Feedback on store results needed.

[jbusecke]

I just generated three different stores in nasa-veda-scratch (one for each scenario, but only a single year for now) and it would be great if folks could give it a go and see how the performance of the store is for some typical workflows.

The instructions how to load the data are in the README.

I did a quick plot with

ds['AvgSurfT_Percentiles'].plot(col='percentile', row='time')

but I am really unsure if this is the output that is expected?