Factors driving bottom salinity variability in the Chesapeake Bay

The long-term variability of bottom salinity in the Chesapeake Bay is examined using observations from 1985 to 2004. Simple linear regression analysis based on annual mean data shows that the bottom salinity is significantly correlated with both the Susquehanna River discharge and the Gulf Stream (GS) index, which represents shelf and slope water salinity. This suggests that local forcings as well as external processes affect the variation of bottom salinity on interannual to decadal time scales. Empirical orthogonal function (EOF) analysis is used to extract patterns among 21 sampling stations along the bay. The first seven EOF modes explain approximately 85% of total variance, and consequently they are combined to represent the variability of bottom salinity anomaly in the bay. The wavelet transform of the bottom salinity anomaly reveals significant areas centered at the periods of 2 and 7.5 years, which is consistent with the results from Fourier spectral analysis. The significant powers in the bottom salinity anomaly exhibit spectral correspondences to the wavelet power spectra of both the river discharge and the GS index, suggesting possible connections between them. Wavelet coherence is employed to determine whether relationship in power spectra between two time series is statistically significant. Our results clearly demonstrate that the interannual variability of the bottom salinity is forced by the freshwater discharge, and the quasi-decadal variability is associated with shelf-slope water condition through horizontal exchange. No evidence is found to indicate that there is influence of the El Nino Southern Oscillation (ENSO) on the variability of the river discharge, but the North Atlantic Oscillation (NAO) may be related to it at decadal or longer time scales. (c) 2008 Elsevier Ltd. All rights reserved.