Preharvest seed composition estimation using satellite imaging provides critical information for food security planning and management at a regional scale. As one of the staple crops, soybean plays important role in U.S. economic development. Estimating soybean seed composition is the precondition for improving seed quality and meal content at scale, therefore, maintaining U.S. soy competence in international markets. Traditionally, soybean seed compositions are measured after harvest via wet chemistry analysis, which is time-consuming and expensive. This study presents very first, to the best of our knowledge, satellite remote sensing of soybean seed composition. We demonstrate that WorldView-3 satellite imagery and machine/deep learning is powerful tool to predict seed composition from standing crops.
High-throughput phenotyping and genotyping have provided a vast source of information for evaluating the genetic merit of different breeding materials, but their implementation has been limited in alfalfa due to the complexity of the genome and the perennial nature of the crop. Vegetative indices (VIs) collected from an unmanned aerial vehicle (UAV) equipped with multi-spectral camera can be used to study forage growth and development throughout each growth cycle. Random regression models could be implemented to fit such longitudinal phenotypes like VIs collected over time to estimate growth curves, to access genetic variation in growth and the relations of VIs to end-use traits like forage yield and quality. The main objectives of this project are (1) to incorporate aerial high-throughput phenotyping to predict performance and genetic merit of the breeding materials, (2) to fit longitudinal random regression model to estimate genotype-specific growth curves, and (3) to develop a genotyping approach to estimate genetic relationships between alfalfa populations. The imaging of the alfalfa experimental trials was done every ~ 4.3 days throughout the growing season. The Vegetative indices (VIs) close to the harvest date were extracted and used to fit multi-traits models to evaluate the genetic correlations between VIs and forage biomass yield. The VIs considered were Normalized Vegetative index (NDVI), Green NDVI, Red Edge NDVI, simple ratio of Near Infrared to Red (NIR), and Digital Surface Map (DSM). The preliminary results showed highest correlation of Green NDVI and biomass yield (0.4053, 0.7875, and 0.6779), followed by Rededge NDVI and biomass yield (0.417, 0.7898, and 0.6417) for the first, second and third cuttings respectively for the experimental trial located at Helfer, Ithaca. Heritability estimates ranging from 0.03 to 0.75 was observed indicating the presence of genetic variation in these VIs. Pairwise Fst values estimated from population-level genotyping approach was found to be efficient estimates of genetic relatedness between populations. Random regression models with a linear spline function and legendre polynomials including other environmental trials are under evaluation to see the potentiality of these models to fit VIs from multiple time points.
Since 1990s, dramatic land-use changes have occurred across mainland China. Large areas of cereal lands have been converted into horticultural crops because of lucrative economic benefits. Fruits and vegetables together in China have consumed more than 30% of N fertilizers. Therefore, understanding the long-term effects of land use change from cereals to orchard on N budgets in croplands at the county scale over long-term is very important for managing N in agricultural systems in China. We selected three Counties (Luochuan (LC), Mixian (MX), and Wugong (WG)) on the Loess Plateau with different land-use changes since 1990s in Shaanxi to compare the changes of N budgets in croplands at the county scales. The main crops in the three Counties were cereals (wheat and corn) before 1990s. After 1990s, the land uses in LC and MX changed dramatically; and LC and MX become the main apple and kiwifruit production county in China, respectively. The main crops in WG are still wheat and corn. It provides an ideal reference to compare the effects of land use changes on N budgets in croplands at the county scale. The N inputs and outputs, N surplus and N use efficiency (NUE, computed as N in harvested crop products divided by N inputs) at the three Counties from 1990 to 2017 were quantified. The annual N inputs and N surplus of the three Counties since 1990s were increased, and NUE decreased significantly. Compared to WG, the annual N input and N surplus of cropland of LC and MX were very high, and NUE was very low. For example, NUE in LC decreased from 64% in 1990 to 12% in 2017; and NUE in WG decreased from 55% in 1992 to 38% in 2017. To understand the fate of surplus N in cropland of LC, we also collected soil profile samples (0-6 m) from cereal lands and apple orchards in different sites of the county. The average nitrate accumulation in 0–6 m soil profiles reached 5611 kg N/ha in 2017. Approximately 67% of the total N surpluses in LC from 1990 to 2017 was accumulated in soil profile as nitrate. Land use change from cereals to orchard result in high surplus N in croplands at the county scale. The nitrate accumulation in the vadose zone is the main fate of surplus N in the intensive agricultural landscape, which should be considered an important component of the soil N budget to optimize production and environmental protection.
Obolodiplosis robiniae was discovered in Eurasia at the beginning of the 21st century and was then spread at an explosive rate. Here, we explore the current and future (in years 2050 and 2070) trends in the potential distribution of O. robiniae in Eurasia under diverse climate change scenarios based on a maximum entropy (MaxEnt) model. Our results showed that the current potential distribution area of O. robiniae is within the range of 21.58°-65.66°N in the Eurasian continent. The total current potential distribution (CPD) area of O. robiniae in Eurasia was 10,896,309.16 km2 , with suitable areas covering a substantial section of Europe. The Annual Mean Temperature (Bio1), Annual Precipitation (Bio12), and the Precipitation of the Driest Month (Bio14) are the most important bioclimatic variables determining the potential distribution of O. robiniae. The future area suitable for habitat of O. robiniae is characterized by a large-scale northward expansion trend with temperature elevation. The marginally suitable and highly suitable areas would thus increase, whereas the southern appropriate areas would shrink. Under the SSP585 scenario, in 2070, the suitable area of O. robiniae would be the largest, up to 14,696,253.77 km2, which is 34.87% more than the current suitable area. This information would facilitate the provision of early warning on the potential distribution areas of O. robiniae issued by the forestry quarantine departments of Asian and European countries and provide a scientific basis for the prevention and control of O. robiniae spread and outbreaks.
The expanding geographic range of Phyllachora maydis, the fungus that induces Tar Spot infection on corn foliage, is increasingly threatening a Michigan industry that contributes over $1 billion to the state’s economy annually. Advances in machine learning now enable quantification of crop infection presence and severity using powerful object detection packages such as Tensorflow, Keras, and more. Tensorflow, specifically, has developed Application Programming Interface (API) tools to connect powerful object detection capabilities with streamlined usability. Foliar infection of maize by P. maydis is often difficult to detect early. Visible lesions initially appear tiny, ambiguous, and sparse, making them difficult to identify with the naked eye. Both farmers and breeders of corn desperately need better tools that allow early, definitive detection of lesions and provide more time for management decisions. This tool must verify presence of P. maydis and quantify infection severity as quickly as possible to allow growers the most options for treatment. I propose a combination of supervised machine learning using Tensorflow for custom object detection, and containerized application-development software such as Docker to create a user interface accessible on desktop or mobile devices. This application will be developed by weaving the transferrable infrastructure of Docker with the powerful machine learning platforms Tensorflow and Tensorflow Lite, thereby allowing users to analyze images using their preferred operating system. By implementing both complementary Tensorflow platforms, farmers and breeders will be afforded the choice of either capturing and analyzing one image at a time, or detecting lesions continuously in real-time.
Cosegmentation is a recent and rapidly emerging and rapidly growing extension of segmentation, which aims to detect the common object(s) in a group of images. Current cosegmentation methods are ideal and effective only for certain dataset types with limited features that still produce errors making it difficult to obtain detailed metrics of object parts. We propose to build a unified, trainable framework that incorporates multiple features of a high-throughput dataset’s segmented images from multiple algorithms using cosegmentation. Specifically, we propose a novel Cosegmentation for Plant Phenotyping Network (CoPPNet) that utilizes a Fully Convolutional Neural Network with a K-Means Clustering feedback loop for optimal temporal loss. The results from this study will set the benchmark for a novel advancement in computer vision segmentation accuracy and plant phenomics to better understand a plant’s environmental interactions for maximal resilience and yield.
Focusing on non-destructive and automated acquisition of plant phenotypic parameters,this extended abstract proposed an end-to-end deep RNN based network structure for single perspective sparse raw point cloud regression task called DRN. It has been proven to achieve accuracy improvements in PointNet++ and PonitCNN when it comes to regression of lettuce plant height. We believe DRN structure is suitable for feature extraction from plant point cloud data and regression of spatial distance related plant phenotypes like plant height.
Long-term spatially explicit information on crop yield is essential for understanding food security in a changing climate. Here we present a study that combines twenty-years of Landsat and MODIS data, climate and weather records, municipality-level crop yield statistics, random forests and linear regression models for mapping crop yield in a multi-temporal, multi-scale modeling framework. The study was conducted for soybean in Brazil, the world’s largest producer and exporter of this commodity crop. Using a recently developed 30 m resolution, annual (2001-2019) soybean classification map product, we aggregated multi-temporal phenological metrics derived from Landsat and MODIS data over soybean pixels to the municipality scale. We combined phenological metrics with topographic features, long-term climate data, in-season weather data and soil variables as inputs to machine learning models. We trained a multi-year random forests model using yield statistics as reference and subsequently applied linear regression to adjust the biases in the direct output of the random forests model. This model combination achieved the best performance with a root-mean-square-error (RMSE) of 344 kg/ha (12% relative to long-term mean yield) and an r2 of 0.69, on the basis of 20% withheld test data. The RMSE of the leave-one-year-out assessment ranged from 259 kg/ha to 816 kg/ha. To eliminate the artifacts caused by the coarse-resolution climate and weather data, we developed multiple models with different categories of input variables. Employing the per-pixel uncertainty estimates of different models, the final soybean yield maps were produced through per-pixel model composition. We applied the models trained on 2001-2019 data to 2020 data and produced a soybean yield map for 2020, demonstrating the predictive capability of trained machine learning models for operational yield mapping in future years. Our research showed that combining satellite, climate and weather data and machine learning could effectively map crop yield at high resolution, providing critical information to understand yield growth, anomaly and food security.
Abstract: Potato is an important commodity in terms of nutrition basket. Lack of knowledge of this situation specially production level of this staple product has led to un-sustainability in supply and demand so that both consumer and producer or farmer subjects to this un-consistency.by considering potato agronomical calendar and its difference with main commodities such as wheat and barley , we tried to identify and estimate area under cropping potato farms in Hamedan province where are the production domain of this staple using remote sensing techniques and Awifs image time series from IRSP6 satellite. For that reason Awifs time series imaging was used for determine potato cropping area. Results of elementary investigation showed that we need three time series images due to difference in planting, maintaining and harvesting of this yield and others regions main crops including alfalfa, wheat and barley because of climate variations. hence in three intervals 27 April, when we see green vegetation of alfalfa and wheat,18 July when potato has been grew completely at Hamedan and Bahar counties but at Razan and Kabudrahang counties we see complete growth of them at 16 Agust ,up-to-date images of region ordered and prepared as organized manner. Necessary processing such as preparation, atmospheric and geometric correction, vegetation index, un-supervised classification were conducted using suitable training sites in different images of supervised classification on the images. each image was classified into 9 class including: plowing classes, water, vegetation, un-arable lands, lime lands, mountains, salty lands, metropolitan and farrow).phase classification method could identify farms and rangelands, but it was unable to distinguish all of potato farms from others crops using image classification due to close similarity between potato spectral refection percentage and other simultaneous crops at greenness period. therefore three time intervals were used for distinguish area under potato cropping using classified maps so that first vegetation classification extracted from three classified maps and then vegetation map resulted from 18 July image which was include alfalfa, potato, orchards and nursery land uses to obtain potato cropping area in Hamedan and Bahar counties using vegetation class of 27 April image which included wheat, barley, orchard and nursery and without potato masked and area under potato cropping in two counties was estimated. To obtain potato cropping area in Razan and Kabudrahang vegetation result from 16 Agust image which included alfalfa, orchard,nursery and potato was masked using vegetation map resulted from April 27 which included wheat, barley, alfalfa, orchardS ,nursery and without potato and area under cropped was determined. Following integration of these two layers, studied area under cropping map prepared using phase classification method. Also by vegetation indices NDVI and SAVI , area under cropping of three ma
Recently, statistical machine learning and deep learning methods have been widely explored for corn yield prediction. Though successful, machine learning models generated within a specific spatial domain often lose their validity when directly applied to new regions. To address this issue, we designed an unsupervised adaptive domain adversarial neural network (ADANN). Specifically, through domain adversarial training, the ADANN model reduced the impact of domain shift by projecting data from different domains into the same subspace. Also, the ADANN model was designed to be trained in an adaptive way, which guaranteed the model can learn the domain-invariant features and perform accurate yield prediction simultaneously. Informative variables including time-series vegetation indices and sequential weather observations were first collected from multiple data sources and aggregated to the county level. Then, we trained the ADANN model with the extracted features and corresponding reported county-level corn yield from the U.S. Department of Agriculture (USDA). Finally, the trained model was evaluated in four testing years 2016-2019. The U.S. corn belt was used as the study area and counties under study were grouped into two diverse ecological regions. The experimental results showed that the developed ADANN model had better performance than three other state-of-the-art machine learning models in both local experiments (train and test in the same region) and transfer experiments (train and test in different regions). As the first study using adversarial learning for crop yield prediction, this research demonstrates a novel solution for improving model transferability on crop yield prediction.
Socio-economic scenarios such as the Shared Socioeconomic Pathways (SSPs) have been widely used to analyse global change impacts, but representing their diversity is a challenge for the analytical tools applied to them. Taking Great Britain as an example, we represent a set of stakeholder-elaborated UK-SSP scenarios, linked to climate change scenarios (Representative Concentration Pathways), in a globally-embedded agent-based modelling framework. We find that distinct model components are required to account for divergent behavioural, social and societal conditions in the SSPs, and that these have dramatic impacts on land system outcomes. From strong social networks and environmental sustainability in SSP1 to land consolidation and technological intensification in SSP5, scenario-specific model designs vary widely from one another and from present-day conditions. Changes in social and human capitals can generate impacts larger than those of technological and economic change, and comparable to those of modelled climate change. We develop an open-access, transferrable model framework and provide UK-SSP projections to 2080 at 1km2 resolution, revealing large differences in land management intensities, provision of a range of ecosystem services, and the knowledge and motivations underlying land manager decision-making. These differences suggest the existence of large but underappreciated areas of scenario space, within which novel options for land system sustainability could occur.
Bayesian inference of the most plausible parameter values during model calibration is influenced by the method used to combine likelihood values from different observation data sets. In the traditional method of combining likelihood values (AND calibration strategy), it is inherently assumed that the model is error-free, and that different data sets are similarly informative for the inference problem. However, practically every model applied to real-world case studies suffers from model-structural errors. Forcing an imperfect model to describe all data sets simultaneously inevitably leads to a compromised solution. As a result, biased and overconfident predictions hinder responsible risk management and any other prediction-based decisions. To overcome this problem, we propose an alternative OR calibration strategy which allows the model to fit distinct data sets, individually. To demonstrate the effect of choosing between the traditional AND and the proposed OR strategy, we present a case study of calibrating a plant phenology model to observations of the maize crop grown in southwestern Germany between 2010 and 2016. We demonstrate that the OR strategy results in conservative but more reliable predictions than the AND strategy when the behaviour of the target prediction does not represent an average of all data sets. Further, an expert knowledge-based combination of AND-OR could be useful; however, selection of representative calibration data sets is not trivial. We expect our proposed strategy to improve the predictive reliability of imperfect, dynamic models in general, by a more realistic formulation of the likelihood function in the “perfect model setting” of Bayesian updating.
Climate change is modifying the conditions of agricultural production. In particular, precipitations are redistributed in time and space. In the Rhine Valley, this results in prolonged and intensified dry and warm periods in summer on the one side, and wetter winters and heavy rain events on the other side (Riach et al., 2019). In agriculture, dry and warm periods can lead to severe loss in yields and revenue (Fuhrer & Jasper, 2009), while heavy rains can cause erosion and mudslides (Heitz, 2009) and excessive humidity can damage soils (Falloon & Betts, 2010) and favors fungal diseases (Rosenzweig et al., 2001). These evolutions of the water cycle will probably get worse as climate change go forth, and cannot anymore be totally prevented (Averbeck et al., 2019). Adaptation is therefore becoming a vital necessity (Darnhofer et al., 2010). Nevertheless, adoption or not of adaptation measures is a choice which depends on several factors: geographical (accessibility of a water resource; spatial, pedological and topographic situation of the farm); technical (equipment, workforce, know-how); economical (financial capacities, possible subsidies); geolegal (according to the rules in place in different territories). But, it can also depend on the perceptions a farmer has of climate change and of the benefits of adaptation, which are partially constructed through networks (interactions with colleagues, customers or agricultural organizations), leading to various trajectories of adaptation. Moreover, the adaptation measures shall not only be considered through their determinants, but also through their consequences, especially in terms of maladaptation, spatial inequalities but also synergies with mitigation and other issues. We base on semi-structured interviews conducted with crops and wine growing actors in the Rhine Valley (shared between France, Germany and Switzerland). Consequently, we can operate an innovative double comparison, between sectors and between countries, which sheds light on the most influential factors. We also observe that some measures are controversial, and promoted or rejected according to the actors, their perceptions and interests, resulting in a heterogeneous landscape where the role of consumers and borders remains significant. And, sometimes, hinders adaptation.
The launches of the Sentinel-1 synthetic aperture radar satellites in 2014 and 2016 started a new era of high-resolution velocity and strain rate mapping for the continents. However, multiple challenges exist in tying independently processed velocity data sets to a common reference frame and producing high-resolution strain rate fields. We analyse Sentinel-1 data acquired between 2014 and 2019 over the northeast Tibetan Plateau, and develop new methods to derive east and vertical velocities with ~100 m resolution and ~1 mm/yr accuracy across an area of 440,000 km^2. By implementing a new method of combining horizontal gradients of filtered east and interpolated north velocities, we derive the first ~1 km resolution strain rate field for this tectonically active region. The strain rate fields show concentrated shear strain along the Haiyuan and East Kunlun Faults, and local contractional strain on fault junctions, within the Qilianshan thrusts, and around the Longyangxia Reservoir. The Laohushan-Jingtai creeping section of the Haiyuan Fault is highlighted in our data set by extremely rapid strain rates. Strain across unknown portions of the Haiyuan Fault system, including shear on the eastern extension of the Dabanshan Fault and contraction at the western flank of the Quwushan, highlight unmapped tectonic structures. In addition to the uplift across most of the lowlands, the vertical velocities also contain climatic, hydrological or anthropogenic-related deformation signals. We demonstrate the enhanced view of large-scale active tectonic processes provided by high-resolution velocities and strain rates derived from Sentinel-1 data and highlight associated wide-ranging research applications.
Surface meteorological conditions in the midlatitudes are embedded within and affected by synoptic-scale systems, including the movement and persistence of air masses (AMs). Changes in the frequencies of different AMs over the past several decades could potentially have large effects on ecosystems: each organism is exposed to the synergistic effects of the entire suite of atmospheric variables acting upon it – an inherently multivariate environment – which is best captured using AMs. Utilizing a global-scale AM classification system and a global network of tree-ring widths, we investigate how variation in AM frequency impacts tree growth at over 900 locations. We find that AM frequencies are well-correlated with tree growth, especially in the 12-month period from July in the year prior to growth through June in the year of growth. The most important AMs are Dry-Warm and Humid-Cool AMs, which exhibit average correlations of ρ=-0.4 and ρ=+0.4 with global tree growth, respectively, among commonly sampled tree species, with correlations at some sites exceeding ρ=+/-0.8 in some seasons. Compared to empirical models based solely on temperature and precipitation, modeling using only AM frequencies proved superior at nearly 60% of the sites and for over 80% of the well-sampled (n≥10) species. These results should provide a foundation for using AMs to improve forecasts of tree growth, tree stress and wildfire potential. Long-term reconstructions of AM frequencies back several centuries may also be feasible using tree-ring data, which will help contextualize and temporally extend multivariate perspectives of climate change that utilize such air masses.
Subseasonal-to-seasonal (S2S) prediction of winter wheat yields is crucial for farmers and decision-makers to reduce yield losses and ensure food security. Recently, numerous researchers have utilized machine learning (ML) methods to predict crop yield using observational climate variables and satellite data. Meanwhile, some studies also illustrate the potential of state-of-the-art dynamical atmospheric prediction in crop yield forecasting. However, the potential of coupling both methods has not been fully exploited. Herein, we aim to establish a skilled ML–dynamical hybrid model for crop yield forecasting (MHCF v1.0), which hybridizes ML and a global dynamical atmospheric prediction system, and apply it to northern China at the S2S time scale. MHCF v1.0 demonstrates that crop yield forecasting with S2S dynamical predictions generally outperforms that with observational climate data. The coupling of ML and S2S dynamical atmospheric prediction provides a useful tool for yield forecasting, which could guide agricultural practices, policy-making and agricultural insurance.
While it remains debated if the mineral deposits mined for phosphorus fertilizer are running out, phosphorus insecurity is an emerging global issue. We explore how it is linked to the current linear phosphorus economy (LPE) and the historic and current implications. The problems are multifold: there are geopolitical concerns over phosphorus deposits held only by a few nations, sharply rising costs of phosphorus fertilizers, heavy metal contaminants affecting soil and food, problematic phosphorus mining wastes, and the widespread environmental degradation caused by fertilizer inefficiencies. A new phosphorus economy can resolve these problems. Transitioning to a sustainable use of phosphorus demands a circular phosphorus economy (CPE). A CPE supports several Sustainable Development Goals and enables countries without phosphorus deposits to achieve greater phosphorus autonomy. We illustrate current problems with case studies and outline opportunities for change. The CPE will feature phosphorus recovery facilities, waste valorisation technologies, and improved fertilizer formulations that are customised to crop systems. We highlight examples of the rapidly advancing CPE that forms an integral part of the bioeconomy and the circular economy.
With the development of large-scale rice cultivation management initiatives in East Asia, there is concern that a reduction in the number of human cultivators per unit area may lead to poor water management, which could result in decreased land productivity, owing to abnormal high- and low-temperature damage to crops. Accurate simulation of paddy field water temperature is important for studying its impact on crops and for providing timely information to aid in decision making for more efficient management under limited resources. We propose a neural-network framework that considers the heat transfer by the vegetation canopy and applies physical-theory constraints in its training. A novel tuning method is proposed to cope with the trade-off between water temperature accuracy and physical consistency during training to ensure that the calculated water temperature variations in a paddy field enjoy high accuracy and physical consistency. In the experiments, the proposed framework outperforms (with RMSE 0.78°C) both physical process models (with RMSE 1.06°C) and pure neural-network models (with RMSE 0.9°C) while maintaining high accuracy in the case of sparse datasets. Furthermore, an attention-mechanism input layer is integrated into the model to rank feature importance, providing global interpretation to the proposed framework. We also perform sensitivity analysis on the physical process and propose models to compare their different strategies of feature ranking. The results show that the two methods have different sensitivities to different types of feature patterns, but they complement each other. In summary, the proposed model is credible and stable for practical applications and has the potential to guide more efficient paddy management.