Disruption of the natural patterns of freshwater flow into estuarine ecosystems occurred in many locations around the world beginning in the twentieth century. To effectively restore these systems, establishing a pre-alteration perspective allows managers to develop science-based restoration targets for salinity and hydrology. This paper describes a process to develop targets based on natural hydrologic functions by coupling paleoecology and regression models using the subtropical Greater Everglades Ecosystem as an example. Paleoecological investigations characterize the circa 1900 CE (pre-alteration) salinity regime in Florida Bay based on molluscan remains in sediment cores. These paleosalinity estimates are converted into time series estimates of paleo-based salinity, stage, and flow using numeric and statistical models. Model outputs are weighted using the mean square error statistic and then combined. Results indicate that, in the absence of water management, salinity in Florida Bay would be about 3 to 9 salinity units lower than current conditions. To achieve this target, upstream freshwater levels must be about 0.25 m higher than indicated by recent observed data, with increased flow inputs to Florida Bay between 2.1 and 3.7 times existing flows. This flow deficit is comparable to the average volume of water currently being diverted from the Everglades ecosystem by water management. The products (paleo-based Florida Bay salinity and upstream hydrology) provide estimates of pre-alteration hydrology and salinity that represent target restoration conditions. This method can be applied to any estuarine ecosystem with available paleoecologic data and empirical and/or model-based hydrologic data.

Current meter time series collected between1995 and 2001 are used to describe the exchanges of water through five tidal channels that connect Lignum Vitae Basin, a large subbasin in eastern Florida Bay, with adjacent subbasins and Atlantic shelf waters. Current data were combined with measured or simulated water levels, and channel geometry measurements were incorporated, to quantify volume transport through the channels. Results indicate a long-term inflow to Lignum Vitae Basin from the northwest through Gopher Keys Cut at an average rate of + 11.9 m3 s-1 during a 13-mo study period and from the west through South Twin Keys Cut at +9.6 m3 s-1 during a 7-mo study period. A quasi-steady long-term outflow was observed to the east through Steamboat Channel, to the southeast through Indian Key Channel, and to the southwest through Bowlegs Cut. Outflow rates averaged -7.1, -25.8, and -6.5 m3 s-1, respectively, through these three channels. Interactions between tidal water-level fluctuations and ebbs or floods resulted in a tide-induced outflow from Lignum Vitae Basin through Gopher Keys and South Twin Keys cuts at rates of -0.06 and -0.09 m3 s-1, respectively. Tidal residual inflows of +5.8, +0.11, and +0.04 m3 s-1 were calculated for Indian Key Channel, Steamboat Channel, and Bowlegs Cut, respectively. Comparisons of local winds with channel transport indicate that winds from any direction will force water through every channel except Steamboat Channel over most time scales greater than about 2 d. Flow through Steamboat Channel was coherent only with winds out of the northeast quadrant over time scales greater than about 1 d. Longterm flow was generally upwind through all channels except Bowlegs Cut.

Near-bottom current meter and dissolved nutrient data are used to describe the advection of nutrients from the Florida Keys to the reef tract, where elevated nutrient concentrations have been shown to adversely affect reef corals. Results indicate a long-term across-shelf flow toward the reef tract that averaged 2 cm s -1 during the 13month study. Combining the average flow rate with a representative dissolved inorganic nitrogen (DIN) concentration of 2.71µM yields an average DIN transport rate of 55 µmole m -2 s -1 seaward past the study site. Lowfrequency, nontidal across-shelf flow was toward the reef 80% of the time and a mean of 11 days was required to transport nutrients 8 km from the Keys to the reef tract (median = 4.7 d). M2 tidal currents transport nutrients seaward 0.38 km during the average ebb. Spectral analysis indicates no cause-and-effect relationship between across-shelf flow and the local wind field.

Current and wind data collected during a 223-d study period from early June 1983 through early January 1984 are used to document net upwind return flow over seasonal time scales in the Indian River Lagoon near Sebastian Inlet on the Atlantic Coast of central Florida. Spectral analysis of wind and current meter data suggests that wind forcing accounts for the majority of nontidal flow. The two time series are highly coherent at periodicities of 2,3–3.5, 4, and 6 d. The serial correlation coefficient of −0.49 is significant at the 99.9% confidence level. Historical water level and wind data recorded simultaneously at Sebastian Inlet and Eau Gallie, 34.5 km to the north, suggest that longitudinal pressure gradients are established in response to seasonal wind patterns. The upwind near-bottom return flow is interpreted as a response to the seasonal set up of the longitudinal pressure gradients. The study documents the importance of windstress in forcing low-frequency transport over seasonal time scales.