Case Study 1: Geography & Sea Level Change in New York City

Modeling Short-Term Sea Level Change & Investigating Coastal Geography

In order to understand sea level change in greater depth, geoscientists need to differentiate individual factors and their potential impacts. So, like any equation in mathematics, geoscientists will use background data as a baseline for developing a predictive algorithm that explains or "fits" the majority of the data. They will then modify the base algorithm by adding and changing one variable at a time to see how that one variable changes the outcome of the composite system. This approach is often called modeling. The ultimate goal of modeling is to produce a predicted or forecasted outcome that is as close to the actual observed outcome as possible.

For an example to illustrate the last point... Let's visit Stony Brook University's Storm Surge Research Group. Click http://stormy.msrc.sunysb.edu/ The website will open in a new window.

As demonstrated on their website, the Stony Brook research group is investigating the factors that might influence actual observed sea levels in the New York Bight region which ranges from Sandy Hook, NJ through the mouth of the Hudson River and around Long Island, NY. When you go to their website, you can click on the observation areas to look at data and model forecasts.

×

Figure 4.8: Hydrographs for Atlantic Beach, New York for January 14, 2014.

Credit: SoMas: screen clipping taken 1/14/2014 11:25 AM

In the screenshot above, taken for Atlantic Beach on Long Island, you will see a hydrograph similar to one you have already studied. In this case, water levels are again plotted relative to mean low-low water (MLLW), and once again red line shows the observed measured readings for 1/14/14. For this date, observed water levels are higher than average for the region. How did this compare to the same interval for San Francisco that we discussed previously? In this case, the forecasted water level was below what was actually measured, suggesting that another but potentially related factor is at play.

In order to unravel these other factors, Stony Brook scientists raise additional ideas. The graph also shows lines for:

1. predicted water level based on historical averages;
2. astronomical factors (more on this shortly);
3. a calculated standard deviation (shaded) line.

Observations and Analysis

The shaded plot is centered around the predicted values which indicates that the majority of historic observations are statistically concentrated around the average; thus, extreme events clearly exceed the standard deviation from time to time.

Clearly, the water levels observed on this date are not explained by normal periodic processes (i.e., tidal variations). The data suggest the observed values are outside of normal "average" events, but given the uppermost plot, the tidal variation obscures the signal of the anomalously high water so that is explored in more detail in the lower plot.

To help visualize the extent to which the observed data is above average, Stony Brook geoscientists have produced the graph labeled "Surge Plot for Atlantic Beach, NY" (the lower of the two plots on Figure 4.8 above).

The concept is similar to the residual purple line on the graph for San Francisco that you looked at earlier. In this plot, they have subtracted out the periodic tidal signal to isolate and reinforce the signal of the observed surge.

In this case on 1/14/2014, the highest water level (surge) occurred at 7:15 a.m. and was 0.31meters (~1 foot) above normal tide level.

Explanation? For the sake of examining this quickly, let's take a look at NOAA's National Data Buoy Center website: http://www.ndbc.noaa.gov/ to get more information about weather in the region. See Figure 4.9. The NOAA NDBC website allows you to zoom into the New York/New Jersey Bight area, where weather readings are being taken at Station 44065.

This weather station is located just offshore from Atlantic Beach, and due to real-time telemetry, the data are made available and updated at set time increments.

In this case, this station's pop-up window shows winds are out of the south-southeast (170 degrees) and are blowing at 7.8 kt (knots) which is approximately 9 miles per hour (mph), and barometric pressure is falling.

This observation confirms that the winds are blowing water toward land and may be contributing to higher than normal water levels, but 9 mph winds are not that strong, so let's look at the last 5 days' worth of data to see if we can substantiate this further (see Figure 4.10).

×

Figure 4.9: Screen clipping from NOAA's NDBC website showing the location of an oceanographic data collection buoy (Station 44065). The map pop‐up window indicates current weather and water conditions for the date and time and provides links to view details and history of weather data taken at that location.
Credit: NOAA: Screen clipping taken: 1/14/2014 12:00 PM

×

Figure 4.10: Wind speed and air pressure at station 44065. Note the increase in wind and decrease in atmospheric pressure (low pressure). These data indicate passage of a storm from 1/11/2014 to 1/13/2014.
Credit: NOAA/NWS/NDBC: Screen clipping taken: 1/14/2014 12:08 PM

In this graph, observations of wind speed, wind gusts, and air pressure are plotted for the five day interval (From 1/9 to 1/14 Eastern Savings Time). So, it is clear that a storm event with high winds and low atmospheric pressure moved through the area on 1/11 and 1/12/14. A short-term higher-pressure event moved through early on January 13, but this appears to be followed by another storm that was developing at the time of observation on 1/14/14.

Given the observations above, the storm-surge level observed in the Stony Brook data for 1/14/14 (Figure 4.8) was most likely produced by the storm event with associated onshore winds and atmospheric low pressure.

In order to understand sea-level change in greater depth, geoscientists need to differentiate individual factors and their potential impacts. Like any equation in mathematics, geoscientists use background data as a baseline for developing predictive algorithms or models that can be used to explain or "fit" the majority of observational data. They will then work to refine their models as new events are added to the dataset.
This data from Long Island, New York and this exercise should illustrate to you just some of the complexity involved in investigating sea level change - at least on periodic, short-term time intervals of a few hours to a few days.