Extended Abstract (96K)P13B.9
Quality information in processing chain—generalized method of documentation based on Unit of Transformation Step (UTS)
Zdzislaw Dziewit, Institute of Meteorology and Water Mangement, Warszawa, Poland
The increasing interest in international exchange of quantitative radar data, leads to the need for developing universal method for documenting, monitoring, and managing information regarding the quality of the data. The purpose of the article is to describe a new generalized universal method of documenting data quality information, independently of the kind of data it concerns. Discussion of the method will be illustrated by specific examples. The method introduces an idea of Unit of Transformation Step (UTS) and, therefore, is especially suitable for describing processing chains such as: calibration of hardware maintenance tools, measurement of reflected power of radar signal, conversion into dBZ, collection of raw 3D data, application of corrections, data creation in form of products, applying more corrections, creation of products of merged data from different sources, assimilation of data into models, forecasting, entering data into decision system, etc.
Many past attempts dealt with the problem of documenting, monitoring and managing quality information for weather radar data. Developed in the past few years the idea of Quality Index gains popularity. Proposed by COST-717 framework the concept focuses on Interface Unit consisting of unlimited sets of Quality Indicators not firmly related to the particular type of radar data (radar product). This leads to user-dependent value of Quality Index. The present article tries to define a generalized point of view focusing on combining radar products (data) with their uncertainties, irrespective of the system where the data will be used.
All modern systems based on remote sensing, such as weather radar, can be treated as "production chain" where information flows through successive modules from physical measurement to indirect results. The full process can naturally be divided into steps (chain links) of information flow. Every of the chain links can be described in an universal manner. A description is proposed in this article and is called UTS (Unit of Transformation Step). The appropriate sections are defined, for both: a) data with error estimation and for b) transformation methods (algorithms) with their parameters. Two main reservations have to be done: 1) only physical or mathematical values (data) are processed in the chain and 2) estimation of errors of the data has to be done at each step (chain link). It is natural for the method to inherit knowledge from previous transformation steps, and to keep it together with output data. The presented method gives a mechanism of reaching information about all details in processing chain even if it is not contained in disseminated unit. Definition of file formats etc. for the information dissemination is not in scope of this article.
It has to be underlined that defined method of error documenting does not require the "true" error values. The best subjective error estimation can be still managed by UTS method to improve the value of the disseminated data for the benefit of the users.
It is planned to test the method in Polish radar network POLRAD.
Poster Session P13B, Wind Profilers / Operational Needs Worldwide, Networks and End to End Forecast Systems I
Thursday, 9 August 2007, 1:30 PM-3:30 PM, Halls C & D
Previous paper Next paper