Explore the vast range of titles available.

Timely yield prediction is crucial for the agri-food supply chain as a whole. However, different stakeholders in the agri-food sector require different levels of accuracy and lead times in which a yield prediction should be available. For the producers, predictions during the growing season are essential to ensure that information is available early enough for the timely implementation of agronomic decisions, while industries can wait until later in the season to optimize their production process and increase their production traceability. In this study, we used machine learning algorithms, dynamic and static predictors, and a phenology approach to determine the time for issuing the yield prediction. In addition, the effect of data reduction was evaluated by comparing results obtained with and without principal component analysis (PCA). Gaussian process regression (GPR) was the best for predicting maize yield. Its best performance (nRMSE of 13.31%) was obtained late in the season and with the full set of predictors (vegetation indices, meteorological and soil predictors). In contrast, neural network (NNET) and support vector machines linear basis function (SVMl) achieved their best accuracy with only vegetation indices and at the tasseling phenological stage. Only slight differences in performance were observed between the algorithms considered, highlighting that the main factors influencing performance are the timing of the yield prediction and the predictors with which the machine learning algorithms are fed. Interestingly, PCA was instrumental in increasing the performances of NNET after this stage. An additional benefit of the application of PCA was the overall reduction between 12 and 30.20% in the standard deviation of the maize yield prediction performance from the leave one-year outer-loop cross-validation, depending on the feature set.

Forecasting crop yields is becoming increasingly important under the current context in which food security needs to be ensured despite the challenges brought by climate change, an expanding world population accompanied by rising incomes, increasing soil erosion, and decreasing water resources. Temperature, radiation, water availability and other environmental conditions influence crop growth, development, and final grain yield in a complex nonlinear manner. Machine learning (ML) techniques, and deep learning (DL) methods in particular, can account for such nonlinear relations between yield and its covariates. However, they typically lack transparency and interpretability, since the way the predictions are derived is not directly evident. Yet, in the context of yield forecasting, understanding which are the underlying factors behind both a predicted loss or gain is of great relevance. Here, we explore how to benefit from the increased predictive performance of DL methods while maintaining the ability to interpret how the models achieve their results. To do so, we applied a deep neural network to multivariate time series of vegetation and meteorological data to estimate the wheat yield in the Indian Wheat Belt. Then, we visualized and analyzed the features and yield drivers learned by the model with the use of regression activation maps. The DL model outperformed other tested models (ridge regression and random forest) and facilitated the interpretation of variables and processes that lead to yield variability. The learned features were mostly related to the length of the growing season, and temperature and light conditions during this time. For example, our results showed that high yields in 2012 were associated with low temperatures accompanied by sunny conditions during the growing period. The proposed methodology can be used for other crops and regions in order to facilitate application of DL models in agriculture.

In this study, we investigate a flexible deep learning approach to forecast crop yield at the provincial administrative level based on deep 1D and 2D convolutional neural networks with small data. Our approach meets the operational requirements (public and global records of satellite data in an application ready format with near real time updates) of the application and can be transferred to any country where yield statistics are available. Three-dimensional histograms of Normalized Difference Vegetation Index (NDVI) and climate data are used as input to the 2D model and administrative level time series averages of NDVI and climate data are inputted to the 1D model. The best model architecture is identified during efficient and extensive hyperparameter optimization. To demonstrate the relevance of our approach, we hindcast (2002-2018) the yields of Algeria’s three main crops (barley, durum and soft wheat) and contrast its performance with machine learning algorithms and conventional benchmark models used in the previous study. We show that simple benchmarks such as peak NDVI remained difficult to beat and that machine learning models outperformed deep learning models for all forecasting months and all tested crops. We attribute this poor performance to the smallness of the dataset available. All the data inputs are free and accessible for download from https://mars.jrc.ec.europa.eu/asap/download.php.

This paper presents a method for mapping the nitrogen (N) status in a maize field using hyperspectral remote sensing imagery. An airborne survey was conducted with an AISA Eagle hyperspectral sensor over an experimental farm where maize (Zea mays L.) was grown with two N fertilization levels (0 and 100 kg N ha-1) in four replicates. Leaf and canopy field data were collected during the flight. The nitrogen (N) status has been estimated in this work based on the Nitrogen Nutrition Index (NNI) defined as the ratio between the leaf actual N concentration (%Na) of the crop and the minimum N content required for the maximum biomass production (critical N concentration (%Nc)) calculated through the dry mass at the time of the flight (Wflight). The inputs required to calculate the NNI (i.e. %Na and Wflight) have been estimated through regression analyses between field data and remotely sensed vegetation indices. MCARI/MTVI2 (Modified Chlorophyll Absorption Ratio Index / Modified Triangular Vegetation Index 2) showed the best performances in estimating the %Na (R2 = 0.59) and MTVI2 in estimating the Wflight (R2 = 0.80). The %Na and the Wflight were then mapped and used to compute the NNI map over the entire field. The NNI map agreed with the NNI estimated using field data through traditional destructive measurements (R2 = 0.70) confirming the potential of using remotely sensed indices to assess the crop N condition. Finally, a method to derive a pixel based variable rate N fertilization map was proposed as the difference between the actual N content and the optimal N content. We think that the proposed operational methodology is promising for precision farming since it represents an innovative attempt to derive from an aerial hyperspectral image a variable rate N fertilization map based on the actual crop N status.

Sun-induced canopy chlorophyll fluorescence in both the red (FR) and far-red (FFR) regions was estimated across a range of temporal scales and a range of species from different plant functional types using high resolution radiance spectra collected on the ground. Field measurements were collected with a state-of-the-art spectrometer setup and standardized methodology. Results showed that different plant species were characterized by different fluorescence magnitude. In general, the highest fluorescence emissions were measured in crops followed by broadleaf and then needleleaf species. Red fluorescence values were generally lower than those measured in the far-red region due to the reabsorption of FR by photosynthetic pigments within the canopy layers. Canopy chlorophyll fluorescence was related to plant photosynthetic capacity, but also varied according to leaf and canopy characteristics, such as leaf chlorophyll concentration and Leaf Area Index (LAI). Results gathered from field measurements were compared to radiative transfer model simulations with the Soil-Canopy Observation of Photochemistry and Energy fluxes (SCOPE) model. Overall, simulation results confirmed a major contribution of leaf chlorophyll concentration and LAI to the fluorescence signal. However, some discrepancies between simulated and experimental data were found in broadleaf species. These discrepancies may be explained by uncertainties in individual species LAI estimation in mixed forests or by the effect of other model parameters and/or model representation errors. This is the first study showing sun-induced fluorescence experimental data on the variations in the two emission regions and providing quantitative information about the absolute magnitude of fluorescence emission from a range of vegetation types.

Livestock plays an important economic role in Niger – especially in the semi-arid regions – while being highly vulnerable due to the large inter-annual variability of precipitation and hence rangeland production. This study aims to support effective rangeland management by developing an approach for mapping rangeland biomass production. The observed spatiotemporal variability of biomass production is utilised to build a predictive model based on ground and remote sensing data for the period 2001 to 2015. The phenology-tuned seasonal cumulative Normalised Difference Vegetation Index (cNDVI), computed from 10-day image composites of the Moderate-resolution Imaging Spectroradiometer (MODIS) NDVI data, was used as proxy of biomass production. A linear regression model was tuned with multi-annual field measurements of herbaceous biomass at the end of the growing season. Besides a general model utilizing all available sites for the calibration, different aggregation schemes of the study area with a varying number of calibration units and different biophysical meaning were tested. Sampling sites belonging to a certain calibration unit of a selected scheme were aggregated to compute the regression. The different aggregation schemes were evaluated with respect to their predictive power. Results gathered at the different aggregation levels were subjected to a cross-validation (cv) applying a jackknife technique (leaving one year out at a time). In general, the model performance increased with increasing model parameterization indicating the importance of additional unobserved and spatially heterogeneous agro-ecological effects (which might relate to grazing, species composition, optical soil properties, etc.) in modifying the relationship between cNDVI and herbaceous biomass at the end of the season. The biophysical aggregation scheme, whose calibration units were derived from an ISODATA classification utilizing 10-day NDVI images from January 2001 to December 2015, showed the best performance in respect of the predictive power (R2cv = 0.47) and the RMSEcv (398 kg ha-1) although not being the model with the highest number of calibration units. The proposed approach can be applied for the timely production of maps of estimated biomass at the end of the growing season before field measurements are made available.

Satellites offer a privileged view on terrestrial ecosystems and a unique possibility to evaluate their status for estimating the reliability of the ecosystem services they provide. We implement an indicator measuring the stability of annual vegetation productivity at the global scale based on the Normalized Differential Vegetation Index (NDVI) that can be used as a proxy of ecosystem services reliability. This indicator, originally developed in agricultural science, is defined as the squared mean of annual crop production divided by its variance. The resilience indicator is proportional to the return periods of system failure by extreme drought events consistently with the ecological definition of resilience. Here, we implement it on annual precipitation and NDVI time-series, as proxies of green water resources and of vegetation primary production, respectively. We find coherent relationships between annual green water resources resilience and vegetation primary production resilience over a wide range of world biomes. Finally, we estimate the changes of green water resources resilience due to climate change using preliminary results from the upcoming Sixth edition of the Coupled Model Inter-comparison Project (CMIP6) and discuss the potential consequences of global warming for ecosystem service reliability.

Agricultural production statistics underpin diverse research efforts and development activities. Yet despite their critical importance, efforts to collate, update, and harmonize detailed sub-national agricultural production statistics are frequently redundant and incomplete due to the substantial time, effort, and resources required. The persisting lack of coordination and standards in the food systems data community wastes valuable resources and hinders advances in action-oriented food systems knowledge. Here we introduce the HarvestStat sub-national data consortium as an open-source, collaborative, and transparent model to overcome these challenges. HarvestStat is collaboratively producing publicly available databases and datasets for the food systems community and the broader environmental and sustainability sciences by moving beyond closed and disjointed data-gathering efforts. We are guided by core principles of complete data openness—prioritizing high standards of quality assurance; active inclusion—emphasizing involvement from local experts; and collaboration—fostering engagement across communities of data producers and users. We extend an open global call to action, inviting organizations and individuals to engage in advancing this critical agenda.

The global spatial extent of croplands is a crucial input to global and regional agricultural monitoring and modeling systems. Although many new remotely-sensed products are now appearing due to recent advances in the spatial and temporal resolution of satellite sensors, there are still issues with these products that are related to the definition of cropland used and the accuracies of these maps, particularly when examined spatially. To address the needs of the agricultural monitoring community, here we have created a hybrid map of global cropland extent at a 500 m resolution by fusing two of the latest high resolution remotely-sensed cropland products: the European Space Agency’s WorldCereal and the cropland layer from the University of Maryland. We aggregated the two products to a common resolution of 500 m to produce percentage cropland and compared them spatially, calculating two kinds of disagreement: density disagreement, where the two maps differ by more than 80%, and absence-presence of cropland disagreement, where one map indicates the presence of cropland while the other does not. Based on these disagreements, we selected continuous areas of disagreement, referred to in the paper as hotspots of disagreement, for manual correction by experts using the Geo-Wiki land cover application. The hybrid map was then validated using a stratified random sample based on the disagreement layer, where the sample was visually interpreted by a different set of experts using Geo-Wiki. The results show that the hybrid product improves upon the overall accuracy statistics in the areas where the underlying cropland layer from the University of Maryland was improved with the WorldCereal product, but more importantly, it represents an improved spatially explicit cropland mask for early warning and food security assessment purposes.