New RadCalNet Instrumented Site at Gobabeb, Namibia: Field Campaign Conclusions and First Absolute Calibration Results Sebastien Marcq 1, Sophie Lacherade 1, Aime Meygret 1, Agnieszka Bialek 2, Claire Greenwell 2, Maxim Lamare 2, Nigel Fox 2, Marc Bouvet 3, Beatrice Berthelot 4 1: CNES, France; 2: NPL, United Kingdom; 3: ESA, Netherlands; 4: Magellium, France
Outline Context - RadCalNet Global search for site Precise location at Gobabeb Results from characterization campaign Conclusion and outlook 2
RadCalNet (Radiometric Calibration Network) CEOS/WGCV IVOS Calibration field campaigns: Atmosphere and surface optical properties are characterized simultaneously to the satellite overpass -> Limitations: cost + availability Baotou Railroad Valley La Crau Gobabeb Automatic field instrumentation Data will be accessible to the public Baotou (AOE, China) La Crau (CNES, France) Railroad Valley (NASA/University of Arizona, USA) Gobabeb (ESA/CNES, Namibia) 3 Already operational To be installed Autumn 2016
Global search for a new site look for a new site (ESA+CNES) as part of RadCalNet (Radiometric Calibration Network) Criteria of global analysis: - Low Cloud coverage - High Spatial homogeneity at several scales (10s of meters to 100s of meters) - Stability (no vegetation) - Atmospheric changes (atmospheric particles, water wapor ) - Pratical reasons (access, transmission) Area of Gobabeb, Namibia 4
Precise location at Gobabeb from Satellite data 1000 m² Site characterized by the instrumentation needs to be homogeneous and representative of the surrounding area, acquired by the satellite sensor (100-10 5 m²) Homogeneity Representativeness Different scales, LxL: - 100x100 m² - 200x200 m² - 500x500 m² - 1000x1000 m² - 2000x2000 m² Different resolutions: 5
Precise location at Gobabeb from Satellite data Spatial homoneneity better than 3% LxL = 1 km² Homogeneity LxL = 200x200 m² Spatial Representativeness LxL = 200x200 m² 6 potential sites 6
Precise location at Gobabeb: visual assessment of Potential sites 1 2 1bis 3 4 5 7 Titre présentation
Characterization GRASS directional reflectance HDRF results from the Gonio RAdiometric Spectrometer System (GRASS NPL)
Characterization ASD Surface reflectance ρρ tt = LL tt LL rrrrrr ρρ rrrrrr ASD measurements Characterized by NPL in the lab Z1 but the wind is getting it dirty Z4 -> needs to be monitored Z3 9
Spectralon panel reflectance monitoring Spectralon reflectance is modeled as ρρ dddddddddddd θθ ss EE dddddd tt + ρρ heeeeeeeeeeeeeeeeeee EE dddddd tt ρρ ssssssss θθ ss, tt = ff tt EE dddddd tt + EE dddddd tt - Direct and diffuse irradiances (EE dddddd, EE ddddff ) given by 6S for each measurement - Directionnal and hemispheric reflectance (ρρ dddddddddddd, ρρ dddddddddddddd ) measured in the lab (NPL) - Day-to-day variations (comparison to «super reference» + cleaning ) -> dimming factor ff tt Day-to-day monitoring using a super reference Spectralon BRDF measured in the lab (NPL) NPL 10
Surface reflectance protocol 10 m 100m ~1.5km Characterize surface reflectance at different resolutions: - < 1m (ASD + CIMEL footprint) - 10m (CIMEL characterized surface) - 100m (potential sensors to calibrate) 11
Surface reflectance results Variability at different scales Note that kilometric variability is not representative (BRDF + poor correction of Spectralon Variation) Absolute Reflectance value Gobabeb site is much more homogeneous than La Crau 12
Surface reflectance results: other potential sites Z1 Z4 Z3 13 Slighly less homogeneous than Z3 Confirmation of satellite data analysis
Surface reflectance results: Comparison with other sites Reflectance 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0.4 0.9 1.4 1.9 2.4 Wavelength (µm) La Crau, Gobabeb reflectance acquired by CNES (in-situ) Algeria spectra acquired by ONERA based on samples La Crau GOBABEB_Z3 Algeria_2 Algeria_3 Algeria_4 14
NPL Measurement Positions Circles represent GRASS measurement locations, and point markers represent ASD measurement locations.
NPL ASD measurements The surface points measured by the NPL team formed four 10 m x 10 m squares. The graph below shows the BRF at 1640 nm for every measurement. -> So significant variations due to solar elevation change
GRASS measurements Gonio Radiometric Spectrometer System The results show homogeneity in the forward scattering direction, with enhanced radiance in a hotspot located centred on the solar principal plane in the backscatter direction
Characterization CIMEL Aerosols Use of the Gobabeb AERONET station (7km away) AOT -> Consistency between station and place of measurements confirmed by Calitoo (handheld sunphotometer) -> Relatively low AOT most of the time 18
Material Analysis From Royal Holloway, University of London During the campaign, samples of the surface materials were collected at all the GRASS measurement sites. Particle characterisation was performed with an SEM. Mineral analysis was undertaken using X-Ray Diffraction. Mineral Weight (%) Quartz 51.4 Calcite 1.6 Feldspar 35.8 Silicate 9.0 Mica 2.2
Impact of the characterization campaign Monitoring using PLEIADES 70cm resolution imagery Before Right After Sept 9th 2015 Dec. 18th 2015 Now July 26th 2016 Footprints Impact: ~6% No visible Footprints Impact 20
Conclusion Location of the site: Determination of the precise location of the future RadCalNet site 23 36 S, 15 7 9 E Confirmation on the ground of what satellite data had shown : very good spatial homogeneity from very high to coarse resolution Installation campaign: currently early October: Mast + photometer Weather station All sky camera Drone measurements Calibration results next year? 21