Determination of the fine scale, temporal pattern of larval release by female blue crabs and application of this information to mathematical models of larval dispersal and recruitment
Date
2006
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
In recent years considerable advances have been made in the mathematical models that simulate larval transport in coastal and estuarine species. One of these models (ECOM3d) coupled to a particle transport model has been used to analyze the dispersal and recruitment of blue crab larvae (Callinectes sapidus) in the Delaware Bay. Larval development in C. sapidus is planktonic and includes seven zoeal stages and a single megalopal stage. Larval dispersal is accomplished by the zoeal stages and occurs in the coastal ocean. Recruitment to juvenile habitat is carried out by the megalopal stage and occurs within the estuary. Thus, successful recruitment requires transport of megalopae from the coastal ocean to the estuary. ☐ The ECOM-3d model is driven by the physical processes of wind, tides, and freshwater outflow. In our application, ECOM-3d has been coupled to a separate particle transport model that treats blue crab larvae as passive particles. Preliminary application of the coupled transport model has resulted in successful reproduction of the occurrence of megalopal settlement events in Delaware Bay. However, to reproduce the magnitude (as well as the timing) of such events, the model requires information concerning the temporal pattern of larval release by ovigerous blue crabs in lower Delaware Bay. ☐ Preliminary sampling was conducted during the summer of 2004 and more extensive sampling in 2005. During both summers, collections were made at two stations in the southern part of the Delaware Bay. Brooding females maintain a benthic distribution during the day and were collected twice a week at a location adjacent to a jetty protecting the Cape May-Lewes Ferry terminal. In addition, swimming females were collected every night from the Cape Henlopen Fishing pier. The study consisted of seining at the Ferry Jetty and dip-netting females at the Pier. An egg sample was collected from each female and returned to the lab for subsequent analysis. ☐ Analysis consisted of laboratory determination of the stage-of-development of each egg sample. Eggs were observed under a dissecting microscope and assigned to one of several developmental stages. To assist in this process, we have developed an index that describes the morphological features of blue crab eggs at each day of egg development. This index was used to predict the hatching date for each ovigerous female collected during the study. These data were then used to construct a time series of the proportion of the total number of ovigerous crabs that released eggs on each day of the study. ☐ Results of microscopic analysis of eggs indicated that release of larvae occurs throughout the summer in the Delaware Bay population of blue crabs. Because the Jetty samples were collected at only a bi-weekly frequency, they were unsuitable for formal time-series analysis. This was due to the low number of samples collected at the Jetty throughout the season. However, data from the Pier resulted from daily sampling over a four-month period and were suitable for further analysis using formal time-series techniques. Spectral analysis indicated a strong temporal variation at a low frequency (~0.014-0.025 cycles/day). This corresponds to a period of 40-70 days, indicating a broad seasonality in larval release. There was only modest variance at higher frequencies. Based on results of the spectral analysis, the time series of larval release was fit to a cosine function of time using regression techniques. ☐ Results from the Pier Site were used in subsequent simulations of larval transport using the coupled transport model described above. Settlement of blue crab megalopae in Delaware Bay was simulated for the spawning seasons of 2002, 2003, and 2004. The model domain included Delaware Bay and the adjacent coastal ocean. Larvae were released at the mouth of Delaware Bay in two basic ways, constant release and cosine-fit release. In the constant-release mode, a total of 651 larvae were released during each tidal cycle throughout the 90-day spawning season that extended from June 29 through October 27. In the cosine-fit release mode, the weighting of the larvae for each day were increased or decreased to reflect variations in spawning. Instead, the temporal pattern of release was defined as the cosine function described above. White noise was also introduced to the time series of larval release in a separate model to represent the high level of aperiodic variation in daily larval release that had been observed during the field investigation. White noise was generated from the residual variance associated with the regression that was used to determine the cosine function for larval release. ☐ Overall, a more realistic simulation of actual spawning patterns in the bay was attained using the cosine-fit mode because this takes into account the observed seasonal changes in larval release. Results of the simulations show that the total number of larvae recruited to Delaware Bay each year can be substantially affected by the use of the cosine-fit simulation. However, the change in settlement from year to year is not predictable. For example in 2002 and 2003, the constant-release settlement was approximately 25% greater than cosine-fit settlement. In contrast, there was virtually no difference in the magnitude of constant-release and cosine-fit settlement in 2004. Adding white noise to the cosine-fit model made the simulations even more realistic by including the observed aperiodic variation in daily release. However, the combined effect of adding cosine-fit release and white noise to the simulations for any given year was relatively small when compared to interannual variation in simulated settlement among the years. This is remarkable because (for any given mode of simulation) both the number of larvae released and the temporal pattern of that release were identical among the three years. Thus, the variation among the three years must have been caused by interannual variation in the physical agents that drive the model. For the ECOM3d model, these agents are primarily wind, fresh-water outflow, and tides—and because tidal flow varies little on long time scales, it appears that variations in wind and freshwater outflow caused the large interannual differences in simulated settlement of blue crab larvae in Delaware Bay. Therefore, the results of this study support recent hypotheses that the supply of blue crab larvae to nursery habitat within the large estuaries of the Middle Atlantic Bight is primarily controlled by physical processes. ☐ Overall the results of my investigation have provided an important new biological parameter to add to the model. The addition of observed seasonal changes in larval release provides a more realistic simulation of actual spawning patterns in the Delaware Bay. This will help scientists to create a better model for predicting the size of the blue crab population in the Delaware Bay each year. This model can then be used by fisheries managers to set limits on the blue crab fishery for each year that will prevent overfishing of this valuable resource.