MorePosted By Deral at Home on 11/9/2008 at 12:36 PM
For others who missed this OPUS (length) submission/post by Loyal , here is the link to some background information on this discussion.
Time Series
Loyal
First thanks for all your time explaining some of the finer points of what you have to contend with in the moving part of the word and also in what I need to be aware of even in my stable little burg.
I read all the information in the links and scroured the NGS for further information as well as some of your sources other links.
Fact. OKLW (local CORS) is showing a very distint easting bias of 1.8cm for the shown period on the 60 day plots.
Fact. We have no time series to look at the overall picture.
Conjecture. Is this a system bias, some shift in the actual mounts recently, a bent support rod or something else. We cannot tell by the 60 day plots. As far as I know, this small building that the antenna is mounted on could have settled or moved that much over a period of time in our soils.
It appears that the time series is in fact a huge improvement over the 60 day plots.
You have systematic biases that will tend to stabilize and using the entire time series can be plotted and graphed as one or more CORS falls out of spec and is need of an upgraded position to better fit in the national network.
You can correct based on the 60 day estimates (and they are estimates) but that does not tell you the full story like the time series.
My interest in this has sparked anew with our very soon to be purchased RTN for Lawton. I expect to tie this station to our network and use our current NSRS derived coordinates for this station.
My concern is other stations that may pop up which could be used in our town. I dont' see this as a very big risk since we are geographically distant from any other large towns and our corrections will be free to anyone that wants to use our network.
Granted that we are chasing very small numbers in a system that is not entirely error free. We have tropo, iono and all the other noise factors that figure into a final solution.
I did a pretty simple test. All reported in US Feet and I just showed the important digits.
Run One with OKLW using OPUS.
N 3.576
E 4.240
Run Two with OKLW Excluded.
N 3.586
E 4.251
So where did the 1.8 cm go, which is around 0.05'. Did it get squished out in the adjustment of OPUS?
Is this a reason that the NGS has decided not to change a coordinate before some threshold? They felt that the difference would not be noticable to 99.9% of the users.
Sometimes I think I may be confusing myself.
Deral
To read the rest of this thread go to www.i-boards.com/bnp/pob/messages.asp?MsgID=1311916&ThreadID=124938&IsResponse=False#1311916
For others who missed this OPUS (length) submission/post by Loyal , here is the link to some background information on this discussion.
Time Series
Loyal
First thanks for all your time explaining some of the finer points of what you have to contend with in the moving part of the word and also in what I need to be aware of even in my stable little burg.
I read all the information in the links and scroured the NGS for further information as well as some of your sources other links.
Fact. OKLW (local CORS) is showing a very distint easting bias of 1.8cm for the shown period on the 60 day plots.
Fact. We have no time series to look at the overall picture.
Conjecture. Is this a system bias, some shift in the actual mounts recently, a bent support rod or something else. We cannot tell by the 60 day plots. As far as I know, this small building that the antenna is mounted on could have settled or moved that much over a period of time in our soils.
It appears that the time series is in fact a huge improvement over the 60 day plots.
You have systematic biases that will tend to stabilize and using the entire time series can be plotted and graphed as one or more CORS falls out of spec and is need of an upgraded position to better fit in the national network.
You can correct based on the 60 day estimates (and they are estimates) but that does not tell you the full story like the time series.
My interest in this has sparked anew with our very soon to be purchased RTN for Lawton. I expect to tie this station to our network and use our current NSRS derived coordinates for this station.
My concern is other stations that may pop up which could be used in our town. I dont' see this as a very big risk since we are geographically distant from any other large towns and our corrections will be free to anyone that wants to use our network.
Granted that we are chasing very small numbers in a system that is not entirely error free. We have tropo, iono and all the other noise factors that figure into a final solution.
I did a pretty simple test. All reported in US Feet and I just showed the important digits.
Run One with OKLW using OPUS.
N 3.576
E 4.240
Run Two with OKLW Excluded.
N 3.586
E 4.251
So where did the 1.8 cm go, which is around 0.05'. Did it get squished out in the adjustment of OPUS?
Is this a reason that the NGS has decided not to change a coordinate before some threshold? They felt that the difference would not be noticable to 99.9% of the users.
Sometimes I think I may be confusing myself.
Deral
To read the rest of this thread go to www.i-boards.com/bnp/pob/messages.asp?MsgID=1311916&ThreadID=124938&IsResponse=False#1311916
