Improve Reading of Fluxtower Data in Site-level Experiments

[Espeer5]

Is your feature request related to a problem? Please describe.

Currently, our site-level experiments assume that the fluxtower data files contain only the data for the year 2010. As a result, the ClimaArtifact for the Ameriflux data must trim the data down to 2010, and the data stored in the cluster is only for the year 2010. This is non-ideal - it means that if someone wants to run the experiments for a different year they must first go get the data, slice it for a different year or set of years, and upload it to the cluster.

The pipeline for reading the fluxtower data also uses Format.jl as a dependency. The single call to format used is the only function from Format.jl used, and it is desirable to remove this call in order to remove Format.jl as a largely superfluous dependency of ClimaLand.

Describe the solution you'd like

Use Dates.jl to read the desired set of years in from the fluxtower CSVs. Update the ClimaArtifact to be the full set of fluxtower data for all available years for each of the experiment sites. Give the desired year of the start and end of the simulation as variables prescribed by the user for any given simulation run. Remove the dependency on Format.jl in the data reading process.

This change will need to be made in both ClimaLand/experiments as well as in ClimaLandSimulations.