1 / 18

Overview

Requirements Consolidation of the Near-Infrared Channel of the GMES-Sentinel-5 UVNS Instrument: Task 1: Initial trade-off: Cloud-characterisation for uv-vis R.Siddans PM1: RAL, 9 July 2013. Overview. Initial assessment based on work in the Eumesat Study for then proposed MTG UVN sounder:

prue
Download Presentation

Overview

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Requirements Consolidation of the Near-Infrared Channel of the GMES-Sentinel-5 UVNS Instrument:Task 1: Initial trade-off:Cloud-characterisation for uv-vis R.SiddansPM1: RAL, 9 July 2013

  2. Overview • Initial assessment based on work in the Eumesat Study for then proposed MTG UVN sounder: • R. Siddans, B.G. Latter, B.J. Kerridge, Study to Consolidate the UVS Mission Requirements for the Oxygen A-band (EumetsatContract No. EUM/CO/05/1411/SAT), 2007 • This provided basis for later studies re height resolved aerosol (ACOR, CAPACITY, CAMELOT), but also studied cloud/trace-gas application in some detail. • Will form basis of new work in T2 of this study. • Basis of approach is to • Define realistic cloud senarios & simulate measurements • Perform imager and A-band spectrometer cloud retrievals • Determine implied errors in the uv-vis trace gases by quantifying air-mass-factor errors stemming from the retrieved cloud representation • Various instrument resolutions tested

  3. Assessment approach • UVS mission – tropospheric trace gas retrievals • O3(trop), BrO, NO2, CH2O give rise to optically thin absorption signatures: DOAS is applicable: • Slant column is fitted to measured spectra • Vertical column estimated by dividing by air-mass factor (AMF) Smwhich is calculated using an RTM (assuming scattering profile) • Errors in modelled scattering only enter via the AMF calculation • Simulate errors due to cloud/aerosol by evaluating error in AMF

  4. Cloud Scenarios • Cloud scenarios based on ground-based radar/lidar data, analyed by CloudNet project (Hogan et al) • Wind fields used to synthesise orbit x-sections from station data provided as fn of time. • Similar data now available from CloudSat/CALIPSO • DARDAR project (Hogan and Delanoe) IWC 1km regridded Distance (km)

  5. Imager retrieval • Based on ORAC, currently used on ATSR, AVHRR, MODIS, SEVIRI etc: • GRAPE, GlobAerosl, CCI-Clouds, CCI-Aerosol, Eumetsat OCA etc: • Uses optimal estimation & retieves • Optical depth at 0.55 µm • Effective radius. • Cloud-top height. • Assumes single, homogenous cloud/aerosol layer of particular type (liquid/ice cloud, maritime / continental / biomass… aerosol) • Cost function can be used to identify type or where single layer assumption not valid (possibly)

  6. Imager retrieval • ORAC currently uses 0.55, 0.67, 0.87, 1.6, 11 and 12µm • only vis/near-ir for aerosol • Extended here to 9 "FDHSI" channels for both cloud & aerosol • 0.55, 0.64, 0.809, 1.63, 3.92, 8.71, 10.8, 11.9 and 13.4 µm • Added aerosol layer height (from ir) to state vector • ORAC Based on RT look-up-tables • Here based on full on-line RT with RAL FM2D to ensure consistency with measurement simulations & facilitate use of additional channels • Here each scene is analysed with 4 particle models: • Liquid cloud, ice cloud, desert aerosol, maritime aerosol • The retrieval with lowest cost function is selected

  7. Imager results true • Single layer cloud / aerosol fits quite well where possible • Mixed layer cloud gives high cost • Generally cost selects correct type • Aerosol noisy (kext & reff) Retrieved (cloud/aerosol only) Cost/type

  8. A-band retrieval scheme • Extinction coefficient profile retrieval as aerosol assessment • Here take scatter type and reff from imager retrieval • Assume spatial resolution 10km cf imager 1km • Divide scenes into cloudy & clear fraction • Cloudmask: Threshold reflectance > 0.2 • In each fraction find type from imager retrieval which is associated with most measured photons. • Then take radiance weighted mean kext and reff for this type as imager representation of cloud/clear fraction. • A-band retrieval run assuming imager type on whole scene if homogenous or "cloudy" fraction if mixed.

  9. A-band retrieval scheme • Standard scheme represents cloud as a scattering profile (CSP) • 2nd scheme implemented: • emulates the approach GOME Operational total column • Cloud as reflecting surface (CRS approach) • Assume cloud is Lambertian surface at elevated atlitude • Assume cloud fraction and type from imager • (operationally fraction from PMDs, type assumed) • Retrieve • Cloud top height • A priori 5±10 km. • Cloud reflectance • A priori 0.05±1 • only applied to 0.6 nm resolution measurements

  10. A-band retrieval results true imager 0.6 nm resolution 3D measurements 1D, SNR=250 1D measurements

  11. A-band retrieval results true imager • Cloud representation reasonable even from low resolution A-band but improves with resolution. 0.6nm resolution 0.06nm resolution

  12. A-band retrieval results true imager • At 1cm-1 much better representation of thin ice cloud + cost OK • SNR 2500 better than 250 3 cm-1 resolution 2km retrieval 1km, SNR=250 1km retrieval

  13. AMFs • AMFs are computed for 3 column amounts • Total • Tropospheric (0-12km) • Boundary- layer (0-2km) • Sub-column AMFs would be used in combination with external info to constrain other layer or profile shape (e.g. model or other wavelengths) • Calculated by perturbing absorber amount & taking ratio of apparent optical depth to actual, vertical optical depth of absorber • AMFs are first at 1km spatial resolution from • True field • Retrieval representation of field (using cloud fraction for sub-pixel representation) • AMF of 10km pixel is given by radiance weighted mean

  14. Imager OK for cloud free & simple cloud • A-band CRS improves results for cloudy conditions

  15. Profile retrieval improves over CRS at 0.6 nm resolution

  16. 0.18 nm slightly better than 0.6 nm

  17. 0.06 nm ~ similar 0.18nm

  18. Conclusions • Realisitic end to end simulations of modelling scattering profile for uv-vis retrievals based on imager and A-band retrievals conducted in Eumetsat A-band study • These indicate • Imager retrieval functions well for simple cloud layers • AMF adequate in cloud-free (aerosol) & simple cloud cases • A-band cloud-as-reflecting-surface improves AMFs • A-band scattering profile retrieval improves further • A-band instrument needed to mitigate scattering profile errors • High resolution, low error instrument demonstrates superior cloud profile retrieval, however AMFs do not improve significantly • For application to characterise AMFs for uv, then A-band with 0.2-0.6nm resolution, signal to noise ~250 (moderate reqs on other instrumental error sources).

More Related