[x] Pick a query that Fernando suggested, and make a concrete instantiation of it. To start, we should pick a time range that either uses the reanalysis OR the prediction dataset.
[x] Save a JSON file to S3 for each HUC8 with the boundary and a list of reach ids (ie. ComID fields). See #48
[ ] Time the query using the Zarr-encoded data that is on S3 and the above HUC8 files.
I will be focusing on this query: "All stream flows within a given set of HUC8's and a date/time range. Samples are then aggregated to daily averages/max/min."
This represents the first few steps in #45