Xcskies

On Sunday, September 4, 2016 at 11:33:52 PM UTC-6, Ramy wrote:
I've been using Dr Jack blipmap and RASP for over 10 years. Never had a problem, and it proved to be the most accurate forecasting tool around. And it 's free. Never used XCSkies, never had a reason to. Unless you live outside of the US I don't see a reason why paying for an inferior service with zero support. Sure the UI is much more modern and you can zoom in on the maps as much as you want, but this does not make it any more accurate. To the contrary.

Ramy

Disclaimer: I am not an expert(i.e. meteorologist), but I have read and researched a fair amount on this topic, so will share some of what I think I've learned.

First, I've been using Dr. Jack and XCSkies (as well as at least 4 other forecasting sites) for more years than I care to count. Since 2010, my subscription period to XCSkies, I have had no subscription issues. Each year it has renewed smoothly. I've also had no issues getting responses to questions when I've asked, but then I haven't asked any for the past couple of years.. Note that these models are all automated. The data are produced and users just have to initiate a hookup to take it in and put it into their website and then display it with whatever representations they think appropriate. There are many Dr. Jack sites set up by others that take the same basic data, but post it in different formats to make it useful for a particular soaring site. XCSkies took a different approach to displaying these data. The websites themselves pretty much run on automatic.

Second, there is no difference in accuracy between Dr. Jack and XCSkies (and any other site that uses the standard GFS, NAM, RAP models. Each model has its own cell resolution, from coarse to fine, and time frame, and each gives the same output. This is a set parameter, regardless of zoom level. They all use the same algorithms and same data inputs, depending on the model.. So they all give the same output for the time of data calculation, the remainder is a matter of presentation.

There is a difference when looking at NWS sites because the meteorologists, using the same models, will put an experiential factor into their forecasts, i.e. they tweak the results regardless of what the models say. Hence NWS forecasts will often deviate substantially from XCSkies or Dr. Jack for basic parameters. It is quite instructive to read the NWS meteorologists reasoning for their forecasts, and also note their comments on model convergence or divergence for a few days out.

The accuracy does depend on where the actual soundings are made, and then the computer model interpolation of the models for the location you are interested in. In some cases, this can be reasonably accurate, for others totally and obviously incorrect because the sites are so far from the actual sounding location.

In comparing between forecast sites, one has to be sure they are comparing the forecasts with the same models. On XCSkies, which enables easily comparing GFS, NAM and RAP for the same day/time there will be considerable differences having to do with cell resolution, time frame, and model differences ranging from no soaring to a great day to soar. Further, in some locations, GFS will be more accurate than RAP (theoretically the most accurate), and v.v. in other locations. You have to build an experience base to know which works best for where you fly. When I go to a new location, I generally average the forecasts of the three models - which is probably about as good as anything. (Note; there have been studies showing that weather forecasting is about 50% accurate, regardless of model. Weather is a dynamic, nonlinear phenomenon. Forecasting is linear extrapolation and interpolation, which leaves a lot of room for things to change in short order.)

XCSkies is the clear UI winner for ease of use (my opinion). As Dan says, you can go anywhere and get a forecast very easily. However, there are things XCSkies does not do and Dr. Jack's blipmaps do. For a wave or convergence forecast, Dr. Jack gives it direct. XCSkies has to be interpreted for convergence (you can get it by looking at winds aloft), but doesn't attempt wave forecasting. Dr. Jack, on the other hand,can give pretty accurate wave forecasts (as I discovered flying in Boulder, CO). If you really want to know the likelihood of wave, there is a separate site that shows a forecast of the jet stream out to seven days, which is a major factor driving wave conditions. You can also look at the whole earth winds map at various altitudes to give an idea of trends.

On the other hand, XCSkies readily gives a route forecast for a task. You can put in a start time and finish time, enter the route, and get a reasonable idea of what the flight will be like for good lift spots, dead spots, and the like as a function of time of day. The model takes into account winds and lift, and shows how the day might progress over the route. I can do this with Dr. Jack as well, but it is more laborious.

Both Dr. Jack and XCSkies have periodic update issues having to do when the model inputs are available. Often, a forecast for current day+1 on Dr. Jack just isn't available. Sometimes it's not available the day of planning to fly. XCSkies has the same issue. If one site doesn't get the model outputs, the other site won't either. Of course, it can be an issue with one site or the other not connecting, but that is invisible to users - it just shows up as "no data."

Bottom line is that Dr. Jack set the bar and the method. Some present data as he did originally, others, like XCSkies, have tried a different means of data presentation. In the end, both (in my opinion) are tools that are helpful, both have strengths and weaknesses - but that's soaring. I haven't found one superior to the other. Both are simply inputs that are factored into deciding whether to fly and what it might be like. A question is whether you get a better idea of soaring conditions by using one model or several. Perhaps the answer is that it depends on what each of us is comfortable using.

My favorite "forecasting" method has been to go to the gliderport, launch and see what develops. This is 100% accurate, and supersedes any forecasting model. Had some of my best days this way.

Sorry for the long post.

Eric -

Eric, thanks for the detailed analysis. Very helpful.

Ramy

On Tuesday, September 6, 2016 at 1:37:10 AM UTC-4, Ramy wrote:
Eric, thanks for the detailed analysis. Very helpful.

Ramy

The key things to keep aware of when using any of these sites (Unisys, Dr. Jack, XCSkies, TopMeteo) a

1. What model data are you looking at (GFS, NAM, RAP, etc.) along with the knowledge of the strengths and weaknesses of each related to your area.

2. What model run are you seeing. Each site occasionally lags on the model run that is being displayed. You do not always see the most recent.

On 9/5/2016 8:58 PM, Eric Bick (ZN7) wrote:
My favorite "forecasting" method has been to go to the gliderport, launch and see what develops. This is 100% accurate, and supersedes any forecasting model. Had some of my best days this way.
Wonderful explanation, Eric, and I wholeheartedly agree with the above
statement! BTW, I enjoyed meeting you at the airport the other day.

Now, if only the monsoon would quiet down...
--
Dan, 5J

On Tuesday, September 6, 2016 at 10:03:36 AM UTC-6, Dan Marotta wrote:
On 9/5/2016 8:58 PM, Eric Bick (ZN7) wrote:
My favorite "forecasting" method has been to go to the gliderport, launch and see what develops. This is 100% accurate, and supersedes any forecasting model. Had some of my best days this way.
Wonderful explanation, Eric, and I wholeheartedly agree with the above
statement! BTW, I enjoyed meeting you at the airport the other day.

Now, if only the monsoon would quiet down...
--
Dan, 5J

Amen re the monsoon stuff. Getting tiresome.

On Tuesday, September 6, 2016 at 5:51:11 AM UTC-7, John Godfrey (QT) wrote:
On Tuesday, September 6, 2016 at 1:37:10 AM UTC-4, Ramy wrote:
Eric, thanks for the detailed analysis. Very helpful.

Ramy

The key things to keep aware of when using any of these sites (Unisys, Dr.. Jack, XCSkies, TopMeteo) a

1. What model data are you looking at (GFS, NAM, RAP, etc.) along with the knowledge of the strengths and weaknesses of each related to your area.

2. What model run are you seeing. Each site occasionally lags on the model run that is being displayed. You do not always see the most recent.

Topmeteo is different from Blipmaps and XCSkies in that they have their own proprietary model; they use data from NWS only to set initial conditions. Consequently the forecast from Topmeteo can be quite different from Blipmap.. Also, Topmeteo gives hourly forecasts; RAP is only available on 3 hour intervals.

I have compared the RAP model soundings to the actual balloon (RAOB) soundings and have found a significant loss in vertical resolution. Sometimes it doesn't matter, but other times the real data shows a distinct temperature inversion that is lost in the models. This results in stronger thermals to significantly higher cloud bases, and clouds vs blue, than what the models predicted. This happened day after day this summer in Ely. If you want the REAL picture you MUST get "under the hood" and understand how the soundings work and do your own real vs model comparisons.

All in all, however, Topmeteo is clearly superior to Blipmaps and XCSkies.

Tom