Investigate 7 July 2010 precip variance

[akrherz]

While reprocessing some old data looking for mismatches in analyzed precip, a simple reprocessing of data from 7 Jul 2010 did not resolve a large data variance.

    huc_12    |   valid    | qc_precip | avg_precip |  diff   
--------------+------------+-----------+------------+---------
 071100010901 | 2010-07-07 |   35.8532 |    97.0687 | 61.2155

Hopefully this is just extreme inner-HUC12 precip variance? Will find out.

[akrherz]

Working theory is that this HUC12 (NE Missouri) is very NW to SE oriented and is intersecting a small band running SW to NE. My zonal stats computation may not be doing a good enough job here or our perhaps our flowpath sampling is not well spread out in this HUC12.

[akrherz]

So down the rabbit hole we go... I thought that this was a problem with not fine enough rasterstats being used, but nope. This plot is of interest.

[akrherz]

Here's another example:

huc_12	valid	qc_precip	avg_precip	max_precip	diff
070300040702	2009-08-19	131.62186	92.47059	93.7	-37.92186

[akrherz]

here's another

huc_12	valid	qc_precip	avg_precip	max_precip	diff
090203040103	2013-07-18	91.73381	50.02	60.3	-31.433811

[akrherz]

And some interesting examples with extremes in the other direction

huc_12	valid	qc_precip	avg_precip	max_precip	diff
071100010901	2010-07-07	36.987823	101.30937	142.2	64.32155
110702010106	2016-07-25	36.395832	92.61579	134.9	56.21996
070300040702	2017-07-06	87.58215	135.77647	148.9	48.19432
070102070107	2016-08-10	55.02648	102.36667	108.8	47.340187