[Chesapeake Quarterly masthead]
Volume 4, Number 3
Table of Contents
Download pdf

A Tale of Two Oysters
Volume 1 . . . Where Will Larvae Settle?
circulation/hdrodynamics of the Bay --> particle tracking and lavarval behaviour --> settlement at each oyster bar

The larval transport model follows the oyster's journey from spawning to settlement. Two circulation models recreate water-driven (hydrodynamic) forces on the larvae, such as currents and tides. These models (top) divide the Bay with a finely meshed grid, each using different geometric rules. To compute fluid motion, the computer solves a system of equations in each of the compartments generated by the grid, in ten-minute intervals of real time. One hydrodynamic model may do a better job predicting currents in the upper estuary and the other a better job in the lower estuary. Using the two together helps quantify potential sources of error in the predictions, according to researcher Elizabeth North.

The particle tracking model (bottom left) uses information from the hydrodynamic models to predict where larvae will go, as though they were passive particles. North will run this model using the hydrodynamic conditions during five different years, 1995-1999, allowing the model to capture the range of flow conditions in the Chesapeake Bay from wet to dry years. Then North builds behavior into the model, making it more realistic. As the larvae grow larger, they swim faster. The model increases their swimming speed from 0 to 3 millimeters per second (based on the scientific literature). As they age, larvae also make behavioral "choices" about their position in the water column. The model provides virtual larvae with behavioral decisions every 30 seconds.

The final output of the larval transport model are maps (bottom right) that show settlement at each oyster bar in the Chesapeake Bay, for both C. virginica and C. ariakensis. These maps will feed directly into the juvenile/adult demographic model.

. . . Volume 2 , Where Will They Thrive?
juvenile/adult demographic model used to predict abundance over time and river flow

The juvenile/adult demographic model builds directly on the larval transport model to predict oyster (native and non-native) populations in the Bay over time, up to the year 2015. This model incorporates estimates of natural oyster mortality, along with mortality from disease and harvesting, and uses equations to describe the growth rate, derived from over 60 data sets from different oyster bars.

"In its simplest form, the demographic model grows them, harvests them, reproduces them, and kills them," says model statistician Mary Christman.

Although it spans many generations, the demographic model is simpler than the larval transport model. The demographic model makes calculations based on whole oyster bars, some more than a kilometer in size, while the larval transport model parcels the Bay into small, one-meter square packages. Whereas the former also runs on a yearly time step, incorporating annual data on growth and mortality, the larval transport model builds in new hydrodynamic data every 10 minutes and updates the behavior of individual larvae every 30 seconds.

So there is a huge difference in computational time between the two models, explains Elizabeth North. The juvenile/adult demographic model can look at a whole year of oyster growth, mortality, and reproduction in the Bay in 10 minutes of computation. The larval transport model takes 24 hours to simulate four days of larval dispersal in the Bay.

Shorter computation time also means that researchers can run the demographic model many times to explore the effects of different environmental scenarios, such as extended periods of either high or low river flow (graph above). (For more on the demographic model, see When Science Meets Policy.) Graph from Elizabeth North.

Top of Page

[Chesapeake Quarterly]
Other Issues

[Chesapeake Quarterly Bar]
[Maryland Sea Grant][NOAA]