Simulate with new forcing data

[officialankan]

Hi and thanks for a great package!

I have, with great pleasure, been playing around with pastas for some Swedish groundwater wells. The results are quite impressive.

I have a use case where I would like to calibrate a model for a well and then use the model to simulate different forecasts (an ensemble of forcing data). I admit that I have not spent many hours looking in the docs but during my few hours I have failed to find an example of this. Basically I was wonder about a modeling workflow which would look like:

1. calibrate and asses a model using some forcing data
2. simulate with completely new forcing data

By forcing data I mean data that I use to define stresses.

I guess I can try adding stresses and selecting the calibration period but I was wondering if there is a preferred method (which would also become cleaner code on my end..)?

Thanks!

[raoulcollenteur]

Hi Andres @officialankan ! Nice use case, I am working on a new module now to use Pastas models for forecasting (i.e. with new (ensemble) data). There is a forecast branch with some WIP code, but I will work and finalize this in March actually related to a manuscript we're preparing now :) . So if you have a couple of weeks you'll be able to do just this with some cleaner code. Up to then you'll have to hack your way into the stresses a bit..

Feel free to reach out if you need it sooner or want to discuss this application :)

Cheers,
Raoul

[officialankan]

Awesome! Will it be opinionated for a certain type of forecast data or more general such that a user can just supply his own forecast? Also, this might be a separate (but linked) topic - will it deal with initial state such that all forecasts use the same initial state/level?

I will happily wait a couple of weeks and read your paper.