Estimating mortality in natural assemblages of oysters

To evaluate methods for calculating mortality in bivalve molluscs, we analyzed historical data from dredge surveys of both planted and natural. oyster (Crassostrea virginica) grounds in Delaware Bay to compare total box-count mortality estimates with those made by accumulating short-term mortality rates obtained from fresh boxes identified by shell condition and degree of fouling. Box-count and cumulative-mortality patterns and values agreed best on grounds with planted oysters, where a cohort with very few dead oysters was broadcast on previously cleaned bottom and was followed over time. This situation is analogous to an artificially created reef with oysters either deployed or naturally set on it. When deaths predominated in late summer and early autumn, the two estimates were similar throughout the following year; when mortality was greatest in spring or early summer, the estimates were similar only through autumn of the same year. Correspondence was much weaker on natural beds, where new individuals constantly recruited to the population and variable numbers of boxes were always present. Nevertheless, total box-count mortality estimates made during autumn stock surveys were significantly correlated with cumulative mortalities calculated for the preceding year. We also estimated disarticulation rates of artificially created boxes by deploying them at three seasons and eight sites in Delaware Bay. Disarticulation time depended on the length of exposure at summer temperatures, with the average time to 50% disarticulation for boxes deployed in spring and in summer being 225 and 345 days, respectively. Disarticulation rates increased with decreasing size, and increasing salinity and temperature. Finally, we compared total box-count and cumulative-mortality estimates with those made using the disarticulation data. Annual averages for the three methods were within 5 percentage points of each other. Our data indicate that total box-count mortality estimates from fall stock surveys can provide a reliable index to total mortality for the previous year, although it cannot describe the seasonal patterns obtainable using the cumulative-mortality method.