V. Demarez

Modeling radiative transfer in heterogeneous 3-D vegetation canopies

Abstract The DART (discrete anisotropic radiative transfer) model simulates radiative transfer in heterogeneous 3-D scenes that may comprise different landscape features, i.e., leaves, grass, trunks, water, soil. The scene is divided into a rectangular cell matrix, i.e., building block for simulating larger scenes. Cells are parallelipipedic. Their optical properties are represented by individual scattering phase functions that are directly input into the model or are computed with optical and structural characteristics of elements within the cell. Radiation scattering and propagation are simulated with the exact kernel and discrete ordinate approaches; any set of discrete direction can be selected. In addition to topography and hot spot, leaf specular and first-order polarization mechanisms are modeled. Two major iterative steps are distinguished: 1) Cell illumination with direct sun radiation: Within cell multiple scattering is accurately simulated. 2) Interception and scattering of previously scattered radiation: Atmospheric radiation, possibly anisotropic, is input at this stage. Multiple scattering is stored as spherical harmonics expansions, for reducing computer memory constraints. The model iterates on step 2, for all cells, and stops with the energetic equilibrium. Two simple accelerating techniques can be used: 1) Gauss Seidel method, i.e., simulation of scattering with radiation already scattered at the iteration stage, and (2) decrease of the spherical harmonics expansion order with the iteration order. Moreover, convergence towards the energetic equilibrium is accelerated with an exponential fitting technique. This model predicts the bidirectional reflectance distribution function of 3-D canopies. Radiation components associated with leaf volume and surface mechanisms are distinguished. It gives also the radiation regime within canopies, for further determination of 3-D photosynthesis rates and primary production. Accurate modeling of multiple scattering within cells, combined with the fact that cells can have different x,y,z dimensions, is well adapted to remote sensing based studies, i.e., scenes with large dimensions. The model was successfully tested with homogeneous covers. Preliminary comparisons of simulated reflectance images with remotely acquired spectral images of a 3-D heterogeneous forest cover stressed the usefulness of the DART model for conducting studies with remotely acquired information.

Modeling BRF and Radiation Regime of Boreal and Tropical Forests: I. BRF

Abstract Monitoring of forest evolution and functioning with remote sensing depends on canopy BRF (bidirectional reflectance factor) sensitivity to biophysical parameters and to canopy PAR (photosynthetically active radiation) regime. Here, we study the canopy BRF of a tropical (Sumatra) and three boreal (Canada) forest sites, with the DART (discrete anisotropic radiative transfer) model. The behavior of PAR regime of these forests is analyzed in a companion article. We assessed the BRF sensitivity to some major experimental parameters (scale of analysis, viewing and illumination directions, sky radiation) and compared it with BRF sensitivity to commonly studied biophysical quantities: Leaf area index (LAI) and leaf optical properties. Simulations showed that BRF directional anisotropy is very large for all forests. For example, maximum relative reflectance difference with view zenith angle less than 25° is around 0.5 in the visible, 0.4 in the short wave infrared, and 0.25 in the near-infrared for tropical forest. We showed that this BRF variability associated with experimental conditions can hamper the remote detection of forest LAI and tree cover change such as deforestation of tropical forest. DART BRFs of the boreal sites were favorably compared with ground (PARABOLA) and airborne (POLDER) measured BRFs. This work stressed 1) the potential of the DART model, 2) the importance of accurate field data for validation approaches, and 3) the very strong influence of canopy architecture on forest BRF; for example, depending on forest sites, a LAI increase may imply that nadir near-infrared reflectance increases or decreases.

A Modeling Approach for Studying Forest Chlorophyll Content

Abstract Imaging spectroscopy from space is a potentially powerful tool for assessing vegetation chemistry with approaches that rely either on empirical relationships or on the inversion of reflectance models. However, this assessment can be erroneous if the 3-D spatial distribution of the vegetation is neglected. Sophisticated radiative transfer models are often required to account for the 3-D canopy architecture. Due to long computation times, however, these models are not well adapted to sensitivity analyses and numerical inversions that require hundred of calls of the merit function. This paper presents a methodology developed to simulate vegetation reflectance spectra quickly and accurately (i.e., without neglecting the 3-D canopy architecture). Canopy reflectance spectra are calculated by linearly interpolating spectra pre-computed with a coupled model: a 3-D canopy model (DART) and a leaf optical properties model (PROSPECT). This approach was successfully tested by studying the influence of forest architecture on the determination of leaf chlorophyll concentration (Chlf) from reflectance measurements. We considered the case of beech stands (Fagus sylvatica L.) of the Fontainebleau Forest, France. The leaf chlorophyll concentration was characterized by the position of the inflection point of the red edge (λi). Apart from Chlf, we considered four other influential factors on λi: the LAI (leaf area index), the viewing direction, the understory reflectance, and the canopy architecture (i.e., a theoretical turbid medium, a pole stand, and a mature stand). Results demonstrated the strong influence of canopy architecture. For example, the λi has larger values for mature stands than for pole stands (δλi>10 nm), whatever the LAI and the viewing directions. Thus, errors on Chlf can be larger than 23 μg/cm2 if canopy architecture is nelected.

Land Cover and Crop Type Classification along the Season Based on Biophysical Variables Retrieved from Multi-Sensor High-Resolution Time Series

With the ever-increasing number of satellites and the availability of data free of charge, the integration of multi-sensor images in coherent time series offers new opportunities for land cover and crop type classification. This article investigates the potential of structural biophysical variables as common parameters to consistently combine multi-sensor time series and to exploit them for land/crop cover classification. Artificial neural networks were trained based on a radiative transfer model in order to retrieve high resolution LAI, FAPAR and FCOVER from Landsat-8 and SPOT-4. The correlation coefficients between field measurements and the retrieved biophysical variables were 0.83, 0.85 and 0.79 for LAI, FAPAR and FCOVER, respectively. The retrieved biophysical variables’ time series displayed consistent average temporal trajectories, even though the class variability and signal-to-noise ratio increased compared to NDVI. Six random forest classifiers were trained and applied along the season with different inputs: spectral bands, NDVI, as well as FAPAR, LAI and FCOVER, separately and jointly. Classifications with structural biophysical variables reached end-of-season overall accuracies ranging from 73%–76% when used alone and 77% when used jointly. This corresponds to 90% and 95% of the accuracy level achieved with the spectral bands and NDVI. FCOVER appears to be the most promising biophysical variable for classification. When assuming that the cropland extent is known, crop type classification reaches 89% with spectral information, 87% with the NDVI and 81%–84% with biophysical variables.

A Generic Algorithm to Estimate LAI, FAPAR and FCOVER Variables from SPOT4_HRVIR and Landsat Sensors: Evaluation of the Consistency and Comparison with Ground Measurements

The leaf area index (LAI) and the fraction of photosynthetically active radiation absorbed by green vegetation (FAPAR) are essential climatic variables in surface process models. FCOVER is also important to separate vegetation and soil for energy balance processes. Currently, several LAI, FAPAR and FCOVER satellite products are derived moderate to coarse spatial resolution. The launch of Sentinel-2 in 2015 will provide data at decametric resolution with a high revisit frequency to allow quantifying the canopy functioning at the local to regional scales. The aim of this study is thus to evaluate the performances of a neural network based algorithm to derive LAI, FAPAR and FCOVER products at decametric spatial resolution and high temporal sampling. The algorithm is generic, i.e., it is applied without any knowledge of the landcover. A time series of high spatial resolution SPOT4_HRVIR (16 scenes) and Landsat 8 (18 scenes) images acquired in 2013 over the France southwestern site were used to generate the LAI, FAPAR and FCOVER products. For each sensor and each biophysical variable, a neural network was first trained over PROSPECT+SAIL radiative transfer model simulations of top of canopy reflectance data for green, red, near-infra red and short wave infra-red bands. Our results show a good spatial and temporal consistency between the variables derived from both sensors: almost half the pixels show an absolute difference between SPOT and LANDSAT estimates of lower that 0.5 unit for LAI, and 0.05 unit for FAPAR and FCOVER. Finally, downward-looking digital hemispherical cameras were completed over the main land cover types to validate the accuracy of the products. Results show that the derived products are strongly correlated with the field measurements (R2 > 0.79), corresponding to a RMSE = 0.49 for LAI, RMSE = 0.10 (RMSE = 0.12) for black-sky (white sky) FAPAR and RMSE = 0.15 for FCOVER. It is concluded that the proposed generic algorithm provides a good basis to monitor the seasonal variation of the vegetation biophysical variables for important crops at decametric resolution.

Early Detection of Summer Crops Using High Spatial Resolution Optical Image Time Series

In the context of climate change, agricultural managers have the imperative to combine sufficient productivity with durability of the resources. Many studies have shown the interest of recent satellite missions as suitable tools for agricultural surveys. Nevertheless, they are not predictive methods. A system able to detect summer crops as early as possible is important in order to obtain valuable information for a better water management strategy. The detection of summer crops before the beginning of the irrigation period is therefore our objective. The study area is located near Toulouse (southwestern France), and is a region of mixed farming with a wide variety of irrigated and non-irrigated crops. Using the reference data for the years concerned, a set of fixed thresholds are applied to a vegetation index (the Normalized Difference Vegetation Index, NDVI) for each agricultural season of multi-spectral satellite optical imagery acquired at decametric spatial resolutions from 2006 to 2013. The performance (i.e., accuracy) is contrasted according to the agricultural practices, the development states of the different crops and the number of acquisition dates (one to three in the results presented here). The detection of summer crops reaches 64% to 88% with a single date, 80% to 88% with two dates and 90% to 99% with three dates. The robustness of this method is tested for several years (showing an impact of meteorological conditions on the actual choice of images), several sensors and several resolutions.

Spatialization of crop leaf area index and biomass by combining a simple crop model SAFY and high spatial and temporal resolutions remote sensing data

The recent availability of high spatial resolution sensors offers new perspectives for terrestrial applications (agriculture, risks). The aim of this work is to develop a methodology for deriving biophysical variables (Green Leaf Area Index — GLAI, phytomass) from multi-temporal observations at high spatial resolution in order to run a crop model at a regional scale. Accurate predictive crop models require a large set of input parameters, which are not easily available over large area. Spatial upscaling of such models is thus difficult. The use of simple model avoids spatial upscaling issues. This study is focused on SAFY model (Simple Algorithm For Yield estimates) developed by [1]. Key SAFY parameters were calibrated using temporal GLAI profiles, empirically estimated from FORMOSAT-2 time series of images. Most of the SAFY parameters are crop related and have been fixed according to literature. However some parameters are more specific and have been calibrated based on GLAI derived from FORMOSAT-2 observations at a field scale. Two calibration strategies are evaluated as a function of sampling (frequency and temporal distribution) of remote sensing data. Spatial upscaling simulations are assessed based on biomass in-situ measurements taken over maize. Good agreement between modelled and measured phytomass have been found on maize (RMSE =3D 20 g.m−2).

Retrieval of forest biophysical parameters from remote sensing images with the DART model

This paper presents a quantitative analysis of the sensitivity of textural information of high resolution remote sensing images of a forest plantation (Les Landes, France) with a number of biophysical characteristics: tree cover, crown diameter, distance between rows and leaf area index (LAI). The influence of spatial resolution and viewing and illumination configurations are also assessed. The work is conducted with directional images simulated with the DART (Discrete Anisotropic Radiative Transfer) model, whereas textural information is quantified by means of variograms. Finally, actual 1.67 m resolution spectral images provide a partial validation of the approach and results.

Contribution of Remote Sensing for Crop and Water Monitoring

Abstract: Agriculture brings with it a number of issues: agricultural production and food security, water and soil resource conservation, limiting the impact of farming on the quality of our environment (water, air, soil). In the context of global changes and the challenges they pose for agriculture sustainability, our ability to characterize how croplands function in terms of water, carbon and particle fluxes is crucial. Developments of agro-ecosystems modeling, including their interactions with the atmosphere and the anthropogenic factors are valuable tools for progress in this direction. Remote sensing, with the high variety of spectral ranges and the fine spatial and temporal resolution currently available, is a tool of great value for various applications in agriculture. The availability of robust inverse methods that allow surface biophysical variables to be assessed, combined with modeling approaches, makes it a high performance tool. The major contributions of remote sensing include: – its ability to cover large stretches of land and to provide information on the various land uses and practices generated by agriculture. These uses and practices are important to know, both for census purposes (agricultural statistics, agri-environmental monitoring, etc.) and for modeling the behavior of agro-hydrosystems; – providing frequent variables characterizing soil and vegetation properties that allows us to monitor the status of crops, their production potential, their irrigation requirements. This monitoring is a highly strategic issue, both for forecasting purposes, food security and good resource management; – the possibility, from the same information, of assessing the contribution of agricultural lands to net emissions of CO 2 and other greenhouse gases (GHGs); this assessment is essential for proposing alternative agricultural scenarios for mitigating the contribution of croplands to climate change; – thanks to the fine spatial and temporal resolution of information, the possibility of providing a decision support for farming activities according to the intra-field variability (precision farming). This dimension represents an important lever for enabling agricultural systems to achieve better efficiency and economical use of inputs for an agriculture that respects the environment.

BRDF behaviour of a tropical forest surveyed from space

Survey of tropical forest evolution and functioning with remote sensing is hampered by the variability of their BRDF (bi-directional reflectance distribution function); e.g. automatic classifications may be totally erroneous if view and illumination conditions are not taken into account. The authors used a new radiative transfer model (DART, Discrete Anisotropic Radiative Transfer) to analyze BRDF behaviour of a tropical forest plot in Central Sumatra, Indonesia. Simulations stressed that BRDF anisotropy, especially for low Sun zenith angles (/spl theta//sub s/), may be large enough to make difficult the study of forest evolution with satellite data time series. Variations were up to 30% for VIS (visible), 20% for NIR (near infrared) and 25% for SWIR (short wave infrared), for viewing zenith angles (/spl theta//sub v/) smaller than 25/spl deg/. Larger variations occurred in the hot spot configuration and for variable Sun zenith angles /spl theta//sub s/: up to 50% for VIS, 30% for NIR, 40% for SWIR, whereas local topography and diffuse atmospheric radiation (SKYL) had a much smaller influence. On the other hand, variations due to a 50% cover degradation were 14% for VIS, 23% for NIR and 18% for SWIR at nadir and /spl theta//sub s/=35.