DSI has released a white paper documenting two approaches for simulating wind waves in EFDC+:
- The internal SMB (Sverdrup-Munk-Bretschneider) method
- The external SWAN (Simulating WAves Nearshore) method
Both were applied to the same wind-wave sediment-transport model: a miniature sloped beach with two submerged offshore sandbars separated by gaps (rip channels), as shown in Figure 1. The bathymetry stems from the numerical rip-current experiment of Svendsen, Haas, and Zhao (“Analysis of Rip Current Systems”). The white paper details the model setup, sediment properties, the differences in how each method computes wave properties, and a discussion of the results and their limitations. This blog post offers a brief preview of that work.
RIP Current Circulation Field
Animation 1 presents the depth-averaged velocity magnitude (color) with the velocity vectors overlaid, for SWAN (left) and SMB (right) simulations. Both models produce the same characteristic rip-current system: longshore feeder currents converge toward the gaps in the sandbars and turn seaward as narrow, high-velocity jets, flanked by counter-rotating circulation cells. The location, structure, and magnitude of these jets are closely in agreement between the two models, indicating that EFDC+ produces a consistent hydrodynamic response regardless of the wave model used.
The simulated wave fields are also in good agreement. Both methods develop a wind-driven wave field that increases along the wind direction toward the shore, and the resulting wave heights are of similar order in the two models.
Morphological Response
Animation 2 shows the change in bed elevation (bed delta; blue denotes erosion, red denotes deposition) with the depth-averaged velocity vectors overlaid, for SWAN (left) and SMB (right) simulations. Although the depth-averaged velocity field remains nearly identical between the two models, the bed-delta fields differ substantially. Under SMB, the erosion and deposition bands are pronounced, with the sandbar crests eroding and sediment moving towards the offshore. Under SWAN, the sediment also moves offshore but the bed changes only modestly over the same period.
Wave-Induced Bed Shear
The origin of the differing morphological response is evident in Animation 3, which pairs cross-shore profiles (left) with the plan-view wave-induced bed shear stress (right); SMB is shown on the top row and SWAN on the bottom. Two differences are apparent.
First, the wave fields diverge near the shore. In the SMB model, wave height continues to increase slightly towards the shoreline. In the SWAN model, wave height levels off offshore and then decreases slightly over the sandbar and again in the surf zone, reflecting depth-induced wave breaking, a process represented by SWAN but not by SMB in this test.
Second, the wave-induced bed shear stress is considerably larger in the SMB model, approaching the upper limit of the plotted scale near the shore, whereas the SWAN simulation produces considerably lower bed shear stress. The higher and longer waves computed by SMB transfer more energy to the bed, producing stronger shear stresses and, in turn, the more pronounced sediment transport seen in Animation 2. This difference is also reflected in the profiles of Animation 3: the SMB bed reworks appreciably over the simulation, while the SWAN bed remains nearly unchanged.
For the full model specifications, methodology, and discussion, please see the complete white paper on the EEMS website.
