Detect and attribute the extreme maize yield losses based on spatio-temporal deep learning.

Providing accurate crop yield estimations at large spatial scales and understanding yield losses under extreme climate stress is an urgent challenge for sustaining global food security. While the data-driven deep learning approach has shown great capacity in predicting yield patterns, its capacity to detect and attribute the impacts of climatic extremes on yields remains unknown. In this study, we developed a deep neural network based multi-task learning framework to estimate variations of maize yield at the county level over the US Corn Belt from 2006 to 2018, with a special focus on the extreme yield loss in 2012. We found that our deep learning model hindcasted the yield variations with good accuracy for 2006-2018 (R2 = 0.81) and well reproduced the extreme yield anomalies in 2012 (R2 = 0.79). Further attribution analysis indicated that extreme heat stress was the major cause for yield loss, contributing to 72.5% of the yield loss, followed by anomalies of vapor pressure deficit (17.6%) and precipitation (10.8%). Our deep learning model was also able to estimate the accumulated impact of climatic factors on maize yield and identify that the silking phase was the most critical stage shaping the yield response to extreme climate stress in 2012. Our results provide a new framework of spatio-temporal deep learning to assess and attribute the crop yield response to climate variations in the data rich era.

Double cropping and cropland expansion boost grain production in Brazil.

Brazilian grain production increased more than fourfold from 1980 to 2016. The grain boom was achieved primarily by soybean-corn double cropping and cropland expansion-both show changing spatiotemporal patterns since the 1980s. Here, we quantified the contributions of these two strategies to corn and soybean production in Brazil using municipality-level data from 1980 to 2016. We found the contribution of double cropping to the grain boom steadily increased to 35% and the largest driving force was the increasing demand for grain export. While double cropping dominated the conventional agricultural regions, cropland expansion was still the major strategy in agricultural frontiers such as the Centre-West and Matopiba. The implementation of double cropping offset the equivalent of 76.7 million ha of Brazilian arable land for grain production from 2003 to 2016. Double cropping in Brazil has the potential to help alleviate land burdens in other pantropical countries with increasing global food demand.

Integrated strategic and tactical biomass-biofuel supply chain optimization.

To ensure effective biomass feedstock provision for large-scale biofuel production, an integrated biomass supply chain optimization model was developed to minimize annual biomass-ethanol production costs by optimizing both strategic and tactical planning decisions simultaneously. The mixed integer linear programming model optimizes the activities range from biomass harvesting, packing, in-field transportation, stacking, transportation, preprocessing, and storage, to ethanol production and distribution. The numbers, locations, and capacities of facilities as well as biomass and ethanol distribution patterns are key strategic decisions; while biomass production, delivery, and operating schedules and inventory monitoring are key tactical decisions. The model was implemented to study Miscanthus-ethanol supply chain in Illinois. The base case results showed unit Miscanthus-ethanol production costs were $0.72L(-1) of ethanol. Biorefinery related costs accounts for 62% of the total costs, followed by biomass procurement costs. Sensitivity analysis showed that a 50% reduction in biomass yield would increase unit production costs by 11%.

Influence of feedstock particle size on lignocellulose conversion--a review.

Feedstock particle sizing can impact the economics of cellulosic ethanol commercialization through its effects on conversion yield and energy cost. Past studies demonstrated that particle size influences biomass enzyme digestibility to a limited extent. Physical size reduction was able to increase conversion rates to maximum of ≈ 50%, whereas chemical modification achieved conversions of >70% regardless of biomass particle size. This suggests that (1) mechanical pretreatment by itself is insufficient to attain economically feasible biomass conversion, and, therefore, (2) necessary particle sizing needs to be determined in the context of thermochemical pretreatment employed for lignocellulose conversion. Studies of thermochemical pretreatments that have taken into account particle size as a factor have exhibited a wide range of maximal sizes (i.e., particle sizes below which no increase in pretreatment effectiveness, measured in terms of the enzymatic conversion resulting from the pretreatment, were observed) from <0.15 to 50 mm. Maximal sizes as defined above were dependent on the pretreatment employed, with maximal size range decreasing as follows: steam explosion > liquid hot water > dilute acid and base pretreatments. Maximal sizes also appeared dependent on feedstock, with herbaceous or grassy biomass exhibiting lower maximal size range (<3 mm) than woody biomass (>3 mm). Such trends, considered alongside the intensive energy requirement of size reduction processes, warrant a more systematic study of particle size effects across different pretreatment technologies and feedstock, as a requisite for optimizing the feedstock supply system.

Adaptation of SUBSTOR for controlled-environment potato production with elevated carbon dioxide.

The SUBSTOR crop growth model was adapted for controlled-environment hydroponic production of potato (Solanum tuberosum L. cv. Norland) under elevated atmospheric carbon dioxide concentration. Adaptations included adjustment of input files to account for cultural differences between the field and controlled environments, calibration of genetic coefficients, and adjustment of crop parameters including radiation use efficiency. Source code modifications were also performed to account for the absorption of light reflected from the surface below the crop canopy, an increased leaf senescence rate, a carbon (mass) balance to the model, and to modify the response of crop growth rate to elevated atmospheric carbon dioxide concentration. Adaptations were primarily based on growth and phenological data obtained from growth chamber experiments at Rutgers University (New Brunswick, N.J.) and from the modeling literature. Modified-SUBSTOR predictions were compared with data from Kennedy Space Center's Biomass Production Chamber for verification. Results show that, with further development, modified-SUBSTOR will be a useful tool for analysis and optimization of potato growth in controlled environments.

Top-level modeling of an ALS system utilizing object-oriented techniques.

The possible configuration of an Advanced Life Support (ALS) System capable of supporting human life for long-term space missions continues to evolve as researchers investigate potential technologies and configurations. To facilitate the decision process the development of acceptable, flexible, and dynamic mathematical computer modeling tools capable of system level analysis is desirable. Object-oriented techniques have been adopted to develop a dynamic top-level model of an ALS system. This approach has several advantages; among these, object-oriented abstractions of systems are inherently modular in architecture. Thus, models can initially be somewhat simplistic, while allowing for adjustments and improvements. In addition, by coding the model in Java, the model can be implemented via the World Wide Web, greatly encouraging the utilization of the model. Systems analysis is further enabled with the utilization of a readily available backend database containing information supporting the model. The subsystem models of the ALS system model include Crew, Biomass Production, Waste Processing and Resource Recovery, Food Processing and Nutrition, and the Interconnecting Space. Each subsystem model and an overall model have been developed. Presented here is the procedure utilized to develop the modeling tool, the vision of the modeling tool, and the current focus for each of the subsystem models.

Modeling and control for closed environment plant production systems.

A computer program was developed to study multiple crop production and control in controlled environment plant production systems. The program simulates crop growth and development under nominal and off-nominal environments. Time-series crop models for wheat (Triticum aestivum), soybean (Glycine max), and white potato (Solanum tuberosum) are integrated with a model-based predictive controller. The controller evaluates and compensates for effects of environmental disturbances on crop production scheduling. The crop models consist of a set of nonlinear polynomial equations, six for each crop, developed using multivariate polynomial regression (MPR). Simulated data from DSSAT crop models, previously modified for crop production in controlled environments with hydroponics under elevated atmospheric carbon dioxide concentration, were used for the MPR fitting. The model-based predictive controller adjusts light intensity, air temperature, and carbon dioxide concentration set points in response to environmental perturbations. Control signals are determined from minimization of a cost function, which is based on the weighted control effort and squared-error between the system response and desired reference signal.

Determination of trace 2,4-dichlorophenoxyacetic acid in fresh produce by gas chromatography with boron trichloride/2-chloroethanol derivatization.

2,4-Dichlorophenoxyacetic acid (2,4-D) residues in fresh produce is officially analyzed as its methyl ester form by gas chromatography with electron capture detection (GC-ECD). Because of safety concerns with diazomethane, the reagent used to form methyl esters, a less toxic and dangerous reagent, BCl3/2-chloroethanol, was considered. With this alternative reagent, the detecting product is a 2-chloroethyl ester. Compared with the methylester, the 2-chloroethylester has a longer retention time and a better signal-to-noise ratio for trace level analysis by GC-ECD. However, the reagent produces too many unwanted background peaks. If peak retention time is the only information available to identify the residue in case of litigation, the presence of too much background noise increases the ambiguity of identification. Therefore, confirmation by interpretation of the mass spectrum and determination of the compound structure is necessary to ensure the validity of the method. Ten commodities were fortified with 2,4-D at 0.1 ppm. Recoveries of 2-chloroethyl esters and methyl esters were 91 and 92%, respectively. The method is safe, simple, and robust.

Off-line high-performance liquid chromatography and solid-phase extraction clean-up for confirmation of pesticide residues in fresh produce by gas chromatography-mass spectrometry.

GC-MS or GC-IT in conjunction with the GC-ECD, GC-FPD and HPLC Post-Column Derivatization for multiresidue scan analyses is a great complementary instrument for identification or confirmation purposes. However, MS in EI mode serves as a non-selective detector that is easily susceptible to interference by produce matrices in trace level residue analysis. The development of an effective clean-up method is essential. In this research, a combination of HPLC/C18 and SPE/Florisil clean-up methods were used to reduce matrices in 10 commodities. Using the PBM system, the reference spectra of the computerized Anaheim data base were fingerprint matched to the fortified residues at the 50 ppb level. For most commodities, the PBM quality value was above 90% with the exception of whole oranges. The study demonstrates that the produce matrix of individual commodities varies considerably. The method met difficulties in the analysis of citrus due to the citrus peels. Although the method is commodity dependent, the clean-up procedure has reduced matrix interference in most cases. Efficiency is approximately in a range of 10(5) to 10(6) times, thus the method is effective in confirming trace level pesticide residues in fresh produce.

Research on flexible automation and robotics for plant production at Rutgers University.

This is an overview of research activities in the areas of flexible automation and robotics (FAR) within controlled environment plant production systems (CEPPS) in the Department of Bioresource Engineering, Rutgers University. In the past thirty years, our CEPPS research has dealt with the topics including structures and energy, environmental monitoring and control, plant growing systems, operations research and decision support systems, flexible automation and robotics, and impact to natural (i.e. surrounding) environment. Computer and modeling/simulation techniques have been utilized extensively. Mechanized systems have been developed to substitute human's physical labor and maintain uniformity in production. Automation research has been directed towards adding, to the mechanized systems, the capabilities of perception, reasoning, communication, and task planning. Computers, because of their programmability, provide flexibility to automated systems, when incorporated with generic hardware devices. Robots are ideal hardware tools to be employed in flexible automation systems. Some technologies developed in our CEPPS research may be readily adaptable to Closed Bioregenerative Life Support Systems (CBLSS).

Gas chromatographic determination of triclopyr in fruits and vegetables.

This research was comprised of two parts: quantitative analyses, and confirmatory test. In the quantitative analyses, five classes of fruits and vegetables comprising 10 individual commodities were fortified with triclopyr herbicide at 0.4 and 0.8 ppm level. Triclopyr was extracted from the matrices and derivatized separately to 2-chloroethylene ester with 2-chloroethanol-BCl3 and methyl ester with diazomethane. The esters were then quantitated by GC-ECD and GC-NPD. The GC-ECD recoveries for 2-chloroethylene ester were 100.0% and 100.7% at 0.4 ppm and 0.8 ppm fortification levels, respectively, whereas methyl ester recovery was 103.9% at 0.4 ppm fortification level. Similarly, the GC-NPD recoveries for 2-chloroethylene ester were 99.0% and 97.9% at 0.4 ppm and 0.8 ppm fortification levels respectively, whereas methyl ester recovery was 102.0% at 0.4 ppm fortification level. In the confirmatory test, the 2-chloroethylene ester was introduced into a GC-ion trap. The EI mass spectrum was then interpreted based on the criteria of molecular ion, isotopes, base ion, characteristic ions and the nitrogen rule. Compared to existing methods, this method has reduced partition solvents to nearly one-tenth. In addition, this method proved to be simple, fast, safe and accurate.

Systems approach to instrumenting and controlling plant growth systems.

Acquisition and analysis of sensory information are foremost for the control and continued operation of any complex system. The sensors and their attributes must be selected by understanding the biological and physical parameters which, first, can describe, and second, when linked to control systems, can modulate, the plant growth system. These parameters are not all understood, or known, and practical sensors may not even exist for their measurement. A systematic analysis of the general plant system would: focus without prejudice on all the descriptive parameters, as well as, their interrelationships within the biophysical system; highlight the significance of each parameter; expose the areas of weakness and strength of current knowledge; expand the knowledge base; provide the platform for the development of operational models for real-time monitoring and control requirements; and support the longer term tactical and strategic planning needs. Components of such a procedure of systematic analysis which is in development for intensive plant production systems within controlled environments will be discussed.

GC/MIP/AED method for pesticide residue determination in fruits and vegetables.

This research describes the results of a gas chromatography/microwave induced plasma/atomic emission detection (GC/MIP/AED) method performed on a Hewlett-Packard 5921A system for pesticide residue analysis in fruits and vegetables. A total of 6 experiments were conducted: (1) sensitivity and linearity studies for elements S, P, Cl, and N by analyzing dursban; (2) a study of instrument response to Cl concentration in pesticide molecules; (3) organochlorinated pesticide recoveries; (4) organophosphate pesticide recoveries; (5) carbamate pesticide recoveries; and (6) investigation of metallic pesticides with plictran and vendex as standards. The rank according to sensitivity and linearity was found to be as follows: S-181 greater than P-178 greater than Cl-479 greater than N-174. Instrument response to the concentration of chlorine atoms in the pesticide molecule was linear, with a correlation coefficient of 0.89. Recoveries of organochlorinated pesticides were 91.7-109.3%, with the exception of citrus, whose recovery was affected by coeluting interferences. Organophosphate recoveries were 73.2% or higher, except for the cygon oxygen analog, which degraded in the GC system under all circumstances. Carbamate recoveries were inconsistent quantitatively; however, the information generated from elements N and S were useful for qualitative confirmation of other methods, such as LC postcolumn derivatization analysis. Overall, the GC/MIP/AED method is powerful for qualitative confirmation in pesticide residue analysis. The instrument's capability of acquiring multi-elements (Cl and P) selectively and accurately is an alternative method for organochlorinated and organophosphate pesticide residue analyses. In addition, the GC/MIP/AED system is easy to use, simple to maintain, and its chromatograms can be interpreted by any chromatography analyst without much prior training.

Gas-liquid chromatographic determination of bromacil residues.

A simple, fast, and accurate method has been developed to determine residues of bromacil (5-bromo-3-sec-butyl-t-methyluracil) herbicide. Following sample extraction, filtration, and concentration, the herbicide is determined using a gas chromatograph equipped with a thermionic nitrogen/phosphorus detector. The method has been used to recover 0.04 ppm bromacil added to citrus, pineapple, soil, and water, and 0.1 ppm bromacil added to alfalfa hay.