Modeling the impacts of climate variability and change on circulation and salinity in Chesapeake and Delaware Bays
Open Access
- Author:
- Ross, Andrew Charles
- Graduate Program:
- Meteorology
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- October 16, 2017
- Committee Members:
- Raymond Najjar, Dissertation Advisor/Co-Advisor
Raymond Najjar, Committee Chair/Co-Chair
Michael Mann, Committee Member
David Titley, Committee Member
Klaus Keller, Outside Member
Ming Li, Special Member - Keywords:
- Chesapeake Bay
Delaware Bay
Tides
Climate change
Uncertainty - Abstract:
- This dissertation discusses the use of numerical models to simulate the effects of climate variability and change on Chesapeake and Delaware Bays, two large coastal plain estuaries in the Mid-Atlantic Region of the United States that are both home to productive ecosystems, important ports, and large concentrations of human population. Estuaries like these bays are uniquely characterized by salinity variation from riverine freshwater to oceanic saltwater, both horizontally and vertically, and by the strength of their tides. From a meteorological perspective, estuaries are interesting because they are influenced by many aspects of weather and climate variability, including runoff and winds. Future climate changes, including changes in river discharge and mean sea level, are also likely to produce significant changes in estuarine salinity and circulation and could alter estuarine ecosystems. To predict the effects these changes may have on estuarine salinity, circulation, and ecosystems, numerical model simulations are often applied. However, the predictive capability of numerical models is unknown due in part to a lack of knowledge about historical trends and whether numerical models can reproduce them. This dissertation is composed of three studies that address this lack of knowledge. The first study of this dissertation analyzes data from tide gauges in Chesapeake and Delaware Bays and compares the results with model simulations to determine what trends are present and whether the numerical model correctly predicts these trends. When using the numerical model for predictions, it is also important to account for uncertainty in the model and its inputs, but doing so is difficult due to the chain of computationally expensive models typically used to simulate estuaries. The second study of this dissertation examines a method to account for uncertainty in future river discharge, and the third study conducts an analysis of which inputs the numerical model is most sensitive to. In the first study, statistical models show negative M2 amplitude trends at the mouths of both bays, while some upstream locations have insignificant or positive trends. To determine whether sea-level rise is responsible for these trends, a term for mean sea level is included in the statistical models and the results are compared with with predictions from numerical and analytical models. The observed and predicted sensitivities of M2 amplitude and phase to mean sea level are similar, although the numerical model amplitude is less sensitive to sea level. The sensitivity occurs as a result of strengthening and shifting of the amphidromic system in the Chesapeake Bay and decreasing frictional effects and increasing convergence in the Delaware Bay. After accounting for the effect of sea level, significant negative background M2 and S2 amplitude trends are present; these trends may be related to other factors such as dredging, tide gauge errors, or river discharge. Projected changes in tidal amplitudes due to sea-level rise over the twenty-first century are substantial in some areas, but depend significantly on modeling assumptions. The second study examines the impacts of methods for model selection on projections of runoff change for the Susquehanna River Basin. The results from an ensemble of 29 global climate models and 29 corresponding hydrological model simulations were compared with the results that would have been obtained by applying five different selection strategies to the climate model data and using only the selected models to drive the hydrological model. Only one method, the KKZ algorithm, produced results that met the objective of the method and were not strongly sensitive to the number of models selected. Regardless of the selection method used, the results for small model subsets (fewer than about 7 models) were highly variable and failed to cover the uncertainty present in the full model dataset. On the other hand, results from the complete model ensemble may be affected by structurally and statistically similar models. This study shows that the methods and models used in similar top-down studies should be carefully chosen and that the results obtained should be interpreted with caution. Finally, estuarine physics and water quality are strongly controlled by climate and oceanographic variability. Climate and oceanographic conditions are likely to change in the future as a result of global climate change, and it is important to consider how these changes, and how uncertainty surrounding these changes, could affect water quality and management. To do so, numerical models are typically used to simulate estuarine physics and water quality under scenarios of future conditions. However, the numerical models typically used for simulating estuaries are computationally demanding, limiting the ability to understand and quantify uncertainty in the model results. In the third study of this dissertation, a computationally inexpensive statistical model, or metamodel, is tested as a surrogate for numerical model simulations. The metamodel is fit to 12 numerical model simulations of the Chesapeake Bay and used to simulate stratification, circulation, and mean salinity under sampled probability distributions of projected future mean sea level, river discharge, and tidal amplitudes along the shelf. The simulations from the metamodel show that future salinity, stratification, and circulation are all likely to be higher than present-day averages. However, the metamodel indicates that model projections of salinity and stratification are highly sensitive to uncertainty about future tidal amplitudes along the shelf. Since previous studies have focused on potential changes in either river discharge or sea level while neglecting any change in tidal amplitude, these results demonstrate the importance of conducting a thorough sensitivity and uncertainty analysis. Future studies should build from this concept by including more sources of uncertainty, such as wind speed and direction and model parameters and structures. The results in this dissertation show that although numerical models are capable of reproducing some past changes, the impacts of future climate and oceanographic changes on these estuaries remain highly uncertain. Salinity and stratification in the Chesapeake Bay are fairly likely to increase as a result of a highly probable increase in mean sea level, although exact changes are especially sensitive to changes in tidal amplitude. Model simulations of future tides in both bays appear to be sensitive to the methods used to model sea-level rise even though the simulations of past tides in this dissertation are not. In future work, it will be important to consider this uncertainty, to consider other uncertainties that were neglected in this dissertation, and to examine the impacts on biogeochemistry and overall ecosystem health.