in the CB watershed can reach 20 years or more (Focazio et al., 1997) and base flow accounts for a major fraction of riverine N load at many CB sites (Bachman et al., 1998). For this region, the legacy stores are comprised primarily of groundwater for N (Bachman et al., 1998; Sanford & Pope, 2013), surface soils and river sediments for P (Ator et al., 2011; Sharpley et al., 2013), and stream corridors and reservoir beds for sediment (Gellis et al., 2008; Pizzuto et al., 2014; Walter & Merritts, 2008). These results suggest the importance of considering lag time between implementation of management actions and achievement of water‐quality improvement. For the NAS, budget estimates indicate the accumulation in river watersheds of inorganic and organic N and P from anthropogenic sources that still negatively affect the quality of freshwater systems (Giani et al., 2012; Viaroli et al., 2018; Volf et al., 2018) and river‐dominated coastal areas (Alvisi & Cozzi, 2016).
2.5.2. Climate Change
Climate change is another major challenge to ecosystem restoration (Charlton et al., 2018; Forber et al., 2018; Meier et al., 2018; Rankinen et al., 2016; Sinha et al., 2017). In general, climate change is expected to result in increased air and water temperature and an acceleration of the water cycle (Bloschl et al., 2017; Milly et al., 2005; Najjar et al., 2010; Rice & Jastram, 2014; Rice et al., 2017), which can alter the volume transport of freshwater and inputs of nutrients and sediments. For example, Sinha et al. (2017) estimated that climate‐change‐induced precipitation changes alone will substantially increase (19 ± 14%) riverine inputs of TN within the continental United States by the end of the century. In addition, the effects of climate change can differ among seasons. For CB, projected acceleration of the water cycle is expected to increase river runoff and associated inputs of nutrients and sediments during winter–spring and to decrease runoff during summer–fall (Wagena et al., 2018). Thus, management strategies for CB need to account for the impact of projected climate change on water quality. In this context, modeling and assessment is underway in the Chesapeake Bay Program partnership to evaluate the effects of climate change on nutrient export, efficacy of best management practices, and water quality in the estuary.
In contrast, climate‐driven changes in the water cycle in the NAS watershed may tend toward persistent periods of low runoff alternating with episodic events of high discharge. Climate change appears to be increasing the frequency of heavy precipitation events (Alcamo et al., 2007). This has yet to induce long‐term changes in the annual discharge of the Po River, which has greatly oscillated over the past three decades without showing clear trends (Cozzi & Giani, 2011). However, flow dynamics of Po River are characterized by a shift towards early spring peaks of runoff (Zampieri et al., 2015) and a decline in summer flows (Cozzi et al., 2019). At the same time, the other NAS rivers have shown a strong reduction in flow (Cozzi et al., 2012). It is important to note that reductions in river flow can result from both a greater anthropogenic use of continental waters as well as from climate‐driven changes. For example, annual runoff to Adriatic rivers of Slovenia were reduced (6%) in 1971–2000 due to increased evapotranspiration of the soils (11%), even in the presence of relatively constant precipitation (Frantar, 2007).
2.5.3. Reservoir Filling
A major challenge that is unique to CB is the filling of the Conowingo Reservoir of the Susquehanna River which has neared its sediment storage capacity after 90 years of operation. As sediment accumulates in this reservoir, the cross‐sectional area available for flow, and the vertical depth from water surface to sediment bed, decreases, thereby increasing the average horizontal flow velocity. Consequently, sediment trapping by the reservoir decreases and sediment load to CB increases. Numerous studies have demonstrated the declining trapping performance of this reservoir in recent decades (Hirsch, 2012; Langland, 2015; Zhang et al., 2013; Zhang, Hirsch, et al., 2016). Moreover, Zhang, Hirsch, et al. (2016) reported that such decline in reservoir trapping has occurred under a wide range of flow conditions. These changes, if not addressed, can hinder the attainment of the Chesapeake Bay Total Maximum Daily Load goals because the reservoir was expected to continue trapping sediments and nutrients at historical rates for another 20–30 years when those goals were established in 2010. Thus, the Chesapeake Bay Program partnership has worked to incorporate recent scientific understanding in upgrading its watershed model to better capture the temporal changes in reservoir function (Linker, Batuik, et al., 2013; Shenk & Linker, 2013), which will be used to adjust the goals of nutrient and sediment reductions by each jurisdiction.
2.6. IMPLICATIONS AND RECOMMENDATIONS
Anthropogenic riverine inputs of N, P, and sediment have led to undesirable consequences in the coastal marine environment, including eutrophication and associated oxygen depletion, declines in water transparency, loss of submerged aquatic vegetation, and shifts in the composition of plankton communities (Boesch et al., 2001; Breitburg et al., 2018; Cloern, 2001; Degobbis, 1989; Diaz & Rosenberg, 2008; Giani et al., 2012; Kemp et al., 2005). Therefore, reduction of watershed inputs has been a management priority for many coastal marine systems, including CB and the NAS. A review of parallel time‐series data on hypoxia and watershed loading rates in coastal ecosystems shows that oxygen conditions tend to improve rapidly and linearly when the primary driver targeted for control is nutrients from wastewater treatment plants (Kemp et al., 2009). In larger more open systems, where nonpoint nutrient loads are more important in fueling eutrophication, responses to remediation tend to be nonlinear with hysteresis and time‐lags. Nonetheless, there have been some signs of ecosystem recovery. For CB, water quality improved with time during 1985–2016, which is statistically linked to the reduction of riverine inputs of TN (Zhang et al., 2018). For the NAS, the reduction of riverine loads of P has been an effective method to alleviate eutrophication, even with high inputs of N and silicates (Djakovac et al., 2012; Giani et al., 2012). However, ecosystem conditions in this posteutrophic phase are still not comparable to those in pristine environments due to the occurrence of hypoxia and degraded benthic habitats in shallow coastal zones (Alvisi & Cozzi, 2016; Stachowitsch, 2014). Thus, continued reduction of watershed loads is indispensable for both CB and the NAS.
After decades of management efforts, the goals of CB and the NAS restoration have not yet been fulfilled (Volf et al., 2018; Zhang et al., 2018). Moving forward, we provide the following recommendations:
continue monitoring river flows and water quality in the major tributaries to CB and the NAS;
improve statistical approaches for quantifying riverine constituent loads and trends, including associated uncertainties;
increase understanding of watershed factors that influence riverborne loads and trends (e.g., land use, hydrology, source controls) and their relative importance;
develop consensus and solutions among stakeholders to address the major challenges that hinder the achievement of restoration goals in a timely fashion (e.g., legacy sources, climate change, and reservoir filling);
increase understanding of the effects of land‐based inputs on downstream water quality and ecological responses (e.g., dissolved oxygen, water clarity, chlorophyll‐a);
enhance public awareness of the impacts of anthropogenic nutrient loading, management goals and actions, progress toward achieving these goals, and major challenges.
In a world with seemingly ubiquitous nutrient enrichment and water‐quality degradation, past and future advancement in our scientific understanding on these two coastal ecosystems can be valuable resources that may guide and facilitate the protection and restoration of estuarine and coastal ecosystems in other geographical locations.
ACKNOWLEDGMENTS
The authors thank many institutions for making the river monitoring data available,
