GRUAN manual open for discussion

Here is a first draft of the GRUAN manual (formerly called the GRUAN guide but now renamed to be consistent with WMO nomenclature). Click here for a MS_doc version.

This is a draft whose main purposes are:

1) To encourage discussion around a number of issues and to push the development of GRUAN on various fronts.

2) To encourage members of the GRUAN community to submit material for assimilation into the manual. As this happens, some discussion is likely to be required, especially around contentious issues. But such discussion is vital to progress the development of GRUAN.

The GRUAN manual must serve the GRUAN community and this will be best achieved when members of the GRUAN community actively participate in its development. To that end we would ask you to:

1) Read through this document.

2) Submit material to Greg Bodeker who is coordinating the development of this manual. Greg is prepared to accept input in any form. It can be as simple as saying “Greg please incorporate the attached paper into the manual” or as detailed as using the MS Word revision tools to make suggested changes to the attached document. The document is also available on google docs if you would like to use that facility for making suggested changes/additions you can do so here.

3) Engage in discussion on some of the issues that might be raised in the manual. The best way to do this is to initiate a new discussion thread on the GRUAN blog and then ask people to comment on that. Greg will also keep an eye on the blog for such activity which will then also provide input to the manual.

The intention is to develop the next version of the manual by ICM-3 (28
February) so the earlier you can get the material to Greg the more likely it will be included before that date.

Any other comments on the manual as it stands would be much appreciated.
But please note that:

1) This is a first draft. If you are unhappy with any aspects of the manual you have the ability to change it.

2) The manual is going to require regular revision. This will not be the last opportunity for you to provide input to the manual.

3) This document must belong to the community. This will only happen if you get involved in its development. Please make the most of this opportunity to do so.


Greg,  Holger,  Franz

4 Responses to GRUAN manual open for discussion

  1. Greg Bodeker says:

    Hello All,

    Over the next few weeks I am going to be asking lots of questions about the requirements for GRUAN measurements of the ECVs. This information will be feeding the development of the GRUAN guide. My first three questions relate to the temperature ECV:

    1) The stated requirement for precision is: ≤0.2 K in measurement repeatability. Is that currently achievable with any instrument? Is it achievable with RS92 radiosondes?

    2) The stated requirement for accuracy is: Uncertainties of ≤0.1 K in the troposphere and ≤0.2 K in the stratosphere. Is that currently achievable with any instrument? Is it achievable with RS92 radiosondes?

    3) The stated requirement for long-term stability is: 0.05K/decade. It is very hard to say something about whether this is achievable or not since it depends on how and how often manufacturers change the instruments. But any comments on that would be much appreciated.

    Thanks for any feedback.

    • Frank Schmidlin says:

      Q. 1 and 2

      As long as radiosondes continue to use fixed corrections in look up tables, I do not feel any radiosonde will meet the ECV requirement.

      Nighttime measurements are a different problem. Many mfr’s claim no IR correction, but this is not necessarily so!

  2. Peter Thorne says:

    Yet again, how to attach to a comment escapes me, so will send the worked up document under seperate cover. Sorry.

    This is clearly the result of substantial efforts on the paart of Greg and the Lead Centre for which they are to be applauded.

    As is usual for me there is lots of virtual ink. In terms of common themes:

    1. There needs to be greater recognition in parts that GRUAN will be a network of networks not a homogeneous network. Sites will have multiple calls on them of which GRUAN will be just one. This has implications on e.g. multiple data serving outlets which require text nuancing. GRUAN is a partnership / confederation not a strictly managed network of uniformity.

    2. The current draft partially re-opens some closed or closed for now issues such as scheduling. Its also not clear whether the data policy is new – we already have a data policy after considerable effort by GCOS secretariat.

    3. Better linking to aspects of existing programs through placing URLs or similar may help.

    4. Because this is going to be WMO documentation I think it needs up front that many of these stations are not currently and are unlikely to be in the future under the direct purview of NMSs . This has practical implications for WMO involvement / managaement which is why it needs spelling out up front.

    There’s much more in specific comments. Sadly I only have time for a single sweep so my apologies if this was not as thorough as it should be.


  3. Kevin Trenberth says:

    The manual is very useful and well put together but there is one major issue that is not adequately addressed at all, and that is the temporal sampling. This pervades the advice in the manual throughout and has major implications for section 7, in particular. The whole philosophy of the GRUAN is flawed unless the context for each observation can be taken into account. I do not believe the GRUAN record itself is useful for trends; what it must be useful for is calibrating variability and trends on all time scales from other observations and analyses and removing biases.
    The only way to deal with variability and trends properly is to have complete temporal sampling over time, preferably every six hours or more, and calibrate those observations using GRUAN. Alternatively one might place the GRUAN observations in the context of the full time series to assess how representative the observations were. Observations are continually being made from satellites and, along with the basic GUAN and radiosonde network, provide the context. These observations are assimilated using four dimensional data assimilation to produce global analyses of temperature and water vapor every six hours, and for some recent reanalyses, every 3 hours. Nowadays each observation has a time attached to it. For GRUAN, the sounding is not almost instantaneous, as a satellite sounding is, and it is not at a fixed location: the balloon goes in different directions each day and differing distances depending on the winds. The top level is not above the bottom level but further downstream. The whole synoptic variability is not addressed in this document.
    It is not sound to use the GRUAN observations alone to create a climate data record without a parallel time series of synoptic variability that places each observation in synoptic context. This could happen in a traditional sense if a GRUAN set of soundings every week were combined with a nearby GUAN every 12 hours. The combination is very powerful. For instance, for a given month with, say, 4 GRUAN obs and 60 12-hourly GUAN obs, the latter can provide the monthly mean while the bias in that mean and rms differences are assessed wrt GRUAN. Over years, the bias may vary, but the monthly values are much more reliable because they do not suffer from sampling error for that month. [A different kind of sampling error comes from whether that month is representative of the population, and that is given by the standard error.] 12 hours is not optimal owing to diurnal cycles and high frequency changes, but it is close to what we have had.
    To me this suggests that either a continual set of soundings at least every 12 hours is mandatory, or the GRUAN soundings must be placed in the context of the global analyses, and thus real time assimilation of the GRUAN observations should be strongly encouraged. It also strongly mandates that observations should be at standard synoptic hours(UTC) not local times, since that is when the analyses apply. The illusion that one can make a sounding coincide with a satellite sounding should be abandoned. In future things will be even more sophisticated: the best record will fully take account of the exact location and time of each observation as part of a global analysis. This is already done in 4D Var (used by ECMWF) and updating incremental techniques (used by MERRA). Monthly biases of the analyses wrt GRUAN (or vice versa) can be readily obtained (cf the work of Haimberger). If the final GRUAN product differs from the real time version, that can readily be built in.
    In the same sense, the need is to have climate data records that establish variability on all time scales, not just “trends” whatever they may be (function of time scale).

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: