With relatively low-altitude flights, say less than 5,000 feet, one only needs to launch one test rocket,
watch where it lands, and assume that future flights will trend in the same direction, at least for a few
hours. But above that, there can be very high wind velocities in unexpected directions, and even the
trend of the rocket can come as a surprise.
NOAA makes the results of their atmospheric models available on the Web as “soundings” at high
resolution. Up to 24 hours in advance, these soundings can be used to get a profile of the speed and
direction of winds aloft at about 500 feet vertical resolution all the way up to about 70,000 feet, far
above the typical waiver altitude. I wrote some simple software to compute the drift of a rocket, given
its descent rate, and predict the final landing point. A few hours before heading to the range, I ran the
software, which predicted a landing to the north of the launch site.
Figure DRIFT shows the predicted landing points using pre-flight soundings (retrieved about 16 hours
prior to the flight), post-flight soundings (which is still based on modeling rather than actual
observations), and the actual landing point as determined by a handhand GPS unit. The distance to the
landing site was fairly well predicted (within 15% or so), while the bearing was about 15-20 degrees
off. While obviously not a substitute for tracking, this is an improvement over not having any idea of
which direction the rocket might go.
It should be noted that my software assumes a perfectly-vertical ascent. A future enhancement would
be to integrate the sounding data into a full 6-degree-of-freedom flight simulation program like
OpenRocket; that would also model the effects of the wind profile on the ascent.