Nowcasting of Precipitation with Dr. Zawadzki

Posted on

Dr. Isztar Zawadzki
Department of Atmospheric and Oceanic Sciences, McGill University
Nowcasting of Precipitation

Precipitation is one of the most important meteorological fields and yet it is the most poorly predicted. This is understandable: an animation of radar images from the US continental network shows a constantly evolving pattern that illustrates the concept of Deterministic Chaos in the clearest manner. NWP does a reasonable job at predicting large time-space scales of precipitation but at the mesoscale we must rely on some heuristic approach. The advent of the continental scale US radar network revealed the full chaotic geometry of precipitation, and its evolution. It became clear that scale invariance is an adequate model to describe precipitation patters (multi-fractal, self-affine, geometries). The challenge was to extend the earlier algorithms to the new view of precipitation. For this, the McGill Algorithm for Precipitation nowcasting by Lagrangian Extrapolation (MAPLE) was developed. The algorithm examines the evolution of the field of precipitation in the last 40 minutes and from this information extracts a set of vectors describing this evolution. This set of vectors is then used by a non-diffusive advection scheme to produce a nowcast and an a posteriori evaluation of its skill when new radar data become available. This leads to nowcasting with better skill than NWP for a lead-time between 3 and 7 hours, depending on the meteorological situation. A more physically meaningful variant of MAPLE, optimized MAPLE, or OMAPLE, is based on scale dependency of predictability of precipitation patterns. O-MAPLE greatly improves the quantitative skill (RMS) of the nowcast. Yet another flavour, P-MAPLE, is designed to nowcast the probability of a given intensity at a point. Blending with NWP leads to minor improvements. The challenge now is  to study the nature of errors of NWP and of MAPLE to understand what techniques must be developed to further improve nowcasting.

Isztar Zawadzki is an Emeritus Professor at McGill University where for many years he was Director of the J. Stewart Marshall Radar Observatory. Prior to that he had studied and taught in Buenos Aires and Montreal (at UQAM and McGill). He is very well known and highly respected as one of the founders of the field of “Nowcasting” and as a leader in the field of radar meteorology. He is a fellow of the Royal Society of Canada, and of the Canadian and American Meteorological  Societies. He was the 2007 recipient of the American Meteorological Sensing Prize and the 1998 CMOS president's prize.

Refreshments will be served at 3:15 p.m. in Room 422 Petrie.

DATE: Wednesday, February 12, 2014
TIME: 3:30 p.m.
LOCATION: Room 422, Petrie Building

Leave Your Comment