I just read a book on atmospheric remote sensing which had discussion throughout on balloon-based sondes, which are the atmospheric counterpart of ocean drifters.
In the early days of remote sensing, it was thought that these sondes, which are neutrally buoyant and can stay aloft for months, could provide enough data to constrain weather models, especially in the southern hemisphere. And then satellite data with all its calibration problems would not be needed.
It turned out that the sondes tended to drift into certain regions and cluster there, leaving large areas without data. Of course, sondes still play a key role (as do the ocean drifters), but the strategy of just using natural drift to gather data everywhere didn't work.
