RE: PROBLEM: water level results with double peaks. Results aren't smooth

Message Boards

Esther Espinoza, modified 6 Years ago.

PROBLEM: water level results with double peaks. Results aren't smooth

Youngling Posts: 19 Join Date: 4/6/17 Recent Posts

Hi, I'm simulating water level for a month period in an estuary and the results in the three monitoring stations have double peaks on almost all the data set, it is not smooth compared to the observed data even though the statistical indicators are fine. I have just one open boundary at the mouth of the estuary, I have used manning values between 0.012 and 0.030 and the eddy viscosity between 10-100. Also, I've done simulations with a reflection parameter alpha of 0 and 100 but it haven't solved the problem.

I'm just simulating water level and I'm just considering what I've mentioned before, the open boundary input data is in a time series format from tide gauge measurements of an institute, I plot that data and it is not a perfect smooth sinusoidal curve but it does not show double peaks (just in two peaks I think but it was small).

I read in another post a recommendation to smooth the bathymetry, which I did once and the results were not better. I will attach the graphics and I hope you can help me with this.

Thanks!

Note: The graphics don't show the complete month just a few days so the peaks can be seen but the whole month shows double peaks.

grid bathymetry and obs points

double peaks

open boundary data plot

open boundary d-flow

Qinghua Ye, modified 6 Years ago.

RE: PROBLEM: water level results with double peaks. Results aren't smooth

Jedi Knight Posts: 168 Join Date: 3/2/11 Recent Posts

Hi Esther Espinoza,

This question belongs more good modelling practice than software problem.

From the results, the amplitude is around 2.5meter, which is close to your boundary conditions, though I cannot judge the time/phase since it is not clear in your figure. Thus I would not think it is a problem yet. You may refer to some other project reports/papers to see how others do their calibration/validation.

Enjoy your modelling,

Qinghua

Richard Measures, modified 6 Years ago.

RE: PROBLEM: water level results with double peaks. Results aren't smooth

Jedi Knight Posts: 143 Join Date: 3/23/11 Recent Posts

Hi Esther,

Interesting problem!

A few thoughts which may be useful or not?

Are the double peaks occuring everywhere through the domain or in a particular area? (you may need to add some more monitoring points to find out? If you can work out where they are happening it may narrow down what is causing the problem?
Are there waves/shocks propogating around the domain - perhaps by looking at the velocity field you may be able to see this? If so it might be possible to solve these with more damping or better initial conditions?
Are water levels near (i.e. within a few cells of) the boundary condition behaving as expected or do these exhibit the same double peaks? I would have thought by looking in detail in this region you might get a better idea of whether you are experiencing some boundary related problems?

In summary I think it might be worth doing some more digging into the results to try and narrow down what is causing the problems.

Good luck!
Richard

tide calibration reflection oscillation

Esther Espinoza, modified 6 Years ago.

RE: PROBLEM: water level results with double peaks. Results aren't smooth

Youngling Posts: 19 Join Date: 4/6/17 Recent Posts

Thanks for your answer. The double peaks were everywhere through the domain. The problem was solved by interpolating the hourly data input for the open boundary to a 10 min interval. The peaks were gone, the curve was smooth and the statistical parameters were a little bit better. I'm not sure why the hourly data was a problem, but it was solved. I think it was on the high and lows because there were not enough points due to the hourly data.