New neural-network-based method to infer total ozone column amounts and cloud effects from multi-channel, moderate bandwidth filter instruments

Lingling Fan, Wei Li, Arne Dahlback, Jakob J. Stamnes, Snorre Stamnes, Knut Stamnes

Research output: Contribution to journalArticlepeer-review

12 Scopus citations

Abstract

A new method is presented based on a radial basis function neural network (RBF-NN) to analyze data obtained by ultraviolet (UV) irradiance instruments. Application of the RBF-NN method to about three years of data obtained by a NILU-UV device, which is a multi-channel, moderate bandwidth filter instrument, revealed that compared to the traditional Look-up table (LUT) method, the RBF-NN method yielded better agreement with a 1% decrease in relative difference and an increase of 0.03 in correlation with total ozone column (TOC) values obtained from the Ozone Monitoring Instrument (OMI). Furthermore, the RBF-NN method retrieved more valid results (daily average values within a meaningful range (200-500 DU)) than the LUT method. Compared with RBF-NN retrievals, TOC values obtained from the OMI are underestimated under cloudy conditions. This finding agrees with conclusions reached by Anton and Loyola (2011).

Original languageEnglish
Pages (from-to)19595-19609
Number of pages15
JournalOptics Express
Volume22
Issue number16
DOIs
StatePublished - 11 Aug 2014

Fingerprint

Dive into the research topics of 'New neural-network-based method to infer total ozone column amounts and cloud effects from multi-channel, moderate bandwidth filter instruments'. Together they form a unique fingerprint.

Cite this