Add plotting utilities for calibration

For calibration, it is useful to be able to plot the observational data and the columns of the G ensemble matrix on the same plot. This is useful to assess how well the fit is and can be used in some cases to determine if the covariance matrix make sense.

I think there are two avenues for adding plotting utilities for ClimaCalibrate. One way is to provide a generic plotting function that plot the G ensemble matrix and observational data as a line plot stacked on top of each other. This is useful for a quick and simple plot, but it is devoid of any other information.

The other way is to use ClimaAnalysis and use all the information there to make a plot of each `OutputVar` and each plot will have all the corresponding `OutputVar`s stacked on top of each other. The difficulty arises from dealing with all the possible edge cases. For example, you would need to handle `OutputVar`s whose dimensions is zero or two or more.

Start by reviewing the existing representations of the observational data, the G ensemble matrix, and ClimaAnalysis OutputVar values. Compare the proposed generic stacked line plot with the ClimaAnalysis-based approach, including zero-dimensional and multidimensional OutputVars. Done means agreeing on an approach and defining plots that handle the relevant cases.

Written by the indexing model from the issue text.