Uncertainties in predicting rice yield by current crop models under a wide range of climatic conditions

Predicting rice (Oryza sativa) productivity under future climates is important for global food security. Ecophysiological crop models in combination with climate model outputs are commonly used in yield prediction, but uncertainties associated with crop models remain largely unquantified. We evaluated 13 rice models against multi‐year experimental yield data at four sites with diverse climatic conditions in Asia and examined whether different modeling approaches on major physiological processes attribute to the uncertainties of prediction to field measured yields and to the uncertainties of sensitivity to changes in temperature and CO₂ concentration [CO₂]. We also examined whether a use of an ensemble of crop models can reduce the uncertainties. Individual models did not consistently reproduce both experimental and regional yields well, and uncertainty was larger at the warmest and coolest sites. The variation in yield projections was larger among crop models than variation resulting from 16 global climate model‐based scenarios. However, the mean of predictions of all crop models reproduced experimental data, with an uncertainty of less than 10% of measured yields. Using an ensemble of eight models calibrated only for phenology or five models calibrated in detail resulted in the uncertainty equivalent to that of the measured yield in well‐controlled agronomic field experiments. Sensitivity analysis indicates the necessity to improve the accuracy in predicting both biomass and harvest index in response to increasing [CO₂] and temperature.     

Expanded thermodynamic model for microbial true yield prediction

Thermodynamic methods to predict true yield and stoichiometry of bacterial reactions have been widely used in biotechnology and environmental engineering. However, yield predictions are often inaccurate for certain simple organic compounds. This work evaluates an existing method and identifies the cause of prediction errors for compounds with low degree of reductance of carbon. For these compounds, carbon, not energy or reducing equivalents, constrains growth. Existing thermodynamically-based models do not account for the potential of carbon-limited growth. The improved method described here consists of four balances: carbon balance, nitrogen balance, electron balance, and energy balance. Two efficiency terms, K1 and K2 are defined and estimated from a priori analysis. The results show that K1 and K2 are nearly the same in value so that only one coefficient, K = 0.41 is used in the modified model. Comparisons with observed yields show that use of the new model and parameters results in significantly improved yield estimation based on inclusion of the carbon balance. The average estimation error is less than 6% for the data set presented.     

Yield prediction and stoichiometry of multi-step biodegradation reactions involving oxygenation

Microorganisms can initiate the degradation of organic compounds by oxygenation reactions that require the investment of energy and electrons. This diversion of energy and electrons away from synthesis reactions leads to decreased overall cell yields. A thermodynamic method was developed that improves the accuracy of cell yield prediction for compounds degraded through pathways involving oxygenation reactions. This method predicts yields and stoichiometry for each step in the biodegradation pathway, thus enabling modeling a multi-step biodegradation process in which oxygenations occur and intermediates may persist. EDTA and benzene biodegradation are presented as examples. The method compares favorably with other yield prediction methods while providing additional information of yields for intermediates produced in the degradation pathway.     

A model for interphase chromosomes and evaluation of radiation-induced aberrations

We have developed a theoretical model for evaluating radiation-induced chromosomal exchanges by explicitly taking into account interphase (G(0)/G(1)) chromosome structure, nuclear organization of chromosomes, the production of double-strand breaks (DSBs), and the subsequent rejoinings in a faithful or unfaithful manner. Each of the 46 chromosomes for human lymphocytes (40 chromosomes for mouse lymphocytes) is modeled as a random polymer inside a spherical volume. The chromosome spheres are packed randomly inside a spherical nucleus with an allowed overlap controlled by a parameter Omega. The rejoining of DSBs is determined by a Monte Carlo procedure using a Gaussian proximity function with an interaction range parameter sigma. Values of Omega and sigma have been found which yield calculated results of interchromosomal aberration frequencies that agree with a wide range of experimental data. Our preferred solution is one with an interaction range of 0.5 microm coupled with a relatively small overlap parameter of 0.675 microm, which more or less confirms previous estimates. We have used our model with these parameter values and with resolution or detectability limits to calculate yields of translocations and dicentrics for human lymphocytes exposed to low-LET radiation that agree with experiments in the dose range 0.09 to 4 Gy. Five different experimental data sets have been compared with the theoretical results. Essentially all of the experimental data fall between theoretical curves corresponding to resolution limits of 1 Mbp and 20 Mbp, which may reflect the fact that different investigators use different limits for sensitivity or detectability. Translocation yields for mouse lymphocytes have also been calculated and are in good agreement with experimental data from 1 cGy to 10 cGy. There is also good agreement with recent data on complex aberrations. Our model is expected to be applicable to both low- and high-LET radiation, and we include a sample prediction of the yield of interchromosomal rejoining in the dose range 0.22 Gy to 2 Gy of 1000 MeV/nucleon iron particles. This dose range corresponds to average particle traversals per nucleus ranging from 1.0 to 9.12.     

Thermodynamic characterization of tandem mismatches found in naturally occurring RNA

Although all sequence symmetric tandem mismatches and some sequence asymmetric tandem mismatches have been thermodynamically characterized and a model has been proposed to predict the stability of previously unmeasured sequence asymmetric tandem mismatches [Christiansen,M.E. and Znosko,B.M. (2008) Biochemistry, 47, 4329–4336], experimental thermodynamic data for frequently occurring tandem mismatches is lacking. Since experimental data is preferred over a predictive model, the thermodynamic parameters for 25 frequently occurring tandem mismatches were determined. These new experimental values, on average, are 1.0 kcal/mol different from the values predicted for these mismatches using the previous model. The data for the sequence asymmetric tandem mismatches reported here were then combined with the data for 72 sequence asymmetric tandem mismatches that were published previously, and the parameters used to predict the thermodynamics of previously unmeasured sequence asymmetric tandem mismatches were updated. The average absolute difference between the measured values and the values predicted using these updated parameters is 0.5 kcal/mol. This updated model improves the prediction for tandem mismatches that were predicted rather poorly by the previous model. This new experimental data and updated predictive model allow for more accurate calculations of the free energy of RNA duplexes containing tandem mismatches, and, furthermore, should allow for improved prediction of secondary structure from sequence.     

Estimation of the carbon dioxide (CO2) fertilization effect using growth rate anomalies of CO2 and crop yields since 1961

The effect of elevated carbon dioxide (CO₂) on crop yields is one of the most uncertain and influential parameters in models used to assess climate change impacts and adaptations. A primary reason for this uncertainty is the limited availability of experimental data on CO₂ responses for crops grown under typical field conditions. However, because of historical variations in CO₂, each year farmers throughout the world perform uncontrolled yield ‘experiments’ under different levels of CO₂. In this study, measurements of atmospheric CO₂ growth rates and crop yields for individual countries since 1961 were compared to empirically determine the average effect of a 1 ppm increase of CO₂ on yields of rice, wheat, and maize. Because the gradual increase in CO₂ is highly correlated with major changes in technology, management, and other yield controlling factors, we focused on first differences of CO₂ and yield time series. Estimates of CO₂ responses obtained from this approach were highly uncertain, reflecting the relatively small importance of year‐to‐year CO₂ changes for yield variability. Combining estimates from the top 20 countries for each crop resulted in estimates with substantially less uncertainty than from any individual country. The results indicate that while current datasets cannot reliably constrain estimates beyond previous experimental studies, an empirical approach supported by large amounts of data may provide a potentially valuable and independent assessment of this critical model parameter. For example, analysis of reliable yield records from hundreds of individual, independent locations (as opposed to national scale yield records with poorly defined errors) may result in empirical estimates with useful levels of uncertainty to complement estimates from experimental studies.     

Estimates of regional work in the canine left ventricle

Assessment of the magnitude of regional myocardial work requires knowledge of regional fiber stress and fiber shortening. The theoretical development and experimental validation of a method is presented which used values of estimated active and passive fiber stress according to a fluid-fiber model, and measured fiber strain values. This enables the construction of regional stress-strain diagrams, a regional analog of the pressure-volume area model by Suga and co-investigators, which can be linked to regional oxygen consumption. In the left ventricle, either normally or asynchronously activated, the method yields reliable data on strain and active and passive fiber stress. The relation between estimated regional work and myocardial oxygen demand is in quantitative agreement with previously reported relations between global oxygen demand and measured pressure-volume area. During coronary artery occlusion, however, these values were less reliable, which might be due to inaqdequate knowledge of the (passive) material properties of the myocardium.     

Temporal robustness of linear relationships between production and transpiration

Seasonal dependence of biomass production on transpiration has been previously reported for a number of crops under salinity and drought. Linear yield (Y) to transpiration (T) relationships have been utilized in plant-growth and water-uptake models to estimate yield based on predicted transpiration values. The relationship is often employed for time steps that are very small compared with the whole season measurements, even though no empirical validation exists for such application. This work tests the hypothesis that linear Y-T relationships are valid throughout the life span of crops under varied natural conditions and levels of environmental stress. Effects of salinity and water supply on growth, water use and yields of tomatoes (Lycopersicon esculentum Mill.) were studied for two distinct conditions of potential transpiration. Linear relationships between relative Y and relative ET were found to be consistent throughout the life span of the crops for both growing seasons. Water-use efficiency increased together with plant growth as a result of changes in the plant's surface area to volume ratio. This empirical validation of linear Y-T relationships for short time periods is beneficial in confirming their usefulness in growth and water uptake models.     

A method for the determination of confidence limits for the P/2e − ratio for chosen values of Y ATP max from the results of continuous culture experiments

A model is described, which allows the determination of 95% confidence limits for the maintenance coefficient and the efficiency of oxidative phosphorylation for chosen values of the growth yield for ATP corrected for energy maintenance (Y _ATP ^max ). As experimental data the specific rates of substrate consumption, product formation and oxygen uptake in chemostat cultures at various growth rates are used.     

Likelihood based observability analysis and confidence intervals for predictions of dynamic models

ABSTRACT: BACKGROUND: Predicting a system's behavior based on a mathematical model is a primary task in Systems Biology. If the model parameters are estimated from experimental data, the parameter uncertainty has to be translated into confidence intervals for model predictions. For dynamic models of biochemical networks, the nonlinearity in combination with the large number of parameters hampers the calculation of prediction confidence intervals and renders classical approaches as hardly feasible. RESULTS: In this article reliable confidence intervals are calculated based on the prediction profile likelihood. Such prediction confidence intervals of the dynamic states can be utilized for a data-based observability analysis. The method is also applicable if there are non-identifiable parameters yielding to some insufficiently specified modelpredictions that can be interpreted as non-observability. Moreover, a validation profile likelihood is introduced that should be applied when noisy validation experiments are to be interpreted. CONCLUSIONS: The presented methodology allows the propagation of uncertainty from experimental to model pre-dictions. Although presented in the context of ordinary differential equations, the concept is general and also applicable to other types of models. Matlab code which can be used as a template to implement the method is provided at http://www.fdmold.uni-freiburg.de/~ckreutz/PPL .     

Modeling soil water regime and corn yields considering climatic uncertainty

Real time estimation of soil moisture and crop yield plays an important role for best irrigation management practices especially in arid and semiarid regions. A simulation model able of real time estimating and forecasting soil water storage and corn yield response to soil moisture was developed by combining two existing models. Soil water storage was estimated through the soil water balance equation considering the uncertainty of evapotranspiration and combing with Kalman filter technique. Crop dry matter and grain yield were simulated by using a functional relationship between yield and soil moisture. Some improvements have been made in the response function by considering different impacts of moisture stress on crop growth and yield for the different growing stages. Four years data sets collected in an experimental station in the North China Plain were used to calibrate and test the model. Results indicate that soil moisture storage in the soil profile estimated and predicted by the model agrees well with the measured data, and the relative error of yield prediction is around 10%, which means that the combined model and the methodology applied are capable of predicting crop yield and soil water storage dynamics.     

Experimental validation of in silico estimated biomass yields of Pseudomonas putida KT2440

Pseudomonas putida is rapidly becoming a microbial cell platform for biotechnological applications. In order to understand genotype‐phenotype relationships genome scale models represent helpful tools. However, the validation of in silico predictions of genome scale models is a task that is rarely performed. In this study the theoretical biomass yields of Pseudomonas putida KT2440 were estimated for 57 different carbon sources based on a genome scale stoichiometric model applying flux balance analysis. The batch growth of P. putida KT2440 with six individual carbon sources covering the range of maximal to minimal in silico biomass yields (acetate, glycerol, citrate, succinate, malate and methanol, respectively) was studied in a defined mineral medium in a fully controlled stirred‐tank bioreactor on a 3 L scale. The highest growth rate of P. putida KT2440 was measured with succinate as carbon source (0.51 h^?1). Among the 57 carbon sources tested, glycerol resulted in the highest estimated biomass yield (0.61 molC_Biomass molC^?1_Glycerol) which was experimentally confirmed. The comparison of experimental determined biomass yields with a modified version of the model iJP815 showed deviations of only up to 10%. The experimental data generated in this study can also be used in future studies to further improve the genome scale models of P. putida KT2440. Improved models will then help to gain deeper insights in genotype‐phenotype relationships.     

Spline modelling electron insert factors using routine measurements

There are many methods available to predict electron output factors; however, many centres still measure the factors for each irregular electron field. Creating an electron output factor prediction model that approaches measurement accuracy – but uses already available data and is simple to implement – would be advantageous in the clinical setting. This work presents an empirical spline model for output factor prediction that requires only the measured factors for arbitrary insert shapes. Equivalent ellipses of the insert shapes are determined and then parameterised by width and ratio of perimeter to area. This takes into account changes in lateral scatter, bremsstrahlung produced in the insert material, and scatter from the edge of the insert. Agreement between prediction and measurement for the 12 MeV validation data had an uncertainty of 0.4% (1SD). The maximum recorded deviation between measurement and prediction over the range of energies was 1.0%. The validation methodology showed that one may expect an approximate uncertainty of 0.5% (1SD) when as little as eight data points are used. The level of accuracy combined with the ease with which this model can be generated demonstrates its suitability for clinical use. Implementation of this method is freely available for download at https://github.com/SimonBiggs/electronfactors.     

Remotely Sensed Rice Yield Prediction Using Multi-Temporal NDVI Data Derived from NOAA's-AVHRR

Grain-yield prediction using remotely sensed data have been intensively studied in wheat and maize, but such information is limited in rice, barley, oats and soybeans. The present study proposes a new framework for rice-yield prediction, which eliminates the influence of the technology development, fertilizer application, and management improvement and can be used for the development and implementation of provincial rice-yield predictions. The technique requires the collection of remotely sensed data over an adequate time frame and a corresponding record of the region''s crop yields. Longer normalized-difference-vegetation-index (NDVI) time series are preferable to shorter ones for the purposes of rice-yield prediction because the well-contrasted seasons in a longer time series provide the opportunity to build regression models with a wide application range. A regression analysis of the yield versus the year indicated an annual gain in the rice yield of 50 to 128 kg ha⁻¹. Stepwise regression models for the remotely sensed rice-yield predictions have been developed for five typical rice-growing provinces in China. The prediction models for the remotely sensed rice yield indicated that the influences of the NDVIs on the rice yield were always positive. The association between the predicted and observed rice yields was highly significant without obvious outliers from 1982 to 2004. Independent validation found that the overall relative error is approximately 5.82%, and a majority of the relative errors were less than 5% in 2005 and 2006, depending on the study area. The proposed models can be used in an operational context to predict rice yields at the provincial level in China. The methodologies described in the present paper can be applied to any crop for which a sufficient time series of NDVI data and the corresponding historical yield information are available, as long as the historical yield increases significantly.     

Global evaluation of a semiempirical model for yield anomalies and application to within‐season yield forecasting

Quantifying the influence of weather on yield variability is decisive for agricultural management under current and future climate anomalies. We extended an existing semiempirical modeling scheme that allows for such quantification. Yield anomalies, measured as interannual differences, were modeled for maize, soybeans, and wheat in the United States and 32 other main producer countries. We used two yield data sets, one derived from reported yields and the other from a global yield data set deduced from remote sensing. We assessed the capacity of the model to forecast yields within the growing season. In the United States, our model can explain at least two‐thirds (63%–81%) of observed yield anomalies. Its out‐of‐sample performance (34%–55%) suggests a robust yield projection capacity when applied to unknown weather. Out‐of‐sample performance is lower when using remote sensing‐derived yield data. The share of weather‐driven yield fluctuation varies spatially, and estimated coefficients agree with expectations. Globally, the explained variance in yield anomalies based on the remote sensing data set is similar to the United States (71%–84%). But the out‐of‐sample performance is lower (15%–42%). The performance discrepancy is likely due to shortcomings of the remote sensing yield data as it diminishes when using reported yield anomalies instead. Our model allows for robust forecasting of yields up to 2 months before harvest for several main producer countries. An additional experiment suggests moderate yield losses under mean warming, assuming no major changes in temperature extremes. We conclude that our model can detect weather influences on yield anomalies and project yields with unknown weather. It requires only monthly input data and has a low computational demand. Its within‐season yield forecasting capacity provides a basis for practical applications like local adaptation planning. Our study underlines high‐quality yield monitoring and statistics as critical prerequisites to guide adaptation under climate change.     

Uncertainty in techno-economic estimates of cellulosic ethanol production due to experimental measurement uncertainty

ABSTRACT: BACKGROUND: Cost-effective production of lignocellulosic biofuels remains a major financial and technical challenge at the industrial scale. A critical tool in biofuels process development is the techno-economic (TE) model, which calculates biofuel production costs using a process model and an economic model. The process model solves mass and energy balances for each unit, and the economic model estimates capital and operating costs from the process model based on economic assumptions. The process model inputs include experimental data on the feedstock composition and intermediate product yields for each unit. These experimental yield data are calculated from primary measurements. Uncertainty in these primary measurements is propagated to the calculated yields, to the process model, and ultimately to the economic model. Thus, outputs of the TE model have a minimum uncertainty associated with the uncertainty in the primary measurements. RESULTS: We calculate the uncertainty in the Minimum Ethanol Selling Price (MESP) estimate for lignocellulosic ethanol production via a biochemical conversion process: dilute sulfuric acid pretreatment of corn stover followed by enzymatic hydrolysis and co-fermentation of the resulting sugars to ethanol. We perform a sensitivity analysis on the TE model and identify the feedstock composition and conversion yields from three unit operations (xylose from pretreatment, glucose from enzymatic hydrolysis, and ethanol from fermentation) as the most important variables. The uncertainty in the pretreatment xylose yield arises from multiple measurements, whereas the glucose and ethanol yields from enzymatic hydrolysis and fermentation, respectively, are dominated by a single measurement: the fraction of insoluble solids (fIS) in the biomass slurries. CONCLUSIONS: We calculate a $0.15/gal uncertainty in MESP from the TE model due to uncertainties in primary measurements. This result sets a lower bound on the error bars of the TE model predictions. This analysis highlights the primary measurements that merit further development to reduce the uncertainty associated with their use in TE models. While we develop and apply this mathematical framework to a specific biorefinery scenario here, this analysis can be readily adapted to other types of biorefining processes and provides a general framework for propagating uncertainty due to analytical measurements through a TE model.     

Prediction of first test day milk yield using historical records in dairy cows

The transition between two lactations remains one of the most critical periods during the productive life of dairy cows. In this study, we aimed to develop a model that predicts the milk yield of dairy cows from test day milk yield data collected in the previous lactation. In the past, data routinely collected in the context of herd improvement programmes on dairy farms have been used to provide insights in the health status of animals or for genetic evaluations. Typically, only data from the current lactation is used, comparing expected (i.e., unperturbed) with realised milk yields. This approach cannot be used to monitor the transition period due to the lack of unperturbed milk yields at the start of a lactation. For multiparous cows, an opportunity lies in the use of data from the previous lactation to predict the expected production of the next one. We developed a methodology to predict the first test day milk yield after calving using information from the previous lactation. To this end, three random forest models (nextMILK_FULL, nextMILK_PH, and nextMILK_P) were trained with three different feature sets to forecast the milk yield on the first test day of the next lactation. To evaluate the added value of using a machine-learning approach against simple models based on contemporary animals or production in the previous lactation, we compared the nextMILK models with four benchmark models. The nextMILK models had an RMSE ranging from 6.08 to 6.24 kg of milk. In conclusion, the nextMILK models had a better prediction performance compared to the benchmark models. Application-wise, the proposed methodology could be part of a monitoring tool tailored towards the transition period. Future research should focus on validation of the developed methodology within such tool.     

Accuracy of genomic selection in European maize elite breeding populations

Genomic selection is a promising breeding strategy for rapid improvement of complex traits. The objective of our study was to investigate the prediction accuracy of genomic breeding values through cross validation. The study was based on experimental data of six segregating populations from a half-diallel mating design with 788 testcross progenies from an elite maize breeding program. The plants were intensively phenotyped in multi-location field trials and fingerprinted with 960 SNP markers. We used random regression best linear unbiased prediction in combination with fivefold cross validation. The prediction accuracy across populations was higher for grain moisture (0.90) than for grain yield (0.58). The accuracy of genomic selection realized for grain yield corresponds to the precision of phenotyping at unreplicated field trials in 3–4 locations. As for maize up to three generations are feasible per year, selection gain per unit time is high and, consequently, genomic selection holds great promise for maize breeding programs.     

Fitting and validating the genomic evaluation model to Polish Holstein-Friesian cattle

The aim of the study was to fit the genomic evaluation model to Polish Holstein-Friesian dairy cattle. A training data set for the estimation of additive effects of single nucleotide polymorphisms (SNPs) consisted of 1227 Polish Holstein-Friesian bulls. Genotypes were obtained by the use of Illumina BovineSNP50 Genotyping BeadChip. Altogether 29 traits were considered: milk-, fat- and protein- yields, somatic cell score, four female fertility traits, and 21 traits describing conformation. The prediction of direct genomic values was based on a mixed model containing deregressed national proofs as a dependent variable and random SNP effects as independent variables. The correlations between direct genomic values and conventional estimated breeding values estimated for the whole data set were overall very high and varied between 0.98 for production traits and 0.78 for non return rates for cows. For the validation data set of 232 bulls the corresponding correlations were 0.38 for milk-, 0.37 for protein-, and 0.32 for fat yields, while the correlations between genomic enhanced breeding values and conventional estimated breeding values for the four traits were: 0.43, 0.44, 0.31, and 0.35. This model was able to pass the interbull validation criteria for genomic selection, which indicates that it is realistic to implement genomic selection in Polish Holstein-Friesian cattle.     

Field-scale rice yield prediction from Sentinel-2 monthly image composites using machine learning algorithms

Machine learning (ML) along with high volume of satellite images offers an alternative to agronomists in crop yield predictions for decision support systems. This research exploited the possibility of using monthly image composites from Sentinel-2 imageries for rice crop yield predictions one month before the harvesting period at the field level using ML techniques in Taiwan. Three ML models, including random forest (RF), support vector machine (SVM), and artificial neural networks (ANN), were designed to address the research question of yield predictions in four consecutive growing seasons from 2019 to 2020 using field survey data. The research findings of yield modeling and predictions showed that SVM slightly outperformed RF and ANN. The results of model validation, obtained from SVM models using the data from transplanting to ripening, showed that the root mean square percentage error (RMSPE) and the mean absolute percentage error (MAPE) values were 5.5% and 4.5% for the 2019 second crop, and 4.7% and 3.5% for the 2020 first crop, respectively. The results of yield predictions (obtained from SVM) for the 2019 second crop and the 2020 first crop evaluated against the government statistics indicated a close agreement between these two datasets, with the RMSPE and MAPE values generally smaller than 11.2% and 9.2%. The SVM model configuration parameters used for rice crop yield predictions indicated satisfactory results. The comparison results between the predicted yields and the official statistics showed slight underestimations, with RMSPE and MAPE values of 9.4% and 7.1% for the 2019 first crop (hindcast), and 11.0% and 9.4% for the 2020 second crop (forecast), respectively. This study has successfully proven the validity of our methods for yield modeling and prediction from monthly composites from Sentinel-2 imageries using ML algorithms. The research findings from this research work could useful for agronomists to timely formulate action plans to address national food security issues.