Ars Technica: A week before Superstorm Sandy struck the coast of New Jersey, the European Centre for Medium-Range Weather Forecasts (ECMWF) predicted the storm would indeed make landfall, whereas the US National Weather Service (NWS) had the storm veering off into the Atlantic Ocean. The NWS model fell into line with the ECMWF model three days later and enabled the NWS to issue timely, life-saving warnings. Still, the apparent shortcoming of the NWS model has highlighted a gap in investment between the US and Europe, writes Scott Johnson for Ars Technica. Not only does the ECMWF model run on a faster supercomputer than the NWS model does, but the European model has finer spatial and temporal resolution. Both advantages enhance accuracy. Comparisons aside, the ECMWF and NWS models rely on data gathered by Earth-observing satellites. If that fleet is not replenished, the accuracy of weather prediction will suffer. Indeed, after Sandy had dissipated, the ECMWF reran its model with data available five days before the storm’s landfall but omitted data from NASA’s polar-orbiting satellites. Without those data, the ECMWF model failed to predict Sandy’s destructive path.