You can derive atmosphere moisture profiles with a GNSS. Trimble has developed a bit of code for that.... cloud thickness might be slightly more difficult.

Thank you for that.

But I guess that if you have a profile, than you can also say that where the moisture is at 100 % you have a cloud, right?

Dear Amir, I have no idea to be honest. But I would assume that the general troposphere/ionosphere situation is spatially quite homogeneous if you compare it to a rising cumulus. Thus I assume that not all the signals coming from different satellites would be affected by the cloud.

I have seen some presentations that were using not only the delay but also S/N ratio. If the cloud is on the line of sight satellite-receiver then S/N is low, otherwise it is not. So using multiple satellite I could estimate at least the position of the cloud only by looking at the S/N. Or am I wrong?

Thank you, I have just found the same paper also myself.

I don't think cloud will have any significant effect on L-band GPS signals.

Thank you, but what about the aerosol clouds - as Amir pointed above, there was this paper of detecting volcanic plumes with GPS. Are aerosols so different than water clouds? The size of water droplets is probably similar to the size of ash particles 1-10 micrometers).

Hi Klemen,

GPS signals are in the L-band, comparable to radar-signals and to the infra-red-signals used in geodesy since the 1960ties (ask Björn Riedel, TU Braunschweig).

The influence of the cloud will change the measured station position (especially the height component), because in one or more satellite-receiver connections the signal-velocity will be changed (ask Prof. Wanninger, TU Dresden).

Nevertheless, you need a model. Maybe a non-parametric method like support vector regression will help: you need general information about the present weather, short-time change of GPS station position, satellite positions and several times a "sky truth" information. A once optimized SVM model should deliver the cloud thickness.

Regards,

Michael

Thank you Michael. I will try to contact Prof. Wanninger in the next days.

http://www.atmos-chem-phys.net/12/5309/2012/acp-12-5309-2012.pdf

It seems Bending angle anomaly will be more helpful to estimate the cloud thickness.

@Amir

Thats true Amir but along with the Temperature profile you can have an idea of the cloud thickness.check Fig3, 4 and 10

Klemen,

While at UCAR many years ago we did some comparisons between liquid and vapor components of pointed water vapor radiometer measurements versus GPS slant-path delays. There is a correlation in GPS slant-path phase delay associated with clouds but the relationship reflects a combination of water vapor delay and hydrometeor scattering effects (e.g., Solheim et al, JGR, 1999). There may be a way to tease these apart-- scattering is dispersive and will reduce the amplitude-- but it would require evaluating delays at multiple frequencies, & sorting those from the ionospheric dispersion, as well as dealing with the rather noisy SNR counts. And, the best you would be likely to get from that would be an integral measure that reflects some combination of thickness along slant path, humidity and scatterer density. Perhaps with a dense enough GPS network that could be turned into a spatial representation... But it would be challenging!

Hope that helps--

Tony

Kleman,

This paper may be helpful.

Tropospheric delay estimation for pseudolite positioning

JJ Wang, J Wang, D Sinclair, L Watts, HK Lee

Journal of Global Positioning Systems 4 (1-2), 106-112

Regards,

Jack

Hi everybody,

I would like to thank you all for the interesting discussion. I have found some points I should consider, papers to read and people to contact in the case we really want to start our project. Although my question was considering a general cloud I am in fact interested mostly into the estimation of volcanic ash cloud thickness. That is a very important parameter for aviation but the only reliable way to determine it today is lidar. And lidar measurments are not as often available as GPS...

So in theory it is possible to estimate the cloud thickness, as Sasha pointed out, as it is possible to correlate the temperature/moisture profile with the cloud. Still, we are talking about L-band, which is pretty insensitive for this purpose. As Amir, Anthony and Michael pointed out it would be necessary to set up a model that would integrate all the possible influences.

So thank you all again!

Hello I'm a bit late on this discussion.

It is definitely possible to detect the cloud top with high accuracy (please see the papers mentioned below) and maybe also the cloud thickness, but it depends what type of clouds and what accuracy you need. I'm actually studying how to detect the thickness using the GPS RO

You may be interested on these papers:

- R. Biondi, S. P. Ho, W. Randel, T. Neubert and S. Syndergaard, “Tropical cyclone cloud-top height and vertical temperature structure detection using GPS radio occultation measurements”, Journal of Geophysical Research (JGR) – Atmospheres, 118, 1-13, DOI: 10.1002/jgrd.50448, 2013

- R. Biondi, W. Randel, S. P. Ho, T. Neubert and S. Syndergaard, “Thermal structure of intense convective clouds derived from GPS radio occultations”, Atmospheric Chemistry and Physics, 12, pp. 5309-5318, DOI:10.5194/acp-12-5309-2012, 2012

Since you where also discussin about volcanic plumes, I'm going to publish a paper in march exactly on it.

Have fun,

Riccardo

Dear Riccardo, that sounds very promising. I would be very interested to see that paper, so please send me the PDF if you still keep this discussion in mind.

I think it is difficult to measure the cloud thickness, because I am using temperature measurements. May be you can try using refractive index profile.

Hi forget to tell you, you can try using bending angle profile

yes you must use the bending angle, since it is not affected by any model ... when you derive the other parameters (e.g. temperature and water vapor) the 1Dvar smooths the small variations. Of course there is a limit at the moment, you can not see water vapor values lower than (about) 100ppm (i.e. high thin cirrus and overshooting) even though you can guess their presence in a different way (I'll submit a paper about it in a couple of weeks to ACP).

You can get some details about the use of bending angle instead of temperature profiles in these papers (especially the first one):

- R. Biondi, T. Neubert, S. Syndengaard, J. Nielsen, “Measurements of the Upper Troposphere and Lower Stratosphere during Tropical Cyclones using the GPS Radio Occultation Technique”, Journal of Advances in Space Research, 47, 348-355, DOI:10.1016/j.asr.2010.05.031, 2011

- R. Biondi, T. Neubert, S. Syndengaard, J. Nielsen, “Radio occultation bending angle anomalies during tropical cyclones”, Atmospheric Measurement Techniques, 4, 1–8, 2011, DOI:10.5194/amt-4-1-2011

Sorry for just citing papers but I do not have much time for xexplaining now. Anyway I think the paper can be quite exaustive :-)

And Klemen of course I'll send you the pdf once it is done, I'll write it in my agenda.

Riccardo

Riccardo, I realize I've some reading to do... It's been awhile since I was active in GPS water vapor work (although I've been told that's going to change). Question: Are you deriving bending angle from single-receiver, multi-satellite refraction? I've bookmarked the papers to review.

Thanks

gerry

Hello Gerry, the measurements are based on GPS signal received by a LEO satellite. The signal is refracted and delayed by the atmosphere, and the bending angle is retrieved from phase and amplitude of the signal. You can get all the details about the different procedures for retrieving bending angle and other atm parameters from this paper: Quantification of structural uncertainty in climate data records from GPS radio occultation, A. Steiner et al., ACP, 2013

Riccardo

OK, so it isn't ground-based. That was my confusion. LEO-occultation would explain the procedure. Reminds me of some of the work by Rocken and Ware on water vapor tomography. Without sufficient receiver density, it's difficult or impossible. Same thing with LEO occultation. Unless we have sufficient LEO birds to relay the data, the upper atmosphere water vapor info is difficult to come by.

Yes LEO occultations. I do not know the ground based GPS receiver density in US, we are trying to create a dense network in EU collecting all the sensors available through the COST Action ES1206 GNSS4SWEC (http://gnss4swec.knmi.nl/) you could be interested on that.

Regarding the water vapor in the upper atmosphere (do you mean UTLS?) using the GPS RO we are able to detect no less than 100 ppm (not overshooting for example).

Regarding the density of GPS RO, we have about 2500 occultations/day at the moment whcih is not much, but we are expecting several thousand (hopefully between 10.000 and 20.000) in the future with COSMIC-2 and other new missions (planned) and with the availability of Galileo, Glonass, Beidou and other GNSS constellations.

Hopefully also the retrieving technique will be improved providing higher sensitivity tot he water vapor, ok this is a dream :-)

Recently i recorded the field strength of GPS signals using simple receiver and it has been observed that there is a noticeable difference during rainy,cloudy days and dry days.But no reference is available calibration could not be carried out.

Thank you Elangovan!

The cloud content of aerosols, water and ice increases delay of passing GPS signal. Estimate on magnitude of increase is reported, for example, here:Article Propagation delays induced in GPS signals by dry air, water ...

The excess delay depends on column denseties of different consituents along the signal path.