Comparison of Mapping Methods for Plumes Using Prior Knowledge from Simulations

Helle Kristina B., Astrup Poul, Raskob Wolfgang, Pebesma Edzer

Research article in edited proceedings (conference)

Abstract

Maps of plumes are required to support decision making in nuclear and radiological emergencies. Such plumes can be modelled with atmospheric transport and dispersion models, but in operational situations the input parameters may not be known sufficiently therefore methods based on measurements may compose a useful complement. We compare contrasting mapping approaches: interpolation, fitting of a parametric trend, and simulations from uncertain parameters. This study is based on simulations. First we derive global properties from a training subset. Using this prior information, maps are produced from some extracted values and compared to the original data. The aim is to minimise absolute residuals and to improve delineation of affected areas. We discovered, that inverse distance weighted interpolation an kriging perform equally well to delineate threshold exceedance, and trend fitting can reduce absolute residuals, especially if few measurements are available.

Details zur Publikation

Book title: Proceedings of the Seventh International Symposium on Spatial Data Quality
Pages: 6
Release year: 2011
ISBN: 978-989-95055-8-2
Language in which the publication is written: English