What was the purpose of SoyFLEX
The previous HYFLEX campaigns (‘Technical Assistance for the Deployment of an advanced hyperspectral imaging sensor during HYFLEX’; Contract No. 4000107143) supported the testing of the novel Hyperspectral Plant Imaging Spectrometer (HyPlant). Within the HYFLEX project maps of sun-induced chlorophyll fluorescence over different agricultural field sites, needle and broadleaf forest were presented. Results demonstrated the capability of HyPlant airborne data to test and evaluate different approaches to model and retrieve top-of-canopy (TOC) fluorescence and the possibility to study the physiological responds between fluorescence and photosynthesis. Within this activity the analyses were extended to measurements of a dedicated experiment with two different soybean varieties. The soybean experiments was set-up in Germany.
In addition, the effect of a shadow and quick light exposure on photosynthesis, fluorescence and canopy temperature was investigated (‘virtual cloud experiment’). This virtual cloud experiment’ experiment was set-up in Germany in an agricultural area and repeated over grassland in a wetland area in Poland. These experiments should be considered to generate specific scientific results to support FLEX activities.
What was the outcome of SoyFLEX
Acquisition of high resolution data by the HyPlant sensor worked fine and high performance reflectance data were acquired on all proposed study sites in Germany and Poland. Both HyPlant sensor modules were fully operational and radiometric quality of the data is high and stable.
The red and far-red fluorescence maps were computed by three retrieval methods, which are available for HyPlant. (1) The ‘singular vector deconvolution’ (SVD) uses solar Fraunhofer lines in the red and far-red spectral region to retrieve fluorescence. (2) The iFLD method was greatly improved in the past years and exploits the two O2 bands in combination with non-fluorescing reference surfaces. (3) The‘Spectral Fitting Method’ (SFM) was for the first time applied operationally on a larger data set of HyPlant and delivers first maps of sun-induced fluorescence that are based on the proposed retrieval method of the FLEX satellite mission.
The Spectral Fitting Method approach was also strongly improved especially on the atmospheric
forward RT module (‘atmospheric correction’), in which atmospheric model inputs are systematically derived from ancillary sun-photometer data. The retrieval algorithms were tested on several flight lines collected on different areas, type of canopy and atmospheric conditions see chapter 6.2 in the final report. The iFLD fluorescence values show best agreement with ground measurements, but this retrieval requires on non-fluorescence pixels and it could suffer from undesired striping effect depending on reference pixels distribution in the image see chapters 6.2.4 and 6.2.5 in the final report. The SVD approach depends on the specific trainingset and the fluorescence values generally are prone to a larger noise level see chapter 6.2.5 in the final report. The SFM relies on physically based retrieval, therefore the overall accuracy strongly depends on the quality of the atmospheric model inputs (sun-photometer) and the precise instrument spectral/radiometric characterization. An extensive testing of SFM retrieval was conducted based on intermediate products (i.e. surface irradiance modelling, SpecCal etc.), image based analysis and comparison with ground based measurements.
|Geographic Site||Agricultural area around Jülich, Germany
Rzecin wetland site, Poland
|Field of Application||Radar altimetry, land and sea-ice|
|Dataset Size||4 TB|
Data Citation Users, who, in their research, use ESA Earth Observation data that have been assigned a Digital Object Identifier (DOI), are asked to use it when citing the data source in their publications:
Digital Object Identifier: https://doi.org/10.5270/ESA-50a3dd4 SoyFLEX 2015