Tulare Lake bed, verticals?

[CBarrett]

So we are surveying out of our area, near lake Tulare bed, near Corcoran, GPS station CRCN.
Are we reading this right, this station is sinking a the rate of almost meters per year?
We didn't use it for vertical but geoid heights got picked up anyway, and we were off 2 feet vertically from somewhat recently published values. I'm only doing a section line retracement west of Terra Bella, and nothing vertical, but now I am very curious to learn what is going on near Chocran.

Anyone surveying in this area more often that can tell a full story?

[Warren Smith]

It's subsidence due to groundwater pumping. There's an infamous photo of a USGS fellow alongside a telephone pole with years marked all the way up to the top.

[CBarrett]

Wow! I'm sneaking in a few moments to read about this.
At first we thought our GPS was broken in some new and interesting way, then as I stared at the greenery and the browns on google earth, the texture west of Cochran looks... different green, so we looked into geologic conditions.
We never see that kind of a movement down south, so I thought like it needed a 'real person with experience' check too. I guess I have my next weekend rabbit hole cut out for me.
Thanks!

[Warren Smith]

Here's an old and a newer illustration.

[SPMPLS]

2010-2018 data on CRCN. I am sure it has gone down substantially since.

CRCN.jpg

[CBarrett]

Apparently, according to several articles, that areas also sits on a 2800 vertical feet thick layer of silt and soft soils... and the mountains to the east seem to be pushed up (at an exponentially slower rate). How interesting!!!!

[CBarrett]

Thanks SPMLS! We looked at the CSRN datasets and the timeframe we were looking at had two meters, in only several years. We looked at a number of details to do a reality check on results and they are pretty mixed up in my memory for the moment, and I didn't write a lot of it down since I was just doing a quick peer review. We learned enough to eliminate this point as an outlier. Luckily since this is a new area we built in a lot of redundancy.

At first we were trying to isolate the vertical blunder.
This seems like a pretty good example in how regional knowledge plays a part in surveying.

[SPMPLS]

The USGS picture of Dr. Joseph Poland that Warren posted was taken on the west side of the valley up near Patterson (the purple area in the picture below). That area of subsidence abated considerably after the Federal and State water projects started delivering surface water for ag use in the area. The more recent subsidence areas have been in the greater Bakersfield area and up in Merced/Madera Counties where the "El Nido Bowl", near Dos Palos, subsided over 2 feet in 24 months - Jan 2008-Jan 2010, before the drought.

El Nido bowl region reduced.jpg

[SPMPLS]

Vertical time series plot for station P565, which is a NGS CORS.

p565 vert only Jan 2021.jpg

[SPMPLS]

A step that I always recommend when planning a network GNSS control campaign is to look at the time series plots of the stations you plan to constrain to (NGS CORS or CSRN). Don't assume they are where the NGS datasheet or CSRC value shows they are on the date(s) you plan to survey. In 2016 station P565 was close to 30 cm lower than where the NGS datasheet said it was on the 2010.00 epoch.

Yes, knowing the areas where you work, whether retracing boundaries, or doing work where the vertical is critical, is powerful professional knowledge.

[SPMPLS]

If you aren't familiar with how to access and evaluate time series plots, you can now use the the SOPAC/CSRC utility SCIP to evaluate the 3D displacement of your area of interest over time, whether from 2017.50 (CSRC) or 2010.00 (NGS) to your survey date.

http://geoapp21.ucsd.edu/

If the magnitude/direction of your preferred primary control is not consistent over time (in a relative sense), your "true of date" observations likely won't play well with 2010.00 (worst case) or 2017.50 constrained positions in a least squares adjustment.

[CBarrett]

This is very interesting, and very good advice, to investigate your stations thoroughly.

P565 was in our dataset also, and thrown out as an outlier. We still had 3 or 4 others to use where we didn't have big, initially unexplained differences.
We didn't need any vertical, I'm just retracing a road centerline to see if a power pole is in the right of way, but still when you see oddities, you want to examine them, find the source, see if anything else got affected... so, thankfully, our equipment didn't spaz out, ephemeris doesn't seem to be in error, I don't have to yell at my field guys for mucking anything up. We actually detected points with drastic vertical movement. P056 had some vertical issues too, but much smaller.

We didn't look at the time series plots for those points, but we did compare their published values vs. the next day ephemeris for the same time slot as our base station setup, and this is where the differences occurred.

I had to go look it up. These are stations we used: CRCN, DLNO, P056, P565, P566, P810, WASG.... to post-process location of three points on a section line. Because - new area - play it safe. As a routine we post-process almost everything with the next day ephemeris.