180 likes | 335 Views
Matt DeLand (for the PEATE team) SSAI 5 December 2013. OMPS LP Release 2 - Status. OMPS Nadir Checkout Measurements. Overview ( since August 2013 ).
E N D
Matt DeLand (for the PEATE team) SSAI 5 December 2013 OMPS LP Release 2 - Status
OMPS Nadir Checkout Measurements Overview (since August 2013) • Created “V2 Pretest” configuration (AS 85391) to evaluate performance of most planned Release 2 changes together. Specific changes listed on next slides. • Reprocessed 38 “golden days” (best orbit alignment with Aura MLS) collected during April-December 2012. • GSFC and SSAI personnel evaluated these results against MLS data (single events, daily zonal mean, matchups over full data set) and lidar data. • Sample results will be shown. • Processing status summarized at end.
L1 Changes for V2 Pretest • Implement tangent height adjustment through OPF (with additional 500 m shift for this test). • Revise wavelength gridding in L1G product to use fixed grid for all events. • Revise reference wavelength scale to use better data set. • Implement intra-orbit and seasonal wavelength scale adjustments. • Eliminate merging of multiple gain/aperture values for determining radiance at each pixel. • Prioritize data selection to use high gain sample for λ < 500 nm, low gain sample for λ > 500 nm. • Revise ancillary data selection to use GMAO products for temperature, pressure, density. Profiles extended to 80 km with constant temperature lapse rate.
L2 Changes for V2 Pretest • Implement new ozone a priori data set created from 2012 MLS data. • Implement SUSIM data for UV portion of high-resolution solar irradiance spectrum. • Exclude OH emission wavelengths (306.5-311 nm) from UV ozone profile retrieval. • Add 1% instrument error term to SNR noise term for retrieval. • Extend reported ozone profile downward to 0 km, upward to 80 km using a priori values. • Report UV and VIS ozone retrieval results separately. • Create mixing ratio profile product on regular pressure grid.
OMPS LP – Individual Event, Multiple Profiles ?? We report 3 profiles Too fast a switch? Combined profile goes to fill values (combined)
OMPS LP vs. Lidar – Individual Stations Agreement with lidar profile data generally within ±10%. Some suggestion of altitude dependence. [44N,5.7E]; [20N, 155W]; [48N,11E]; [34N, 118W]
OMPS LP vs. MLS – Zonal Mean Average All latitudes • Agreement with MLS profile data generally within ±5% over 20-55 km altitude region. • LP ozone values smaller below 20 km (with much more variability).
OMPS LP vs. MLS – Tangent Height Evaluate vertical shift needed to get best alignment between OMPS LP ozone profile and MLS ozone profile
OMPS LP vs. MLS – Merging of Retrievals • Only Center slit is considered • In general, UV retrieval is better compared to MLS. • Combined average of UV-Vis weighted by the retrieval error (always toward UV) • 28km is the cut off alt • The quality and agreement of UV/Vis vary by season and location • Notice at overlapping altitudes, the UV retrieval agree more with MLS than the Combined profile (~1%) Cut off altitude
1-178 hPa Column Ozone: OMPS Minus MLS OMPS stratospheric ozone column consistently shows lower values in the Northern Hemisphere
OMPS LP vs. MLS – Latitude Dependence Sept 14: DZM ozone comparisons between OMPS-LP Center slit and MLS
OMPS LP – Comparison Between Slits Center slit – Left slit Center slit – Right slit
OMPS Nadir Checkout Measurements Conclusions and Decisions [1] • Retrieval algorithm has more difficulty near normalization altitude Limit public ozone product to zmax = 60.5 km. • Do not extend retrieved profiles using climatology data. • Ozone differences between slits are comparable to external comparisons Provide data from all three slits. • No separate reporting of UV and VIS retrieval results. • Combined ozone profile will not merge data. Use UV retrieval values from 60.5 km down to 27.5 km. Use VIS retrieval values from 26.5 km down to retrieval cutoff. • VIS retrieval can be adjusted to use 510 nm and 673 nm as “guard” wavelengths for triplet formation, 549-633 nm as range to sample Chappuis band. This reduces aerosol effects within retrieval range, and avoids possible crossover into high gain data due to L1G wavelength selection.
Impact of Aerosol Retrieval on Residuals 530 nm 40 km 30 km 20 km
OMPS Nadir Checkout Measurements Conclusions and Decisions [2] • Ozone differences with MLS below ~25 km are affected by aerosol correction used in VIS retrieval. • Interpretation of these data using radiance residuals is currently complicated, because aerosol correction is based on full extinction coefficient profile retrieval. • L2 algorithm can also use only aerosol climatology data for determining this correction. • Initial tests using this approach indicate ozone changes are within current uncertainty range. • Release 1 processing uses SAGE-based aerosol climatology. New aerosol climatology based on CALIPSO data during 2012 has also been developed, but needs refinement.
OMPS Nadir Checkout Measurements Conclusions and Decisions [3] • NEXT STEP: Reprocess “golden days” with no aerosol retrieval, using only aerosol correction based on climatology data set. Compare new ozone products with MLS data. Compare new residuals with AS 85391 results. • If agreement is comparable to previous results (or better), then Release 2 ozone processing will use this approach. • Hope to begin forward processing once this evaluation is complete (2 weeks?). • Modified L2 code will be created to generate separate aerosol product data set using current retrieval algorithm. • May revise reference data set used to assign South Atlantic Anomaly flag in L1B. Additional event-based SAA parameter derived from L2 residuals is being tested.
OMPS Nadir Checkout Measurements New Sample Table • Add wavelengths to fill some gaps in current sampling (e.g. 312-315 nm, 331-340 nm). • Simplify pixel sampling based on revised gain selection procedure. • Sample table uploaded on November 27. Operational processing not successful. Currently working to re-establish forward processing (AS 60000). Release 1 processing (AS 61002) will not be resumed.