Open ththelen opened 2 years ago
Summary presentation with model run background, measured water levels from Yacht Basin, and modeled maximum water levels. Timing of peak water levels does not seem to match between ADCIRC and measured values
CD: To investigate your questions, we will want to look at time series of the water levels and winds at observation stations, to see how the quality of the input tides and atmospheric forcing. I checked the NAM site, and I am confident that their atmospheric products are relative to UTC. But they are low resolution -- (about 0.1 degrees in space, 3 hr in time), so it is possible that they are not representing (or 'smearing') an atmospheric driver.
CD: I found the error -- I mistakenly used 2011 (instead of 2021) when setting up the tidal constituents for the ADCIRC simulation. This explains the errors in magnitude and phase that you found. Let me re-run the simulation -- it is an easy change in the input files, but it may take a few days to push the simulation through.
CD: In the meantime, if you have time to process the bad simulation, then I will be curious to see time series comparisons for the winds -- the error in the tidal forcing should not affect the winds.
Thomas, I put an updated set of model results on the cluster here: /share/jcdietri/For/Thomas/220610-NC9-Nov2021/ Can you please (a) copy these files into your archive, and then (b) investigate whether they are a better match to the extreme water levels during Nov 2021?
Peak water levels from the new ADCIRC run match peak water levels in the Yacht Basin to within half a foot (4.4 ft per ADCRIC vs 4.74 ft per Yacht Basin sensor). Temporally, the peak modeled water level is within one hour of what we would expect based on the measured values. We also start to see some normally dry mesh cells flooded with the new simulation results.
I will pull time series results at the CB, Wilmington, and Wrightsville Beach stations to see how well we match water levels and winds measured at these locations. Moving forward it still evident that we will need to update both the Yacht Basin / tidal inlet bathymetry and surrounding topography to get better results. It may also be necessary to improve the resolution of the mesh in these locations. I will start tasks for these two items
What datum are the modeled water levels relative to? @caseydietrich will follow up on this
ADCIRC misses long term drivers like the gulf stream. May have to adjust for this, eg use a constant offset
CB Nodes used for WL timeseries interpolation circled in blue. Not convinced that these are the 3 closest nodes to the sensor (circled in red), but they are in the right general area
CD: For this simulation, I used atmospheric forcing from the NCEP North American Mesoscale (NAM) Forecast System, specifically the NAM Analysis products from Oct-Nov 2021. I ran a spin-up simulation for the entire month of Oct, and then these results are from a 14-day simulation starting 0000 UTC 1 Nov and ending 0000 UTC 15 Nov. ADCIRC writes its first output after the first interval in the simulation, so the first snap in the result files will correspond to 0100 UTC 1 Nov. Then they are hourly after that.
I'm curious how well this simulation will represent the king tide and the offshore low. The king tide should be represented, because we are reconstructing the tides from their harmonic constituents at the ocean boundary in our model. Not sure about the offshore low -- this should be included in the NAM products, but it will be at low resolution (0.1 degree in space, 3-hr in time). ADCIRC does not include the Gulf stream or other oceanographic processes, not sure how important they are for your application.
Can you please load these files (especially the water levels in the maxele.63.nc and fort.63.nc files) into SMS and visualize the results near Carolina Beach? Then we should talk this week about how to pull time series for hydrographs at specific locations.