OMNO2 Release Notes
Contents
1. Ozone Monitoring Instrument level 1b data products J Geophys Res in press Changes that have been made in the Level 1B to Level 2A processing are as follows 1 The order of the polynomial that is used in the DOAS fit to remove the background has been increased from 3 to 5 Analyses have shown that this increase in polynomial order decreases the error in the NO slant column density computed using a nonlinear fit The solar spectrum used to compute the Sun normalized radiance in Version 1 0 5 is static Version 1 0 0 had used a the daily measured solar spectrum The static solar spectrum in version 1 0 5 is an average over all the daily Sun observations from OMI in the year 2005 Analyses have shown that using a static irradiance spectrum based on many observations resulted in much smaller dark current related stripes In Version 1 0 5 a water vapor spectrum has been added to the set of basis functions used in the DOAS fit Analyses have shown that the water vapor columns derived from the DOAS fit were physically meaningful and improvedthe quality of the retrieved NO slant columns The same analyses showed that no physically meaningful O were obtained when the O spectrum was included in the DOAS fit and there was generally no improvement in the retrieved NO columns For that reason O is not included in the fit The reference cross sections of NO O3 O4 H2O and the Solar and Ring spectra have been updated based o2. Vp obs V Vohost obs Virop Virop Vohost V yes Vohost In the above expressions A is an air mass factor Vis a vertical column amount Sis a slant column amount fis the effective cloud fraction wis the cloud radiance fraction and tis the tropospheric fraction of the unpolluted column about 0 05 The subscripts are p for polluted u for unpolluted and init for initial The air mass factor superscripts are clear for clear sky and cloudy for overcast sky The column superscripts are above cloud indicating integration from cloud top to infinity and obs for the observed column the column above cloud or ground that is visible to the satellite Column amounts labeled a priori were calculated by integrating an a priori NO profile Vs V and Vro are retrieved observed column amounts and V V Vro are the corresponding amounts above ground computed by adding estimates of the below cloud column to the visible column V is the smooth unpolluted field from the planetary wave analysis Version 0 9 1 Released 2006 05 15 1 A bug was identified and fixed The cloud reflectivity assumed when reading and interpolating the dAMF file was not consistent with the cloud reflectivity assumed in the Level 2A cloud parameters from the 02 02 cloud algorithm Prior to this version a cloud reflectivity of 85 had been used in the OMNO2B algorithm rather than the value of 80 assumed in the O2 O cloud algorithm A value of 80 is no
3. F_OUTPIXEL_EARTH_RAD_WARNING bit due to the row anomaly Because of the wavelength registration shift the slant columns are still rather good as DOAS does not depend too strongly on the absolute radiance level Note however that the cloud product can still be heavily affected by the row anomaly so further processing may still need to be changed to take the flags into account 1 2 7 Further documentation of the changes can be found in tech note TN_OMIE KNMI 921 Version 1 0 5 Collection 3 Released 2007 11 09 In this release changes have only been made to the Level 1B processing GDPS and PGE A OMNO2A The Collection 3 Level 1B data have significant improvements over the previous version leading to better quality fit results The striping seen in Version 1 0 0 even after a destriping algorithm was applied has been greatly reduced due to an improved dark current correction and improved QA flags The DOAS fit residuals are much smaller and the precision of the NO slant column is in line with estimates from the instrument noise Specific changes to the Level 1B processing are detailed elsewhere See the following references 1 Dutch Space Ozone Monitoring Instrument for EOS AURA OMI GDPS Software User Manual Document No MA OMIE 7000 DS 459 Issue 6 5 October 2007 M Dobber Q Kleipool R Dirksen P Levelt G Jaross S Taylor T Kelly L Flynn G Leppelmeier and N Rozemeijer Validation of
4. OMNO2 Release Notes Document Version 3 0 February 2009 Data Product Version 1 1 0 2 These Release Notes describe changes that have been made to the OMI NO data set OMNO2 Only changes made since the provisional release 2005 1 1 21 are listed Version 1 1 0 2 Starting with orbit number 24349 2009 02 11 In previous versions the wavelength registration check was executed for up to 10 wavelengths and was done on rows 0 30 and 59 It is now run on all rows but on a smaller wavelength window An extra parameter was added to the OPF for PGE A to indicate the desired fitting windows for the wavelength registration The wavelength registration is no longer a check it is applied as a correction to the L1B wavelength grid The reason to do this now is the row anomaly observed in several cross track positions The row anomaly has an influence on the radiance level However such a multiplicative error should not have an effect on the DOAS fit and we observe a strong influence The obstruction in the field of view causes an inhomogeneous illumination of the slit in the spectrometer which in turn shifts the wavelength registration in OMI Since the wavelength assignment in L1B does not use the position of various Fraunhofer lines in the radiance spectrum this shift is not taken into account in the assigned L1B wavelength grid For unaffected rows the fitted shift is in the order of 0 002 nm on average This is about 1 100 of a pixel For a
5. ffected rows the shift can be up to 0 01 or larger 1 20 of a pixel Including the shift reduces the error on the fit significantly and the values for the affected rows are flat compared to the surrounding lines In the northern hemisphere there is a second effect of the row anomaly Here sunlight is coupled directly into the spectrometer This is an additional error and this is not taken into account or corrected for In addition a new flag field XTrackQualityFlags was added to the to carry forward the L1B flagging of the row anomaly The meaning of this field is detailed in the following table Value Meaning Ss FitQualityFlags set Ba a a Affected by row anomaly QAF_OUTPIXEL_EARTH_RAD_ERROR bit 2 Pixel not corrected Do not use pixel Slightly affected by row anomaly QAF_OUTPIXEL_EARTH_RAD_WARNING bit 3 Pixel not corrected Pixel can be used with caution Affected by row anomaly QAF_OUTPIXEL_EARTH_RAD_WARNING bit 3 Pixel corrected but correction is not optimal 3 Use pixel with caution 4 Affected by row anomaly QAF_OUTPIXEL_EARTH_RAD_WARNING bit 3 Pixel is corrected for the row anomaly and the correction is optimal Pixel is less accurate than pixels that are not affected by the row anomaly Error during correction for row QAF_OUTPIXEL_EARTH_RAD_ERROR bit 2 anomaly e Do not use pixel Note that OMNOA2A still processes all pixels even those marked with a QAF_OUTPIXEL_EARTH_RAD_ERROR or QA
6. n the latest version of the OMI on ground measured slit functions Include a reference citation In addition the water vapor spectrum has been updated to remove artifacts from the reference spectra Several smaller changes and bug fixes have been implemented Version 1 0 0 Released 2006 09 08 Public Release 1 Level 2 files now contain three main types of vertical column NO and 2 3 estimates of their uncertainties These are the observed total column ColumnAmountNO2 the observed tropospheric ColumnAmountNO2Trop column and observed polluted column ColumnAmountNO2Polluted The fields for the unpolluted column ColumnAmountNO2Unpolluted and its uncertainty ColumnAmountNO2UnpollutedStd now contain only fill values As before the smooth field is defined by the Fourier coefficients which are included in the file The files also contain an estimate of the column amount hidden any clouds in the field of view ColumnAmountNO2BelowCloud This quantity also called the ghost column and its uncertainty are given for each OMI pixel with measurable pollution and a defined cloud fraction greater than zero It is not a fixed quantity from a look up table but is computed from on the measured column amount and the a priori tropospheric profile shape ColumnAmountNO2BelowCloud can be added to the total tropospheric or polluted columns to obtain the respective complete columns which are independent of cloud amou
7. nt However the ghost column amount becomes less certain as the cloud fraction increases particularly for cloud fractions gt 40 Therefore the user Should carefully consider cloud information when selecting data for use in validation studies or other investigations A summary flag is now included with each observation to alert the user to the presence of one or more significant data quality issues 4 Cloudy and clear components of the AMFs both polluted and unpolluted are now defined relative to the above ground vertical column previously they were relative to the visible column and thus contained cloud fraction information The values of the radiance weighted sums of the clear and cloudy AMF components are unaffected by this change Uncertainty calculations for all quantities have been improved since the provisional release Air mass factor uncertainties are now dependent on cloud parameters terrain reflectivity and estimated a priori profile variability In all calculations cloud fraction uncertainty is fixed at 0 05 The relationships among the column amounts and air mass factors in the current OMI algorithm are as follows Ay w Aur 1 w A A W AP 1 w A Ae SA Vid Vey aio Vee epion FV 14 a 1A Vol apron Vint S A Unpolluted Case V Vint Viop Vine t Polluted Case Ves Vint Vu Au Ap Vrp eV Vut v VO V Vara V Vrp Vonost Vni Vu Au 1 Ap 1 A
8. w assumed throughout the calculation The effect of this change has not been quantified but is expected to be minor
Download Pdf Manuals
Related Search