Search

A high‐resolution wave model is crucial for accurate modeling of sediment and organic material transports, but its computational costs hinder direct coupling to an ecosystem model. We developed a machine learning model using long short‐term memory to simulate large‐scale, high‐resolution waves. Trained with numerical wave model (NWM) outputs and wind data from nine locations, our model successfully replicates NWM results for daily mean significant wave height and period in Chesapeake Bay with identical spatial resolution. Compared to the NWM, the data‐driven model has root‐mean‐square errors below 6 cm for daily mean significant wave height and 1 s for the wave period in the bay. It demonstrates excellent model skills and can accurately forecast daily mean significant wave height and period at NOAA wave stations comparable to NWMs. Using minimal wind data and having a short runtime, our data‐driven model shows promise as an alternative for wave forecasting and coupling with sediment and ecological models. Plain Language Summary: Accurately modeling sediment and organic transport necessitates the use of a high‐resolution numerical wave model (NWM). However, the computational expenses of running an NWM within an ecosystem model pose significant challenges. We conducted a study using a machine learning approach based on long short‐term memory to replicate the results of an NWM. The data‐driven wave model (DWM) was trained with the NWM outputs and wind data from nine locations. The trained DWM can replicate the NWM results of daily mean significant wave height and period in Chesapeake Bay with identical spatial resolution and the same predictive skill with a root‐mean‐square error of less than 6 cm and 1 s, respectively for the model domain. The model successfully forecasted daily mean significant wave height and period at NOAA wave stations, showing great forward prediction capability. Notably, the DWM only relied on a minimal amount of wind data from selected locations and had a short runtime. These findings indicate that machine learning is promising as an alternative approach to NWM for wave forecasting. Moreover, it can serve as an alternative wave model and be coupled to ecological models for simulating sediment transport, light attenuation, and resuspension of organic materials. Key Points: A machine learning‐based wave model reproduces the numerical wave model (NWM) results at the same spatial resolution with good accuracyThe data‐driven model, trained on daily significant wave height and period using NWM outputs, has comparable wave prediction skillThe fast‐running data wave model, driven by wind data from nine locations, can forecast waves and couple with sediment and ecological models [ABSTRACT FROM AUTHOR]

Feedbacks between sediment and plants in the submersed aquatic vegetation (SAV) beds of the Susquehanna Flats help modulate sediment delivery into the upper Chesapeake Bay from its major tributary (Susquehanna River). Recent modeling work has shown that SAV can steer the river plume, directly controlling sediment transport and fate. However, transport mechanisms likely differ between flood and non‐flood conditions and depend on event timing. An opportunity to evaluate these insights in the field occurred with several flood events in summer 2018. Sediment and biomass samples were collected after the flood in the large, continuous main SAV bed of the Flats and in smaller patches to the west of the main bed to evaluate the effect of SAV bed size (width). Sediment characteristics (grain size, organic content) and deposition rates were compared to previous observations during non‐flood conditions in the main bed in 2014–2015 and in the smaller patches in 2019. Overall, sediment deposition rates could be predicted by an empirical model including sediment load, vegetation biomass, and SAV bed width, reflecting the role of vegetation in steering the Susquehanna River plume and modulating sediment input to the upper Chesapeake Bay. Understanding spatiotemporal patterns of sedimentation is essential for realistic estimates of sediment and nutrient (especially nitrogen and carbon) storage in freshwater SAV systems, which have not received as much attention as their saltwater counterparts. [ABSTRACT FROM AUTHOR]