Search

Flavescence dorée is a grapevine disease affecting European vineyards which has severe economic consequences and containing its spread is therefore considered as a major challenge for viticulture. Flavescence dorée is subject to mandatory pest control including removal of the infected vines and, in this context, automatic detection of Flavescence dorée symptomatic vines by unmanned aerial vehicle (UAV) remote sensing could constitute a key diagnosis instrument for growers. The objective of this paper is to evaluate the feasibility of discriminating the Flavescence dorée symptoms in red and white cultivars from healthy vine vegetation using UAV multispectral imagery. Exhaustive ground truth data and UAV multispectral imagery (visible and near-infrared domain) have been acquired in September 2015 over four selected vineyards in Southwest France. Spectral signatures of healthy and symptomatic plants were studied with a set of 20 variables computed from the UAV images (spectral bands, vegetation indices and biophysical parameters) using univariate and multivariate classification approaches. Best results were achieved with red cultivars (both using univariate and multivariate approaches). For white cultivars, results were not satisfactory either for the univariate or the multivariate. Nevertheless, external accuracy assessment show that despite problems of Flavescence dorée and healthy pixel misclassification, an operational Flavescence dorée mapping technique using UAV-based imagery can still be proposed. [ABSTRACT FROM AUTHOR]

To mitigate impacts of climate-related reduced productivity of French grasslands, a new insurance scheme bases indemnity payouts to farmers on a Moderate Resolution Imaging Spectroradiometer (MODIS)-derived forage production index (FPI). The objective of this study is to compare several approaches for deriving FPI from satellite data to assess whether better relationships with forage productivity can be attained. The approaches assess pasture productivity using as five input factors estimated from remote sensing and ancillary data, i.e.: (1) fraction of absorbed photosynthetically active radiation (fAPAR); (2) radiation use efficiency estimates; (3) PAR estimates; (4) leaf senescence modelling; and (5) growing season modelling. All the possible combinations from these five factors, including different modalities to estimate some of them, lead to 768 models. Model outputs are compared to reference grassland production estimates provided by a mechanistic model (Information et Suivi Objectif des Prairies - ISOP - system) for a sample of 25 forage regions across France for the years 2003, 2007, 2009, 2011, and 2012 (containing one humid, two normal, and two dry years). Results revealed that: (1) the baseline model based on the fraction of green vegetation cover (fCover) seasonal integral has a reasonable linear relationship to production estimates (standardized root mean square error - SRMSE = 0.57 and coefficient of determination - R2 = 0.68); (2) performance of the baseline model improved with a quadratic function (SRMSE = 0.54 and R2 = 0.71); (3) 34 models outperform the baseline model. We, therefore, suggest to replace the baseline model with the best-performing model (SRMSE = 0.42 and R2 = 0.83) in the insurance product. This model integrates daily fCover with a water stress index and sums these over a variable monitoring period in space and time characterized by the phenological indicators start of season and end of season derived from the fCover annual profile. [ABSTRACT FROM AUTHOR]

An index-based insurance solution was developed to estimate and monitor near real-time forage production using the indicator Forage Production Index (FPI) as a surrogate of the grassland production. The FPI corresponds to the integral of the fraction of green vegetation cover derived from moderate spatial resolution time series images and was calculated at the 6 km × 6 km scale. An upscaled approach based on direct validation was used that compared FPI with field-collected biomass data and high spatial resolution (HR) time series images. The experimental site was located in the Lot and Aveyron departments of southwestern France. Data collected included biomass ground measurements from grassland plots at 28 farms for the years 2012, 2013 and 2014 and HR images covering the Lot department in 2013 (n = 26) and 2014 (n = 22). Direct comparison with ground-measured yield led to good accuracy (R2 = 0.71 and RMSE = 14.5%). With indirect comparison, the relationship was still strong (R2 ranging from 0.78 to 0.93) and informative. These results highlight the effect of disaggregation, the grassland sampling rate, and irregularity of image acquisition in the HR time series. In advance of Sentinel-2, this study provides valuable information on the strengths and weaknesses of a potential index-based insurance product from HR time series images. [ABSTRACT FROM AUTHOR]

Precise accounting of carbon stocks and fluxes in tropical vegetation using remote sensing approaches remains a challenging exercise, as both signal saturation and ground sampling limitations contribute to inaccurate extrapolations. Airborne LiDAR Scanning (ALS) data can be used as an intermediate level to radically increase sampling and enhance model calibration. Here we tested the potential of using ALS data for upscaling vegetation aboveground biomass (AGB) from field plots to a forest-savanna transitional landscape in the Guineo–Congolian region in Cameroon, using either a design-based approach or a model-based approach leveraging multispectral satellite imagery. Two sets of reference data were used: (1) AGB values collected from 62 0.16-ha plots distributed both in forests and savannas; and (2) an AGB map generated form ALS data. In the model-based approach, we trained Random Forest models using predictors from recent sensors of varying spectral and spatial resolutions (Spot 6/7, Landsat 8, and Sentinel 2), along with biophysical predictors derived after pre-processing into the Overland processing chain, following a forward variable selection procedure with a spatial 4-folds cross validation. The models calibrated with field plots lead to a systematic overestimation in AGB density estimates and a root mean squared prediction error (RMSPE) of up to 65 Mg.ha−1 (90%), whereas calibration with ALS lead to low bias and a drop of ~30% in RMSPE (down to 43 Mg.ha−1, 58%) with little effect of the satellite sensor used. Decomposing bias along the AGB density range, we show that multispectral images can (in some specific cases) be used for unbiased prediction at landscape scale on the basis of ALS-calibrated statistical models. However, our results also confirm that, whatever the spectral indices used and attention paid to sensor quality and pre-processing, the signal is not sufficient to warrant accurate pixelwise predictions, because of large relative RMSPE, especially above (200–250 t/ha). The design-based approach, for which average AGB density values were attributed to mapped land cover classes, proved to be a simple and reliable alternative (for landscape to region level estimations), when trained with dense ALS samples. [ABSTRACT FROM AUTHOR]