Developing a Standardized Testing Procedure for Cloud Tracking Wind Measurement Methods

We present preliminary results of our effort to develop a standardized benchmark test for cloud tracking wind measurement methods. Various algorithms have been developed over the years to measure wind speeds in planetary atmospheres through Earth- and space- based remote sensing. However, unlike satellite-based cloud-tracking measurements of Earth, these planetary measurements cannot easily be validated against in-situ data, which makes the interpretation difficult when different cloud-tracking schemes do not agree on their results. To address the issue of data validation, we run multiple automated cloud-tracking algorithms independently developed at multiple institutions on synthetic wind data generated using a General Circulation Model. Our simulations calculate the advection of tracer distributions to represent cloud motions as done by Sayanagi and Showman (2007, Icarus 187, p520-539). The motions of tracers are measured using cloud-tracking software to derive wind vector fields, which will be compared against the model "truth." In our synthetic wind/cloud fields, the tempo-spatial scales of the winds and clouds are separately controlled so that the robustness of cloud tracking tools can be assessed against various conditions. Our setup enables measuring the performance of cloud-tracking software using two metrics. The first metric is the ratio between characteristic length scale of cloud morphology L_cloud and the size of smallest eddies successfully resolved by a cloud-tracking method L_Leddy, λ = L_{cloud/L_{Leddy}}. The second performance metric is the ratio between the temporal interval between image acquisitions T_imaging and the characteristic lifetime of clouds T_cloud, τ = T_{imaging/T_cloud}. These metrics are designed to compare the abilities of tracking algorithms to resolve cloud motions against the absolute theoretical limit; note that both metrics have the maximum value of 1.0 as cloud tracking methods cannot resolve features that change in less than the temporal and spatial scales of the clouds. Our study is supported by a grant from the NSF Planetary Astronomy program.