Trajectory predictions for nHADT1

This page provides a number of ways of visualising the weather predictions for balloon flights over the coming 16 days. The code is based on the CUSF balloon predictor software.

The predictions make use of the NOAA GFS weather model. This generally provides excellent predictions several days in advance. Actual landing positions are generally within a few kilometers of the launch-time predictions. The accuracy of predictions has been assessed by comparing the 0 hour prediction for a given launch time with the predictions for the same time which were generated several days before. The results allow the addition of uncertainty circles to future predictions.

In practice, the trajectories can be modified by modulating the rate of ascent (the model assumes a constant rate). This can provide significant improvement to the trajectories in certain cases; when the wind is blowing in different directions at different altitudes. However, when there is no wind component in a given direction (e.g. North), there is no way of piloting the balloon in that direction.

The GFS model provides predictions up to 16 days in the future. Predictions are available for the full 16 days here. The 95%ile error at 10 days is approximately 110 km (i.e. 95% of landing positions will be within 110 km of the point predicted 10 days beforehand).

Predictions for the forthcoming week are available here in the same format as the 16 day predictions. Versions with the median uncertainty and 95%ile uncertainty are also available.