The spatial distribution of aerosols influences the cooling efficiency, and all aspects of the surface climate response. Part of the resulting uncertainty can be minimized by control systems which alter injection locations in response, There are limited observations of stratospheric circulation, and poor model performance in matching existing observations.
Metric
The average spatial pattern obtained in releasing a hypothetical passive tracer is outside the multi-model range.
Uncertainty
Stratospheric circulation is not currently well observed, and there is significant model-observation disagreement. Note that defining this for a passive tracer excludes uncertainty arising from differences in sedimentation from varying aerosol size distribution, and from feedbacks from stratospheric heating. The uncertainty here has contributions from both errors in the mean circulation, and uncertainty in mixing.
Decision relevance
Tests and early-stage deployment will rapidly resolve this uncertainty in transport, and choice of injection location likely removes much of its consequence. That being said, the multi-model range is too large to not list decision relevance as medium here as this already allows for significant variance.