European Facility For Airborne Research Dec. 22, 2024, 18:07
AHSPECT
Agriculture-Health-SPECTrometry
Scientific project
Hyperspectral measurements over agro-forestry areas for landscape assessment of agricultural health, ecophysiology, and satellite product validation
None
ROUJEAN Jean-louis
Modelling and measuring aerosol and surface parameters from space, airborne and ground measurements
DO228 - ARF
The measurements of plant physiology and structure (leaf area index, water content, plant pigment content, canopy architecture and density) is at the core of the investigation extensively. Note that such topics have been extensively addressed over the past decade (D’Urso et al., 2004; Schlerf et al., 2005; Stimson et al., 2005; Chun-Jiang et al., 2006) using hyperspectral imagery. The targeted objective is the identification of key spectral wavebands in relationship with specific plant physiology and structural characteristics. In this respect, different approaches are followed : the derivation of more or less empirical spectral vegetation indices (VIs) is found useful as they can serve for proxy to perform extension of the measured quantities around instrumented tower fluxes (that measure continously leaf pigment and structure of the canopy). Note VIs also permit to avoid non-destructive measurements at the patch scale. Radiative transfer modeling is also found useful since all vegetation properties can be retrieved by same time. In this respect, the participants of the project have developed recently the model CANOPUS (e.g. Carrer et al., JGR, 2013), which is specifically designed to handle photon fluxes pathway in discontinuous canopies. Validation with forest measurements for radiation regime has proven to be successful. The next step will be a qualification of the model with crops. In this regard, the chlorophyll pigments will be encoded in CANOPUS modeling in 2014 and the airborne campaign would sustain a model qualification. Another approach will consist to develop a 3D modeling of vegetation canopies over selected sites with DART, which will require collecting architecture measurements. The performances of the new CANOPUS version will be intercompared against PROSPECT-4 & -5 serving for reference to the scientific community in regard to the inclusion of photosynthetic pigments, and also DART at measurement sites. Worth mentioning that PROSPECT is designed to handle the scattering processes from the leaf scale whereas CANOPUS is more specific to the canopy scale. Bridging the two models will be part of the challenging investigation based on the collection of APEX pixels at sub-metric resolution.Simultaneously to airborne hyperspectral measurements, it will be acquired a collection of mandatory ground measurements for the project by partners. The measurement protocol will follow stringent rules of high resolution footprint. Incidentally, at such resolution this will exclude the concommittant presence of ground-based teams during the flights in order to avoid any pixel contamination. Usually, measurements are achieved after aircraft passage. In situ measurements will concern the chlorophyll using state-of-art instrument DUALEX Force-A. (http://www.force-a.eu/fr/dualex.html ). It is also well designed to measure other pigment pools like xanthophyll and anthocyanin that are central for detecting abiotic stress factors. Also the proposing team owes a TRAC instrument for characterizing the clumping effects of the canopy with ranges from leaf scale to plant stand. This is crucial for measuring upscaled leaf area index (LAI) parameter as any inaccuracy in determining properly the structural properties has an impact on the retrieved biochemical leaf properties. In complement, LICOR measurements will be performed to get efficient LAI values. The leaf angle inclination would go also under determination. Importantly, all aforementioned instruments are already properties of the proponents.Biophysical parameters (LAI, fAPAR, fCover) can be derived from optical high and very high resolution images from sensors such as DMC/Deimos (22m), Landsat (30m), SPOT (10m) and RapidEye (5m) using a neural network approach (ref. INRA, Baret and Weiss). In the frame of AHSPECT these biophysical parameters at various spatial resolutions could be compared with each other and validated against field measurements in southwestern France for the main agricultural crops growing in this area. The field work would consist of taking Digital Hemispherical Photographs (DHP) to measure gap fractions from which biophysical parameters can be computed using CanEye software. The proposed work would be highly relevant in preparation of Sentinel-2.If a sufficiently long time series of high resolution images would be available over the 2014 growing season phenological information could be derived from the time series by coupling profile metrics with observed development stages. Hence, the spectral and spatial information content of the APEX data will be exploited to valiÂdate canopy reflectance models such as PROSPECT, SAILH (D’Urso et al., 2004) and CANOPUS in the inverse mode in order to retrieve vegetation parameters such as chlorophyll content, dry matter and canopy geometrical characteristics. Some attempts to retrieve the sun-indiced chlorophyll fluorescence will be performed according previous work by Damm et al. (2011). The accurate estimation of plant water status and plant water stress is essential to the integration of remote sensing into precision agricultural and forestry management. The potential to spectrally estimate plant physiological properties over relatively large areas, and to predict plant water status and plant water stress offer possibility that would need further investigation. Their results indicate the potential use of spectral VIs derived from various scales of remote sensing data for determining plant physiological properties and characteristics. These studies amongst others clearly indicate the improved estimates of plant physiological and structural characteristics from hyperspectral data, allowing for much more detailed spectral analyses and hence more accurate estimates. This will benefit from continuous PRI (Photochemical Reflectance Index).A main objective is also to support the validation of several satellite products (spectral albedo, LAI, FAPAR, etc.) at 333m and km scale issued from PROBA-V first (flying since May 2013). This will serve to prepare the future fusion with S3/OLCI albedo products from 2015. Also benefit will undergo the validation of Venµs and Sentinel-2. Thus, clearly APEX measurements will serve to simulate landscape scenes of S3/OLCI with the help of ACRI-ST (www.acri-st.fr) that are currently developing simulators. The context is the project FP7/ImagineS from Copernicus programme. A separation between bare soil and vegetation albedo will be achieved in considering a method of ‘devegetalisation’ applied to the 21 spectral bands of S3/OLCI. As a health of the vegetation is a key concern, the reflectance products and its derivatives (surface parameters) will be crossed compared with maps of ozone and VOC (volatile organic components) from anthropogenic sources. Hence, quantitative effects of pollutants on the plant physiology during growing stage willl be assessed. For such, an existing landscape classification for southwestern France available at CESBIO will be considered.After cross-calibration of airborne APEX data against ground truth samples, data integration will bring support to simulations of water and carbon flux exchanges at the boundary layers soil-vegetation-atmosphere within the SURFEX platform (http://www.cnrm.meteo.fr/surfex/) using the upgraded radiative transfer code for vegetation CANOPUS. This latter is now incorporated in ISBA-MEB (multiple energy budget) SVAT (Soil Vegetation Atmospheric Transfer) model recently developed at Météo France to allow for a clear separation between soil and vegetation energy budget. This requires of course now a certain level of qualification, which is the aim of AHSPECT also. In this respect, carbon assimilation and primary production are outputs to be analyzed. It will rely on a stratification of the landscape using ECOCLIMAP land cover product consolidated by a higher resolution land cover map of 30m based on Landsat imagery and prepared by CESBIO.
The cross-comparison of the in situ data sets generally depends to a large extent on the meteorological conditions such as spatial variability of the water vapor content, and the orientation of the aircraft with respect to the mean wind. In case of the availability of several flights, at least one flight willl be dedicated to data inter-comparison under some diffuse illumination versus dominant direct illumination as this may yield differences in the photosynthetic activity. In this respect, the scenario of a pollution peak with unusual aerosol load must be envisaged. One flight will be dedicated to such investigation. Note that only day time measurements are required. Note that direct and diffuse incoming light components (global, PAR) will be continuously measured at the CESBIO experimental sites and over SMOSMANIA stations on flight days thanks to a large set of sunphotometers operated manually.
The time schedule for the specific flights requested in the present EUFAR project are the months of September and October 2014. The base of operations for the DO228-101/D-CODE – DLR will be near Toulouse (Francazal airport). Likely, coincident times during morning time with the overpasses of PROBA-V satellite and/or Sentinel-3 scenes is desirable, if weather conditions offer such possibility.
Most confidence sites are located in the vicinity of Toulouse. The two ICOS-like sites of Lamasquère and Auradé maintained by CESBIO propose complete instrumentation, in particular continuous measurements of incident and reflected PAR measurements to be further compared with radiative transfer. Moreover, along a transect between Toulouse and Atlantic ocean, it can be found 7 SMOSMANIA stations measuring soil moisture and key meteorological variables. SMOSMANIA network encompasses a variety of ecosystems representatitve of southern Europe. Also a similar station at the Meteopole (Toulouse) is maintained that should be overflown on the way back to Francazal airport.Besides, it is requested to make some acquisitions over the 1 additonal station (INRA, Avignon) during the flight driveaway from Grenoble to Toulouse.
Four flights of time duration about 2 hours each for a total of 8 or 9 hours would permit to better sample the time component of vegetation growth. Preferably, 1 flight each week is desirable.The aircraft will departure from Grenoble (LFLS) to join Toulouse (LFBO) in about 1:30 (optimized time). During the AHSPECT campaign, the aircraft will be based in Toulouse (France) at Francazal. The flight protocol will aim at over flying the ground stations of reference (Figure and also Table with WGS84 & UTM coordinates were established and are made available)
The aircraft will departure from Grenoble (LFLS) to join Toulouse (LFBO) in about 1:30 (optimized time). During the AHSPECT campaign, the aircraft will be based in Toulouse (France) at Francazal airport. The flight protocol will aim at over flying the ground stations of reference (see Figure 1 & Table 1)
Along the flight, APEX will provide geolocated hyperspectral measurements at regional scale.
APEX
APEX hyperspectral instrument.
Data will be put on LTO-tapes and provided to APEX PAF (Processing and Archive Facility) for treatment.
Starting date: 15-09-2014Ending date: 15-10-2014
Yes
The applicant’s institution has acquired some experience with airborne research. Thus, this project would help to build up an expertise in this area, both among researchers, post-doctoral fellows and Ph.D. students. An approval of this programme may lead to funding for extended stays of graduate students and visiting scientists in the European institutes involved in the project in regard to data analysis and exploitation.
This demand has several objectives amongst which to provide a strong support to the validation and besides enhanced pre-processing of PROBA-V sensor data in the frame of the funded Copernicus FP7 project ImagineS (Implementing Multi-scale AGricultural Indicators Exploiting Sentinels). The PROBA-V project delivers surface albedo products and vegetation parameters (leaf area index, fraction of vegetation, absorbed PAR). Also, still in the frame of ImagineS, it is foreseen to merge PROBA-V and forthcoming Sentinel-3 (S3) data. The present proposal will serve to simulate S3 but also S2 scenes over agro-forested areas in order to prepare the future data fusion and S2-S3 exploitation.In this respect, it is clustered with ISOTHERM (Ice SnOw vegetation HypERspectral Measurements) proposal as this latter yields also the same targeted objectives in regard to PROBA-V (333meters pixels) and S2-S3 exploitation over mountained areas with snow/ice pixels in order generate consolidated remotely sensed surface albedo products.
30