Discussing some GPS problems with Mark Silver yesterday and he mentioned some height discrepancies. I am blessed with so many CORS in this area that I can get 2 sets or more of OPUS-RS solutions using at least 18 CORS. I downloaded an hour of data from Mark's SGU1 CORS and submitted to OPUS-RS excluding SHU1 CORS. Resubmitted this morning for extended output. The RINEX file is from 6/18/2014, 0.00 UTC.
I am at a loss to understand how all residuals can be of the same sign. Any mathematical expert that can satisfactorily explain it would be appreciated.
Residuals from he Extended Output solution:
"Residuals of position determined by individual baselines from the final position
CORS X Y Z East North Up
fred -0.013 -0.005 0.006 -0.010 -0.001 0.011
echo -0.010 -0.001 -0.001 -0.009 -0.004 0.004
nvws -0.005 -0.020 0.020 0.004 0.003 0.028
nvpi -0.009 -0.020 0.011 -0.000 -0.005 0.025
azpg -0.011 -0.008 0.018 -0.007 0.008 0.020
azds -0.002 -0.015 0.014 0.004 0.003 0.020
nvbm -0.012 -0.025 0.020 -0.001 -0.001 0.034
p009 -0.025 -0.061 0.057 0.002 0.006 0.087"
I have resubmitted excluding SGU1 and P001 and will post those residuals.
Paul in PA
Resubmitted Results
his solution is not much better, OPUS-RS is really reaching out and in fact the prior solution used 8 CORS and after one additional exclusion this solution used 9 CORS.
"Residuals of position determined by individual baselines from the final position
CORS X Y Z East North Up
fred -0.015 -0.020 0.018 -0.006 -0.001 0.031
echo -0.014 -0.014 0.012 -0.007 -0.002 0.022
nvws -0.007 -0.029 0.027 0.005 0.004 0.040
nvpi -0.013 -0.033 0.024 0.001 -0.002 0.043
nvlm -0.008 -0.006 0.011 -0.005 0.004 0.013
azpg -0.014 -0.025 0.031 -0.002 0.007 0.041
azds -0.006 -0.030 0.024 0.006 0.001 0.038
nvbm -0.016 -0.037 0.028 0.000 -0.002 0.049
kgmn -0.019 -0.015 0.011 -0.011 -0.004 0.023
I will now exclude 4 more CORS with Up residuals of 0.04+
Paul in PA
Not Getting Better Yet
"Residuals of position determined by individual baselines from the final position
CORS X Y Z East North Up
fred -0.008 -0.017 0.013 -0.001 -0.001 0.023
echo -0.005 -0.010 0.005 -0.001 -0.002 0.012
azds 0.006 -0.018 0.012 0.013 0.001 0.018
nvtp -0.018 -0.013 0.021 -0.011 0.005 0.028
kgmn -0.006 -0.001 -0.001 -0.005 -0.003 0.002
nvsv -0.026 -0.058 0.045 -0.001 -0.002 0.078
nvpo -0.016 -0.031 0.030 -0.002 0.003 0.046
fern -0.035 -0.057 0.058 -0.009 0.006 0.088
I will dump 2 more and give it a go.
Paul in PA
Looking at the numbers you post, I do not see where all the residuals are of the same sign. In the geocentric system the triplet of coordinates, XYZ, show different signs. In the local geodetic horizon system the N and E are small magnitudes as expected with larger height (U) residuals as expected. I see sign variations.
As for cases where all residuals ARE of the same sign, this is a case of a systematic bias. Least squares is to be applied to data after the removal of biases as it relies on errors being random.
Finally, Results I Would Accept
REF FRAME: NAD_83(2011)(EPOCH:2010.0000) IGS08 (EPOCH:2014.46033)
X: -2036611.867(m) 0.010(m) -2036612.701(m) 0.010(m)
Y: -4668218.356(m) 0.016(m) -4668217.043(m) 0.016(m)
Z: 3827958.191(m) 0.013(m) 3827958.092(m) 0.013(m)
LAT: 37 6 47.48164 0.003(m) 37 6 47.49611 0.003(m)
E LON: 246 25 46.97671 0.008(m) 246 25 46.92449 0.008(m)
W LON: 113 34 13.02329 0.008(m) 113 34 13.07551 0.008(m)
EL HGT: 895.537(m) 0.021(m) 894.783(m) 0.021(m)
ORTHO HGT: 920.032(m) 0.026(m) [NAVD88 (Computed using GEOID12A)]
Residuals of position determined by individual baselines from the final position
CORS X Y Z East North Up
fred -0.005 -0.015 0.009 0.002 -0.002 0.018
echo 0.001 0.005 -0.006 -0.001 -0.002 -0.008
nvlm 0.007 0.012 -0.008 0.002 0.002 -0.016
azds 0.005 -0.020 0.010 0.013 -0.002 0.019
nvtp -0.015 -0.005 0.012 -0.012 0.003 0.016
kgmn -0.008 -0.007 -0.001 -0.005 -0.007 0.008
nvpo -0.014 -0.025 0.023 -0.003 0.001 0.036
I will drop NVPO and take another look.
Paul in PA
The Solution Gets More Precise, But Not More Accuarate
I am just not happy with how the early residuals fall.
[inlinecode]REF FRAME: NAD_83(2011)(EPOCH:2010.0000) IGS08 (EPOCH:2014.46033)
X: -2036611.866(m) 0.010(m) -2036612.700(m) 0.010(m)
Y: -4668218.355(m) 0.014(m) -4668217.042(m) 0.014(m)
Z: 3827958.189(m) 0.011(m) 3827958.090(m) 0.011(m)
LAT: 37 6 47.48162 0.004(m) 37 6 47.49608 0.004(m)
E LON: 246 25 46.97673 0.008(m) 246 25 46.92451 0.008(m)
W LON: 113 34 13.02327 0.008(m) 113 34 13.07549 0.008(m)
EL HGT: 895.534(m) 0.018(m) 894.781(m) 0.018(m)
ORTHO HGT: 920.029(m) 0.024(m) [NAVD88 (Computed using GEOID12A)]
Residuals of position determined by individual baselines from the final position
X Y Z East North Up
fred -0.004 -0.015 0.010 0.002 -0.001 0.018
echo 0.000 0.003 -0.003 -0.001 -0.001 -0.005
nvlm 0.005 0.009 -0.004 0.001 0.002 -0.011
azds 0.003 -0.023 0.013 0.012 -0.002 0.023
nvtp -0.018 -0.009 0.016 -0.013 0.003 0.022
kgmn -0.010 -0.010 0.002 -0.005 -0.007 0.012
[/inlinecode]
Seems I learned something today. Thanks Loyal.
Paul in PA
The Solution Gets More Precise, But Not More Accuarate
I suspect that (at least) part of what you are "seeing," is the DIFFERENCE between the predicted Epoch 2014.xxxx IGS08 position, and the "actual" IGS08 position of the several CORS (relative to each other).
The Short Term Time Series (90day) will give you a pretty good handle on how the several CORS are currently behaving, both relative to the predicted positions, and to each other (as well as a "feel" for the day-to-day behavior).
OPUS Solution Short Term Time Series
East North Up East North Up
fred 0.002 -0.001 0.018 -0.0011 0.0060 0.0104
echo -0.001 -0.001 -0.005 0.0049 0.0037 0.0183
nvlm 0.001 0.002 -0.011 0.0052 0.0059 0.0103
azds 0.012 -0.002 0.023 -0.0040 0.0074 -0.0009
nvtp -0.013 0.003 0.022 0.0055 0.0080 0.0197
kgmn -0.005 -0.007 0.012 0.0103 0.0057 -0.0082
I think I see a little bit of correlation here...
Remember... OPUS uses the PREDICTED IGS08 Position!
Loyal
The Solution Gets More Precise, But Not More Accuarate
It is interesting how small this world is. I talked to Paul yesterday for an hour and a half and I am going to have a beer with Loyal Sunday afternoon. I mean, really this world now seems to be about three blocks wide.
The trick in that area is to use OPUS static. If you look at where those stations are, with respect to SGU1 (St George Utah) you will see that typically you end up with a bunch of vectors piled up on each other going down to Las Vegas.
FRED is East and ECHO is west, but you always end up with a bunch of vectors pointing to Vegas and nothing balancing to the north. (For some reason, I don't like ECHO but I can't remember why. Which is interesting because I can remember why I don't like all of the other CORS that I don't like.)
It probably works out in the wash, but I prefer to force PUC1 or P009 into my solutions.
Paul, I have to say that you easterners are kind of OPUS spoiled. It appears that you have a CORS station on every single corner. OPUS-RS must be the bomb in PA, I can't even count how many sites are in your circle:
vs. my pathetic circle:
actually where I am today, I don't even have a circle.
🙂
M
