GEWEX-GPCP Workshop on Objective Analysis of Precipitation

Subject

ECMWF will host the GEWEX sponsored Workshop on Objective Analysis of Precipitation from 11 to 13 March 2003.

The primary goals of the workshop are to improve our understanding of the issues involved in the objective analysis of precipitation using the many sources of information available (e.g., gauges, satellite-derived estimates, radar observations and model output data) and to make recommendations for GPCP to advance its efforts to provide global analyses of precipitation.

While the current GPCP precipitation analyses, as well as others that have been constructed, have proven useful for many purposes, it has become clear that efforts to improve their accuracy and consistency are required. This is partly because of the desire for higher spatial and temporal resolution, which necessitates extracting more information from the available data, and partly because the methods currently in use were developed 5-10 years ago and the state of the science has advanced. This workshop will bring together members of the GPCP community with experts in the theory and practice objective analysis, including in particular, members of the numerical weather prediction (NWP) community, to describe the current state of the science and to make recommendations to GPCP.

Some of the questions that will be considered at the workshop are:

• What is the best (in terms of efficiency, accuracy and physics) analysis procedure for merging and interpolating (spatially and temporally) the various precipitation estimates?
• Are different analysis procedures needed for different spatial and temporal scales?
• What analysis procedures should be used in complex terrain?
• What is the best way to validate results of analyses? How does one validate over the open ocean?
• Since objective analysis techniques deal best with unbiased input data:
  • How can biases in the various input data be removed?
  • Are these biases function of satellite type, geographical location, season and even observation time or date?
  • Can they be done once for all or need to be done at every analysis time, i.e., are they stationary or time dependent?
  • Is bias removal best done during the retrieval process or during the analysis of by a combination of both?

Presentations are planned for the first two days and the remainder is reserved for discussion and preparation of workshop recommendations.

Timetable

Abstracts of presentations

Any request for information regarding this meeting should be directed to the organizers: Masao Kanamitsu or Arnold Gruber.