Physically-based atmospheric correction of optical Earth Observation satellite data is used to accurately derive surface biogeophysical parameters free from the atmospheric influence. While water vapor or surface pressure can be univocally characterized, the compensation of aerosol radiometric effects relies on assumptions and parametric approximations of their properties. To determine the validity of these assumptions and approximations in the atmospheric correction of ESA’s FLEX/Sentinel-3 tandem mission, a systematic error analysis of simulated FLEX data within the O ${}_2$ absorption bands was conducted. This paper presents the impact of key aerosol parameters in atmospherically-corrected FLEX surface reflectance and the subsequent Sun-Induced Fluorescence retrieval (SIF). We observed that: (1) a parametric characterization of aerosol scattering effects increases the accuracy of the atmospheric correction with respect to the commonly implemented discretization of aerosol optical properties by aerosol types and (2) the Ångström exponent and the aerosol vertical distribution have a residual influence in the atmospherically-corrected surface reflectance. In conclusion, a multi-parametric aerosol characterization is sufficient for the atmospheric correction of FLEX data (and SIF retrieval) within the mission requirements in nearly 85% (70%) of the cases with average aerosol load conditions. The future development of the FLEX atmospheric correction algorithm would therefore gain from a multi-parametric aerosol characterization based on the synergy of FLEX and Sentinel-3 data. Full article

In the last decade, significant progress has been made in estimating Solar-Induced chlorophyll Fluorescence (SIF) by passive remote sensing techniques that exploit the oxygen absorption spectral regions. Although the O₂–B and the deep O₂–A absorption bands present a high sensitivity to detect SIF, these regions are also largely influenced by atmospheric effects. Therefore, an accurate Atmospheric Correction (AC) process is required to measure SIF from oxygen bands. In this regard, the suitability of a two-step approach, i.e., first an AC and second a Spectral Fitting technique to disentangle SIF from reflected light, has been evaluated. One of the advantages of the two-step approach resides in the derived intermediate products provided prior to SIF estimation, such as surface apparent reflectance. Results suggest that errors introduced in the AC, e.g., related to the characterization of aerosol optical properties, are propagated into systematic residual errors in the apparent reflectance. However, of interest is that these errors can be easily detected in the oxygen bands thanks to the high spectral resolution required to measure SIF. To illustrate this, the predictive power of the apparent reflectance spectra to detect and correct inaccuracies in the aerosols characterization is assessed by using a simulated database with SCOPE and MODTRAN radiative transfer models. In 75% of cases, the aerosol optical thickness, the Angstrom coefficient and the scattering asymmetry factor are corrected with a relative error below of 0.5%, 8% and 3%, respectively. To conclude with, and in view of future SIF monitoring satellite missions such as FLEX, the analysis of the apparent reflectance can entail a valuable quality indicator to detect and correct errors in the AC prior to the SIF estimation. Full article

Physically-based radiative transfer models (RTMs) help understand the interactions of radiation with vegetation and atmosphere. However, advanced RTMs can be computationally burdensome, which makes them impractical in many real applications, especially when many state conditions and model couplings need to be studied. To overcome this problem, it is proposed to substitute RTMs through surrogate meta-models also named emulators. Emulators approximate the functioning of RTMs through statistical learning regression methods, and can open many new applications because of their computational efficiency and outstanding accuracy. Emulators allow fast global sensitivity analysis (GSA) studies on advanced, computationally expensive RTMs. As a proof-of-concept, three machine learning regression algorithms (MLRAs) were tested to function as emulators for the leaf RTM PROSPECT-4, the canopy RTM PROSAIL, and the computationally expensive atmospheric RTM MODTRAN5. Selected MLRAs were: kernel ridge regression (KRR), neural networks (NN) and Gaussian processes regression (GPR). For each RTM, 500 simulations were generated for training and validation. The majority of MLRAs were excellently validated to function as emulators with relative errors well below 0.2%. The emulators were then put into a GSA scheme and compared against GSA results as generated by original PROSPECT-4 and PROSAIL runs. NN and GPR emulators delivered identical GSA results, while processing speed compared to the original RTMs doubled for PROSPECT-4 and tripled for PROSAIL. Having the emulator-GSA concept successfully tested, for six MODTRAN5 atmospheric transfer functions (outputs), i.e., direct and diffuse at-surface solar irradiance ( $E_{dif}$ , $E_{dir}$ ), direct and diffuse upward transmittance ( $T_{dir}$ , $T_{dif}$ ), spherical albedo (S) and path radiance ( $L_0$ ), the most accurate MLRA’s were subsequently applied as emulator into the GSA scheme. The sensitivity analysis along the 400–2500 nm spectral range took no more than a few minutes on a contemporary computer—in comparison, the same analysis in the original MODTRAN5 would have taken over a month. Key atmospheric drivers were identified, which are on the one hand aerosol optical properties, i.e., aerosol optical thickness (AOT), Angstrom coefficient (AMS) and scattering asymmetry variable (G), mostly driving diffuse atmospheric components, $E_{dif}$ and $T_{dif}$ ; and those affected by atmospheric scattering, $L_0$ and S. On the other hand, as expected, AOT, AMS and columnar water vapor (CWV) in the absorption regions mostly drive $E_{dir}$ and $T_{dir}$ atmospheric functions. The presented emulation schemes showed very promising results in replacing costly RTMs, and we think they can contribute to the adoption of machine learning techniques in remote sensing and environmental applications. Full article

Physically-based radiative transfer models (RTMs) help in understanding the processes occurring on the Earth’s surface and their interactions with vegetation and atmosphere. When it comes to studying vegetation properties, RTMs allows us to study light interception by plant canopies and are used in the retrieval of biophysical variables through model inversion. However, advanced RTMs can take a long computational time, which makes them unfeasible in many real applications. To overcome this problem, it has been proposed to substitute RTMs through so-called emulators. Emulators are statistical models that approximate the functioning of RTMs. Emulators are advantageous in real practice because of the computational efficiency and excellent accuracy and flexibility for extrapolation. We hereby present an “Emulator toolbox” that enables analysing multi-output machine learning regression algorithms (MO-MLRAs) on their ability to approximate an RTM. The toolbox is included in the free-access ARTMO’s MATLAB suite for parameter retrieval and model inversion and currently contains both linear and non-linear MO-MLRAs, namely partial least squares regression (PLSR), kernel ridge regression (KRR) and neural networks (NN). These MO-MLRAs have been evaluated on their precision and speed to approximate the soil vegetation atmosphere transfer model SCOPE (Soil Canopy Observation, Photochemistry and Energy balance). SCOPE generates, amongst others, sun-induced chlorophyll fluorescence as the output signal. KRR and NN were evaluated as capable of reconstructing fluorescence spectra with great precision. Relative errors fell below 0.5% when trained with 500 or more samples using cross-validation and principal component analysis to alleviate the underdetermination problem. Moreover, NN reconstructed fluorescence spectra about 50-times faster and KRR about 800-times faster than SCOPE. The Emulator toolbox is foreseen to open new opportunities in the use of advanced RTMs, in which both consistent physical assumptions and data-driven machine learning algorithms live together. Full article

Quantitative methods for mapping sub-pixel land cover fractions are gaining increasing attention, particularly with regard to upcoming hyperspectral satellite missions. We evaluated five advanced regression algorithms combined with synthetically mixed training data for quantifying urban land cover from HyMap data at 3.6 and 9 m spatial resolution. Methods included support vector regression (SVR), kernel ridge regression (KRR), artificial neural networks (NN), random forest regression (RFR) and partial least squares regression (PLSR). Our experiments demonstrate that both kernel methods SVR and KRR yield high accuracies for mapping complex urban surface types, i.e., rooftops, pavements, grass- and tree-covered areas. SVR and KRR models proved to be stable with regard to the spatial and spectral differences between both images and effectively utilized the higher complexity of the synthetic training mixtures for improving estimates for coarser resolution data. Observed deficiencies mainly relate to known problems arising from spectral similarities or shadowing. The remaining regressors either revealed erratic (NN) or limited (RFR and PLSR) performances when comprehensively mapping urban land cover. Our findings suggest that the combination of kernel-based regression methods, such as SVR and KRR, with synthetically mixed training data is well suited for quantifying urban land cover from imaging spectrometer data at multiple scales. Full article

Regression models based on spectral indices are typically empirical formulae enabling the mapping of biophysical parameters derived from Earth Observation (EO) data. Due to its empirical nature, it remains nevertheless uncertain to what extent a selected regression model is the most appropriate one, until all band combinations and curve fitting functions are assessed. This paper describes the application of a Spectral Index (SI) assessment toolbox in the Automated Radiative Transfer Models Operator (ARTMO) package. ARTMO enables semi-automatic retrieval and mapping of biophysical parameters from optical remote sensing observations. The SI toolbox facilitates the assessment of biophysical parameter retrieval accuracy of established as well as new and generic SIs. For instance, based on the SI formulation used, all possible band combinations of formulations with up to ten bands can be defined and evaluated. Several options are available in the SI assessment: calibration/validation data partitioning, the addition of noise and the definition of curve fitting models. To illustrate its functioning, all two-band combinations according to simple ratio (SR) and normalized difference (ND) formulations as well as various fitting functions (linear, exponential, power, logarithmic, polynomial) have been assessed. HyMap imaging spectrometer (430–2490 nm) data obtained during the SPARC-2003 campaign in Barrax, Spain, have been used to extract leaf area index (LAI) and leaf chlorophyll content (LCC) estimates. For both SR and ND formulations the most sensitive regions have been identified for two-band combinations of green (539–570 nm) with longwave SWIR (2421–2453 nm) for LAI (r2: 0.83) and far-red (692 nm) with NIR (1340 nm) or shortwave SWIR (1661–1686 nm) for LCC (r2: 0.93). Polynomial, logarithmic and linear fitting functions led to similar best correlations, though spatial differences emerged when applying the functions to HyMap imagery. We suggest that a systematic SI assessment is a strong requirement in the quality assurance approach for accurate biophysical parameter retrieval. Full article

Abstract: Lookup-table (LUT)-based radiative transfer model inversion is considered a physically-sound and robust method to retrieve biophysical parameters from Earth observation data but regularization strategies are needed to mitigate the drawback of ill-posedness. We systematically evaluated various regularization options to improve leaf chlorophyll content (LCC) and leaf area index (LAI) retrievals over agricultural lands, including the role of (1) cost functions (CFs); (2) added noise; and (3) multiple solutions in LUT-based inversion. Three families of CFs were compared: information measures, M-estimates and minimum contrast methods. We have only selected CFs without additional parameters to be tuned, and thus they can be immediately implemented in processing chains. The coupled leaf/canopy model PROSAIL was inverted against simulated Sentinel-2 imagery at 20 m spatial resolution (8 bands) and validated against field data from the ESA-led SPARC (Barrax, Spain) campaign. For all 18 considered CFs with noise introduction and opting for the mean of multiple best solutions considerably improved retrievals; relative errors can be twice reduced as opposed to those without these regularization options. M-estimates were found most successful, but also data normalization influences the accuracy of the retrievals. Here, best LCC retrievals were obtained using a normalized “L₁ -estimate” function with a relative error of 17.6% (r² : 0.73), while best LAI retrievals were obtained through non-normalized “least-squares estimator” (LSE) with a relative error of 15.3% (r² : 0.74). Full article

River floodplains in the Netherlands serve as water storage areas, while they also have the function of nature rehabilitation areas. Floodplain vegetation is therefore subject to natural processes of vegetation succession. At the same time, vegetation encroachment obstructs the water flow into the floodplains and increases the flood risk for the hinterland. Spaceborne pointable imaging spectroscopy has the potential to quantify vegetation density on the basis of leaf area index (LAI) from a desired view zenith angle. In this respect, hyperspectral pointable CHRIS data were linked to the ray tracing canopy reflectance model FLIGHT to retrieve vegetation density estimates over a heterogeneous river floodplain. FLIGHT enables simulating top-of-canopy reflectance of vegetated surfaces either in turbid (e.g., grasslands) or in 3D (e.g., forests) mode. By inverting FLIGHT against CHRIS data, LAI was computed for three main classified vegetation types, ‘herbaceous’, ‘shrubs’ and ‘forest’, and for the CHRIS view zenith angles in nadir, backward (−36°) and forward (+36°) scatter direction. The −36° direction showed most LAI variability within the vegetation types and was best validated, closely followed by the nadir direction. The +36° direction led to poorest LAI retrievals. The class-based inversion process has been implemented into a GUI toolbox which would enable the river manager to generate LAI maps in a semiautomatic way. Full article

ESA’s upcoming satellite Sentinel-2 will provide Earth images of high spatial, spectral and temporal resolution and aims to ensure continuity for Landsat and SPOT observations. In comparison to the latter sensors, Sentinel-2 incorporates three new spectral bands in the red-edge region, which are centered at 705, 740 and 783 nm. This study addresses the importance of these new bands for the retrieval and monitoring of two important biophysical parameters: green leaf area index (LAI) and chlorophyll content (Ch). With data from several ESA field campaigns over agricultural sites (SPARC, AgriSAR, CEFLES2) we have evaluated the efficacy of two empirical methods that specifically make use of the new Sentinel-2 bands. First, it was shown that LAI can be derived from a generic normalized difference index (NDI) using hyperspectral data, with 674 nm with 712 nm as best performing bands. These bands are positioned closely to the Sentinel-2 B4 (665 nm) and the new red-edge B5 (705 nm) band. The method has been applied to simulated Sentinel-2 data. The resulting green LAI map was validated against field data of various crop types, thereby spanning a LAI between 0 and 6, and yielded a RMSE of 0.6. Second, the recently developed “Normalized Area Over reflectance Curve” (NAOC), an index that derives Ch from hyperspectral data, was studied on its compatibility with simulated Sentinel-2 data. This index integrates the reflectance curve between 643 and 795 nm, thereby including the new Sentinel-2 bands in the red-edge region. We found that these new bands significantly improve the accuracy of Ch estimation. Both methods emphasize the importance of red-edge bands for operational estimation of biophysical parameters from Sentinel-2. Full article