Ozone column density determination from direct irradiance measurements in the ultraviolet performed by a four-channel precision filter radiometer

Ultraviolet light was measured at four channels (305, 311, 318, and 332 nm) with a precision filter radiometer (UV-PFR) at Arosa, Switzerland (46.78 d egrees, 9.68 degrees, 1.850 m above sea level), within the instrument trial phase of a cooperative venture of the Swiss Meteorological Institute (Mete oSwiss) and the Physikalisch-Meteorologisches Observatorium Davos/World Rad iation Center. We retrieved ozone-column density data from these direct rel ative irradiance measurements by adapting the Dobson standard method for al l possible single-difference wavelength pairs and one double-difference pai r (305/311 and 305/318) under conditions of cloud-free sky and of thin clou ds (cloud optical depth <2.5 at 500 nm). All W-PFR retrievals exhibited exc ellent agreement with those of collocated Dobson and Brewer spectrophotomet ers for data obtained during two months in 1999. Combining the results of t he error analysis and the findings of the validation, we propose to retriev e ozone-column density by using the 305/311 single difference pair and the double-difference pair. Furthermore, combining both retrievals by building the ratio of ozone-column density yields information that is relevant to da ta quality control. Estimates of the 305/311 pair agree with measurements b y the Dobson and Brewer instruments within 1% for both the mean and the sta ndard deviation of the differences. For the double pair these values are in a range up to 1.6%. However, this pair is less sensitive to model errors. The retrieval performance is also consistent with satellite-based data from the Earth Probe Total Ozone Mapping Spectrometer (EP-TOMS) and the Global Ozone Monitoring Experiment instrument (GOME). <(c)> 2001 Optical Society o f America.