Thanks for the overview! Sounds like field of hydrology has pretty decent scientific standards!
Just a random thought, do you think it'd make sense to test the integrated world models by using agent-based simulations? Or maybe using those simulations for setting some initial parameter ranges for key variables, before fitting them on real data?
Yeah hydrology is quite good when it comes to such thing. Only other field I know that is even better is astronomy.
I am not sure how you mean "test the integrated world models by using agent-based simulations". Their underlying architecture is kinda incompatible. Or I might not get what you mean.
Also your parameter ranges are usually already constrained by real world considerations. For example, you might not now the exact water infiltration rate for the area you are looking at, but you know what values it could possibly have. In addition, good calibration algorithms are usually able to find the right parameter values, even if you have wide ranges.
Thanks for the overview! Sounds like field of hydrology has pretty decent scientific standards!
Just a random thought, do you think it'd make sense to test the integrated world models by using agent-based simulations? Or maybe using those simulations for setting some initial parameter ranges for key variables, before fitting them on real data?
Yeah hydrology is quite good when it comes to such thing. Only other field I know that is even better is astronomy.
I am not sure how you mean "test the integrated world models by using agent-based simulations". Their underlying architecture is kinda incompatible. Or I might not get what you mean.
Also your parameter ranges are usually already constrained by real world considerations. For example, you might not now the exact water infiltration rate for the area you are looking at, but you know what values it could possibly have. In addition, good calibration algorithms are usually able to find the right parameter values, even if you have wide ranges.