I have written several posts lately on weather and each has mentioned GRIBs. In this post I’m going to get into the nuts and bolts of the software we use to view GRIB data.
What You Need Is Free
It seems like hardly a month goes by without yet another premium (that means you pay) weather site being announced. But as far as I can see, all of these sites, despite their claims of adding value, are doing nothing but repackaging government generated content that you can get for free. (Actually, particularly if you are an American, you have already paid for it in your taxes, so I think that companies that try to make you pay again are being more than a little cheeky!)
If you think I’m wrong about this, wash your mouth out with soap. Just kidding, please leave a comment with the web address and details of the value added over and above the underlying government provided data and I will take a look.
Viewers We Use
So let’s look at the GRIB viewer included with our electronic navigation system and the two free viewers that we use:
Nobeltec Admiral
Nobeltec Admiral, our electronic navigation software, has an integrated GRIB viewer as well as GRIB email request feature.
Strengths
- Overlays the wind and waves on the chart so that we can see the upcoming weather in relation to our course.
- Adjusts the forecasts to local time from GMT, which saves us doing it in our heads.
Weaknesses
- The email GRIB request feature is clunky and hard to use.
- There is no way to request 500 mb data.
- We find the display poor and difficult to read when we add pressure and precipitation.
Our Use
We request and download the data using other tools, and display wind only as it relates to our immediate surroundings and projected course.
UGRIB
UGRIB is a viewer offered for free by GRIB US.
Strengths
- As long as you have internet access, it is very easy to select the area you are interested in and download a GRIB that covers it.
Weaknesses
- Does not generate an email request to a GRIB server so that you can receive GRIBs as email attachments, which is a much more efficient method if you are using a satellite phone like an Iridium, as we are, and the only option if you are using SailMail over HF radio.
- Has no way to download or display 500 mb forecasts. See this post for why that’s a real drawback.
- Only downloads seven days of data, but we find that looking out as much as 10 days, particularly at the 500 mb forecast, can be useful.
- Does not annotate the surface pressure to indicate what the central pressure of a system is or even whether it is a low or a high. Yes, I know you can mouse over the system to see this information, but I don’t find that very elegant.
Our Use
When we have internet access, we use UGRIB for a quick and easy download of the winds for the immediate area around us.
ViewFax
ViewFax was written by Jim Corenman of Sailmail fame and is available from the associated Saildocs web site for free.
Strengths
- Displays all the data available in GRIB files (wind, waves, surface pressure, precipitation, surface temperature, and 500 mb) .
- By far the most intuitive to read of the three viewers, particularly when multiple data types are being displayed at once.
- Annotates the surface pressure systems with L or H and central pressure.
- Highlights the 564 line in bold on the 500 mb forecast display. (The 564 contour is the so called “storm line” that depressions (lows) tend to track just north of. It is also generally the southern limit of gale force winds.)
- You can enter your boat’s position, course and speed, and the software will advance your position as you step through the forecasts and display the forecast wind speed and direction at each step for your projected location.
- Includes an elegant utility to request GRIBs and other information from the SailDocs server as long as you are connected to the internet.
Weaknesses
- Does not have an integrated email GRIB request feature, I think because Jim, quite justifiably, assumes that you are using Sailmail, which does. So if you, like us, are not using SailMail, you will need to learn how to formulate a raw request to the SailDocs server, as we have.
- The most feature rich version 5.0.56 is still in beta test and does have a couple of bugs including the tendency to crash occasionally (gracefully, without hanging your computer) and the boat advance feature seems too fast, at least in the high latitudes.
Our Use
As I’m sure you have guessed by now, ViewFax is our preferred viewer, particularly for looking way out into the future and getting the big picture.
Bob Hinden October 15, 2011 at 1:56 pm
I use ViewFax along with SailMail and like them both.
The beta version of ViewFax (ver 5.0.056) also includes a direct download capability. From: http://www.siriuscyber.net/wxfax/
“Gribs, fax images, and text forecasts can be downloaded directly from the Saildocs server (if an internet connection is available), see the File menu”
Bob
John October 15, 2011 at 5:23 pm
Hi Bob,
Oops I missed the new internet download feature in the latest version of ViewFax, and very elegant it is too. I think that will mean that we won’t use UGRIB at all in the future. As far as I can see, Viewfax still does not have the ability to generate a request email to the Saildocs server. I guess because, as I understand it, that feature is within the SailMail client? Do I have that right?
Bob Hinden October 15, 2011 at 9:20 pm
John,
That is correct. There is a similar interface in SailMail client that sends email requests to Saildocs.
SailMail also includes the features to send/retrieve email directly over an Internet connection and/or Iridium phone, as well as SSB email. I haven’t used the Iridium features, but would think it’s very efficient in its use of bandwidth and connect time.
Kudos to Jim Corenman for all of this!
Bob
Alex October 16, 2011 at 9:06 am
Does anyone knows the free software Zygrib? What to think about it? http://www.zygrib.org/
Alan Teale January 6, 2012 at 3:44 am
Alex, Zygrib has one very useful feature that the other viewers I use don’t have, including Viewfax. This is the capability to see CAPE data. CAPE data give an indication of the probability of squalls and thunderstorms generated by temperature differences between different levels of the atmosphere. I am not a meteorologist, but that at least is my understanding. If you have CAPE data you can make better decisions on whether you need to run your radar overnight to warn of squalls. Alan
Alan Teale January 6, 2012 at 4:46 am
Alex, A correction. Viewfax can also do CAPE. Alan
Alan Teale January 7, 2012 at 3:29 am
But … the Viewfax Beta version 5.05.6 doesn’t do CAPE well. If you want CAPE data you must request CAPE as a parameter in an email. It doesn’t appear to be an option within the programme. It is also not mentioned as an option in the Saildocs guidance. Third, the display does not quantify the parameter numerically, i.e., does not give the raw J/kg data. Zygrib on the other hand does provide the raw data, but only bundled with other atmospheric data which makes it quite costly in download time. The fourth difficulty with Viewfax is that it can take some persuading to actually show CAPE, although persistence eventually pays off.
On balance I find Viewfax has the edge in terms of overall capabilities and user-friendliness, but it would be nice if it were to improve on CAPE, which I think many folk would find useful. Alan
Ernie October 16, 2011 at 11:44 am
John…thanks for the posting on GRIBs. We have been using a simple reader from Xaxero Marine Software called Gribplot 9.05. I don’t remember whether we paid for this or not. Have had it for quite a few years. What this allows us to do is download GRIB files via email which we access by sending an email to the robot at globalmarinenet.net….much smaller files that don’t require internet connection if you get your email via SSB…thus the beauty.
Also…recently downloaded to my iPhone an app from Weathertrack that allows GRIBs to be read. This is a really cool app that is very well thought out by the designer. I think it cost me 9.99 or something but well worth it as it works inland as well when away from the coast…
Thanks so much once again for all of posts from you and Phyllis. Bette and I read them with a magnifier as we ready our boat for future adventures…
Ernie
David October 16, 2011 at 3:37 pm
Hi John,
Thanks for another very helpful post. Can you comment on availability of these solutions for Apple users? Viewfax seems to only have the options to download a .exe file, which if I’m not mistaken will only work on Windows machines. Will I need a Windows machine to get good weather data?
Cheers,
David
John October 17, 2011 at 8:00 am
Hi David,
I don’t have a Mac, and so I’m not in a position to evaluate GRIB viewers for Mac. I have always loved Macs and in fact a division of my old company, when I was in the computer business, sold them. However, when I went cruising I found that there was just more choices in marine software for Windows.
Can anyone else help with GRIB viewers for Mac?
Jeff March 20, 2013 at 6:45 pm
You may want to look into OCENS WeatherNet and GRIB Explorer packages for Mac.
Chris October 18, 2011 at 3:23 pm
GRIB x 2
John, a trick I learned a few years ago was to sample the GRIBs at a regular time of day. Then I would step through them in side by side windows (for example, allowing 4 am data from the 12th to be compared with 4 am data from the 13th). Headers and lifts became quickly apparent as did N/S changes to H or L tracks.
I did the same with predicted vs observed (via GRIB) and predicted vs observed (via text…a human in the loop forecast).
Lo and behold, I found spring and fall GRIB data and text forecasts became statistically dodgy beyond four days. The likelihood that the weather observed on day “5″ was what had been depicted on day “0″ was less than 50%. Summer and winter forecasts did better with wind direction than strength.
Interestingly when I asked people how they felt about the “0″ day predicted weather, they couldn’t remember that far back and so were inclined to give the forecasts a much higher or much lower score than statistics. This seemed to be based on how well the weather supported or interfered with their planned activities — picnics, sailing, flying, etc. People with no planned activities generally had no idea what the predicted weather was even from day “3.”
The other thing I do is as I step through the 500mb GRIB data, I print each increment to a GIF with all but the isobars transparent. These I paste as geo-referenced layers to an appropriate map. Then I draw a line from the closest point of each 564 line to our projected position. I sometimes make animated GIFs of the lines to illustrate a point. Depending on the jet stream, planned scalloped courses can make for better sailing than straight ones.
—————————
When I was learning to fly, the location, speed and track of the 570mb iso determined whether I would be allowed to make long solo flights across Texas.
John October 19, 2011 at 9:43 am
Hi Chris,
Thanks for the great comment, as always. Sounds like you have a very good system. I particularly like your scientific approach to accessing forecast accuracy. Bottom line, as you have found, forecast accuracy suffers, no matter how good the model or how smart the forecaster, after 3-4 days.
One of the great things about developing an understanding of the 500 mb forecasts, as you have done, is that it gives us tools to assess how difficult the prevailing pattern is to forecast and therefore we can assign a level of comfort to what we are being told.
Chris October 19, 2011 at 3:16 pm
Just want to be clear I’m amazed by how well the Weather Folks do given the semi-chaotic nature of their business.
Where I live they have a cauldron of effects to deal with. The Atlantic, the Chesapeake, the Appalachians, continental air from Canada and the maritime air from the Gulf (not to mention the special features like tropical and extra-tropical lows).
Four wins out of seven gets you the Stanley Cup, and the Weather folks do it over and over again.
John October 19, 2011 at 5:08 pm
Hi Chris,
I could not agree more, as we say on “Morgan’s Cloud” “it’s a forecast, not a prophecy”. I think many people do not understand the fundamental chaos that meteorologists are dealing with and that to be accurate most of the time out to 4 days is an incredible achievement. I can easily remember when 48 hours of accuracy 50% of the time was the best one could hope for.
David Head October 19, 2011 at 5:20 am
So much here that is useful. I use ugrib as and when we need some confidence, but as others feel the longer term forecasts are less accurate. In the UK the majority of our weather has had an Atlantic crossing and predictability does suffer. Generally the systems choose an upper or lower path from about 500nm out in the Atlantic, thereafter some predictability does appear.
I am so fed up with the desperate inaccuracy and generalising offered by our UK Met Office that in 2012 I am subscribing to Theyr (http://theyr.com). I have confidence in them and find they offer an excellent accurate service.
John October 19, 2011 at 9:37 am
Hi David,
I agree completely that the weather on the east side of the Atlantic is much more difficult to forecast than that on the west, as we found when we were in Europe on “Morgan’s Cloud” for three years. For a start, by the time they get to you, lows tend to be occluded and spawning secondary systems, which makes them inherently more difficult to forecast.
I took a look at Theyr and their tools do look very slick and the price seems more reasonable than some. Having said that, I still suspect that the underlying models that they use are those produced by government agencies and therefore you can get access to the same information for free using SailDocs. I can’t see anywhere where they are actually adding forecast accuracy other than by providing a higher level of detail than government agency forecasts.
We have found that we can significantly improve accuracy over the Canadian forecasts too, by going to the original data using free GRIB download. I think the reason is that the government forecasters simply don’t bother to extract the maximum available detail from their own models. There also seems to be a 24 hour delay in a significant change to a model making it to the text forecast.
The exception, in my experience, is the NOAA (USA) forecasts, which are excellent.
Frank Singleton February 20, 2012 at 5:00 am
John
A few comments about GRIBs.
I have reviewed some free viewers at http://weather.mailasail.com/Franks-Weather/Grib-File-Viewers. It is very much a personal matter but my choice is zyGrib which has better map projections than Ugrib and options for detail such as town name, rivers etc. I have used Viewfax a great t deal and have always found it easy to use.
My preferred way of getting GRIBs is via Saildocs. This has a repeat request facility so that one email from me gets a daily GRIB until I say stop. Saildocs is also the most flexible in that, having chosen grid spacing and length of forecast, I can use a variable time interval. For example, I can ask for a 3 hour interval for the first 36 hours and 1 hour intervals out to 8 days. Having got a GRIB file, I can open it whatever viewer I choose – eg zyGrib.
I have used GRIBs 5 months each year over the past 12 years around Western Europe and the Mediterranean. Over that period, I have seen a general improvement so the now I get good guidance usually to 5 to 7 days ahead. My technique is to look at GRIBs on a daily basis, or 12 hourly if it is critical, and compare successive forecasts looking for consistency. This is a poor man’s ensemble.
I doubt that here is any significant systematic difference between the various GRIBs. Of course, you have to remember that any grid can only represent weather and topography on a scale of about 4 or 5 grid lengths. This, for global models, is about 100 to 130 km at best although we only get the output on a 50 km grid.
I am rather sceptical about the value of meso-scale models run by private firms. As far as I am aware they do not start with an analysis as detailed as those used by National Met Services. They may improve on the GFS through areas such as the Gibraltar and Dover Straits. They may represent some large scale sea breeze effects. However, they are dependent on the global models in which they are nested. If the GFS is not good then the meso-scale model will not improve matters.
For many sailing purposes, I doubt that any meso-scale model will be of much benefit. Looking at output from models such as the US Navy COAMPS and the UK NAE (both using grid lengths of around 10 km), I find it difficult to see real benefit. You get more data but little more information. With a smaller grid length such as the UK Met Office 1.5 km model, weather features described will be small. About 7 km. Such sizes of weather feature will have a short lifetime and will have disappeared before we can get the output. For round the cans racing fleets, wanting short period detailed forecasts it might well be different.
I thank Alan Teale for pointing out that Saildocs can provide CAPE. This parameter can be obtained from zyGrib either individually or in conjunction with other parameters. Using Saildocs, I have only been able to get CAPE in a single email attachment and have to use a separate message to ask for, say, wind and rain. Having got CAPE from Saildocs, the best way to display it is probably by using the zyGrib viewer.
John February 20, 2012 at 10:25 am
Hi Frank,
Once again, thank you for the useful information packed comment. It’s great to have a professional meteorologist confirm my gut feeling that most of the GRIB for fee services are more sizzle than sausage.
Also, thanks for the tip on zyGrib, I will take a look at it.
Frank Singleton February 21, 2012 at 10:41 am
John
John
Some people have problems in unpacking the zyGrib software. You might like to see
http://weather.mailasail.com/Franks-Weather/Zygrib-Ftp-Grib-Service and
http://weather.mailasail.com/Franks-Weather/Zygrib-Ftp-Grib-Software
Sujit January 16, 2013 at 10:16 am
Hi,
Thanks for the detailed information above. I am looking at a GRIB viewer that displays wave height and direction inclduing wind/pressure details.
I need to download the old Grib files directly from NCEP site and upload on a viewer. Is there any software that will allow me to do that ?
Frank Singleton March 21, 2013 at 8:53 am
I think zyGrib is your best bet. It can handle .grb as well as .grb.bz2 files.
Frank Singleton March 21, 2013 at 12:04 pm
If you are using Windows and having problems in opening the zyGrib file from the zyGrib ie the .tgz file then either try
http://weather.mailasail.com/Franks-Weather/Zygrib-Ftp-Grib-Software
or go to Tony Cross’ public Dropbox site where you will find a Winzip version
http://dl.dropbox.com/u/25934908/zyGrib_win.zip.
or go to
http://www.ziltmagazine.com/software/
For an uncompressed version — about 125 mb as opposed to about 27 in compressed form.