Update topography and bathymetry at / around yacht basin (YB) and tidal inlet (TI); visualize results from updated topobathy

[ththelen]

This dataset looks like a good option for topo/bathy elevation data. It is a new topobathy dataset hosted on digital coast

[ththelen]

CoNED dataset has what looks to be pretty good topography and bathymetry data at 1 m resolution. Looks like an ideal dataset to use for mesh enhancement.

[ththelen]

@caseydietrich how difficult will it be for me to learn how to update elevation data in the NC9 mesh? Is this something that we should cover in the next mesh bros or SunnyD modeling meeting? Alternatively, should I seek out help from Jenero and Carter on this?

[caseydietrich]

Thomas, let's discuss tomorrow in one of our meetings. You want to update an existing mesh, so I think it will be better if you learn and use SMS. It will allow you to be more precise in selecting an area to update. OceanMesh2D is better for generating meshes from scratch, but it is not as interactive for what you need. Thus it will be helpful to stay connected with Jenero and Carter, but their experience will be different from yours.

I'm attaching a tutorial from the modeling course about how to update a patch within an existing mesh. This tutorial is detailed, with step-by-step instructions and screenshots, but it is also out-of-date, so it will not match the latest SMS version you are using. Please look over this PDF, and then let's discuss in our meeting tomorrow.

[Uploading Tutorial5-SMS-Willapa-Bay.zip…]()

CE596-Unit6-4.pdf

[ththelen]

I investigated the topo and bathy data from the CoNED raster that I proposed to use to update mesh elevations. My conclusion is that both the topographic and bathymetric data is reasonable. We can confidently use this raster to update our mesh, with the exception of areas around dockage. Full analysis with screen captures is available in the linked powerpoint.

CoNEDRasterInvestigation.pptx

[ththelen]

My follow up question for @caseydietrich is what is the vertical datum used in the NC9 ADCIRC mesh run? This mesh has a depth of -6 m in the YB channel and 0 to +2 m east of the islands in the YB, but I do not know what datum this is relative to. See below for a screen capture of the mesh elevations.

[caseydietrich]

Thomas, the NC9 mesh has ground surface elevations relative to NAVD88. Hopefully you will be able to get the CoNED data in the same datum.

Let's discuss today about how to update the mesh in SMS. I may need to join you at your computer in the lab, or we might be able to do a remote desktop from my office. My first suggested method will be to:

1. Leave the horizontal layout of the mesh as-is.
2. Load the CoNED DEM into SMS. I don't know how to do this step, but we can look at it together, and Jenero/Johnathan may have ideas.
3. Hand-select the vertices to update near the YB. You can draw a polygon in the Map module and use it to select mesh nodes, so the process is repeatable.
4. Use SMS to interpolate the bathymetry from the CoNED DEM. This will do a nearest-neighbor interpolation, which is not ideal.
5. Hand-correct the bathymetries at vertices in channels.

[ththelen]

Hi @caseydietrich, I have corrected vertices in the yacht basin (YB), tidal inlet (TI), and intracoastal (IC) as we discussed yesterday. The fort.14 file with corrected vertices is uploaded to a google drive folder that should be shared with you: https://drive.google.com/drive/folders/1OXwUXLjxFMOyY1R8jCaO8RGp-E9vwGsB?usp=sharing

Before I leave for Canada at the end of next week I hope to have comparisons of winds at non-CB locations and average water levels comparison for Nov 9-15. This way you should have everything you need to inform the next ADCIRC run. As we discussed, I will not have time to do any work with the new ADCIRC results before I get back from Canada on July 25, so there is no big rush on completing the run.

[caseydietrich]

@ththelen Thanks for sharing this fort.14 file. However, when I load it into SMS and compare it with the existing NC9, I do not see any differences in the bathymetry. Can you please verify that the bathymetry has been updated in this mesh?

[ththelen]

Hi @caseydietrich, apologies for not checking the fort.14 file before sending it your way. It seems like the export functionality for SMS 13.1 is glitchy - updates to the mesh are not reflected in the fort.14 file exported from SMS 13.1. I downloaded the beta version of SMS 13.2 and the ADCIRC export functionality seems to be fixed. Please find the updated mesh file in the google drive: https://drive.google.com/drive/folders/1OXwUXLjxFMOyY1R8jCaO8RGp-E9vwGsB?usp=sharing . Let me know if you find any other issues.

[caseydietrich]

The first leg of the simulation (start-up for the first 16 days in Oct 2021) is running now, no obvious problems. Good job to edit the bathymetry!

I'm attaching screenshots to compare the bathymetry in the original and revised NC9 meshes. The channels got deeper, but maybe also narrower in spots. In the revised NC9 mesh, note the 'stair-step' pattern in the connection on the west side toward the Cape Fear River, and the narrower Snow's Cut in a few locations including on its east side toward the IWW. It will be interesting to see the effect of these changes.

[ththelen]

That is an interesting comparison. Mesh depths certainly look less idealized in this second run. I can go back and check the area around the stair step pattern to make sure my cherry-picking of nodes to update did not cause the strange bathymetry there. Thanks for starting the run!

[caseydietrich]

@ththelen, this simulation finished over the weekend. I posted the files on the cluster here:

/share/jcdietri/For/Thomas/220707-NC9rev-Nov2021

Can you please (a) copy them into your archive, and then (b) post-process and visualize them? It will be interesting to see the effects of the deeper channels!

[caseydietrich]

@ththelen ... when you return from your travels, of course! :)

[ththelen]

@caseydietrich I have some insight into the 'stair-step' pattern in the connection on the west side toward the Cape Fear River. The mesh nodes are pretty sparse in this location. The 'stair-step' pattern results because only a couple of the nodes actually coincide with the middle of the channel. This may be a location where we either have to add more nodes or move the existing nodes so that they are located in the center of the channel. Note that in the figures below the solid black lines indicate the boundaries of where I have updates mesh Z values with elevations from the CoNED DEM

The narrower Snow's Cut in a few locations including on its east side toward the IWW that you noted does appear to be accurate to be an accurate reflection of the CoNED DEM I used to updated the mesh. I am just starting to visualize your most recent results. I should have some results to share in the next couple days.

[ththelen]

Hi @caseydietrich and @anardek, I have a time series plot of water levels in the YB with the updated bathymetry to share with you. WLs from the latest ADCIRC run with updated topobathy are displayed in pink, FIMAN YB gauge WLs are shown in blue, and WLs from the original unedited NC9 mesh are shown in yellow for comparison. Some observations:

1. The updated bathymetry improved the agreement of the timing of high and low tides compared to the unedited mesh. This seems reasonable since we deepened the channels that lead into and out of the YB, making the connectivity between the modeled ocean and intracoastal/YB system more similar to the real world system.
2. We see higher maximum water levels in the YB for the days with the highest winds (Nov 6-8). I suspect this may lead to us overpredicting maximum WLs once we add in the fixed ADCIRC WL offset.
3. While the maximum WLs on high wind days are higher than the unedited mesh, high tides on non-windy days are lower in the revised mesh compared to the original run.

It will be interesting to compare these results to the ADCIRC run on the original mesh that CD is doing with fixed WL offset. We can discuss these results further in our 2:00 pm meeting today.

[ththelen]

@anardek and @caseydietrich here are some answers to questions that arose in yesterday's modeling discussion:

Q. Are the water level measurement spikes on Nov 2 and 5 due to rain, or are they noise from interference with the ultrasonic water level sensor? A. There was no rain recorded at the CB gage Nov 1-5 (see FIMAN snip). This leads me to believe the hypothesis that the jumpy measurements are noise associated with an object coming between the sensor and the water level. Also, I cannot think of a natural phenomenon short of a hurricane that could cause a 1.5 foot increase in water level in the YB in 20 or so minutes.

Q. Does the model show a water level gradient in the YB when we see peak winds out of the north on Nov 7? A. Kind of, the ADCIRC results show a relatively consistent water level gradient of 0.02 m from the north to south end of the YB during the morning of Nov 7. The screen capture below showing time = 14:00 UTC on Nov 7 below has been annotated to highlight the location of the 1.40 to 1.42 WL contours. Note that the scale in the top left shows water level in meters relative to NAVD88.

Q. How do tides propagate through the Cape Fear, Snow's Cut, intracoastal, tidal inlet, yacht basin system? A. It seems like the first avenue for the rising tide to reach the yacht basin is via the tidal inlet, but it is close between the tidal inlet / intracoastal route and the Cape Fear River / Snow's Cut propagation pathway. Recall that this model does not yet have a BC forcing flow down the Cape Fear river, so the modeled tidal propagation up the Cape Fear may not be quite accurate. That said, feel free to watch the tidal propagation videos (one at really fast speed and one slowed down) in the folder linked here and draw your own conclusions: https://drive.google.com/drive/folders/1P9aYPhVBz6Z_pMrhNebQvxY3quPLSy_v?usp=sharing

[ththelen]

Hi @anardek and @caseydietrich, I have some time series that should put the resolve our tidal propagation questions. The map below shows the locations from which all the data were pulled.

The three plots below are, from top to bottom:

1. Tidal propagation through the CB tidal inlet to the Yacht Basin entrance. Intermediate time series between the two bounding locations are the confluence of the inlet with the intracoastal (TI - Confluence) and halfway between the tidal inlet and the yacht basin along the intracoastal (TI - Intracoastal). This propagation pathway controls water levels at the YB entrance, although water levels at the YB entrance are attenuated by about 0.1 ft compared to open ocean conditions. This is likely a result of shallow water depths in the tidal inlet slowing transmission of tidal flows. However, it is evident from the plot the when water levels at the ocean side of the mouth fall below YB water levels, YB water level also start to recede. This indicates that the CB tidal inlet is the primary pathway for tides to propagate into the YB. I also note that water levels in the YB (see gray line 20220727-NC9rev-Nov2021 pulled from a location bumped just off the FIMAN gauge) are identical to water levels at the mouth of the YB for the duration of the tidal cycle. This is in line with our finding that there is not much of a WL gradient across the YB, and will be an important assumption in coupling the ADCIRC and 3Di model.

2. Tidal propagation up the Cape Fear River and through Snow's Cut to the Yacht Basin entrance. Intermediate points are denoted on the map. We see that water levels at the YB entrance actually peak before water levels at the Hohonu located on Snow's cut bridge. Tidal propagation coming up the Cape Fear is not the main propagation pathway at the YB.

3. Tidal propagation through the Masonboro channel that is armored with jetties and presumably dredged for navigation. Intermediate points are denoted on the map. Tidal propagation from the Masonboro channel up the intracoastal from CB is actually a more significant contributor to the tidal signal in the YB than the Cape Fear Snow's cut propagation pathway. However, we still observe that the tide at the CB yacht basin peaks after the water levels further north in the intracoastal start to recede. That said, we see that the rate at which the tide is going out at the YB accelerates once water levels further north in the intracoastal go below YB water levels.

Final takeaway: The propagation pathways have the following order of influence on contributions to WLs in the CB yacht basin (most influence to least influence):

1. CB Tidal Inlet
2. Masonboro Jettied Channel
3. Cape Fear River / Snow's Cut

[ththelen]

@anardek @caseydietrich, as promised, here is a follow up to our questions about how the CB tidal inlet is represented in the NC9rev mesh.

Figure 1. 2019 imagery of the tidal inlet. ADCIRC vertices are denoted with yellow dots. Elevations at vertices within the green polygon were updated with bathymetries from CoNED

Figure 2. Partially transparent CoNED elevations overlaid on the tidal inlet. Note that the color scale indicated in the box at the upper left goes from -16 m to 12 m NAVD88.

Figure 3. Partially transparent NC9rev mesh elevations overlaid on the tidal inlet. Note that the color scale is the same as Figure 1. I have tried to match the transparencies as closely as I can, but the transparency on the raster layer was not specified so they may not match exactly.

https://drive.google.com/file/d/1IDkoTPW3KijOcH4zlLvx9mx-u0kTi0c7/view?usp=sharing Video 1. Modeled tidal propagation through the CB tidal inlet on November 1 and 2. Note that the color scale shows water surface elevation from -1 m to 1 m NAVD88.

There is one location in the inlet where the inlet is only a single element wide at low tidal levels. Maybe this is an inaccurate representation of the real world? That said, it is hard to say if water is getting stuck at this narrowing of the channel or before as shoaling occurs on the sandbar. I am interested to hear your thoughts on this.

[caseydietrich]

@ththelen Thanks, these figures and video are interesting. I agree that the one-element-wide section is not ideal, but it seems like the choke point is closer to the ocean, as the tides are moving over/past the sandbar (roughly at the coordinates -77.874676, 34.080753). Figures 2 and 3 are labeled as being elevations from CoNED -- is this correct? Do you have an figure to show the interpolated ground elevations in your revised NC9 mesh? I'm curious to know if we are aliasing the bathymetries in that part of the inlet. If so, then you could hand-edit the bathymetries at a few vertices, to deepen the inlet.

[ththelen]

@caseydietrich agreed that the choke point is closer to the ocean. Figure 3 actually shows the NC9 mesh elevations - I just forget to update that part of the caption when I copied and pasted, oops. I have edited the caption such that Figure 3 is now correctly labeled as NC9rev elevations. I am not convinced that we are dramatically aliasing bathymetries, but if you think it is worth hand editing a few vertices I could do that when I add the Cape Fear River BC and fix the stairstep pattern in the channel connecting to the Cape Fear River.

[ththelen]

Hi @caseydietrich and @anardek I was able to hone in on the tidal propagation sticking point in the inlet. The thalweg through the channel with CoNED bathymetries between -3 and -4 m NAVD88 for most of the channel length is shown in Figure 1. However, there is one location circled in red in Figure 1 where the bathymetry jumps up to about -1.3 m NAVD88. This jump in the bathymetry is reflected in the NC9rev mesh (Figure 2), and is likely responsible for hindering tidal propagation. It is difficult to say if the bar in the channel is an accurate reflection of real world conditions, but I see nothing in the latest Google Earth imagery to suggest that this jump in bathymetry actually exists.

If we remesh and add more vertices using elevations the CoNED DEM, we will still have the bar in the middle of the channel because it is present in CoNED. To achieve more tidal propagation through our inlet, we will have to deepen the channel beyond what is reflected in CoNED. I think this is a reasonable step to take because increasing tidal propagation through the inlet seems like it might address the low peaks and high valleys we were seeing in our modeled WL results compared to FIMAN data.

I propose hand-editing a couple of the shallowest elements along the channel to bring them to -3 m, a depth consistent with the rest of the tidal channel. This seems like a good first pass at improving tidal connectivity before we put in the effort to remesh the whole area. I can also bump up a few of the elevations along the section of beach that was being overwashed in our ADCIRC model to block off this flow route. We could look at whether these simple hand-edits improve the agreement with our tidal propagation in the Yacht Basin, then decide whether a more extensive remeshing effort is warranted. Let me know what you think about this proposal.

Figure 1. CoNED bathymetries from -12 m to 3 m NAVD88.

Figure 2. NC9rev mesh bathymetries from -12 m to 3 m NAVD88.