Autoregressive statistical models of monthly salinity variations in the mai
nstem of Chesapeake Bay are developed for use in climate change application
s. Observations of salinity and Susquehanna River flow from 1984 to 1994 ar
e used to calibrate the models. Up to 93% of the variance in salinity is ca
ptured by these models, with the best fits occurring; in the middle of the
bay, where submonthly fluctuations due to river flow and oceanic currents a
re damped. Our salinity models use output from a climate-forced hydrologica
l model under a doubling of present-day atmospheric CO,. This coupling sche
me predicts salinity changes between +3.5% and -27.5% near the mouth of the
Susquehanna River, to between +0.1% and -0.7% near the ocean. These ranges
demonstrate the uncertainty in climate model predictions, although three o
f the four scenarios used indicate increased Susquehanna River flow and the
refore decreased salinity. In the highest flow scenario (a 32% annual incre
ase) our results show that the bay's isohalines may recede by approximately
6.3 km (about 2% of the length of the bay) near the Susquehanna River, to
as much as 55 km (about 17% of the length of the bay) near the middle bay.
This shift implies that climate change may have consequences for organisms
with low-salinity thresholds, including oysters and crabs.