Introduction
Feasibility studies are being performed numerically, using NCEP/NCAR Reanalysis Project data, to find out what effort is required to maintain balloons in stratospheric orbits by managing the height of each balloon. A year's worth of reanaysis data was used to simulate the effects of stratospheric winds at any given balloon altitude for the life of each ballon. We are in the process of developing a balloon management strategy for the Global Atmosphere-Ocean IN-situ System, (GAINS) project.  The results presented here are of an initial numerical experiment in automatically controlling the balloon's orbiting latitude.

Fig. 1. A plot of a year's drift simulation of a balloon released from Tillamook, OR and steered to 20 N.

Application of reanalysis data
Reanalysis data are available at 17 levels, globally on a 2.5 by 2.5 degree grid in longitude and latitude, every six hours for a full year per file. The six hour kinematic data are interpolated linearly to the particular time and location of the balloon, and these data are used to model the drift of the balloon. The relase point for the balloon is Tillamook, OR (42.38 N  122.87 W). The integration time step to compute the balloon's trajectory is 0.1 h.

Model description
The balloon is allowed to drift in the wind at the same pressure level for a fixed period before any correction is made in flight level.
Here a 12 h fixed period (the interval between synoptic soundings) has been chosen. This option corresponds to the use of no
Numerical Weater Prediction (NWP) model output. At each synoptic time, a correction in flight level is introduced
instantaneously by selecting the maximum or minimum value of the meridional component of the wind which most quickly would
restore the flight track to the desired latitudinal circle. This simple steering model is able to keep the balloon on track for most of the
year except for a few deep circulations that the ballon encounters trowing it well off course. In the next iteration, a ramp-up time will be introduced for flight level changes, but in this series, an instantaneous change was chosen because we wanted quick, approximate results.

Results
The above model was used in several numerical experiments for various specified, latitudinal orbits. The orbit for a full year about the specified latitude of 20 N is shown in Fig. 1. These results show that the ballon management model generally is able to hold the orbit about the specified latitude but suffers some large excursions from which it takes time to recover. A complimentary way of viewing these results is in Fig. 2 which shows the latitude of the balloon as a function of time.

Fig. 2.  A plot of the balloon's latitude as a function of time in days.

Discussion
The results presented here, and others not shown, indicate that it is possible to keep balloons orbiting in the stratosphere at specified latitudes, using a computer automated management scheme. Sometimes, however, a balloon can become trapped in a strong circulation from which it cannot readily escape by changing flight levels. The result is that the balloon trajectory is held close to the specified latitude a large percentage of the time, but sometimes undergoes large displacements from which it takes a long time to recover.

These results suggest a two step balloon management strategy. First, devise an automate flight management program that does not need human intervention to steer the ballon. This automated flight management program would not depend on model forecasts; but rely only analysis of observations, or initial model data. Second, have a manual override that allows human intervention to steer the baloon away from a strong circulations that are likely to take the balloon very far off course. At this point, model forecasts become an important tool for making flight level decisions.
  ÿ