A hydrodynamic model-based approach to assess sampling approaches for dissolved oxygen criteria in the Chesapeake Bay

Technological advances in water quality measurement systems have provided the potential to expand high-frequency observations into coastal monitoring programs. However, with limited resources for monitoring budgets in natural waters that exhibit high temporal and spatial variability in water quality, there is a need to identify the locations and time periods where these new technologies can be deployed for maximum efficacy. To advance the capacity to make quantitative and objective decisions on the selection of monitoring locations and sampling frequency, we combined high-resolution numerical model simulations and multi-frequency water quality measurements to conduct a power analysis comparing alternative sampling designs in the assessment of water quality in the Chesapeake Bay. Specifically, we evaluated candidate monitoring networks that deployed both conventional long-term fixed station monitoring in deep channel areas and short-term continuous monitoring technologies in near-shore, shallow areas to assess 30-day dissolved oxygen criteria in two Bay tributaries. We conducted a cumulative frequency diagrams analysis to quantify the accuracy of each monitoring scheme in evaluating compliance with respect to the model. We used a Monte Carlo simulation to incorporate the spatial and temporal uncertainty of criteria failure. We found that additional long-term biweekly channel and short-term continuous shallow sampling efforts can lead to statistically unbiased and improved assessments at local spatial extents (less than 0.2 proportion of the assessed water body), especially when additional sampling is added at stations representing hypoxic water areas. Stations that represented seaward regions of the tributaries were more valuable in maintaining unbiased assessments of dissolved oxygen criteria attainment. This analysis highlights the importance of statistical evaluation of ongoing monitoring programs and suggests an approach to identify efficient deployments of monitoring resources and to improve assessment of other water quality metrics in estuarine ecosystems.