Extratropical Water Level Forecast User Manual

To use this website, click on the appropriate REGION. This will update the list of STATIONS and show a "Status Map" for that region. Click on your desired station, either on the map or in the list of STATIONS. This will bring up a graph of the total water level, as well as a text file that contains the numbers used in the graph.

The graph combines several sources of data to produce a total water level prediction. To do so, it graphs the observed water levels in comparison to the predicted tide and predicted surge before the current time. This allows it to compute the "Anomaly". The "Anomaly" is the amount of water that was not predicted by either the tide or the storm surge model. This "Anomaly" is averaged over 5 days, and is then added to the future predictions of the tide and storm surge to predict the Total Water Level.

The first thing one notices is that there are two magenta vertical lines. The earlier one is when the storm surge model was run. It is run at 0Z and 12Z every day and the text form is available at: https://www.nws.noaa.gov/mdl/marine/etsurge.htm. The later magenta line is when the graph was generated. It is currently being generated 15 minutes after the top of every hour. (This is also the date that follows the label.)

The next thing one notices are the horizontal lines labeled MLLW, MSL, MHHW, and HAT. These stand for the Mean Lower Low Water, Mean Sea Level, Mean Higher High Water, and Highest Astronomical Tide. HAT was computed using our tide model, by computing the maximum of the predicted value for every hour (on the hour) for 19 years. The thought is that there is probably flooding if the total water level crosses HAT. The other datums came from https://www.co-ops.nos.noaa.gov/data_res.html.

One might next notice the red observation line. This is based on data attained from Tides Online . Please see their Disclaimer for information as to the quality of these observations. If there is no red line, then either Tides Online does not have data for that station, or there has been a communications break down. In this case, the graph computes an anomaly based on what data it has, or sets it to 0. Then it predicts the total water level for all hours, or after the last of any observations it does have.

The next thing of interest is the green Tide line. This is the astronomical tide at every hour. The Harmonic Constants used were obtained from https://www.co-ops.nos.noaa.gov/data_res.html.

We then note the gold storm surge curve, which is created by "pasting" one 48 hour prediction to the next 48 hour prediction. That is, using 12 hours from each prediction until the last prediction where we use 48 hours. The result is that we may generate kinks in the curve every 12 hours, where the model adjusted its prediction based on new data from the GFS wind model.

Next we note the blue curve, which is the "Anomaly" referred to above. This is simply the observation - (tide + storm surge). Preferably it is constant. The amount of deviation from a constant is an approximation of our error. Since we add the 5 day average of this value to our prediction, the perfect forecast does not have to have a zero Anomaly.

Finally we see the black forecast curve. This is what we are really interested in, which is the total water level created by adding the 5 day average anomaly to the predicted tide, and the predicted storm surge.

• We plot a green dot if in the particular time span (usually next 48 hours) the total water level stays below Mean Higher High Water. This will only happen for half the month as Mean Higher High water is an average, so half the time it should be above this.
• We plot a yellow dot if the total water level stays below Highest Astronomical Tide (HAT). Assuming there are no storms, then it should be in this region for the other half of the month. However the Anomaly may be large enough at a particular station to force it into our next grouping.
• We plot a red dot if the total water level crosses HAT somewhere in the time span. These stations are where we are predicting potential flooding. To see how far above HAT it is, click on the station, and then in the comment section of the text file there is a number for feet above HAT.

The history button allows one to see how the model has done over the last day or so. It displays 3 graphs. The first one is the current graph based on the current model run, and the current observations. The second graph is the last graph generated using the last model run. The third graph is the last graph generated using the next to last model run. This gives a view of the model over the last 24 to 36 hours depending on when the current time is.

To print this page out (Netscape instructions) it is recommended that you right click on the history frame and choose "Open Frame in New Window". Then choose page setup, and set the top and bottom margins to 0. Then choose print, and preferably send it to a color printer, (although a black and white does work). The result should be 3 graphs on the same page.

We would like to thank the following people/organizations:

• Web page development:
  • Arthur Taylor, Anne Kramer, Wilson Shaffer, Mark Stram
• Graph Program:
  • Arthur Taylor, Anne Kramer
• Extratropical Storm Surge Model Development:
  • Jye Chen, Chester Jelesnianski, Sung Kim
  • GFS Model.
• Observations:
  • National Ocean Service (Tides Online) https://tidesonline.nos.noaa.gov/
• Tide Model.
  • Art Pore, Kurt Hess, A. Gilad
  • National Ocean Service (for tidal constituents)