The AATSR LST retrieval: State of knowledge and current developments

Transcription

1 The AATSR LST retrieval: State of knowledge and current developments Darren Ghent, Ed Comyn-Platt, Gary Corlett, David Llewellyn-Jones, Harjinder Sembhi, Karen Veal, Christopher Whyte and John Remedios Earth Observation Science Department of Physics and Astronomy University of Leicester July 6, 2012 GlobTemperature User Consultation Meeting 1

2 Outline Background Product improvements LST validation Output Requirements for LST July 6, 2012 GlobTemperature User Consultation Meeting 2

3 AATSR LST retrieval Single algorithm Two sets of auxiliary data files LST_AUX_1: Dorman and Sellers biome map + SiB/ISLSCP fractional vegetation (Dorman and Sellers, 1989) LST_AUX_2: Biome map based on Globcover (Arino et al., 2007) + CYCLOPES fractional vegetation (Barat et al., 2007) Three sets of LST retrieval coefficients LST_COEF_1: Generated by Fred Prata using MODTRAN and NCEP profiles LST_COEF_2: Generated using RTTOV-10.1 and Chevallier profiles (Chevallier et al., 2006) LST_COEF_3: Generated using RTTOV-10.2 and ~ ECMWF ERA-Interim profiles Current algorithm: Emissivity dependence wrapped up in coefficients which are dependant on biome and FV Dependant on accurate auxiliary data No precipitable water influence or square-dependence in the nadir Prata (2002) July 6, 2012 GlobTemperature User Consultation Meeting 3

4 AATSR LST algorithm Auxiliary Data: LST_AUX_1: o Spatial resolution (0.5 ) o Temporal resolution - monthly climatology o Outdated biome mask and climatology LST_AUX_2: o Improved spatial resolution (1 / 120 ) o Improved temporal resolution for fractional vegetation - 10 day climatology o More current: 2006 / 2009 Globcover; 1999 to 2007 CYCLOPES climatology Products: LST v1: LST_AUX_1 + LST_COEF_1 LST v2: LST_AUX_2 + LST_COEF_2 LST v3: LST_AUX_2 + LST_COEF_3 July 6, 2012 GlobTemperature User Consultation Meeting 4

5 Improving auxiliary datasets AATSR LST v1 (right) using operational biome (left) and fractional vegetation data (centre) AATSR LST v2 (right) using Globcover biome (left) and CYCLOPES fractional vegetation data (centre) July 6, 2012 GlobTemperature User Consultation Meeting 5

6 Improving cloud masking Operational cloud mask Improved cloud mask July 6, 2012 GlobTemperature User Consultation Meeting 6

7 Improving ice / snow detection RGB image (1 st column); operational snow masking orange (2 nd column); operational cloud masking green (3 rd column); updated snow masking using a Bayesian approach (modification on Istomina et al. (2010) and OSISAF (Eastwood and Andersen, 2007)) - Water, Sea Ice, Land Ice, Sea Cloud, Land Cloud, Clear Land (4 th column) over Ostrov Sakhalin, Russia, Jan July 6, 2012 GlobTemperature User Consultation Meeting 7

8 Geolocation accuracy Current CH1 data: image grid BT 11 (centre); instrument grid BT 11 (right) Updated CH1 data: image grid BT 11 (centre); instrument grid BT 11 (right) July 6, 2012 GlobTemperature User Consultation Meeting 8

9 2 nd $June$2009$ Geolocation accuracy July 6, 2012 GlobTemperature User Consultation Meeting 9

10 Geolocation accuracy July 6, 2012 GlobTemperature User Consultation Meeting 10

11 Geolocation accuracy July 6, 2012 GlobTemperature User Consultation Meeting 11

12 LST v3 coefficient generation The new AATSR biome map incorporates 27 land and inland water classes LST retrieval coefficients were generated for each biome and time-of-day combination The fast radiative transfer model RTTOV-10.2 is used to simulate brightness temperatures from profile and emissivity data ECMWF ERA-Interim daily and invariant fields are used to generate the necessary profile data for radiative transfer A uniform random sampling distribution is used to select profile data for each biome class Linear regression is applied to find the a,b, and c retrieval coefficients for both bare soil and fully vegetated states Latitudinally banded water vapour profiles for biome class 19 (urban) from left to right: 50S-30S, 30S-10S, 10S-10N, 10N-30N, 30N-50N July 6, 2012 GlobTemperature User Consultation Meeting 12

13 Algorithm verification Scatterplots of simulated LST profile skin temperature with respect to precipitable water for different combinations of the retrieval parameters for simulated edge-of-swath brightness temperatures July 6, 2012 GlobTemperature User Consultation Meeting 13

14 LST validation Four different approaches to LST validation have been carried out conforming to the four different categories defined in the LST Validation Protocol Report to ESA (Schneider et al., 2012) Temperature-based Temperature method represents the conventional method for satellite LST validation. It involves the direct comparison of concurrent ground measurements acquired at a field site with LST acquired from the satellite overpass Creation of matchup database (MDB) equivalent to that for SST Radiance-based Radiance method consists of calculating the ground LST from (TOA) brightness temperatures simulated using surface emissivity and temperature data plus atmospheric profile data in a radiative transfer model Intercomparison Multi-sensor intercomparison enables an evaluation of the relative consistency of a product Time series Time series analysis provides a platform for identifying time-dependent biases but also forms the final climate product from an individual sensor July 6, 2012 GlobTemperature User Consultation Meeting 14

15 Temperature-based validation AATSR LST vs. Gobebeb in situ measurements for Day (left) and night (right) during 2009; AATSR LST v1 validation (top), AATSR v3 validation (bottom). In situ measurements courtesy of Frank Goettsche. July 6, 2012 GlobTemperature User Consultation Meeting 15

16 Radiance-based validation For each instrument top-of-atmosphere brightness temperatures (BTs) are simulated for the 11 and 12 µm channels using the fast radiative transfer model RTTOV-10.2 Utilised the ECMWF ERA-Interim profiles and 11 and 12 µm channel emissivities derived from CIMSS emissivity database (Seemann et al., 2008). Satellite-retrieved LST is taken as the input skin temperature to the model Perturbations applied to the input skin temperature until the simulated BTs match the satellite retrieved BTs as per the methodology of Wan and Li (2008) LST error is the difference between satellite-retrieved LST and inverted skin temperature Quality of the atmospheric profiles assessed through the difference Δ(T 11 T 12 ) between the satellite-retrieved BT differences (T 11 T 12 ) and the simulated BT differences (T 11 T 12 ) Δ(T 11 T 12 ) should be close to zero when the atmospheric temperature and water vapour profiles used in simulations represent the real atmospheric conditions and effect of the surface emissivities for the satellite observations. Mean ΔLST is 0.63 K for AATSR LST v2 July 6, 2012 GlobTemperature User Consultation Meeting 16

17 Intercomparison Monthly daytime LST comparisons for 2006: AATSR LST v1 vs. SEVIRI LandSAF LST (left); AATSR LST v2 vs. SEVIRI LandSAF LST (centre); Terra-MODIS v5 (MOD11_L2) vs. SEVIRI LandSAF LST (right) July 6, 2012 GlobTemperature User Consultation Meeting 17

18 UK LST anomalies Day% anomaly% AATSR%LST%v3%day,me%anomalies%above;%night9,me%anomalies%below;%warm%events%highlighted%% Night% anomaly% July 6, 2012 GlobTemperature User Consultation Meeting 18

19 Outputs Generating SLSTR Level-2 Auxiliary Data Files based on experiences learned during development of AATSR LST v3 ESA LST data portal (developed by NILU and UoL) AATSR v3 Level-3 product (above) at user-defined spatial resolution. Example here is for daytime July 2007 at 0.05 AATSR LST v3 time series (left). Example here is for British Isles July 6, 2012 GlobTemperature User Consultation Meeting 19

20 Other work at Leicester Testing other LST algorithms (consistency, multisensor) Alternative coefficient algorithms 1-d variational assimilation (with Met. Office) LST product intercomparisons (AATSR-MODIS- SEVIRI) Polar Surface Temperature datasets Ice Surface tenperature Improved cloud clearing and ice identification Emissivity product intercomparisons July 6, 2012 GlobTemperature User Consultation Meeting 20

21 Requirements for LST Consultation with users from several scientific fields which utilise LST data, both across Europe and the US elicited the following recommendations Availability Provision of level-3 gridded data at a variety of resolutions (0.5, 1.0 for example) Commonly used file formats are important to ensure maximum exploitation of LST products The processing of long time-series of LST Accuracy Improved cloud screening algorithms Provision of uncertainty estimates Multiple observation angles to improve atmospheric correction Validation Longer validation studies over a variety of surface regimes An increased emphasis on radiance-based validation An increase in the number of intercomparison studies Resolution A better resolution of the LST diurnal cycle Higher resolution products for resolving agricultural fields and urban features for example Combined LST products from instruments on Low Earth Orbit (LEO) satellites and geostationary satellites Combined LST products from LEO satellites for polar regions. July 6, 2012 GlobTemperature User Consultation Meeting 21

22 Acknowledgements We wish to thank the funding agencies for supporting this work ESA NERC / NCEO DECC We also wish to acknowledge our colleagues for their input and provision of data Fred Prata and Philipp Schneider (NILU) Isabel Trigo (Eumetsat and Landsaf) Frank Goettsche (KIT) Lizzie Good (now at Hadley Centre) July 6, 2012 GlobTemperature User Consultation Meeting 22