Does PyGnome provide any functionalities for "blending" environmental data from multiple sources?

[oliskir]

Hello,

I am new to PyGnome and have not used it yet. Having studied the documentation, I have the following question:

Does PyGnome include any functionalities for "blending" environmental data from multiple sources? Say, I have gridded wind speeds and directions from an atmospheric model, and additionally I have direct measurements performed at a few locations with weather stations attached to buoys or something like that. Can I add both of those data sources to my Model and PyGnome somehow magically finds out how to best combine them, or do I have to "blend" the two datasets myself?

Thanks,
Oliver

[ChrisBarker-NOAA]

Sorry, it does not have anything like that built in. However, there are some things you can do that might get you partway there:

• you can have any number of movers, for example, more than one wind mover.
• the effects of each of those winds will be added up -- so double (or triple counting)
• However, you can set the "active_start" and "active stop" times for different movers, so that the model can switch between them at a certain time
  • we use that for when we have observations up to some point and then a forecast beyond that point, for instance.
  • you can also have different gridded winds (or currents) that do not overlap in space, and then at any given location, the elements would be on only one grid (or zero at the velocities at parts of the same grid).

You can also create a custom Wind object that does the blending on the fly, and then pass that to a WindMover for use in PyGNOME.

But if I understand what you want to do, the best way is to probably "blend" them yourself, ahead of time, and produce a gridded wind file with the blended results.

One of the reasons we haven't done that ourselves is that it's very hard to figure out how one might "blend" different winds in a way that would work in the general case.

Note that combining winds (or currents) from multiple sources can create very strong, and spurious, convergences and divergences -- it's tricky to do.

[oliskir]

Hi @ChrisBarker-NOAA ,

Thanks for responding so quickly - and for confirming my suspicion that there is no generally applicable solution for combining multiple data sources that overlap in space and time. I figured this might be the case.

Still, I'm impressed by the versatile of PyGNOME and how it seamlessly handles temporally or spatially separated data inputs. Thanks for your efforts!

Oliver