Quantifying Nutrient and Sediment Export from the Chesapeake Bay Watershed: Retrospective Analyses and Method Improvements
MetadataShow full item record
Toward Chesapeake Bay restoration, management programs have focused for decades on reducing nutrient and sediment loadings from the Chesapeake Bay watershed (CBW). To assess progress and shape future strategies, a critical need is to better understand historical loading changes from different regions of the CBW, using best current methods and data. In this regard, investigators at USGS have developed the “Weighted Regressions on Time, Discharge, and Season (WRTDS)” method for better loading and trend estimation. Motivated by the above need, this dissertation research focused on applying the WRTDS method to long-term monitoring records for various major tributaries, exploring the method’s uncertainties, and improving the method’s estimation performance. Specific contributions include: (1) analysis of long-term seasonal trends of riverine nutrient and sediment loadings from major tributaries to Chesapeake Bay; (2) evaluation of decadal-scale changes in sediment and nutrient processing within the Lower Susquehanna River Reservoir System (LSRRS); (3) analysis of uncertainty and sensitivity of results to sample availability and storm events when estimating LSRRS input and output loadings; (4) investigation of the temporal and spatial patterns of nutrient and sediment export from the Susquehanna River basin and major factors affecting such patterns; (5) development of an improved method for making robust interpretations of riverine concentration-discharge relationships and application of this method toward Chesapeake tributaries for a top-down synthesis of export patterns; (6) development of improved methods for estimating riverine concentration and loading through incorporation of antecedent discharge conditions into WRTDS and evaluation of the methods’ performance under various sampling strategies; and, (7) comparison of alternative approaches for quantifying long-range dependence in irregularly sampled water-quality data through Monte-Carlo simulations. Overall, this research has demonstrated the utility of statistical modeling approaches toward large-scale analysis and synthesis of decadal-scale water-quality data collected in river systems. The applications to the CBW have provided new evidence on the decreasing trapping performance of the LSRRS and new understanding of nutrient and sediment export from various locations in the watershed. In addition, this work has made important methodological advancements with respect to WRTDS estimation performance, interpretation of riverine concentration-discharge relationships, and quantification of long-range dependence in irregularly sampled data.