Uncertainty-Aware Visualization for Analyzing Heterogeneous Wildfire Detections.

There is growing interest in using data science techniques to characterize and predict natural disasters and extreme weather events. Such techniques merge noisy data gathered in the real world, from sources such as satellite detections, with algorithms that strongly depend on the noise, resolution, and uncertainty in these data. In this study, we present a visualization approach for interpolating multiresolution, uncertain satellite detections of wildfires into intuitive visual representations. We use extrinsic, intrinsic, coincident, and adjacent uncertainty representations as appropriate for understanding the information at each stage. To demonstrate our approach, we use our framework to tune two different algorithms for characterizing satellite detections of wildfires.