Search for: All records

Abstract AI-based algorithms are emerging in many meteorological applications that produce imagery as output, including for global weather forecasting models. However, the imagery produced by AI algorithms, especially by convolutional neural networks (CNNs), is often described as too blurry to look realistic, partly because CNNs tend to represent uncertainty as blurriness. This blurriness can be undesirable since it might obscure important meteorological features. More complex AI models, such as Generative AI models, produce images that appear to be sharper. However, improved sharpness may come at the expense of a decline in other performance criteria, such as standard forecast verification metrics. To navigate any trade-off between sharpness and other performance metrics it is important to quantitatively assess those other metrics along with sharpness. While there is a rich set of forecast verification metrics available for meteorological images, none of them focus on sharpness. This paper seeks to fill this gap by 1) exploring a variety of sharpness metrics from other fields, 2) evaluating properties of these metrics, 3) proposing the new concept of Gaussian Blur Equivalence as a tool for their uniform interpretation, and 4) demonstrating their use for sample meteorological applications, including a CNN that emulates radar imagery from satellite imagery (GREMLIN) and an AI-based global weather forecasting model (GraphCast). 

Abstract. This paper describes and analyzes the Reed–Jablonowski (RJ) tropical cyclone (TC) test case used in the 2016 Dynamical Core Model Intercomparison Project (DCMIP2016). This intermediate-complexity test case analyzes the evolution of a weak vortex into a TC in an idealized tropical environment. Reference solutions from nine general circulation models (GCMs) with identical simplified physics parameterization packages that participated in DCMIP2016 are analyzed in this study at 50 km horizontal grid spacing, with five of these models also providing solutions at 25 km grid spacing. Evolution of minimum surface pressure (MSP) and maximum 1 km azimuthally averaged wind speed (MWS), the wind–pressure relationship, radial profiles of wind speed and surface pressure, and wind composites are presented for all participating GCMs at both horizontal grid spacings. While all TCs undergo a similar evolution process, some reach significantly higher intensities than others, ultimately impacting their horizontal and vertical structures. TCs simulated at 25 km grid spacings retain these differences but reach higher intensities and are more compact than their 50 km counterparts. These results indicate that dynamical core choice is an essential factor in GCM development, and future work should be conducted to explore how specific differences within the dynamical core affect TC behavior in GCMs.