How high could a weather balloon ascend while under control. Can we control the ascent and decent speeds? Could we program it to respond to various climate inputs? How is the monitoring of the atmospheric conditions presently done?
The standard NWS weather balloons are totally uncontrolled. They fill 'em with helium and turn them loose. The rate of climb and the lateral path they take after that are whatever they turn out to be. Meanwhile the instrument package is sending back data continuously.
Here’s more about the instrument package:
And here is the NWS’s explanation for laymen which may give you some more search terms. Radiosondes | National Oceanic and Atmospheric Administration (noaa.gov)
So that’s what’s routinely done all day every day. It’s unclear whether you’re trying to ask what could be done, rather than what is done. These devices are disposable, so there’s a strong incentive to minimize capabilities that aren’t essential.
What could be done is more elaborate, but would need a reason to be done before anyone would bother. What problem do you suppose a more maneuverable (or at least more controllable) balloon would solve?
And of course, that is in itself part of the data that they yield.
Regarding blimps (ie airships), they can do a pretty good job of controlling where they go and aren’t always forced to go wherever the wind blows. More modern ones have a lot more props than just one or two at the back of the gondola. They still have to do what they can stay in calm areas, but we’ve come a long way since the Hindenburg and the Good Year Blimp.
This doesn’t entirely answer the questions in the OP, but it does address parts of it.
I don’t know exactly what I would be looking for, I just think that more control might produce better info on how heat and humidity move around. A little more detail.
More so than letting a GPS receiver track exactly how the heat and humidity are moving it? Wouldn’t that be easier than trying differentiate between movement due to air currents vs movement due to it’s own controls?
Like, I could figure out how far my car was drifting due to bad alignment by hooking up sensors and servos to the steering wheel. The servos could make sure it tracks straight and the sensors could determine how much work the servos are doing to maintain a straight line. Or I could just mark a starting point, take my hands of the wheel and see how far it drifts over a given distance.
If you want more detail, launch 10,000 from 10,000 spots on the ground, not 1,000 from 1,000 spots on the ground. Likewise do it hourly, not once every 12 hours.
What you will find however, is that people who are experts in this stuff have discovered that when predicting large scale phenomena like weather fronts, finer resolution doesn’t help much. Beyond a certain point more sampling just increases the volume of data to crunch, but doesn’t improve the fidelity of the outcome.
The mathematical field is called chaos theory. The atmosphere is not a clockwork gizmo where if you knew every molecule individually you could predict exactly the weather for next Tuesday at 1:23 pm at your house and separately for your neighbor across the street at 1:25pm. Even armed with perfect knowledge the system must evolve in ways that can only be fuzzily predicted.