A new scale‐selective measure of curvature for meteorological fields

Abstract

We introduce a new diagnostic called curvature using radial variation (CURV) that objectively quantifies spatially varying cyclonic or anticyclonic curvature from fields of mean‐sea‐level pressure (MSLP) or geopotential height Z . CURV enables patterns of “lows”, “highs”, “troughs”, and “ridges” to be objectively identified and differs from methods that detect singular features, such as cyclone centres or trough axes, or from mathematical operators such as vorticity, that act more locally. It is conceptually simple, providing a discrete set of values that are increasingly negative/positive for increasingly cyclonic/anticyclonic curvature with defined minima/maxima for closed cyclones/anticyclones. CURV can be computed over specified distances, thus enabling scale separation of weather systems, and is applicable at any latitude. The method for computing CURV, along with its attributes, is described with examples provided using medium‐range ensemble control forecasts from the European Centre for Medium Range Weather Forecasts (ECMWF). The relationship between CURV and MSLP is examined and comparisons are made with the Laplacian of Z , vorticity, and circulation before discussing some practical considerations. As an example of its application, CURV is used to verify a year of 15‐day ECMWF medium‐range and extended‐range ensemble forecasts. The verification reveals greater predictability of synoptic patterns at larger scales, with anticyclonic curvature more predictable than cyclonic. Slightly higher skill is seen for the first week over Europe compared with the Northern Hemisphere. Skill in the Tropics transitions from lower to higher than the Northern Hemisphere, with the crossover 2 days earlier at large mesoscale compared with synoptic scale. A range of further CURV applications, such as regime‐based verification, ensemble member selection, cyclone objects, and climate time‐scales, are discussed at the end.