Search

In support of cropland monitoring, operational Copernicus Sentinel-2 (S2) data became available globally and can be explored for the retrieval of important crop traits. Based on a hybrid workflow, retrieval models for six essential biochemical and biophysical crop traits were developed for both S2 bottom-of-atmosphere (BOA) L2A and S2 top-of-atmosphere (TOA) L1C data. A variational heteroscedastic Gaussian process regression (VHGPR) algorithm was trained with simulations generated by the combined leaf-canopy reflectance model PROSAILat the BOA scale and further combined with the Second Simulation of a Satellite Signal in the Solar Spectrum (6SV) atmosphere model at the TOA scale. Established VHGPR models were then applied to S2 L1C and L2A reflectance data for mapping: leaf chlorophyll content (Cab), leaf water content (Cw), fractional vegetation coverage (FVC), leaf area index (LAI), and upscaled leaf biochemical compounds, i.e., LAI ∗ Cab (laiCab) and LAI ∗ Cw (laiCw). Estimated variables were validated using in situ reference data collected during the Munich-North-Isar field campaigns within growing seasons of maize and winter wheat in the years 2017 and 2018. For leaf biochemicals, retrieval from BOA reflectance slightly outperformed results from TOA reflectance, e.g., obtaining a root mean squared error (RMSE) of 6.5 μμg/cm2 (BOA) vs. 8 μμg/cm2 (TOA) in the case of Cab. For the majority of canopy-level variables, instead, estimation accuracy was higher when using TOA reflectance data, e.g., with an RMSE of 139 g/m2 (BOA) vs. 113 g/m2 (TOA) for laiCw. Derived maps were further compared against reference products obtained from the ESA Sentinel Application Platform (SNAP) Biophysical Processor. Altogether, the consistency between L1C and L2A retrievals confirmed that crop traits can potentially be estimated directly from TOA reflectance data. Successful mapping of canopy-level crop traits including information about prediction confidence suggests that the models can be transferred over spatial and temporal scales and, therefore, can contribute to decision-making processes for cropland management.

With an upcoming unprecedented stream of imaging spectroscopy data, there is a rising need for tools and software applications exploiting the spectral possibilities to extract relevant information on an operational basis. In this study, we investigate the potential of a scientific processor designed to quantify biophysical and biochemical crop traits from spectroscopic imagery of the upcoming Environmental Mapping and Analysis Program (EnMAP) satellite. Said processor relies on a hybrid retrieval workflow executing pre-trained machine learning regression models fast and efficiently based on training data from a lookup table of synthetic vegetation spectra and their associated parameterization of the well-known radiative transfer model (RTM) PROSAIL. The established models provide spatial information about leaf area index (LAI), average leaf inclination angle (ALIA), leaf chlorophyll content (Cab) and leaf mass per area (Cm). In contrast to using site-specific training data, the approach facilitates a universal application without the need to integrate a priori information into the processor. Four machine learning algorithms, namely artificial neural networks (ANN), random forest regression (RFR), support vector machine regression (SVR), and Gaussian process regression (GPR), were found to estimate biophysical and biochemical variables of unseen targets with high performance (relative error scores

The inference of functional vegetation traits from remotely sensed signals is key to providing efficient information for multiple plant-based applications and to solve related problems [...]

Conventional methods of plant nutrient estimation for nutrient management need a huge number of leaf or tissue samples and extensive chemical analysis, which is time-consuming and expensive. Remote sensing is a viable tool to estimate the plant’s nutritional status to determine the appropriate amounts of fertilizer inputs. The aim of the study was to use remote sensing to characterize the foliar nutrient status of mango through the development of spectral indices, multivariate analysis, chemometrics, and machine learning modeling of the spectral data. A spectral database within the 350–1050 nm wavelength range of the leaf samples and leaf nutrients were analyzed for the development of spectral indices and multivariate model development. The normalized difference and ratio spectral indices and multivariate models–partial least square regression (PLSR), principal component regression, and support vector regression (SVR) were ineffective in predicting any of the leaf nutrients. An approach of using PLSR-combined machine learning models was found to be the best to predict most of the nutrients. Based on the independent validation performance and summed ranks, the best performing models were cubist (R2 ≥ 0.91, the ratio of performance to deviation (RPD) ≥ 3.3, and the ratio of performance to interquartile distance (RPIQ) ≥ 3.71) for nitrogen, phosphorus, potassium, and zinc, SVR (R2 ≥ 0.88, RPD ≥ 2.73, RPIQ ≥ 3.31) for calcium, iron, copper, boron, and elastic net (R2 ≥ 0.95, RPD ≥ 4.47, RPIQ ≥ 6.11) for magnesium and sulfur. The results of the study revealed the potential of using hyperspectral remote sensing data for non-destructive estimation of mango leaf macro- and micro-nutrients. The developed approach is suggested to be employed within operational retrieval workflows for precision management of mango orchard nutrients.

The current exponential increase of spatiotemporally explicit data streams from satellite-based Earth observation missions offers promising opportunities for global vegetation monitoring. Intelligent sampling through active learning (AL) heuristics provides a pathway for fast inference of essential vegetation variables by means of hybrid retrieval approaches, i.e., machine learning regression algorithms trained by radiative transfer model (RTM) simulations. In this study we summarize AL theory and perform a brief systematic literature survey about AL heuristics used in the context of Earth observation regression problems over terrestrial targets. Across all relevant studies it appeared that: (i) retrieval accuracy of AL-optimized training data sets outperformed models trained over large randomly sampled data sets, and (ii) Euclidean distance-based (EBD) diversity method tends to be the most efficient AL technique in terms of accuracy and computational demand. Additionally, a case study is presented based on experimental data employing both uncertainty and diversity AL criteria. Hereby, a a simulated training data base by the PROSAIL-PRO canopy RTM is used to demonstrate the benefit of AL techniques for the estimation of total leaf carotenoid content (Cxc) and leaf water content (Cw). Gaussian process regression (GPR) was incorporated to minimize and optimize the training data set with AL. Training the GPR algorithm on optimally AL-based sampled data sets led to improved variable retrievals compared to training on full data pools, which is further demonstrated on a mapping example. From these findings we can recommend the use of AL-based sub-sampling procedures to select the most informative samples out of large training data pools. This will not only optimize regression accuracy due to exclusion of redundant information, but also speed up processing time and reduce final model size of kernel-based machine learning regression algorithms, such as GPR. With this study we want to encourage further testing and implementation of AL sampling methods for hybrid retrieval workflows. AL can contribute to the solution of regression problems within the framework of operational vegetation monitoring using satellite imaging spectroscopy data, and may strongly facilitate data processing for cloud-computing platforms.

Non-photosynthetic vegetation (NPV) has been identified as priority variable in the context of new spaceborne imaging spectroscopy missions. In this study we provide a first attempt to quantify NPV biomass from these unprecedented data streams to be provided by multiple recently launched or planned instruments. A hybrid workflow is proposed including Gaussian process regression (GPR) trained over radiative transfer model (RTM) simulations and applying active learning strategies. A soybean field data set including two dates with NPV measurements on yellow and senescent (brown) plant organs was used for model validation, resulting in relative errors of 13.4%. This prototype retrieval model was then applied over a resampled Copernicus Hyperspectral Imaging Mission for the Environment (CHIME) scene, resulting in trustful estimates of NPV biomass for some areas with crop residue cover and senescent vegetation. In view of these results, the proposed workflow may show a promising path towards operational delivery of next-generation global NPV products.

Leaf nitrogen content is key information for ecological and agronomic processes. A number of studies aiming at estimation of leaf nitrogen content used chlorophyll content as a proxy due to a moderate to strong correlation between chlorophyll and nitrogen content during vegetative growth stages. Since leaf nitrogen content is directly linked to leaf protein content, the capacity to accurately estimate leaf protein content may improve robustness of an operational nitrogen monitoring. In the past, the introduction of proteins - as an absorbing input constituent of the PROSPECT leaf model - has been attempted numerous times. Yet, the attempts suffered from a certain number of shortcomings, including limited applicability to both fresh and dry vegetation, inaccurate definition of the specific absorption coefficients, or incomplete accounting for different constituents of leaf dry matter.Here, we introduce PROSPECT-PRO, a new version of the PROSPECT model simulating leaf optical properties based on their biochemical properties and including protein and carbon-based constituents (CBC) as new input variables. These two additional chemical constituents correspond to two complementary constituents of LMA. Specific absorption coefficients for proteins and CBC were produced splitting LOPEX dataset into 50% for calibration and 50%for validation. Both data sets included fresh and dry samples. Our objective is to keep compatibility between PROSPECT-PRO and PROSPECT-D, the previous version of the model, and to ensure the same performances for the estimation of LMA even through its decomposition into two constituents. Therefore, the full validation consisted of two steps:1) PROSPECT-PRO inversion using an iterative optimization approach to retrieve proteins and CBC from LOPEX data2) Testing the compatibility with PROSPECT-D by estimating LMA as the sum of protein and CBC content from independent datasetsThe capacity of PROSPECT-PRO for the accurate estimation of leaf proteins and CBC on LOPEX could be evidenced, with slightly higher performances for the estimation of fresh leaf proteins (NRMSE = 17.3%, R2 = 0.75) than of dry leaf proteins (NRMSE =24.0%, R2 = 0.62). Good overall performances were obtained for the estimation of CBC (NRMSE2>0.90). Based on these results, the carbon/nitrogen ratio of leaves could be modelled accurately.The indirect estimation of LMA through PROSPECT-PRO inversion led to similar or slightly improved results when compared to the estimation of LMA with PROSPECT-D. Hence, PROSPECT-PRO might be of particular interest for precision agriculture applications in the context of nitrogen sensing using observations of current and forthcoming satellite imaging spectroscopy missions.

Monitoring cropland phenology from optical satellite data remains a challenging task due to the influence of clouds and atmospheric artifacts. Therefore, measures need to be taken to overcome these challenges and gain better knowledge of crop dynamics. The arrival of cloud computing platforms such as Google Earth Engine (GEE) has enabled us to propose a Sentinel-2 (S2) phenology end-to-end processing chain. To achieve this, the following pipeline was implemented: (1) the building of hybrid Gaussian Process Regression (GPR) retrieval models of crop traits optimized with active learning, (2) implementation of these models on GEE (3) generation of spatiotemporally continuous maps and time series of these crop traits with the use of gap-filling through GPR fitting, and finally, (4) calculation of land surface phenology (LSP) metrics such as the start of season (SOS) or end of season (EOS). Overall, from good to high performance was achieved, in particular for the estimation of canopy-level traits such as leaf area index (LAI) and canopy chlorophyll content, with normalized root mean square errors (NRMSE) of 9% and 10%, respectively. By means of the GPR gap-filling time series of S2, entire tiles were reconstructed, and resulting maps were demonstrated over an agricultural area in Castile and Leon, Spain, where crop calendar data were available to assess the validity of LSP metrics derived from crop traits. In addition, phenology derived from the normalized difference vegetation index (NDVI) was used as reference. NDVI not only proved to be a robust indicator for the calculation of LSP metrics, but also served to demonstrate the good phenology quality of the quantitative trait products. Thanks to the GEE framework, the proposed workflow can be realized anywhere in the world and for any time window, thus representing a shift in the satellite data processing paradigm. We anticipate that the produced LSP metrics can provide meaningful insights into crop seasonal patterns in a changing environment that demands adaptive agricultural production.

Models of radiative transfer (RT) are important tools for remote sensing of vegetation, as they facilitate forward simulations of remotely sensed data as well as inverse estimation of biophysical and biochemical properties from vegetation optical properties. The remote sensing estimation of foliar protein content is a key to monitoring the nitrogen cycle in terrestrial ecosystems in particular to better understand photosynthetic capacity of plants and improve nitrogen management in agriculture. However, no physically based leaf RT model currently allows for proper decomposition of leaf dry matter into nitrogen-based proteins and carbon-based constituents (CBC), estimated from optical properties of fresh or dry foliage. We developed a new version of the PROSPECT model, named PROSPECT-PRO, which separates nitrogen-based constituents (proteins) from CBC (including cellulose, lignin, hemicellulose and starch). PROSPECT-PRO was calibrated and validated on subsets of the LOPEX dataset, accounting for both fresh and dry broadleaf and grass samples. We applied an iterative model inversion optimization algorithm to identify optimal spectral subdomains for retrieval of leaf protein and CBC contents, with 2125-2174 nm optimal for proteins and 2025-2349 nm optimal for CBCs. PROSPECT-PRO inversions revealed a better performance in estimating proteins from optical properties of fresh than dry leaves. We further tested the ability of PROSPECT-PRO to estimate leaf mass per area (LMA) as the sum of proteins and CBC using independent datasets acquired for numerous plant species. Results showed that PROSPECT-PRO is fully compatible and comparable with its predecessor PROSPECT-D in indirect estimation of LMA. We can conclude from findings of this study that PROSPECT-PRO has a high potential in establishing the carbon-to-nitrogen ratio based on the retrieved CBC-to-proteins ratio., Comment: 47 pages, 8 figures Submitted to Remote Sensing of Environment, March, 20th, 2020

Decades after release of the first PROSPECT + SAIL (commonly called PROSAIL) versions, the model is still the most famous representative in the field of canopy reflectance modelling and has been widely used to obtain plant biochemical and structural variables, particularly in the agricultural context. The performance of the retrieval is usually assessed by quantifying the distance between the estimated and the in situ measured variables. While this has worked for hundreds of studies that obtained canopy density as a one-sided Leaf Area Index (LAI) or pigment content, little is known about the role of the canopy geometrical properties specified as the Average Leaf Inclination Angle (ALIA). In this study, we exploit an extensive field dataset, including narrow-band field spectra, leaf variables and canopy properties recorded in seven individual campaigns for winter wheat (4x) and silage maize (3x). PROSAIL outputs generally did not represent field spectra well, when in situ variables served as input for the model. A manual fitting of ALIA and leaf water (EWT) revealed significant deviations for both variables (RMSE = 14.5°, 0.020 cm) and an additional fitting of the brown leaf pigments (Cbrown) was necessary to obtain matching spectra at the near infrared (NIR) shoulder. Wheat spectra tend to be underestimated by the model until the emergence of inflorescence when PROSAIL begins to overestimate crop reflectance. This seasonal pattern could be attributed to an attenuated development of ALIAopt compared to in situ measured ALIA. Segmentation of nadir images of wheat was further used to separate spectral contributors into dark background, ears and leaves + stalks. It could be shown that the share of visible fruit ears from nadir view correlates positively with the deviations between field spectral measurement and PROSAIL spectral outputs (R² = 0.78 for aggregation by phenological stages), indicating that retrieval errors increase for ripening stages. An appropriate model parameterization is recommended to assure accurate retrievals of biophysical and biochemical products of interest. The interpretation of inverted ALIA as physical leaf inclinations is considered unfeasible and we argue in favour of treating it as a free calibration parameter.

Information about pigment and water contents provides comprehensive insights for evaluating photosynthetic potential and activity of agricultural crops. In this study, we present the concept of using spectral integral ratios (SIR) to retrieve three biochemical traits, namely chlorophyll a and b (Cab), carotenoids (Ccx), and water (Cw) content, simultaneously from hyperspectral measurements in the wavelength range 460−1100 nm. The SIR concept is based on automatic separation of respective absorption features through local peak and intercept analysis between log-transformed reflectance and convex hulls. The algorithm was tested on two synthetically established databases using a physiologically constrained look-up-table (LUT) generated by (i) the leaf optical properties model PROSPECT and (ii) the canopy radiative transfer model (RTM) PROSAIL. LUT constraints were realized based on natural Ccx-Cab relations and green peak locations identified in the leaf optical database ANGERS. Linear regression between obtained SIRs and model parameters resulted in coefficients of determination (R²) of 0.66 (i and ii) for Ccx, R2 = 0.85 (i) and 0.53 (ii) for Cab, and R2 = 0.97 (i) and 0.67 (ii) for Cw, respectively. Using the model established from the PROSPECT LUT, leaf level validation was carried out based on ANGERS data with reasonable results both in terms of goodness of fit and root mean square error (RMSE) (Ccx: R2 = 0.86, RMSE = 2.1 μg cm−2; Cab: R2 = 0.67, RMSE = 12.5 μg cm-2; Cw: R2 = 0.89, RMSE = 0.007 cm). The algorithm was applied to airborne spectrometric HyMap data acquired on 12th July 2003 in Barrax, Spain and to AVIRIS-NG data recorded on 2nd July 2018 southwest of Munich, Germany. Mapping of the SIR results as multiband images (3-segment SIR) allows for intuitive visualization of dominant absorptions with respect to the three considered biochemical variables. Barrax in situ validation using linear regression models derived from PROSAIL LUT showed satisfactory results regarding Cab (R2 = 0.84; RMSE = 9.06 μg cm-2) and canopy water content (CWC, R2 = 0.70; RMSE = 0.05 cm). Retrieved Ccx values were reasonable according to Cab-Ccx-dependence plausibility analysis. Hence, the presented SIR algorithm allows for computationally efficient and RTM supported robust retrievals of the two most important vegetation pigments as well as of water content and is ready to be applied on satellite imaging spectroscopy data available in the near future. The algorithm is publicly available as an interface supported tool within the 'Agricultural Applications' of the EnMAP-Box 3 hyperspectral remote sensing software suite.

Quantitative equivalent water thickness on canopy level (EWTcanopy) is an important land surface variable and retrieving EWTcanopy from remote sensing has been targeted by many studies. However, the effect of radiative penetration into the canopy has not been fully understood. Therefore, in this study the Beer-Lambert law is applied to inversely determine water content information in the 930 to 1060 nm range of canopy reflectance from measured winter wheat and corn spectra collected in 2015, 2017, and 2018. The spectral model was calibrated using a look-up-table (LUT) of 50,000 PROSPECT spectra. Internal model validation was performed using two leaf optical properties datasets (LOPEX93 and ANGERS). Destructive in-situ measurements of water content were collected separately for leaves, stalks, and fruits. Correlation between measured and modelled water content was most promising for leaves and ears in case of wheat, reaching coefficients of determination (R2) up to 0.72 and relative RMSE (rRMSE) of 26% and in case of corn for the leaf fraction only (R2 = 0.86, rRMSE = 23%). These findings indicate that, depending on the crop type and its structure, different parts of the canopy are observed by optical sensors. The results from the Munich-North-Isar test sites indicated that plant compartment specific EWTcanopy allows us to deduce more information about the physical meaning of model results than from equivalent water thickness on leaf level (EWT) which is upscaled to canopy water content (CWC) by multiplication of the leaf area index (LAI). Therefore, it is suggested to collect EWTcanopy data and corresponding reflectance for different crop types over the entire growing cycle. Nevertheless, the calibrated model proved to be transferable in time and space and thus can be applied for fast and effective retrieval of EWTcanopy in the scope of future hyperspectral satellite missions.

The retrieval of quantitative equivalent water thickness on canopy level $(\mathrm{EWT}_{\mathrm{c}})$ is an agriculturally important task for hyperspectral remote sensing. In this study the Beer-Lambert law is applied to inversely determine water content from measured winter wheat spectra collected in 2015 and 2017. The spectral model is calibrated using a look-up-table (LUT) of 50.000 PROSPECT spectra. Validation was performed using two leaf optical properties datasets (LOPEX93 and ANGERS) and in-situ data acquired in Southern Germany. After considering destructive in-situ water content measurements separately for leaves, stems, and fruits, results indicate optically active plant water by plant component in the 930 to 1060 nm range of canopy reflectance. Results for spectrally derived $\mathrm{EWT}_{\mathrm{c}}$ were most promising for leaves and ears reaching coefficients of determination up to 0.75 and a normalized RMSE (nRMSE) of 24% between measured and estimated canopy water content.

The proper management of nitrogen (N) is a pre-requisite for sustainable fertilization in modern agriculture. Methods for N - retrieval from Earth Observation (E.O.) data have been mainly based on empirical algorithms. In the present study, two methods (physically based / hybrid) for the assessment of crop nitrogen content $(\mathrm{N}_{\mathrm{area}})$ and concentration $(\mathrm{N}_{\mathrm{mass}})$ were tested. Data from a hyperspectral field campaign in the framework of the future satellite mission Environmental Mapping and Analysis Program (EnMAP) were exploited using a recalibrated PROSPECT model coupled with the canopy reflectance model 4SAIL. The physically based algorithm achieved relative errors (rRMSE) of 72% for $\mathrm{N}_{\mathrm{area}}$ with $\mathrm{R}^{2}=0.92$ . The hybrid approach obtained higher accuracies with rRMSE lower than 16% for the retrieval of $\mathrm{N}_{\mathrm{mass}}$ . Uncertainties of the predictor variables have to be taken into account. Both algorithms represent interesting techniques for global agricultural monitoring from hyperspectral satellite data but further analysis is required.

We introduce the Interactive Visualization of Vegetation Reflectance Models (IVVRM) tool as a sandbox environment for the PROSAIL family of radiative transfer models. Every interaction with the Graphical User Interface (GUI) invokes a new model run of the updated parameter set and the results are instantly plotted on the screen. The quasi-simultaneous response allows easy hands-on practice with PROSAIL for education and training as well as straightforward inversions of biophysical variables from spectra by manual curve fitting. It is shown that expert knowledge can improve the quality of parameter retrieval and reveal sources of uncertainties in the field data and the models. IVVRM is free of charge and available as an application through the EnMAP-Box 3.0.

Upcoming satellite hyperspectral sensors require powerful and robust methodologies for making optimum use of the rich spectral data. This paper reviews the widely applied coupled PROSPECT and SAIL radiative transfer models (PROSAIL), regarding their suitability for the retrieval of biophysical and biochemical variables in the context of agricultural crop monitoring. Evaluation was carried out using a systematic literature review of 281 scientific publications with regard to their (i) spectral exploitation, (ii) vegetation type analyzed, (iii) variables retrieved, and (iv) choice of retrieval methods. From the analysis, current trends were derived, and problems identified and discussed. Our analysis clearly shows that the PROSAIL model is well suited for the analysis of imaging spectrometer data from future satellite missions and that the model should be integrated in appropriate software tools that are being developed in this context for agricultural applications. The review supports the decision of potential users to employ PROSAIL for their specific data analysis and provides guidelines for choosing between the diverse retrieval techniques.

The future German Environmental Mapping and Analysis Program (EnMAP) mission, due to launch in late 2019, will deliver high resolution hyperspectral data from space and will thus contribute to a better monitoring of the dynamic surface of the earth. Exploiting the satellite’s ±30° across-track pointing capabilities will allow for the collection of hyperspectral time-series of homogeneous quality. Various studies have shown the possibility to retrieve geo-biophysical plant variables, like leaf area index (LAI) or leaf chlorophyll content (LCC), from narrowband observations with fixed viewing geometry by inversion of radiative transfer models (RTM). In this study we assess the capability of the well-known PROSPECT 5B + 4SAIL (Scattering by Arbitrarily Inclined Leaves) RTM to estimate these variables from off-nadir observations obtained during a field campaign with respect to EnMAP-like sun–target–sensor-geometries. A novel approach for multiple inquiries of a large look-up-table (LUT) in hierarchical steps is introduced that accounts for the varying instances of all variables of interest. Results show that anisotropic effects are strongest for early growth stages of the winter wheat canopy which influences also the retrieval of the variables. RTM inversions from off-nadir spectra lead to a decreased accuracy for the retrieval of LAI with a relative root mean squared error (rRMSE) of 18% at nadir vs. 25% (backscatter) and 24% (forward scatter) at off-nadir. For LCC estimations, however, off-nadir observations yield improvements, i.e., rRMSE (nadir) = 24% vs. rRMSE (forward scatter) = 20%. It follows that for a variable retrieval through RTM inversion, the final user will benefit from EnMAP time-series for biophysical studies regardless of the acquisition angle and will thus be able to exploit the maximum revisit capability of the mission.

Spatial and temporal variability of meteorological variables across urban areas due to differences in land surface characteristics is a common phenomenon. Most pronounced is the effect of land cover on air temperature. In this study, parametric and non-parametric statistical approaches (stepwise multiple linear regression, random forests) were applied in order to model sub-daily and daily spatial patterns of the urban heat island intensity in the major city of Augsburg, Southern Germany, and its rural surroundings. A large number of model setups utilizing variables from different land surface data sets as predictors and taking into account different seasonal, daily and meteorological situations was examined. The results were compared concerning different measures of model performance (mean squared skill score, mean squared error, explained variance). For individual setups and situations considerable skill with a mean squared skill score of up to 0.85 was reached. The best performing models were obtained from multiple linear regression for situations with low wind speeds and cloud cover in the morning and evening. Selected models were utilized to derive continuous spatial distributions of the air temperature deviations from a rural reference station. The resulting maps can be useful for various applications, e.g. in the context of urban planning.

Hyperspectral acquisitions have proven to be the most informative Earth observation data source for the estimation of nitrogen (N) content, which is the main limiting nutrient for plant growth and thus agricultural production. In the past, empirical algorithms have been widely employed to retrieve information on this biochemical plant component from canopy reflectance. However, these approaches do not seek for a cause-effect relationship based on physical laws. Moreover, most studies solely relied on the correlation of chlorophyll content with nitrogen, and thus neglected the fact that most N is bound in proteins. Our study presents a hybrid retrieval method using a physically-based approach combined with machine learning regression to estimate crop N content. Within the workflow, the leaf optical properties model PROSPECT-PRO including the newly calibrated specific absorption coefficients (SAC) of proteins, was coupled with the canopy reflectance model 4SAIL to PROSAIL-PRO. The latter was then employed to generate a training database to be used for advanced probabilistic machine learning methods: a standard homoscedastic Gaussian process (GP) and a heteroscedastic GP regression that accounts for signal-to-noise relations. Both GP models have the property of providing confidence intervals for the estimates, which sets them apart from other machine learners. Moreover, a GP-based sequential backward band removal algorithm was employed to analyze the band-specific information content of PROSAIL-PRO simulated spectra for the estimation of aboveground N. Data from multiple hyperspectral field campaigns, carried out in the framework of the future satellite mission Environmental Mapping and Analysis Program (EnMAP), were exploited for validation. In these campaigns, corn and winter wheat spectra were acquired to simulate spectral EnMAP data. Moreover, destructive N measurements from leaves, stalks and fruits were collected separately to enable plant-organ-specific validation. The results showed that both GP models can provide accurate aboveground N simulations, with slightly better results of the heteroscedastic GP in terms of model testing and against in situ N measurements from leaves plus stalks, with root mean square error (RMSE) of 2.1 g/m². However, the inclusion of fruit N content for validation deteriorated the results, which can be explained by the inability of the radiation to penetrate the thick tissues of stalks, corn cobs and wheat ears. GP-based band analysis identified optimal spectral settings with ten bands mainly situated in the shortwave infrared (SWIR) spectral region. Use of well-known protein absorption bands from the literature showed comparative results. Finally, the heteroscedastic GP model was successfully applied on airborne hyperspectral data for N mapping. We conclude that GP algorithms, and in particular the heteroscedastic GP, should be implemented for global agricultural monitoring of aboveground N from future imaging spectroscopy data.

Nitrogen (N) is considered as one of the most important plant macronutrients and proper management of N therefore is a pre-requisite for modern agriculture. Continuous satellite-based monitoring of this key plant trait would help to understand individual crop N use efficiency and thus would enable site-specific N management. Since hyperspectral imaging sensors could provide detailed measurements of spectral signatures corresponding to the optical activity of chemical constituents, they have a theoretical advantage over multi-spectral sensing for the detection of crop N. The current study aims to provide a state-of-the-art overview of crop N retrieval methods from hyperspectral data in the agricultural sector and in the context of future satellite imaging spectroscopy missions. Over 400 studies were reviewed for this purpose, identifying those estimating mass-based N (N concentration, N%) and area-based N (N content, Narea) using hyperspectral remote sensing data. Retrieval methods of the 125 studies selected in this review can be grouped into: (1) parametric regression methods, (2) linear nonparametric regression methods or chemometrics, (3) nonlinear nonparametric regression methods or machine learning regression algorithms, (4) physically-based or radiative transfer models (RTM), (5) use of alternative data sources (sun-induced fluorescence, SIF) and (6) hybrid or combined techniques. Whereas in the last decades methods for estimation of Narea and N% from hyperspectral data have been mainly based on simple parametric regression algorithms, such as narrowband vegetation indices, there is an increasing trend of using machine learning, RTM and hybrid techniques. Within plants, N is invested in proteins and chlorophylls stored in the leaf cells, with the proteins being the major nitrogen-containing biochemical constituent. However, in most studies, the relationship between N and chlorophyll content was used to estimate crop N, focusing on the visible-near infrared (VNIR) spectral domains, and thus neglecting protein-related N and reallocation of nitrogen to non-photosynthetic compartments. Therefore, we recommend exploiting the estimation of nitrogen via the proxy of proteins using hyperspectral data and in particular the short-wave infrared (SWIR) spectral domain. We further strongly encourage a standardization of nitrogen terminology, distinguishing between N% and Narea. Moreover, the exploitation of physically-based approaches is highly recommended combined with machine learning regression algorithms, which represents an interesting perspective for future research in view of new spaceborne imaging spectroscopy sensors.