Thorough sensitivity analysis varying as many param as possible #7
Description
It would be great to conduct an automated experiment to explore the parameter space.
- Experiment with emulkit.
- Calculate weighted distance along the data between the model and the reanalysis for the reanalysis inputs.
- Optimise to minimise this distance.
Hypothesis: There exists a set of parameters that could reproduce the same results for nino3.4 (i.e. the solution is not unique)
H0: The model only has one set of parameters that could reasonably reproduce the model results.
Hypothesis: A different set of parameters from a reasonable range could more closely fit the observed data over the whole tropical Pacific.
H0: The model is tuned to the best set of parameters to replicate the behaviour of the tropical Pacific.