crab yield models established by Wang et al. (2006) were based on a 12-month investigation during Dec. 2001-Dec. 2002 in 20 crab culture sub-areas in four lakes in the middle Changjiang Basin. The ratio of Secchi depth to water depth (ZSD/ZM) was used as the independent variable in these models because these two parameters are easy to measure and closely related to submerged macrophytes, which are important for crab growth. ZSD/ZM during crab stocking season (Dec.-May) was used instead of the annual mean ZSD/ZM to establish the models (Wang et al., 2006). The model based on the ZSD/ZM for those four months gave the strongest correlation between ZSD/ZM and crab yield (CY, calculated as the weight of caught adult crabs divided by area, kg/ha). However, in this study we tested the models (P < 0.001) over a time scale of 1 month, because this requires less effort and hence is more practical for local farmers.

    (1)
    (2)
    (3)
    (4)

Because crab stocking has led to serious deterioration in these lakes, the above models were regarded as maximal yield (CYMax) models. According to the MSY (Maximum Sustainable Yield) theory (Larkin, 1977), 50% of the maximal yield was assumed as the maximal sustainable yield, a level that achieves both optimum profit for farmers and environmental sustainability. Accordingly, the optimum-stocking rates (SROpt, ind./ha) with a combination of adult crab body size (BW) and recapture rate (RR) were estimated as follows:

    (5)
2 MATERIAL AND METHOD 2.1 Study sites and sampling

During 2003-2005, 26 lakes (sub-areas) in the mid-lower Changjiang Basin (114°08'-116°53'E, 30°07'-42'N) were investigated from Feb. to May (juvenile crab stocking season) (Table 1). The number of sampling sites in each lake ranged from 3 to 16, depending on lake size. Submerged macrophyte biomass (BMac), water depth (ZM), and Secchi depth (ZSD) were measured during the investigation. ZM and ZSD were measured by a sounding lead and a Secchi Disc, respectively. In terms of submerged macrophytes, 2-4 replicates were sampled by a scythe-type sampler at each sampling site, mixed, cleaned, drained, and then weighed for wet biomass. The adult crab sizes were measured for all of the lakes during the harvest season (autumn, Sep.-Nov.). For each lake, at least 20 crabs were randomly selected from the harvest obtained by the local fish farms. Body mass (weight) was measured with an electronic balance. Because this species of crab is a valuable product, the local farms take detailed notes during the culture process. Therefore, we obtained the stocking and harvest data directly from the local farm's records for all lakes in all years.

Table 1 Stocked Chinese mitten crab populations and main environmental variables in the lakes investigated during 2003-2005

The dominant submerged macrophytes in these lakes were Potamogeton crispus, P. maackianus, Vallisneria spp., Hydrilla verticillata, Ceratophyllum oryzetorum, and Myriophyllum spicatum.

2.2 Data analysis

The 2003-2005 data given in Table 1 were used to test the crab yield models established by Wang et al. (2006). When calculating the predicted crab yield (CYP), ZSD/ZM measured in Feb. was entered into Model 1 and those measured in Mar., Apr. and May were entered into Models 2, 3, and 4, respectively. A percentage error (PE) was calculated for each case as the difference between observed (CYO) and predicted (CYP) crab yield, |CYP/CYO-1|×100.

To modify the optimum-stocking model by Wang et al. (2006), a Spearman's rank correlation was used to explore the most important factors affecting adult crab body mass (BW) and recapture rate (RR). Culture practice and lake conditions are considered potentially important factors. They are crab juvenile (SR) stocking rate representing crab culture intensity, area and ZM representing lake dimensions, ZSD and ZSD/ZM representing water quality, and BMac representing food resources and information on water quality.

To compare the harvested biomass (i.e., CY) with the initial (i.e., stocked) biomass (SB), their ratio (CY/SB) and difference (CY-SB) were calculated. SB was calculated by multiplying SR by 10 g, an averaged weight of stocked juveniles (Wang et al., 2006).

Microsoft Excel® 2010 and STATISTICA 8.0 were used for data processing and analysis.

3 RESULT 3.1 Relationship between net harvest and submerged macrophytes

When submerged macrophyte biomass (BMac) in 2001 was higher than 1 000 g/m2, the ratios of initial biomass (SB) and end biomass (reflected by crab yield, CY) in 2002 were >3.5 and average=4.1 (between 3.8 and 4.7), respectively (Fig. 1). When BMac < 1 000 g/m2, the ratio was 2.3 on average (between 0.7 and 4.1) and five of the nine lakes had a ratio around one. A highly significant positive relationship was found between BMac in 2001 and CYSB in 2002. CY-SB was much higher in lakes with BMac>1 000 g/m2 (average=66 and ranging from 44- 84) than in lakes with BMac < 1 000 g/m2 (average=17 and ranging from -7-53).

Figure 1 Relationship between submerged macrophyte biomass (BMac) in 2001 and the ratio and difference between initial biomass (SB, kg/ha) and end biomass (reflected by crab yield, CY, kg/ha) in 2002
3.2 Predictive ability of the established crab yield models

The percentage predicted errors (PEs) of the four crab yield models (Models 1-4) were analyzed based on the difference between observed (CYO) and predicted (CYP) crab yields (Table 1). Although a large variation was observed in PE (0.2%-259.0%), the mean (47%) and median (36%) values were relatively low (Fig. 2).

Figure 2 Distribution of the percentage errors in the crab yield models established by Wang et al. (2006) tested with the independent dataset given in Table 1

When comparing the predicted errors among the four models, Model 1 corresponding to Feb. had the highest mean value (57.2%), followed by Model 2 of Mar. (40.8%), Model 3 of Apr. (37.0%), and Model 4 of May (17.6%). When analyzing the PE Spearman's rank correlations of the potential affecting factors (SR, area, ZM, ZSD, ZSD/ZM, and BMac), no significant correlation was found (P=0.21-0.59) (Table 2).

Table 2 Spearman's rank correlations of percentage errors in the crab yield models (PE), adult crab body mass (BW) and recapture rate (RR) with crab stocking rate (SR), lake area (area), water depth (ZM), Secchi depth (ZSD), ratio of Secchi depth to water depth (ZSD/ZM), and submerged macrophyte biomass (BMac)
3.3 Factors affecting adult crab harvest

Adult crab body mass (BW) was only significantly correlated with juvenile crab stocking rates (SRs) and lake surface area (area) (Table 2). However, variations from the fitted lines in the scatterplots were quite large for the BW with SR and area regressions (Fig. 3a, b).

Figure 3 Relationship between adult crab body mass (BW) and lake surface area (area) (a) and juvenile crab stocking rate (SR) (b) and between adult crab recapture rate (RR) and submerged macrophyte biomass (BMac) (c) (n= 43)

Similarly, adult crab recapture rate (RR) was only significantly correlated with submerged macrophyte biomass (BMac) (Table 2). The variations from the fitted line in the plots were also quite large (Fig. 3c). Further analysis revealed that RR in the lake groups with BMac < 10 g/m2, 10-100 g/m2, and from 100- 1 000 g/m2 (average 17%, 16%, and 20%, respectively) did not differ significantly from each other (P≥0.86) (Fig. 4). RR in the lake group with BMac>1 000 g/m2 (averaged 32%), however, was significantly higher than all of the other three groups (average 18%) (P < 0.006).

Figure 4 Adult crab recapture rate in lakes with submerged macrophyte biomass < 10 g/m2, 10-100 g/m2, 100- 1 000 g/m2, and >1 000 g/m2
3.4 Optimum-stocking model modification

Because no close relationship was found between adult crab body mass (BW) and the potential affecting factors, the average value of 135 g/ind. was defined for BW to be included in the optimum-stocking model. For the crab recapture rates (RRs), two specific values were defined. They were 18% for lakes with BMac < 1 000 g/m2 and 32% for lakes with BMac>1 000 g/m2. Therefore, when Models 1-4 were included, the Model 5 optimum-stocking rates (SROpt) for lakes with BMac < 1 000 g/m2 were further defined as:

    (6)
    (7)
    (8)
    (9)

for lakes with BMac>1 000 g/m2 as:

    (10)
    (11)
    (12)
    (13)
4 DISCUSSION 4.1 Reliance of crab culture sustainability on macrophytes

In this study, a highly significant positive relationship was found between crab biomass accumulation and macrophyte biomass at the time of crab stocking (Fig. 1). This demonstrates the importance of macrophytes in supporting crab culture, especially when macrophyte biomass is >1 000 g/m2. Submerged macrophytes may benefit crabs by providing food either directly or indirectly (Dvorăk and Bestz, 1982; Ju and Shu, 1999; Jin et al., 2003) and providing appropriate habitats to avoid predators, particularly during molting (Pan, 2002). Our results also suggest that these are the dominant mechanisms based on the significantly higher recapture rate in lakes with abundant macrophytes (Fig. 4).

Furthermore, the lower, even negative crab biomass accumulation in lakes with BMac < 1 000 g/m2 further demonstrates that fewer macrophytes could not support sustainable crab culture. The negative impacts of crabs on macrophytes have been widely reported in studies of various scales (Jin et al., 2001; Xu et al., 2003; Li et al., 2007; Wang et al., 2016), implying that continuous high-density crab stocking may in turn prevent sustainable crab culture. Low macrophyte density is also weak in its resilience to crab disturbance. Therefore, reasonable crab stocking rates and sufficient macrophyte abundance are the two fundamental factors for both sustainable crab culture and healthy ecosystems.

4.2 Yield model test and optimum-stocking model modification

Average and median percentage errors of 47% and 36%, respectively, were obtained for the crab yield models established by Wang et al. (2006). No comparable result can be found in the literature because no similar prior research has been carried out. Out of approximately 150 biogeochemical models, Arhonditsis and Brett (2004) obtained a median percentage error of 44% for phytoplankton, 70% for zooplankton, and 36% for bacteria. A much higher percentage error was obtained by phytoplankton chlorophyll a models, e.g., 95% (Canfield Jr, 1983) and 92% (Wang et al., 2008).

In Wang et al. (2006), BW was preliminarily suggested as 150 g/ind. However, such a value is not ideal for these lakes because BW (75%) was < 150 g/ind. in most cases. In this study, no close relationship was found between BW and the other factors. BW was only slightly significantly related to lake surface area and stocking rate. It is, therefore, impossible to define specific BW classes relating to either different culture practices or lake conditions. The fact that BW only varied slightly might explain the poor relationships. The BW coefficient of variation among these lakes was as low as 16%. It is safe, however, to set the mean as the specific value for BW because of its small variation.

Wang et al. (2006) proposed that RR=30%, a value calculated from lakes with abundant macrophytes and zoobenthos. However, only 20% of the lakes in this study approached an RR of 30%. In this study, no reliable relationship was found to define RR. However, RR can be defined specifically for lake groups with BMac less and greater than 1 000 g/m2 because of their significantly different RRs (Fig. 4). The significantly higher RR in lakes with BMac>1 000 g/m2 is mainly because macrophytes provide an ideal habitat for the crabs. The crabs may hide among macrophytes to avoid either predators or attacks from other crabs. They may also forage for food among the macrophytes.

4.3 Application of the optimum-stocking models

The actual stocking rate and ZSD/ZM (Fig. 5a) scatterplots revealed that the fishermen did not consider lake conditions when stocking crab seed. The stocking intensities were similar in lakes with different ZSD/ZM values. When using the newly defined Models 6-13 to calculate the optimum-stocking rates in these lakes, the actual stocking rates were much higher than the calculated rates in 86% of cases (37 of 43) (Fig. 5b). Therefore, current stocking rates should be greatly reduced in most of the lakes. An alternative strategy to maintain healthy lake ecosystems is a rotation of crab stocking. One option is to separate the lake into two parts and rotate every 2 or 3 years, only stocking one section in any given year. Another option is to stock mitten crabs every 2 or 3 years.

Figure 5 The relationships between the actual stocking rates (SRs) and the ratio of Secchi depth to mean depth (ZSD/ZM) (a) and the calculated optimum-stocking rates (SROpt) (b) (n= 43)
5 CONCLUSION

We evaluated the crab yield models established by Wang et al. (2006) by an independent dataset. These models performed well, obtaining an average and median percentage error of 47% and 36%, respectively. The optimum-stocking model by Wang et al. (2006) was modified by defining specific values for adult crab body mass (BW) and recapture rate (RR) to be included in the model. BW was defined as 135 g/ind. and RR was defined as 18% for lakes with BMac < 1 000 g/m2 and 30% for lakes with BMac>1 000 g/m2.

When using the modified models to calculate the optimum-stocking rates in these lakes, the actual stocking rates were much higher than the calculated rates in most cases. Therefore, we suggest a reasonable approach to crab culture based on our results to maintain healthy lake ecosystems and sustain profit. In lakes with low underwater light conditions, poor food resources and submerged vegetation, crab stocking rates should be massively reduced and possibly even ceased.

6 ACKNOWLEDGEMENT

We thank CUI Yongde, LIU Xueqin, and ZHAO Weihua for their help with fieldwork. Special thanks to YE Shaowen for providing crab data for Niushan Lake and LI Yan for help with graph preparation.

References
Arhonditsis G B, Brett M T, 2004. Evaluation of the current state of mechanistic aquatic biogeochemical modeling. Marine Ecology Progress Series, 271: 13–26. Doi: 10.3354/meps271013
Canfield Jr D E, 1983. Prediction of chlorophyll a concentrations in Florida lakes:the importance of phosphorus and nitrogen. Jawra Journal of the American Water Resources Association, 19(2): 255–262. Doi: 10.1111/jawr.1983.19.issue-2
Dittel A I, Epifanio C E, 2009. Invasion biology of the Chinese mitten crab Eriochier sinensis:a brief review. Journal of Experimental Marine Biology & Ecology, 374(2): 79–92.
Dvorăk J, Bestz E P H, 1982. Macro-invertebrate communities associated with the macrophytes of Lake Vechten:structural and functional relationships. Hydrobiologia, 95(1): 115–126. Doi: 10.1007/BF00044479
FAO (Food and Agricultural Organization of the United Nations). 2013. Global Aquaculture Production 1950-2011 (online query). Fisheries and Aquaculture Information and Statistics Service, Food and Agricultural Organization of the United Nations.
Jin G, Li Z J, Xie P, 2001. The growth patterns of juvenile and precocious Chinese mitten crabs, Eriocheir sinensis(Decapoda, Grapsidae), stocked in freshwater lakes of China. Crustaceana, 74(3): 261–273. Doi: 10.1163/156854001505505
Jin G, Xie P, Li Z J, 2003. Food habits of 2-year-old Chinese mitten crab, Eriocheir sinensis, stocked in Lake Bao'an. Acta Hydrobiologica Sinica, 27(2): 140–146.
Ju C M, Shu S W, 1999. Requirement for and damage to Vallisneria spiralis by Eriocheir sinensis. Acta Hydrobiologica Sinica, 23(6): 700–704.
Larkin P A, 1977. An epitaph for the concept of maximum sustained yield. Transactions of the American Fisheries Society, 106(1): 1–11. Doi: 10.1577/1548-8659(1977)106<1:AEFTCO>2.0.CO;2
Li W D, Dong S L, Lei Y Z, Li Y H, 2007. The effect of stocking density of Chinese mitten crab Eriocheir sinensis on rice and crab seed yields in rice-crab culture systems. Aquaculture, 273(4): 487–493. Doi: 10.1016/j.aquaculture.2007.10.028
Pan H Q, 2002. Ecological Culture of Chinese Mitten Crab. China Agriculture Science and Technology Press, Beijingp.221.
Rudnick D A, Hieb K, Grimmer K F, Resh V H, 2003. Patterns and processes of biological invasion:the Chinese mitten crab in San Francisco Bay. Basic & Applied Ecology, 4(3): 249–262.
Wang H J, Liang X M, Jiang P H, Wang J, Wu S K, Wang H Z, 2008. TN:TP ratio and planktivorous fish do not affect nutrient-chlorophyll relationships in shallow lakes. Freshwater Biology, 53(5): 935–944. Doi: 10.1111/j.1365-2427.2007.01950.x
Wang H J, Wang H Z, Liang X M, Wu S K, 2014. Total phosphorus thresholds for regime shifts are nearly equal in subtropical and temperate shallow lakes with moderate depths and areas. p. Biology, 59(8): 1659–1671.
Wang H J, Xu C, Wang H Z, Kosten S. 2016. Long-term density dependent effects of the Chinese mitten crab(Eriocheir sinensis (H. Milne Edwards, 1854)) on submersed macrophytes. Aquatic Botany, http://dx.doi.org/10.1016/j.aquabot.2016.02.001.
Wang H Z, Wang H J, Liang X M, Cui Y D, 2006. Stocking models of Chinese mitten crab (Eriocheir japonica sinensis) in Changjiang lakes. Aquaculture, 255(1-4): 456–465. Doi: 10.1016/j.aquaculture.2006.01.005
Xu Q Q, Wang H Z, Zhang S P, 2003. The impact of overstocking of mitten crab, Eriocheir sinensis, on lacustrine zoobenthos community. Acta Hydrobiologica Sinica, 27(1): 41–46.