- 1 Basic Information
- 2 Catchwords
- 3 Short Description of Functionality
- 4 Program(s) to run before this Program
- 5 Program(s) to run after this Program
- 6 Additional Information
Name of Program
time series of high water and low water events
coupled band-pass filter
filter considering fresh water flow
desired and actual value
Short Description of Functionality
In order to analyse time series of high water or low water events you usually calculate their annual mean values. These values are influenced by meteorologic processes e.g. wind or fresh water flow in an estuary. This makes it difficult to compare the mean values of several years. The program EVENTFILTER reduces the meteorologic influence by filtering the events and generates means you can better compare with each other.
The input values are filtered by a coupled band-pass and / or under consideration of the fresh water flow.
- general input data (filetype eventfilter.dat)
- The reference files 2. and 3. have to contain high water events as well as low water events.
- time series of measured actual high water and low water events at a reference station (file of type boewrt.dat)
- (optional) time series of synthetic, desired values of high water and low water at a reference, generated for example by the
- harmonic process (file of type boewrt.dat)
- measured time series of high water and low water events at further stations e.g. along an estuary (files of type boewrt.dat)
- (optional) time series of fresh water flow into an estuary (file of type boewrt.dat), needed for the filter considering fresh
- water flow. The timestep has to be 24 h.
- water flow. The timestep has to be 24 h.
- file(s) containing a table with the statistics of a single station (type eventstatistic.dat (ASCII) or eventstatistic.excel (excel compatibly))
- informative printer file (eventFilter.sdr)
- (optional) trace of program execution (filetype eventFilter.trc)
The reference station should be situated in an area with few meteorological influence on high water and low water. In the German Bight Helgoland would be a suited place. At first EVENTFILTER has to identify a reference event as high water or low water. You get the best results by analyzing the synthetic values. EVENTFILTER compares date and time of a measured event with date and time of a neighboured synthetic event and identifies the actual high and low water events. (All events in the following pictures are linear connected and therefore not realistically represented.)
The picture above shows that the identification works well even in the case of a storm flood. The water level differs between the synthetic red curve and the measured green curve up to 2 meters, whereas the time always differs less than an hour.
The second way to identify high and low water is to use the mean flood period at the reference station. It should only be used if there are no synthetic values available.
The coupled band-pass processes the measured high and low water at the reference station. At first all events inside a scope around the mean value are marked. Example high water: all high water events higher than mean high water - 0,1m and lower than mean high water + 0,1m are marked. The events that have not been marked are examined. In the example above a high water higher than the upper high water limit is marked when the low water before or after is lower than the lower low water limit. This is typical for a spring tide. This phenomenon is caused by astronomical forces and the events should be considered. The algorithm for nipp tides works analogously.
The diagram shows the described case. Thw1 has been marked by the filter though it has a high value. The tide calender confirms the supposition that it takes place during a spring tide period. Thw2 and Thw3 are marked because they are inside the limits. Thw4 is without astronomical cause to low and has not been marked. Tnw1 has been marked because Thw1 is to high. Tnw2 to Tnw4 are to low.
Next mean values of the marked events are evaluated and the upper and lower limits are adapted to them.
The process of marking and evaluating is so often repeated until both mean values converge.
Then eventFilter identifies high and low water at the other stations by using the known event runtime between reference and station. The events corresponding the marked ones at the reference are marked as well and last not least their mean values get evaluated.
Filter considering fresh water flow
The water level in the upstream region of an estuary is highly influenced by the amount of fresh water flow. But in other areas the water level is influenced by extremly high or low amounts, too. Therefore water flow measurements between a lower and upper limit are marked. The corresponding high and low water events are marked as well and their mean values get calculated.
Example: The daily mean fresh water flow at Neu Darchau in the Elbe estuary is shown in the upper part of the picture above. The values 200m**3/s and 800m**3/s limit the filter. E2 is the first event that lays inside these boundaries and that gets marked. The lower diagram shows the high and low water events at Hamburg St. Pauli 87 km downstream from Neu Darchau. The events Tnw3, Thw3, Tnw4 and Thw4 are marked because they take place within 24 h after the marked event E2.
Combination of both filters
If you combine both filters the data is at first processed by the coupled band-pass and then by the fresh water flow filter. The combination works similar to a series connection. As a consequence the results in eventstatistic.dat/.excel named '*_Q' are based on events that have been marked in both filters.
Program(s) to run before this Program
Program(s) to run after this Program
back to Program Descriptions