GFS & NOGAPS: Massive disagreement

As I mentioned in an earlier post NOGAPS and GFS do not always agree... actually a lot of time they don't agree. However, sometimes the differences in them are really quite massive.

The charts below are for the same day and time and are off of the same model run. The coloured areas show the amount of precipitation that is expected to have fallen in the 12 hour period prior to the listed time - so in this instance between 0600-1800 on Saturday 31 October. As you can see they don't look very much alike. My gut instinct says NOGAPS is closer to being right and I really hope that's the case!

Camera and phone tests complete

The mobile phone is working perfectly having over 50% battery life even after well over 2 hours of tracking. The £5 credit loaded is going to be more than enough for our purposes.

Our camera ran for 1 hour and 23 minutes before running out of space on the card - but still with 25% battery remaining. I'm going to adjust the script to take 5 pictures a minute instead of 6 as we'd like to get a runtime of more than 100 minutes.

T minus 4 days

A "Pay As You Talk" Vodafone SIM has now been acquired for our Windows Mobile device and is now having a full retest. At the same time the camera is also being retested with a 4Gb SDHC card (actually its an 8Gb card formatted to 4Gb!). The reason for dropping to 4Gb from 8 is that our CHDK firmware will not boot on the camera from a FAT32 partition.

Prediction on upper level winds have changed a little, showing an increase of 20 knots. We'll be monitoring the situation, but we are currently still 'go for launch'.

Zephyrus 1 Launch: T minus 5 days

We've been looking at the predicted tracks and examining wind and rain predictions for this coming weekend and we're fairly happy with what we are seeing. No rain is predicted during daylight hours (top chart), upper winds are showing as being no higher than 60 knots (middle chart) and surface winds, although not ideal should still be below 15 knots (bottom chart). Upper and surface winds are showing as being SW on the NOGAPS model and WSW on the GFS model - and either would suit us.

We're now looking at a Saturday launch. Stay tuned.

Track looks promising for coming weekend...

...but heavy rain may prevent launch:

Weather decides whether (to launch)

With all the equipment pretty much ready to go, we are now at the mercy of the weather. Whilst rain is an issue - the wind is even more so. Our gas has to be ordered at least 24 hours before launch as the canister has to be hired by the week - which means we need to predict when we will have the right conditions. Just to make things that little bit more tricky, standard weather reports are of no use to us as surface winds give no indication of upper wind strength, or even direction!

We have access to a balloon flight prediction website (provided by Cambridge University) which gives forecasts for up to 120 hours ahead - but we cannot place complete faith in just this. Although many weather prediction models exist, none are accurate all of the time and often they models can disagree quite substantially.

Only a couple of models are available free to the public (most are expensive subscriptions): GFS, used by the aforementioned prediction website, and NOGAPS. The idea will be to combine the information from both models (which can be analysed for free on the US Navy's weather website) and using the website provided by CU.

At the moment the weekend doesn't look too promising, but I'll be checking later on as the models are re-run every six hours and drastic changes are not uncommon.

Camera test complete

We currently only have a 2Gb card installed in our camera, which we now know needs to be replaced with an 8Gb for launch. Our test shows that we will use around 2Gb every 50 minutes - so we could theoretically get 200 minutes of footage with the higher capacity card. In reality though the batteries would not last for 200 minutes (and we expect our payload to have returned to Earth within 100 minutes).

Although not a 100% realistic test, for one I had to delete the contents of the memory card and restart the script, the battery lasted two full cycles. The CHDK/SDM software suggested there was still around 20-25% battery life left at the end of this test - a good indicator that we will have power for the entire duration of our flight.