Message Boards

Disk write frequency in delft-part

Stephen Green, modified 4 Years ago.

Disk write frequency in delft-part

Youngling Posts: 3 Join Date: 12/1/14 Recent Posts

I would like to use delft-part to calculate residence times in an estuary. The idea is to run a hydro model to simulate tides etc. Then I would like to use delft-part to fill up the estuary with conservative tracers and either:

a) remove the particles from the model as they leave the estuary mouth (noting their age and original position) or
b) save the particles tracks by checking the '3D tracks' box in the output window and analyse the particle tracks as a post processing step

I can't see a way to do a). Am I missing something? Can particles be removed when they enter a certain area of the model?

I can't do b) for a large number of particles because it saves the particle positions to disk every time step and the file will get prohibitively big very quickly. Is there a way to get delft-part to output particle positions to disk every nth time step? It seems like a weak reason to not be able to use this method.


Frank Kleissen, modified 4 Years ago.

RE: Disk write frequency in delft-part (Answer)

Youngling Posts: 2 Join Date: 7/15/13 Recent Posts
Hi Stephen,
In PART it is not possible to remove particles. Particles are only removed from the model simulation once they exit the model across an open boundary. The track output at the moment will write the particle coordinates at every timestep. You can only change that when getting into the code and change it there. If you are interested in the 'age; of the particles, then what you may consider is to use the conservative tracer and a decayable tracer at the same time. Decay is a first order process. The ratio of the concentrations of the two determines the 'age'. The 'age' of particles is not included explicitely in the model's output. (In Delwaq age can be derived automatically). So if you would introduce a continuous discharge at a location with both tracers, then after some time, when the situation has become more or less stable then then you can derive the average age distribution in the model domain. You could repeat this for a number of different locations in your model domain.
Alternatively, have you considered using Delwaq to derive residence times?

Best regards