Climate change and public policies restricting freshwater use for agricultural irrigation are compelling farmers to maintain production with limited water. Advanced irrigation scheduling tools that combine data and computer simulations are needed to optimize water use and maximize crop productivity. Limited studies have evaluated the performance of data assimilation and model-based simulation optimization of irrigation scheduling under field conditions. The objective of this study was to evaluate model-based irrigation scheduling with and without assimilation of LAI in processing tomatoes. The treatments included two DSSAT CropGro-Tomato models. The treatments were T1 (TM0005 with LAI data assimilation), T2 (TM0030 with LAI data assimilation), and T3 (Control: TM0005 without LAI data assimilation). This study was conducted near Davis, California. Model performance was evaluated using applied water, soil water content, growth, yield, and fruit quality. Results showed no significant yield differences between treatments that assimilated LAI and the control. All the models accurately predicted LAI and yield within one standard deviation of measured values, suggesting that model-based optimization was effective with or without data assimilation. The framework reduced applied water by 26% compared to current irrigation recommendations for processing tomatoes. The average applied irrigation was 391 mm, compared to the recommended (533 to 762) mm. No significant differences in fruit quality were observed between the treatments. Overall, the model-based simulation-optimization irrigation scheduling approach maintained the desired yield and fruit quality while reducing water use. A well-calibrated crop model did not benefit from LAI data assimilation, implying that model-based irrigation scheduling could be easily implemented without the need for monitoring and additional computation costs of assimilating LAI during the season which also include labor and instrumentation costs. The model-based irrigation scheduling framework proposed in this study could be applied to other crops to help growers cope with limited water supplies.

The increasing availability of remote sensing (RS) data and advancements in data assimilation (DA) techniques facilitate the non-destructive calibration of mechanistic crop models but necessitate a framework that digitally represents cropping systems and their spectral properties. This study implemented a coupling scheme linking the outputs of a crop model (DSSAT-CROPGRO) with a radiative transfer model (RTMo module in SCOPE). Reflectance data acquired from a multispectral camera mounted on a UAV were assimilated into the coupled model. The DA scheme was tested in an irrigation and fertilization trial with processing tomatoes, a row crop, requiring the adjustment of the model in order to reflect the vegetation and soil pixel proportions. Examining the relative contribution of dynamically updating specific RTMo parameters showed that the coupled model performed better when parameters were adjusted than when using their nominal values. Applying the DA scheme improved the normalized root mean square error (NRMSE) of the Leaf Area Index (LAI) from 59% to 42% and yield from 64% to 35%. The best performance was achieved when the most water-stressed treatment was excluded, resulting in NRMSE of 34% for LAI and 16% for yield. Since the DA scheme presented here performed well at low to moderate water stress, it should be further tested in assimilating space-borne RS data into simulations of large-scale, commercial fields.

Crop models can serve as decision-support tools, but their uncertainty must be accounted for. While previous research has shown effective calibration of crop models using remote sensing (RS) data, the remaining uncertainty is rarely quantified. This study investigated the propagation of errors associated with RS data in a coupled crop-radiative transfer model in two steps. First, the results of a Particle Filter (PF) process were examined to assess the uncertainty of the model parameters and outputs. Next, the Winding Stairs (WS) method was used to quantify the contribution of crop model parameters uncertainty to the total model uncertainty. The results show that parameters related to crop growth rate contribute more to the variance of simulated Leaf Area Index (LAI) and yield than the phenology-related parameters. These findings can guide future research to improve the model reliability by focusing on calibrating the parameters with a higher impact on model outcome uncertainty.

Accurate monitoring of nitrogen (N) levels, while accounting for spatiotemporal variability is crucial for optimizing fertilization in citrus orchards. Traditional methods, such as frequent leaf and soil sampling followed by laboratory analysis, are costly, labor-intensive, and prone to human error. Remote sensing (RS) technologies, including unmanned aerial vehicles (UAVs) and satellite platforms, offer scalable and precise alternatives for N management. However, integrating these platforms poses challenges due to significant differences in spatial, temporal, and spectral resolution. This study presents a novel approach incorporating multispectral and temporal data from UAVs and Sentinel-2 satellites to estimate canopy N content (CNC) in citrus orchards. This method captures spatiotemporal variability across multiple citrus cultivars, aiming to enhance nitrogen use efficiency (NUE) while reducing environmental impact, ultimately promoting sustainable orchard management practices. The study was conducted in commercial citrus plots in the Hefer Valley, Israel, and spanned two phases. The first phase (May 2019 to April 2022) focused on four plots of the 'Newhall' cultivar, while the second phase expanded to twelve additional plots featuring five different citrus cultivars. The methodology consisted of six key steps: (1) Leaf samples from the study area were collected for laboratory nitrogen (N) analysis. (2) Acquiring and preprocessing bimonthly UAV multispectral images and Sentinel-2 satellite images to ensure data quality and consistency. (3) Segmenting individual trees using UAV imagery and extracting structural features through Structure-from-Motion (SfM) photogrammetry. (4) Processing images and extracting spectral and structural features relevant to N estimation. (5) Developing Random Forest (RF) models to estimate CNC using UAV-derived vegetation indices (VIs) and SfM data and combining these with Sentinel-2 VIs to generate canopy-scale CNC heatmaps. (6) Analyzing the relationship between CNC and yield to understand nitrogen dynamics and their impact on productivity. The integrated RF model, which combined UAV-VIs, Sentinel-2 VIs, and SfM-derived structural data, achieved superior performance (R² = 0.80, RMSE = 0.17 kg/m²) compared to models relying solely on UAV-VIs (R² = 0.68, RMSE = 0.23 kg/m²) or Sentinel-2 VIs (R² = 0.48, RMSE = 0.30 kg/m²). Additionally, CNC expressed as mass per tree demonstrated a strong positive correlation with yield (R² = 0.66), highlighting the relationship between nitrogen dynamics and orchard productivity. These results underscore the robustness of the integrated model and the clear advantage of multi-platform data fusion over single-source approaches. The study provides compelling evidence for the potential of combining UAV and Sentinel-2 data to improve CNC estimation and its correlation with yield in citrus orchards. The findings contribute to advancements in precision agriculture by offering a scalable, data-driven framework to enhance nutrient management and support sustainable orchard practices.

Multi-year planning of allocation of agricultural land and irrigation water remains a major challenge, which is exacerbated by decreasing arable land and increasing water scarcity in many regions. This paper presents a model-based framework to address this challenge. One of the key elements of the proposed framework is that it takes into account explicitly the need to rotate crops according to some agronomically-based sequences. The framework consists of three nested optimizations: Innermost: Optimize water allocation assuming pre-divided fields and pre-determined crop rotations. Middle: Optimize crop rotation sequences within each field. Outermost: Optimize fields geometry to maximize net income. These computations leverage crop- and soil-specific 'Yield value vs. Irrigation' functions derived from an auxiliary multi-objective optimization problem, namely maximizing yield and minimizing water use. In this manner, all the planning is based on the knowledge contained in complex crop growth models (rather than simplistic models), without having to actually run these models a prohibitively high number of times. The procedure is illustrated on two 67 ha areas near Davis, CA, that altogether contained seven types of soil. Planning was performed for a 10-year planning horizon, assuming seven crops (&fallowing) and six crop rotation patterns were available to choose from. Several scenarios that differed in terms of water availability are presented. The results demonstrate the strong impact that crop rotation requirements have on the overall performance.

The increasing availability of remote sensing (RS) data and the advancement of computation abilities, combined with the demands for enhancing crop production, encourages the creation of a framework in which crop growth simulation can be updated sequentially to serve as a yield predictor and be part of a decision support system. However, crop model outputs and RS data must be linked via a radiative transfer model (RTM), which simulates the interaction between the crop and the intercepted radiation. In this study, a comprehensive coupling scheme between a crop model (DSSAT-CROPGRO-tomato) and an RTM (SCOPE-RTMo) was formulated and investigated through global sensitivity analysis (SA) and by testing the coupled model in a synthetic data assimilation (DA) experiment. The DA experiment utilized a sensitivity-based particle filter (PF) in which the SA results were used to enhance the PF convergence rate and accuracy. The SA results provide the sensitivity of simulated reflectance at different wavelengths to DSSAT-CROPGRO parameters throughout the season. This information can help guide future data assimilation experiments by choosing imaging instruments with appropriate spectral bands and timing the measurements to enhance model calibration. The results of the synthetic DA experiment showed a good convergence of the particle filter towards the ground truth. The results also demonstrated the strong relation between LAI and reflectance, as several model runs with different initial values of DSSAT-CROPGRO parameters all converged and predicted the synthetic LAI observations very well. The convergence of DSSAT-CROPGRO parameters to their ground truth values was only partial, and phenology-related parameters tended to converge better than growth-related parameters.

Irrigation forecasting is essential for improving water management in agriculture. This study proposed a novel and practical framework for irrigation forecasting for paddy rice. Public weather forecasts in China were selected to forecast ETo and quantitative rainfall. For the latter, a simple deterministic model and a probabilistic model based on a Pearson-III representation of the data were considered. A modified Python version of the AquaCrop model (ACOP-Rice model) was adopted to generate the probability distribution of forecasted irrigation events and rule-based irrigation recommendations. The performance of weekly irrigation forecasting was then evaluated. The analysis of the weather forecasts revealed that the accuracy of the temperature and ETo predictions was acceptable, whereas the quality of the rainfall forecasts was poor. The performance of the scenarios that using probabilistic rainfall forecasts outperformed scenarios that relied on deterministic rainfall forecasts, despite the lower quality of the probabilistic rainfall forecasts compared to the deterministic forecasts. The irrigation recommendations exhibited a systematic bias (i.e., the inclination toward overirrigation or underirrigation) when using weather forecasts, and this bias was lower when using probabilistic forecasts. Additionally, compared to ETo forecasts, the inaccuracy of rainfall forecasts had a much greater impact on irrigation forecasts.

Sustainability in our food and fiber agriculture systems is inherently knowledge intensive. It is more likely to be achieved by using all the knowledge, technology, and resources available, including data-driven agricultural technology and precision agriculture methods, than by relying entirely on human powers of observation, analysis, and memory following practical experience. Data collected by sensors and digested by artificial intelligence (AI) can help farmers learn about synergies between the domains of natural systems that are key to simultaneously achieve sustainability and food security. In the quest for agricultural sustainability, some high-payoff research areas are suggested to resolve critical legal and technical barriers as well as economic and social constraints. These include: the development of holistic decision-making systems, automated animal intake measurement, low-cost environmental sensors, robot obstacle avoidance, integrating remote sensing with crop and pasture models, extension methods for data-driven agriculture, methods for exploiting naturally occurring Genotype x Environment x Management experiments, innovation in business models for data sharing and data regulation reinforcing trust. Public funding for research is needed in several critical areas identified in this paper to enable sustainable agriculture and innovation.

Fire blight disease causes significant losses in pear orchards. Fire blight infection is accompanied by visual symptoms that can easily be recognized by a grower or adviser, but such visual inspection is time consuming. The present work focused on the use of convolutional neural networks (CNNs) to identify one type of visual symptom (cankers on the main trunk of dormant trees) as well as autumn blooming, which in Israel plays an important role in the epidemiology of Erwinia amylovora—the bacterium responsible for fire blight. Images of dormant trees were acquired with a tripod-mounted DSLR camera while, for autumn blooming detection, the images were acquired using a small unmanned aerial vehicle flying a few meters above the trees. In both cases, several Faster R-CNNs were trained and tested with several datasets acquired at various locations and over several years. Overall, the CNNs for canker detection achieved precision and recall rates that exceeded 90% while, for autumn blooming detection, the precision and recall rates exceeded 80% in all but one case. These trained CNNs were used to analyze automatically geo-references images, hence generating infection/blooming maps. Such maps could be one of the information layers used by growers for managing the orchard, for instance to determine whether winter sanitation is needed and/or if it was carried out properly, or to decide when and where costly manual removal of autumn flowers or copper application is required.

Crop simulation models are essential tools in supporting sustainable agricultural management. However, due to uncertainty in model parameters, the model predictions may not be sufficiently accurate. Data assimilation (DA) is a common approach to improve dynamic crop modeling by combining it with observation data. Among DA approaches, particle filter (PF) is a popular choice. Since conventional PF (CPF) may suffer from sample impoverishment problems, various filter modifications have been proposed in the literature, including the application of genetic operators (arithmetic cross-over and mutations). In this study, a novel PF approach inspired by the gene's recombination process has been developed. In this new filter, named “recombination” PF (RPF), particle diversity is increased via information exchange between surviving particles and intermediate particles which are located close to existing particles. In turn, increased particle diversity reduces the chances of sample impoverishment and thus improves filter performance. The proposed method was tested on two synthetic study cases using the open-source AquaCrop model (v5.0a) and assuming weekly observations of canopy cover and soil water content. When CPF was implemented, the overall average normalized root mean square error (NRMSE), combining state and parameter estimations, and yield forecasts, all performed throughout the season, ranged from 4.0 to 5.1 % for ensemble sizes ranging from 150 to 500 particles. When RPF was implemented with a similar number of particles, the overall average NRMSE decreased to 3.6–3.7 %, corresponding to a 7–26 % improvement. Furthermore, higher stability of the results was observed, and the final parameter estimations improved in all the ensemble sizes investigated by approximately 40 %, which would be very beneficial for predicting crop growth in the next season.

Estimating crop nitrogen status to optimize production and minimize environmental pollution is a major challenge for modern agriculture. The study objective was to develop a multivariate spatiotemporal dynamic clustering approach to generate Nitrogen (N) Management Zones (MZs) in a citrus orchard during the growing season. The research was conducted in four citrus plots in the coastal area of Israel. Five variables were selected to characterize each plot’s spatiotemporal variability of canopy N content. These were split into constant (i.e., elevation, northness, and slope) and non-constant (i.e., canopy N content and tree height) variables. The non-constant data were obtained via bi-monthly imaging campaigns with a multispectral camera mounted on an unmanned aerial vehicle (UAV) throughout the growing season of 2019. The selected variables were then standardized to define the clusters by applying the Getis-Ord Gi* z-score. These were used to develop a spatiotemporal dynamic clustering model using Fuzzy C-means (FCM). Four input variables were investigated in this final stage, including the constant variables only and different combinations of constant and non-constant variables. The support vector machine regression model results for estimating canopy N-content from multispectral images were R² = 0.771 and RMSE = 0.227. This model was used to predict monthly canopy-level N content and classify the N content levels based on the October N-to-yield content envelope curve. Delineating MZs was followed by the comparison of spatial association among cluster maps. This process may support site-specific and time-specific nitrogen management.

Due to climate change, increased regulation of water resources and competition from other beneficial uses, the agricultural sector is under pressure to use water more efficiently. This paper reports the field evaluation of two model-based simulation-optimization approaches for irrigation scheduling: deterministic optimization and stochastic optimization. The field experiments were conducted at the University of California Davis research farm with a processing tomato crop. The crop growth simulation model used in the study was DSSAT-CROPGRO processing tomato. In order to mitigate the impact of weather forecasts inaccuracies, irrigation schedules were updated every 5 to 10 days, depending on operational constraints. These updates were performed via custom graphical user interfaces that enabled the user to visualize the expected outcomes of various decisions or scenarios before choosing which irrigation schedule to implement. An irrigation treatment based on manual monitoring of soil water content with a field-calibrated neutron probe served as benchmark. The model-based treatments achieved yields and water use efficiencies that were not significantly different from those obtained in the neutron probe-based treatment. These results demonstrate the high potential of model-based simulation-optimization approaches for in-season adaptive irrigation scheduling, especially since neutron probes, which are considered one of the most accurate indirect methods of measuring soil water content, are not commonly used by growers.

Crop growth simulation models are important components in agricultural management. Such models include parameters that should be calibrated locally, which can be achieved by assimilating observations using a particle filter. The number of model parameters is usually high while the number of observations is limited and adjusting parameters that are non-influential under the specific environmental conditions and growth stages should be avoided. This study suggests a novel particle filter-based framework in which sensitivity analysis (SA) is embedded in the filter so that at each data assimilation step only a subset of influential parameters is adjusted. The proposed framework was implemented in two synthetic study cases with the open-source AquaCrop model (v5.0a), assuming weekly observations of canopy cover and soil water content, and in some cases biomass. In the first case study, the adjustment of a subset of influential parameters identified by SA was compared with the adjustment of all candidate parameters, the impact of the SA screening threshold and the variance of the parameter perturbation included in the filter were investigated, and the addition of biomass measurements was examined. In the second case study, different irrigation treatments were implemented to demonstrate the impact of the growing conditions on the subset of parameters that could be calibrated. The performance of the proposed framework was evaluated by computing the root mean square error (RMSE) and normalized RMSE (NRMSE) of the states and parameters estimations, and of the final biomass and yield forecasts. Overall, all state predictions were very accurate. Estimation of the model parameters was more challenging and not all the parameters converged toward their true values. This result is not surprising considering the low sensitivity of certain parameters and correlations that exist within the model. Nonetheless, and more importantly, after the assimilation of relatively few observations, the model was able to forecast final biomass and yield quite accurately. Compared to standard data assimilation in which all parameters were adjusted, the SA-embedded filter performed better according to all the indicators considered:NRMSE of canopy cover and soil water content decreased from 2.4% to 1.3% and from 3.2% to 2.4%, respectively, NRMSE of parameter estimations decreased from 13.5% to 11.4%, and NRMSE of forecasted yield decreased from 5.9% to 5.2%. Overall, a relative improvement of 16% was obtained in the average NRMSE. Assimilation of additional biomass measurements improved only the ability of the model to forecast biomass but had no positive impact on yield forecasting. The next step should be to test such an approach thoroughly with experimental data.

Crop models can be combined with optimization algorithms in order to develop management tools. However, such model-based tools are inherently affected by the imperfectness of the model on which they are based. In this paper we describe a procedure in which data assimilation and partial re-parametrization of the model embedded in the optimization procedure used to determine irrigation scheduling are performed before each optimization run. Furthermore, sensitivity analysis is performed before performing data assimilation, which ensures that only influential parameters are adjusted. The procedure was tested via simulation with DSSAT-CROPGRO for a hypothetical processing tomato crop in Davis, CA, in 2010–2019. Several scenarios that differed in terms of the measurements assumed to be available for assimilation (leaf area index, biomass and/or soil water content) were simulated. The results were compared to a benchmark scenario involving a perfect model, as well as a scenario in which data assimilation was not performed. The analysis focused on the overall performance of the irrigation schedule (yield vs. irrigation amount) derived using the model rather on the accuracy of the estimated model parameters. Assimilating weekly measurements of leaf area index led to overall performance that was within 3% of the benchmark performance. Adding weekly measurements of biomass or daily measurements of soil water content did not improve the performance. On the other hand, assimilating only daily soil water content measurements led to poorer results (5% decrease compared to benchmark) and also affected the repeatability of the results. Defining dynamically the subset of parameters for calibration via sensitivity analysis rather than calibrating a fixed subset of parameters or all parameters was beneficial, both in terms of overall performance and repeatability of the results. Overall, concurrent data assimilation and model-based optimization has potential to enhance irrigation scheduling decision making, particularly in water limited environments.

Real-time in situ measurements are increasingly being used to improve the estimations of simulation models via data assimilation techniques such as particle filter. However, models that describe complex processes such as water flow contain a large number of parameters while the data available are typically very limited. In such situations, applying particle filter to a large, fixed set of parameters chosen a priori can lead to unstable behavior, i.e., inconsistent adjustment of some of the parameters that have only limited impact on the states that are being measured. To prevent this, in this study correlation-based variable selection is embedded in the particle filter, so that at each step only a subset of the most influential parameters is adjusted. The particle filter used in this study includes genetic algorithm operators and Monte Carlo Markov Chain for alleviating filter degeneracy and sample impoverishment. The proposed method was applied to a water flow model (Hydrus-1D) in which soil water content at various depths and soil hydraulic parameters were updated. Two case studies are presented. Overall, the proposed method yielded parameters and states estimates that were more accurate and more consistent than those obtained when adjusting all the parameters. Furthermore, the results show that the higher the influence of a parameter on the model output under the current conditions, the better the estimation of this parameter is.

Current rice production in China is associated with low rainfall use efficiency. In order to increase rainfall use efficiency and develop simple water-saving irrigation modes that could be readily implemented by farmers, a new multi-objective optimization framework for irrigation modes of paddy rice was developed, based on a modified version of the AquaCrop model called ACOP-Rice model. The optimization focused on the water level at which irrigation was triggered for five growth periods and the irrigation frequency, rainfall use efficiency and yield were optimized. The procedure was tested on nine rice production cases in China for which over 60 years of historical meteorological data and irrigation guidelines were available. Analysis of the weather data showed that rainfall distribution varied greatly between the different locations and growth periods. The results obtained by following the current guidelines were compared to three optimal solutions that corresponded to minimum number of irrigation events, maximum rainfall use efficiency and “balanced” performance in which equal attention was given to rainfall use efficiency and irrigation frequency, respectively. Overall, the optimization led to lowering the water depth at which irrigation was triggered. The optimal water level after irrigation varied between the different cases, depending on the combined effects of rainfall distribution, operation constraints and length of growth period. Compared to the current guidelines, the optimized irrigation modes reduced the proportion of drainage caused by rainfall after irrigation. For optimal solutions with minimum number of irrigation events, maximum rainfall use efficiency and “balanced” performance, the number of irrigation events was reduced by 57 %, 18 % and 44 % on average (9.4, 3.0 and 7.4 fewer irrigation events per year) while on average rainfall use efficiency improved by 5 %, 19 % and 17 % without significant yield loss.

Precision drip irrigation of horticultural crops is receiving interest, mostly in applications in large orchards and vineyards where spatial variability results in costs to yields, quality and water productivity. We review the issues and status of precision and variable rate drip irrigation and summarize advancements and issues regarding opportunity to consider spatial irrigation management in orchards and vineyards. Topics discussed include: the conflict between "smart" and "precise" irrigation; challenges and advancement in variable-rate drip irrigation application technologies and; the use of data, including that acquired from sensors and remote sensing, for both the delineation of management zones and decision making for irrigation scheduling within zones. Prospects for future work and progress for drip irrigated horticultural application of precision water management include variable rate dripper technologies, utilization of big data from sensors and remote sensing, and unique multivariate data processing including spatial-temporal modeling.

A challenge in modern farming is to find a sustainable way of achieving sufficient production. Precision in dosage, timing and allocation of water, biocides, fertilizer and other inputs is essential, as are such management actions as harvesting, pruning and weeding. Despite the increasing availability of sensor and actuator technologies, decision-making is still largely left to the farmer. This is creating a strong demand for support in operational management. This paper presents an overview of methods involving the use of technology and data to develop model-based management support and automation for productive and input-efficient farming. For each method, the main advantages and drawbacks relating to typical farm characteristics are discussed and summarized. Three case studies are presented, to illustrate the design steps involved in developing a model, observer and controller. The overall design procedure is summarized in a flowchart, and serves as a basic guide for method selection and model development.

Site-specific agricultural management (SSM) relies on identifying within-field spatial variability and is used for variable rate input of resources. Precision agricultural management commonly attempts to integrate multiple datasets to determine management zones (MZs), homogenous units within the field, based on spatial characteristics of environmental and crop properties (i.e., terrain, soil, vegetation conditions). This study compared several multivariate spatial clustering methods to determine MZs for precision nitrogen fertilization in a citrus orchard. Six variables, namely normalized difference vegetation index, crop water stress index, digital surface model, slope, elevation and aspect, were used to characterize spatial variability within four plots. Six clustering model composites were compared, each including some or all of the following components: (1) spatial representation of the data (e.g., Getis Ord Gi*); (2) variable weights based on their relative contribution; and (3) clustering methods, including different extensions of K-means and hierarchical clustering algorithms. The fuzzy K-means algorithm applied to the weighted spatial representation was found to generate MZs with similar numbers of trees, while the K-means algorithm applied over the spatial representation generated MZs that were more continuous over space, with minimum fragmentation. Spatial variability was not constant across the orchard and among the different variables. Management of the sub-units, or plots, using spatial representation rather than the measured values, is proposed as a more suitable platform for agricultural practices. SSM is dependent upon available variable rate application technologies. Future development of fertilizer application for individual trees will require adjusting the statistical approach to support tree-specific management. The suggested model composite is flexible and may be composed of different models for delineating plot-specific MZs.

Delineation of management zones for applying Variable Rate Irrigation (VRI) in drip irrigation presents unique challenges due to the directionality of the drip lines and the fact that the water flow rate can not be varied along a drip line. This work presents a method for optimal delineation of irrigation zones under such constraints. It is assumed that historical weather records, a soil map, a simulation model that predicts the crop development in response to weather and irrigation, as well as crop and water prices, are available. The first step consists of estimating the water productivity function of the crop for each soil type if managed optimally, using multi-objective optimization. The second step consists of estimating the water productivity function of the crop on each soil if the irrigation is managed according to another soil. Optimal delineation of the irrigation zones is then obtained by solving a non-linear optimization problem in which the decision variables are the positions of the zones boundaries. This involves an intermediate step in which the area of each soil in each irrigation zone is calculated, and the soil according to which the irrigation should be managed in each zone is determined. The procedure is illustrated on two examples involving three soils and a cotton crop in Greece. In both cases, optimal management zones delineation and management was obtained using the average weather of years 1982–1991 and performance was estimated for years 1992–2016. The results show that splitting the field into several management zones increases profit slightly via yield increase and/or water savings, depending on the relative price of water.

This paper presents a scheme for applying two-stage explicit stochastic optimization to seasonal irrigation scheduling. It is assumed that an ensemble of N_s weather forecasts (scenarios) is available. At each decision point during the season up to N_s*N_smulti-objective optimization problems are solved by assuming a specific scenario for the immediate decision period and all possibleN_s scenarios for the subsequent periods. The irrigation schedule selected for implementation during the immediate decision period is the one that produces the highest worst-case yield, which mimics the traditional risk-adverse farmers’ strategy. The procedure is illustrated for a maize crop at Davis, CA, modeled with DSSAT. The optimization was performed for ten years, using as forecasts the weather recorded on the previous 15 years. The proposed approach yielded consistently results that were very close to truly optimal, i.e. results that could have been obtained if perfect weather forecasts were available at the beginning of the season. These results were better than those obtained with a deterministic approach that relied on the same data and decision rules but used only a single forecast that consisted of the average weather of the 15 previous years. However, these improved results came at the expense of a significant increase of the computation burden. In addition to the overall improved performance in terms of yield, a main advantage of the stochastic approach is that, since the solution for implementation is selected from an ensemble of solutions, it is possible to develop a selection strategy that mimics farmers’ traditional selection strategy. This could prove a key factor toward the adoption of decision support tools that involve model-based optimization.

Particle filter has received increasing attention in data assimilation for estimating model states and parameters in cases of non-linear and non-Gaussian dynamic processes. Various modifications of the original particle filter have been suggested in the literature, including integrating particle filter with Markov Chain Monte Carlo (PF-MCMC) and, later, using genetic algorithm evolutionary operators as part of the state updating process. In this work, a modified genetic-based PF-MCMC approach for estimating the states and parameters simultaneously and without assuming Gaussian distribution for priors is presented. The method was tested on two simulation examples on the basis of the crop model AquaCrop-OS. In the first example, the method was compared to a PF-MCMC method in which states and parameters are updated sequentially and genetic operators are used only for state adjustments. The influence of ensemble size, measurement noise, and mutation and crossover parameters were also investigated. Accurate and stable estimations of the model states were obtained in all cases. Parameter estimation was more challenging than state estimation and not all parameters converged to their true value, especially when the parameter value had little influence on the measured variables. Overall, the proposed method showed more accurate and consistent parameter estimation than the PF-MCMC with sequential estimation, which showed highly conservative behavior. The superiority of the proposed method was more pronounced when the ensemble included a large number of particles and the measurement noise was low.

Apple trees tend to bear an excess of young fruits (fruitlets). Chemical thinning with an application of bioregulators early in the season is currently the most common solution to adjust fruit load. However, the results of this approach are highly inconsistent between seasons. In the present study, we identified the most informative bands in the visible-near infrared range for forecasting fruitlet destiny in response to the thinner application, which could be used to support grower decisions. “Golden Delicious” apple trees were monitored during two consecutive seasons, following the application of synthetic auxins commonly used in commercial orchards. Simple bands, band differences, and band ratios were investigated as features for forecasting fruitlet destiny from the first week of monitoring. Optimal threshold values and best wavelengths combinations were determined via the receiver operating characteristic (ROC) curve. At the early stage of fruitlet development (4–6 days after treatment, DAT), applying thresholding to the band difference R₉₇₃−R₄₀₄ was preferable (accuracies ranging from 66% to 87%), while applying thresholding to the band ratio R₆₉₃/R₆₇₄ was preferable at 7–12 DAT (accuracies ranging from 76% to 95%). The present results indicate that the development of a simplified portable device for forecasting apple fruitlet drop seems feasible. Complimentary measurements performed on intact vs. denuded fruitlets revealed that the spectral differences observed can be explained by substantial changes in trichomes density that occur naturally during fruitlet development and affect the apparent pigment absorption.

The paper presents a model-based optimization scheme for allocation of cropping areas and water. The novelty of the scheme is that rather than using highly simplified models of crop response to deficit irrigation, detailed dynamic models of crop/soil/atmosphere interactions are used to determine the crop water productivity function. While the use of such models has traditionally been considered as prohibitive in terms of computation time, the current scheme circumvents this limitation by using each model independently in a simple multi-objective optimization procedure outside of the main optimization procedure. Once the optimization problem dealing with land and water allocation has been solved, the same models are used to compute optimal irrigation schedules for each crop at each location. During the season, a simplified version of the optimization scheme can be used to update water allocation in response to discrepancies between the actual and forecasted weather or factors such as changes in water quota or crop prices. The proposed scheme is illustrated for a hypothetical farm with four fields near Davis, CA. The model AquaCrop is used to determine optimal cropping and water allocation for simultaneous cultivation of maize and sunflower at that farm.

Apple trees (Malus domestica Borkh.) tend to exhibit a biennial cycle: a heavy-flowering year with an excessive amount of low-quality fruits is followed by a year with scarce flowering and low fruit load. Chemical thinning is currently the only viable solution in large commercial operations to ensure adequate yield. However, most thinners are effective only in the first few weeks following bloom and thinning efficiency depends on numerous factors and is difficult to predict. Forecasting the expected fruitlet drop after an initial thinner application would help perform corrections with the subsequent application. In this study, we used in-situ spectroscopy in the visible and near-infrared (Vis-NIR) range to forecast fruitlets drop rate. The study was carried out on “Golden Delicious” apple trees during two growing seasons – in April 2017 and April 2018. As commonly done in commercial orchards, the fruitlet drop was amplified by the application of synthetic auxins 1-naphthaleneacetic acid (NAA) and its amide (NAD). Fruitlets were tagged and monitored in situ every 2–4 days by measuring reflectance over the 400–1000 nm range. Special care was taken to assess sunlight interference during the measurement and correct it using a custom post-processing procedure. Measurements at 4–12 days after NAA/NAD treatment (days after treatment - DAT) were used to forecast fruitlet drop by 20–26 DAT (prediction dates). Principal component analysis (PCA(was carried out, followed by a classification algorithm (linear or quadratic discriminant analysis). Performing measurements on 4 DAT proved too early to predict fruitlet drop with satisfactory reliability (forecast accuracy 65%). Measurements at 6–12 DAT resulted in forecast accuracies of 80–97%, depending on the selected dates. The method offers a non-destructive prediction of apple fruitlet drop rate, which could lead to the development of a low-cost device that could help manage chemical thinning.

The ensemble Kalman filter (EnKF) is a widely used data assimilation method in soil hydrology. However, underestimation of the modeling errors and of the sampling errors may cause systematic reduction of state variances and rejection of the observations. Inflation methods are used to alleviate this phenomenon. Here, we suggest a novel inflation method based on confidence intervals constructed using the collected ensemble of the measurements. The proposed method is illustrated via two synthetic examples of a three-layer soil with (i) precipitation and evaporation boundary condition and (ii) irrigation boundary condition. We present a comparison of two existing inflation methods and discuss the advantages and limitations of the proposed method. Basically, the suggested method behavior is superior to the behavior of the existing methods.

Vines for wine production are commonly cultivated under regulated deficit irrigation, i.e. the vines (Vitis vinifera) are maintained in a state of mild-sever, controlled drought stress in order to enhance the quality of the yield and the wine. The plant water status can be estimated via measurements of the Stem Water Potential (SWP) using Scholander pressure chamber. The objective of the present work was to demonstrate the application of Model Predictive Control (MPC) for managing grapevine irrigation via midday Stem Water Potential at two levels: (1) MPC was used to determine SWP reference values for the whole season, and (2) MPC was used to estimate twice a week the irrigation required to achieve the desired SWP.

Agricultural engineering is moving towards a future that contains autonomous robots similar to other fields such as the automotive and aerospace industry. Planning systems based on autonomous robots is associated with many challenges in various areas such as navigation, guidance, communication, software, logic, hardware and more. The present work focused on the challenges related to guidance, navigation, and control of an autonomous robot inside a vineyard or an orchard. Unreliable GPS signal, constantly changing three-dimensional obstacles, changing lighting conditions, and winds are just a few examples of the challenges posed by the agricultural environment. Previous research has addressed this issue using sophisticated and expensive robots that contain high-quality sensors and powerful computers. Despite extensive research, it is difficult to find systems that have become commercial, mainly due to the high cost of these systems. This research offers an alternative, more affordable solution, working with a group of small unmanned ground vehicles and quad copter that work together to achieve a common goal. The ground robots use a small set of affordable sensors which are combined with information produced by processing the quadcopter images. In addition to reduced cost, the autonomous multi-robots system approach also provides other advantages such as increased system reliability. This work included numerical experiments and a series of field experiments that verified the performance of the system under real conditions.

The paper aims to develop an approximated analytical solution to model the bending moment profile in a sewage pipe, buried within an unsaturated soil, which occurs as a result of a leak. The solution involves evaluation of the greenfield displacements due to a buried point source, and its use as an input to a soil-pipeline interaction problem. The solution is extended for a general wetted sphere (having different degree of saturation with the radial distance). The final model is tested against finite element simulations of the coupled problem without the simplified assumptions and approximations, and is found to be satisfactory. The work may be considered a first step towards realization of a distributed fiber optic sensing system that, together with an appropriate spatial signal analysis, could identify leaks at their early stage. The current analysis indicates that the developed strain signal (and its profile) could be detectable for leaks having liquid loss as little as 300 to 500 liters.

This study examines the impact of weekly and seasonal weather forecasts on the optimal irrigation scheduling problem. We compare stochastic and deterministic simulation-optimization frameworks for irrigation scheduling using perfect weekly (deterministic) and imperfect seasonal (stochastic) forecasts. The analysis is performed on a case study of irrigated chickpeas in Kibbutz HaZore'a in northern Israel. The results demonstrate that, for the case study area and crop, the optimization with seasonal stochastic forecasts outperforms the deterministic optimization that uses perfect weekly forecast.

The development of secondary cell wall and final composition affects the quality of mature cotton fiber. The previous studies reported that the biosynthesis of phenolic compounds occur throughout the development of cotton fiber secondary wall. This study investigated the accumulation and structural unit of wall-linked phenolics in the cotton fiber secondary wall. The accumulation of phenolics in cotton fibers within a single boll increased along with the fiber development from 20 days postanthesis (DPA) to maturity. The method, which the wall-linked phenolics was observed thoroughly by fourier transform infrared spectroscopy (FTIR) after treated by thioglycolic acid (TGA), was established. The result showed that the cell wall phenolics of cotton fiber consisted predominantly of guaiacyl units (G) with some syringyl units (S). The results of derivatization followed by reductive cleavage (DFRC) confirmed G and some S units existed in the cell wall of cotton fiber. The nuclear magnetic resonance (NMR) showed connection ways between G and S were C-C bond of β-5 and β-β, and ether linkage of β-O-4. Gel permeation chromatography (GPC) indicated the average molecular weight (Mw) distribution range of wall-linked phenolics with 1023–2169 g mol^-1, while number-average molecular weight (Mn) was 900–1970 g mol^-1 and the polydispersity coefficient was 1.09–1.74.

This paper analyses the performance of sub-optimal irrigation schedules obtained daily by solving a multi-objective optimization problem with updated weather measurements and forecasts. The approach was tested using five crops at four European locations with contrasting weather conditions. Four- and 6-day Global Forecast System (GFS) forecasts were used at all locations, and comparison with a down-scaled locally tuned model was conducted at one location. Accurate GFS temperature forecasts were observed at all four locations, but the accuracy of the potential evapotranspiration calculated from the GFS forecasts was not as consistent. Precipitations forecasts were very poor at all locations. In Greece, the down-scaled locally tuned forecasts were only marginally better than the GFS ones. In most cases, recomputing the sub-optimal irrigation schedule daily greatly reduced the impact of the imperfect weather forecasts on the final results. Using 4- or 6-day actual forecasts did not yield results appreciably better than those obtained using only historical averages as surrogate forecasts. The main consequence of the imperfect forecasts was that the final yield differed from the target one, but the (yield, irrigation) combination remained close to optimal, unless the target yield was set too high and water availability was not the main factor limiting crop development.

This paper presents a modeling framework for real-time decision support for irrigation scheduling using probabilistic seasonal weather forecasts which are incorporated into a simulation-optimization framework. The simulation of the field processes is performed by the Soil Water Atmosphere Plant (SWAP) model, whereas the optimization is performed by three different stochastic programming methods: implicit approach, explicit single-stage approach and explicit two-stage approach. To evaluate the benefit of the probabilistic forecasts, the irrigation schedules from the different stochastic methods are compared with the best benchmark of perfect forecasts as well as with the real field and the Agriculture Extension Service of Israel schedules. The analysis is performed on a real case study of irrigated chickpeas field in Kibbutz Hazorea, Northern Israel. The results show that incorporating stochastic weather forecasts could lead to substantial improvements compared with current irrigation practices.

Estimating pesticide spray drift, which is a part of total drift loss, is complex as airborne pesticide concentrations are low and depend on multiple factors. The aim was to measure and compare vertical profiles of spray drift generated by different sprayers using Open Path Fourier-Transform-Infra-Red (OP-FTIR) spectrometer. Field tests included three types of commercial agricultural sprayers. The OP-FTIR was placed at the edge of an apple orchard with the line of sight parallel to tree rows. The OP-FTIR and its reflector were mounted on platform lifts to allow measurements at 4 heights: 3 (canopy height), 4, 5, and 6 m above ground. The sprayers sprayed water within the three tree rows closest to the OP-FTIR as well as outside each tree row in order to estimate the spray drift as function of distance with and without tree interference. The results of the experiments showed that, under the meteorological conditions prevailing, there were substantial differences between the sprayers in terms of spray drift of droplets with diameter > 5 μm. Additionally, the results showed that spray drift can be reduced substantially (by up to 50%) by using a tree-line barrier or a buffer zone.

A procedure for identifying apples in night-time orchard images was developed and tested on two datasets totalling over 550 images of Golden Delicious trees captured on two years with different cameras and lighting systems. The analysis started by detecting specular reflection highlights and extracting sub-images (101 by 101 pixels) centred at those local maxima. Each sub-image was reduced to a 676 Upright Speeded Up Robust Features (U-SURF) vector. Close to 20,000 sub-images from one dataset were manually labelled as “apple” or “not apple”. The latter group included parts of leaves, branches and other objects which exhibited strong specular reflection. Seventy-two classifiers were trained with the number of “apple” and “not apple” training samples ranging from 500 to 2000 and from 5000 to 10,000, respectively, and with vocabulary size ranging from 500 to 10,000. Misclassifications occurred mostly in dark and low contrast regions, which led to developing alternate models based on the posterior probability that the classification result was correct taking into account the sub-image entropy or intensity. Yield models were calibrated for each dataset, using 20 random trees. For both datasets the overall yield estimate was within 10% of the actual yield, and the standard deviation was around 30% of the average tree yield. These results are similar to those reported in previous studies, but while these previous studies used procedures calibrated and tested with images from the same dataset, in the present study the classifier trained with images from one dataset was successfully applied to the second dataset.

Rapid, easy, and frequent prediction of gross or potential nitrogen mineralization rates (GNMR and PNMR, respectively) is desirable for improved understanding and quantification of soil N dynamics and for enabling advanced sustainable N management. Our goal was to extend the use of excitation-emission matrix (EEM) fluorescence spectroscopy to characterize constituents of soluble organic matter (OM) pools in agricultural soils and to couple these characterizations with advanced chemometric techniques to improve prediction of PNMR and GNMR. To achieve this, EEM-based predictions must be valid across diverse soils, climates, and management systems. Accordingly, we analyzed soil water extracts spanning a broad range of OM contents from midwest United States (MUS) and Israeli (ISL) agroecosystems under organic- and mineral-based N management strategies. Parallel factors analysis, a multiway data analysis method, was used to quantify meaningful EEM spectral components, which were used to detect changes in labile soil OM pools and to predict N mineralization rates. N-way partial least squares regression (NPLS) was also applied to EEM data to obtain spectral factors correlated to total organic carbon concentration potential and gross N mineralization rates. This NPLS analysis led to reliable estimation of all three tested properties for ISL and MUS soils.

Measurements indicative of crop development, such as leaf area index, canopy cover or biomass are typically performed only a few times throughout the season at irregular time intervals. Furthermore, due to the inherent spatial variability that exists in the field, combining measurements taken at different locations in the field usually leads to large uncertainty around the mean value. These factors, together with the fact that crop-soil models are strongly non-linear, render assimilation of measurements in crop-soil models non-trivial. This work presents procedures for performing such data assimilation, using the crop model AquaCrop as specific example. The procedures are based on Extended Kalman Filter, with some heuristic adjustments, and enable re-initialisation of state variables and/or adjustments of selected parameters of the model. The uncertainties of the measurements are taken into account explicitly in the proposed assimilation scheme. The procedures were tested with data obtained from experiments conducted with potato in Denmark and cotton in Greece. In both cases the data available consisted of canopy cover and biomass (average and standard deviation on 5–10 days), and a locally-calibrated AquaCrop model was used as starting point for the assimilation process. The results demonstrate the soundness of the approach but also emphasise the inherent limitations associated with data assimilation. In particular, assimilation of easy-to-obtain canopy cover measurements did not always improve the predictions of biomass.

Gross N mineralization is a fundamental soil process that plays an important role in determining the supply of soil inorganic N, highlighted by recent research demonstrating that plants can effectively compete with microbes for inorganic N. However, predictions of the supply of plant available N from soil have largely neglected gross N mineralization. As soil organic matter (SOM) is the substrate that microbes use in the process of N mineralization, characteristics of SOM fractions that are relatively easy to measure may hold value as predictors of gross N mineralization. To improve understanding of predictive relationships between SOM fraction properties and gross N mineralization, we assessed 32 measures of SOM quality and quantity, including physically, chemically, and biologically defined SOM fractions, for their ability to predict gross N mineralization across a wide range of soil types (Aridisols to Mollisols) and crop management systems (organic vs. inorganic based fertility) in Israel and the United States. We also assessed predictions of a commonly employed indicator of soil N availability, potentially mineralizable N (PMN, determined by 7-d anaerobic incubation). Organic fertility management systems consistently enhanced gross N mineralization and PMN compared with inorganic fertility management systems. While several SOM characteristics were significantly correlated with both gross N mineralization and PMN, other characteristics differed in their relationships with gross N mineralization and PMN, highlighting that these assays are controlled by different factors. Multiple linear regressions (MLR) were utilized to generate N mineralization predictions: Five (gross N mineralization) or six (PMN) predictor models explained >80% of the variation in both gross N mineralization and PMN (R² > 0.8). The MLR models successfully predicted gross N mineralization and PMN across diverse soil types and management systems, indicating that the relationships were valid across a wide range of diverse agroecosystems. The ability to develop predictive models that apply across diverse soil types can aid soil health assessment and management efforts.

Precise movement of a rigid body with jerk as the control signal, and with friction compensation is considered. The optimal solution for the minimization of the electrical energy consumption is obtained. The optimal solution was tested numerically.

Machine vision technologies hold the promise of enabling rapid and accurate fruit crop yield predictions in the field. The key to fulfilling this promise is accurate segmentation and detection of fruit in images of tree canopies. This paper proposes two new methods for automated counting of fruit in images of mango tree canopies, one using texture-based dense segmentation and one using shape-based fruit detection, and compares the use of these methods relative to existing techniques:—(i) a method based on K-nearest neighbour pixel classification and contour segmentation, and (ii) a method based on super-pixel over-segmentation and classification using support vector machines. The robustness of each algorithm was tested on multiple sets of images of mango trees acquired over a period of 3 years. These image sets were acquired under varying conditions (light and exposure), distance to the tree, average number of fruit on the tree, orchard and season. For images collected under the same conditions as the calibration images, estimated fruit numbers were within 16 % of actual fruit numbers, and the F1 measure of detection performance was above 0.68 for these methods. Results were poorer when models were used for estimating fruit numbers in trees of different canopy shape and when different imaging conditions were used. For fruit-background segmentation, K-nearest neighbour pixel classification based on colour and smoothness or pixel classification based on super-pixel over-segmentation, clustering of dense scale invariant feature transform features into visual words and bag-of-visual-word super-pixel classification using support vector machines was more effective than simple contrast and colour based segmentation. Pixel classification was best followed by fruit detection using an elliptical shape model or blob detection using colour filtering and morphological image processing techniques. Method results were also compared using precision–recall plots. Imaging at night under artificial illumination with careful attention to maintaining constant illumination conditions is highly recommended.

Informative spectral bands for green leaf area index (LAI) estimation in two crops were identified and generic models for soybean and maize were developed and validated using spectral data taken at close range. The objective of this paper was to test developed models using Aqua and Terra MODIS, Landsat TM and ETM+, ENVISAT MERIS surface reflectance products, and simulated data of the recently-launched Sentinel 2 MSI and Sentinel 3 OLCI. Special emphasis was placed on testing generic models which require no re-parameterization for these species. Four techniques were investigated: support vector machines (SVM), neural network (NN), multiple linear regression (MLR), and vegetation indices (VI). For each technique two types of models were tested based on (a) reflectance data, taken at close range and resampled to simulate spectral bands of satellite sensors; and (b) surface reflectance satellite products. Both types of models were validated using MODIS, TM/ETM+, and MERIS data. MERIS was used as a prototype of OLCI Sentinel-3 data which allowed for assessment of the anticipated accuracy of OLCI. All models tested provided a robust and consistent selection of spectral bands related to green LAI in crops representing a wide range of biochemical and structural traits. The MERIS observations had the lowest errors (around 11%) compared to the remaining satellites with observational data. Sentinel 2 MSI and OLCI Sentinel 3 estimates, based on simulated data, had errors below 8%. However the accuracy of these models with actual MSI and OLCI surface reflectance products remains to be determined.

The sudden collapse of sinkholes in the Dead Sea area represents a serious threat to infrastructure in the area. The formation of these sinkholes has been shown to be directly correlated with the drop in the Dead Sea water level which is accompanied by a corresponding lowering of the groundwater level and permits the penetration of low-salinity groundwater into coastal areas. This water causes dissolution of the salt layers which results in the formation of subsurface voids that develop into collapse sinkholes. Various tools and measurement methods have been investigated in order to attempt to detect the formation of sinkholes, but to date, there is no method capable of providing early warning of possible collapse. This paper investigates the use of fiber-optic Brillouin optical time-domain reflectometry (BOTDR) or Brillouin optical time-domain analysis (BOTDA) for such detection. Brillouin optical time-domain reflectometry or analysis (BOTDR/A) is an optical measurement technique that provides distributed measurements of strain along tens of kilometers of conventional optical fibers, based on the Brillouin frequency shift of backscattered light. The rationale for this approach is that the formation of an underground cavity causes strains in the soil which can be detected using a fiber-optic cable buried at a shallow depth. A closed-form solution for the expected surface sinkholeinduced strain profile attributable to spherical voids in elastic-plastic soil is developed, validated, and evaluated against more realistic conditions. The model is then used to develop a procedure that can differentiate between signals induced by a sinkhole and signals caused by disturbances. The suggested procedure uses wavelet decomposition to filter out the disturbances and extract the sinkhole contributions. This procedure is evaluated with signals measured in the field during a 50-day period on which were superimposed theoretical strains based both on the closed-form model and finite-difference analysis of more realistic cases. This analysis showed that in the depth range investigated (up to 50 m), detection can be achieved while the subsurface cavities are still stable and early enough to enable the implementation of countermeasures.

A procedure for estimating the number of mature apples in orchard images captured at night-time with artificial illumination was developed and its potential for estimating yield was investigated. The procedure was tested using four datasets totaling more than 800 images taken with cameras positioned at three heights. The procedure for detecting apples was based on the observation that the light distribution on apples follows a simple pattern in which the perceived light intensity decreases with the distance from a local maximum due to specular reflection. Accordingly, apple detection was achieved by detecting concentric circles (or parts of circles) in binary images obtained via threshold operations. For each dataset, after calibration of the procedure using 12 images, the estimates of the number of apples were within a few percent of the number of apples counted by visual inspection. Yield estimations were obtained via multi-linear models that used between two and six images per tree. The results obtained using all three cameras were only slightly better than those obtained using only two cameras. Using images from only one side of the tree did not worsen the results significantly. Overall, the yield estimated by the best models was within ± 10 % of the actual yield. However, the standard deviation of the yield estimation errors corresponded to ~26–37 % of the average tree yield, indicating that improvements are still needed in order to achieve accurate yield estimation at the single-tree level.

In this paper, we tested the operational capacity of an interoperable model coupling system for the irrigation scheduling (IMCIS) at an experimental cotton (Gossypium hirsutum L.) field in Northern Greece. IMCIS comprises a meteorological model (TAPM), downscaled at field level, and a water-driven cultivation tool (AquaCrop), to optimize irrigation and enhance crop growth and yield. Both models were evaluated through on-site observations of meteorological variables, soil moisture levels and canopy cover progress. Based on irrigation management (deficit, precise and farmer’s practice) and method (drip and sprinkler), the field was divided into six sub-plots. Prognostic meteorological model results exhibited satisfactory agreement in most parameters affecting ET_o, simulating adequately the soil water balance. Precipitation events were fairly predicted, although rainfall depths needed further adjustment. Soil water content levels computed by the crop growth model followed the trend of soil humidity measurements, while the canopy cover patterns and the seed cotton yield were well predicted, especially at the drip irrigated plots. Overall, the system exhibited robustness and good predicting ability for crop water needs, based on local evapotranspiration forecasts and crop phenological stages. The comparison of yield and irrigation levels at all sub-plots revealed that drip irrigation under IMCIS guidance could achieve the same yield levels as traditional farmer’s practice, utilizing approximately 32% less water, thus raising water productivity up to 0.96 kg/m³.

Fungi can modify the pH in or around the infected site via alkalization or acidification, and pH monitoring may provide valuable information on host-fungus interactions. The objective of the present study was to examine the ability of two fungi, Colletotrichum coccodes and Helminthosporium solani, to modify the pH of potato tubers during artificial inoculation in situ. Both fungi cause blemishes on potato tubers, which downgrades tuber quality and yield. Direct visualization and estimation of pH changes near the inoculation area were achieved using pH indicators and image analysis. The results showed that the pH of the area infected by either fungus increased from potato native pH of approximately 6.0 to 7.4 to 8.0. By performing simple analysis of the images, it was also possible to derive the growth curve of each fungus and estimate the lag phase of the radial growth: 10 days for C. coccodes and 17 days H. solani. In addition, a distinctive halo (an edge area with increased pH) was observed only during the lag phase of H. solani infection. pH modulation is a major factor in pathogen-host interaction and the proposed method offers a simple and rapid way to monitor these changes.

Maize is the dominant irrigated crop in Kansas. In recent years, as a result of declining groundwater levels in the Ogallala aquifer and diminished well capacities, farmers are turning to deficit irrigation strategies. This study demonstrates the potential of model-based optimization for determining adequate soil water depletion levels. CERES-Maize was used as surrogate crop, while the AquaCrop model was used in an optimization procedure that determined the optimal water depletion levels. A multi-objective optimization framework was used to determine several combinations of optimal water depletion levels based on ten years of historical weather, and these combinations were tested using an additional 50 years of historical weather. The results show that, although imperfect modeling and weather fluctuations caused the actual yield to be different from the target yield, the fluctuations around the multi-year averages were not significantly larger when testing the irrigation schedule with the CERES-Maize model than when testing it with the AquaCrop model that had been used to develop the irrigation schedule.

This work describes an efficient model-based procedure for computing seasonal sub-optimal irrigation schedules. The optimization problem is formulated as a multi-objective one, with the objective function consisting of the end-of-season yield and total irrigation. In order to compute the sub-optimal irrigation schedules, a hybrid formulation is presented, according to which a single irrigation event is optimized for the five subsequent days, while irrigation during the rest of the season is assumed to be triggered at some soil water levels which are determined as part of the optimization. This hybrid formulation minimizes the number of decision variables so that solving the optimization problem requires less than one minute on an i5-3470 PC. Such an efficient procedure can be executed whenever new weather forecasts become available and could be integrated in a web-based decision support system. In order to test the proposed approach we used nine years of climatic data from Northern Greece and a locally-calibrated AquaCrop model for cotton crop. We considered three extreme cases with respect to the weather forecasts: perfect weather forecasts available for the whole season; perfect short-term weather forecasts available daily for the five subsequent days and historical weather data available as forecasts for the rest of the season; and only historical weather available as forecasts. The results show that re-computing the sub-optimal solution minimizes the negative effect of imperfect weather forecasts. In the present case-study, even in the worst-case scenario, the multi-year average deviation from optimum was less than 35 mm and 75 kg/ha for irrigation and yield, respectively.

Archaeological waterlogged wood objects exposed on the Dead Sea shore exhibit little visual evidence of degradation when first exposed, and after prolonged exposure and dehydration. An investigation on the state of preservation of this material was recognised as a necessary step towards its long-term conservation. Micromorphological observations, ATR FTIR, ash content, and physical tests showed that deterioration is limited and is mostly non-biological in nature. Natural bulking and impregnation with lake minerals and salts appear to play a significant role in the physical stability of these woods when dried, and apparently inhibit microbial colonization and subsequent degradation. In contrast, archaeological wood examined from a typical Mediterranean marine environment showed advanced stages of degradation by bacteria, with the wood structure extensively compromised.

The use of pesticides is important to ensure food security around the world. Unfortunately, exposure to pesticides is harmful to human health and the environment. This study suggests using active Open Path Fourier Transform Infra-Red (OP-FTIR) spectroscopy for monitoring and characterizing pesticide spray drift, which is one of the transfer mechanisms that lead to inhalation exposure to pesticides. Experiments were conducted in a research farm with two fungicides (Impulse and Bogiron), which were sprayed in the recommended concentration of ∼0.1%w in water, using a tractor-mounted air-assisted sprayer. The ability to detect and characterize the pesticide spray drift was tested in three types of environments: fallow field, young orchard, and mature orchard. During all spraying experiments the spectral signature of the organic phase of the pesticide solution was identified. Additionally, after estimating the droplets' size distribution using water sensitive papers, the OP-FTIR measurements enabled the estimation of the droplets load in the line of sight.

A mobile melon robotic harvester consisting of multiple Cartesian manipulators, each with three degrees of freedom, is being developed. In order to design an optimal robot in terms of number of arms, manipulator capabilities, and robot speed, a method of allocating the fruits to be picked by each manipulator in a way that yields the maximum harvest has been developed. Such a method has already been devised for a multi-arm robot with 2DOF each. The maximum robotic harvesting problem was shown there to be an example of the maximum k-colorable subgraph problem (MKCSP) on an interval graph. However, for manipulators with 3DOF, the additional longitudinal motion results in variable intervals. To overcome this issue, we devise a new model based on the color-dependent interval graph (CDIG). This enables the harvest by multiple robotic arms to be modeled as a modified version of the MKCSP. Based on previous research, we develop a greedy algorithm that solves the problem in polynomial time, and prove its optimality using induction. As with the multi-arm 2DOF robot, when simulated numerous times on a field of randomly distributed fruits, the algorithm yields a nearly identical percentage of fruit harvested for given robot parameters. The results of the probabilistic analysis developed for the 2DOF robot was modified to yield a formula for the expected harvest ratio of the 3DOF robot. The significance of this method is that it enables selecting the most efficient actuators, number of manipulators, and robot forward velocity for maximal robotic fruit harvest.

Aerosols have a leading role in many eco-systems and knowledge of their properties is critical for many applications. This study suggests using active Open-Path Fourier Transform Infra-Red (OP-FTIR) spectroscopy for quantifying water droplets and solutes load in the atmosphere. The OP-FTIR was used to measure water droplets, with and without solutes, in a 20 m spray tunnel. Three sets of spraying experiments generated different hydrosols clouds: (1) tap water only, (2) aqueous ammonium sulfate (0.25-3.6%wt) and (3) aqueous ethylene glycol (0.47-2.38%wt). Experiment (1) yielded a linear relationship between the shift of the extinction spectrum baseline and the water load in the line-of-sight (LOS) (R² = 0.984). Experiment (2) also yielded a linear relationship between the integrated extinction in the range of 880-1150 cm^-1 and the ammonium sulfate load in the LOS (R² = 0.972). For the semi-volatile ethylene glycol (experiment 3), present in the gas and condense phases, quantification was much more complex and two spectral approaches were developed: (1) according to the linear relationship from the first experiment (determination error of 8%), and (2) inverse modeling (determination error of 57%). This work demonstrates the potential of the OP-FTIR for detecting clouds of water-based aerosols and for quantifying water droplets and solutes at relatively low concentrations.

Green leaf area index (LAI) provides insight into the productivity, physiological and phenological status of vegetation. Measurement of spectral reflectance offers a fast and nondestructive estimation of green LAI. A number of methods have been used for the estimation of green LAI; however, the specific spectral bands employed varied widely among the methods and data used. Our objectives were (i) to find informative spectral bands retained in three types of methods, neural network (NN), partial least squares (PLS) regression and vegetation indices (VI), for estimating green LAI in maize (a C4 species) and soybean (a C3 species); (ii) to assess the accuracy of the algorithms estimating green LAI using a minimal number of bands for each crop and generic algorithms for the two crops combined. Hyperspectral reflectance and green LAI of irrigated and rainfed maize and soybean were taken during eight years of observations (altogether 24 field-years) in very different weather conditions. The bands retained in the best NN, PLS and VI methods were in close agreement. The validity of these bands was further confirmed via the uninformative variable elimination PLS technique. The red edge and the NIR bands were selected in all models and were found the most informative. Identifying informative spectral bands across all four techniques provided insight into spectral features of reflectance specific for each species as well as those that are common to species with different leaf structures, canopy architectures and photosynthetic pathways. The analyses allowed development of algorithms for estimating green LAI in soybean and maize with no re-parameterization. These findings lay a strong foundation for the development of generic algorithms which is crucial for remote sensing of vegetation biophysical parameters.

Water shortage is the main limiting factor for agricultural productivity in many countries and improving water use efficiency in agriculture has been the focus of numerous studies. The usual approach to limit water consumption in agriculture is to apply water quotas and in such a situation farmers should use an irrigation schedule that maximizes the yield and abides to the quota constraints. In contrast to the widespread use of irrigation scheduling based on agronomy practices, irrigation scheduling may be considered as a constrained optimization problem. When drip irrigation is used, the decision variables are the irrigation amounts for each day of the season. The objective function is the expected yield calculated with the use of a model. In the present work we solved this optimization problem for three crops modeled by the model AquaCrop. This optimization problem is non-trivial due to the non-smooth behavior of the objective function and the fact that it involves multiple integer variables. We developed an optimization scheme for generating sub-optimal irrigation schedules that take implicitly into account the response of the crop to water stress, and used these as initial guesses for a full optimization of daily irrigation. Performing this optimization with various values of water quotas produced the function that expresses the relationship between water quota and yield.

The goal of a melon harvesting robot is to maximize the number of melons it harvests given a progressive speed. Selecting the sequence of melons that yields this maximum is an example of the orienteering problem with time windows. We present a dynamic programming-based algorithm that yields a strictly optimal solution to this problem. In contrast to similar methods, this algorithm utilizes the unique properties of the robotic harvesting task, such as uniform gain per vertex and time windows, to expand domination criteria and quicken the optimal path selection process. We prove that the complexity of this algorithm is linearithmic in the number of melons and can be implemented online if there is a bound on the density. The results of this algorithm are demonstrated to be significantly better than the standard heuristic solution for a wide range of harvesting robot scenarios.

Food security and reducing malnutrition of the growing world population is a permanent issue in agricultural research. Wheat (Triticum aestivum L.) is an important part of the world food market. The Canadian Prairies comprise the provinces of Alberta, Saskatchewan, and Manitoba which produce very large quantities of wheat, mostly for export. While in many other countries, and in particular in the Middle East and North Africa (MENA) region, wheat is grown with supplemental irrigation, the common practice in Canadian Prairies is to grow wheat as a rain-fed crop. Taking into account the growing pressure on fresh water resources demand and factors of the profitability the possibly optimal use of the irrigation water should be determined according to the its deficit in the region. Thus the set of the optimization problems must be solved: find maximal wheat yield with respect to the limited irrigation water quota and given weather and hydrological data. Systematically solving this problem for different values of the water quota (W) allows to create an irrigation water use efficiency function IWUE = Y (W) which presents the yield as a function of the total irrigation water applied optimally during the season. We demonstrate this approach using the FAO model AquaCrop in conjunction with the TOMLAB optimization library. The results of the model-based optimization show that for this specific case study (Carman,Manitoba 2006) wheat yield could be roughly doubled with limited amount of irrigation water. The average increase of the yield in the range of 0 − 100 mm of irrigation water was more than 20 kg/ha per mm of water.

Remote sensing of atmospheric aerosols is of great importance to public and environmental health. This research promotes a simple way of detecting an aerosol cloud using a passive Open Path FTIR (OP-FTIR) system, without utilizing radiative transfer models and without relying on an artificial light source. Meteorological measurements (temperature, relative humidity and solar irradiance), and chemometric methods (multiple linear regression and artificial neural networks) together with previous cloud-free OP-FTIR measurements were used to estimate the ambient spectrum in real time. The cloud detection process included a statistical comparison between the estimated cloud-free signal and the measured OPFTIR signal. During the study we were able to successfully detect several aerosol clouds (water spray) in controlled conditions as well as during agricultural pesticide spraying in an orchard.

Detection of fruit in tree images has been the focus of numerous studies. Although most studies considered approaches based primarily on color analysis, the major drawback of such approaches is that the fruit apparent color depends not only on variety or physiological stage but also on illumination, which is inherently non-uniform within the canopy, even if artificial lighting is used. In the present work we developed a novel approach to detect apples in nighttime images by analyzing the spatial distribution of the light around highlights ("bright spots"). The approach is based on the observation that, under the artificial illumination used, apples exhibit strong specular reflection so that a small, but very bright, spot is visible on almost all apples. Each of these highlights serves as the center of a region of interest and is the seed of the investigated light patch. This patch is initially very small but its size is increased iteratively by annexing pixels with predefined decreasing gray level intensities. The evolution of the patch geometry is used to determine whether it corresponds to an apple. The approach was tested with two datasets containing over 360 images (close to 13,000 apples) acquired in the same 'Golden Delicious' orchard in July 2012 and August 2013. Twenty images from the 2012 dataset were randomly selected to develop and calibrate the procedure. The results of these 20 images were used to establish a linear relationship between the number of detected objects and the actual number of apples visible in the images (R²~0.75). Applying the calibrated procedure to the remaining images of this dataset led to an estimate of 6739 apples compared to a visual count of 6195 apples (~9% overestimate). Analysis of the 2013 dataset, in which the apparent size of the apples was smaller, required only adjustment of the two parameters related to apple size. Following this adjustment, 12 images were randomly selected to determine the relationship between the number of detected objects and the actual number of apples (R²~0.74). Using this relationship, the estimated number of apples was 6687, compared to the visual count of 6713 fruits.

Leaf pigment content provides valuable insight into the productivity, physiological and phenological status of vegetation. Measurement of spectral reflectance offers a fast, nondestructive method for pigment estimation. A number of methods were used previously for estimation of leaf pigment content, however, spectral bands employed varied widely among the models and data used. Our objective was to find informative spectral bands in three types of models, vegetation indices (VI), neural network (NN) and partial least squares (PLS) regression, for estimating leaf chlorophyll (Chl) and carotenoids (Car) contents of three unrelated tree species and to assess the accuracy of the models using a minimal number of bands. The bands selected by PLS, NN and VIs were in close agreement and did not depend on the data used. The results of the uninformative variable elimination PLS approach, where the reliability parameter was used as an indicator of the information contained in the spectral bands, confirmed the bands selected by the VIs, NN, and PLS models. All three types of models were able to accurately estimate Chl content with coefficient of variation below 12% for all three species with VI showing the best performance. NN and PLS using reflectance in four spectral bands were able to estimate accurately Car content with coefficient of variation below 14%. The quantitative framework presented here offers a new way of estimating foliar pigment content not requiring model re-parameterization for different species. The approach was tested using the spectral bands of the future Sentinel-2 satellite and the results of these simulations showed that accurate pigment estimation from satellite would be possible.

Increasing water demand, dwindling fresh water resources and extended periods of drought underscore the urgent need of improving water use efficiency in agriculture. The EU-funded project FIGARO aims to improve irrigation water use efficiency via the development and implementation of irrigation strategies that take into account in real-time soil water availability, local weather forecasts, crop physiological status and water needs. As part of this project we have developed two model-based procedures to compute optimal and sub-optimal irrigation schedules. The advantage of the sub-optimal approach is that it can be implemented in real-time to recalculate the irrigation schedule every few days based on the estimated soil water content, crop status and weather forecasts. In the present work we used 10 years of climatic data to simulate the performance of both approaches for a hypothetic maize crop in Kansas. In particular, we investigated the influence of the accuracy of short-term and long-term weather forecasts on the results. When perfect weather forecasts were assumed to be available for the whole season the sub-optimal approach produced results which were within a few percent of the optimal ones. Assuming that perfect weather forecasts were available only for the next five days did not change the results significantly on most years. Relying solely on historical weather led to poorer results, but the sub-optimal approach still achieved a yield which, on average, was less than 10% below the yield which could have been obtained with the same amount of irrigation water if perfect forecasts had been available.

Modern greenhouses include a large number of control actuators. Such a situation leads to the question of how to divide the control effort between these actuators. This question can be answered using optimal control theory and more particularly the system co-states. By modifying the value of these co-states the grower can control the greenhouse climate and correct deviations due to imperfect modelling or unexpected weather. This work presents a simulation study of this concept for a greenhouse with a tomato crop. For the specific scenarios investigated, by adjusting the co-state two to five times during the season, it was possible to maintain the yield within 6% of its target value. The control costs were on average 10% higher than the control costs that could be achieved if perfect weather forecasts were available at the beginning of the season.

This work investigates the potential of night time imaging for estimating apple orchard yield. Forty two trees were photographed from two sides with cameras mounted at three heights. Each image was analyzed and the results of the six images associated with each tree were summed up to provide a 'tree count'. Fourteen trees were selected randomly in order to calibrate a relationship between the 'tree count' estimate and the actual tree yield. This relationship was then applied to the 'tree count' results of the remaining trees. Although the yield estimate error for a single tree was sometimes large, the overall yield estimate was within 10% of the actual yield.

A hybrid approach for optimizing irrigation was developed and tested via simulation of cotton cultivation in Northern Greece. At any given day during the season, the remaining cultivation days were conceptually split into two parts: The five coming days and the rest of the season. Accurate weather forecasts were assumed to be available for the first period, during which only one irrigation event could take place. For the second period, average historical weather data were used as long-term forecasts, and rather than optimizing directly the irrigation events, the decision variables corresponded to a number of soil water levels at which irrigation would be triggered and the corresponding irrigation amounts. This hybrid approach ensured that the number of decision variables remained acceptably small so that the procedure can be computed repeatedly in real time during the season.

Water stress is one of the most influential factors contributing to crop yield loss. The importance of the irrigation constantly increases because of water scarcity and growing demand for agricultural production worldwide. Previously, an approach using empirical water production functions and analytic optimal control methodology has been developed for optimal irrigation scheduling. Such an approach based on numerical optimal control is an alternative to common irrigation scheduling based on agronomy practice. Nowadays, more complex dynamic crop simulation models, such as the FAO AquaCrop model, predict crop responses to different irrigation strategies and climates. The state variables of the AquaCrop model include crop characteristics, such as biomass, and soil water content in up to 12 soil layers. In this paper the numerical optimal control scheme for irrigation scheduling and crop water production function development is described and demonstrated using this model and the TOMLAB optimization library. Maize crop in Foggia, Italy, for season of the year 2000, is used as an illustrative case study.

This paper describes an analytical procedure to calculate the time-optimal trajectory for a mobile Cartesian manipulator to traverse between any two fruits it picks up it. The goal is to minimize the time required from the retrieval of one fruit to that of the next while adhering to velocity, acceleration, location, and endpoint constraints. This is accomplished using a six stage procedure, based on Bellman's Principle of Optimality and nonsmooth optimization that is completely analytical and requires no numerical computations. The procedure sequentially calculates all relevant parameters, from which side of the mobile platform to place the fruit on to the velocity profile and drop-off point, that yield a minimum time trajectory. In addition, it provides a time window under which the mobile manipulator can traverse from any fruit to any other, which can be used for a globally optimal retrieving sequence algorithm.

This paper details a procedure for detecting apples in tree images using shape analysis are presented. The core of the procedure consists of a so-termed convexity test that identifies edges that could correspond to three-dimensional convex objects of a given size range from a much larger set of edges. This is achieved by analysing a number of intensity profiles that originate at each edge and determining whether they have a shape that is suitable with a 3D convex object of the correct size. We show that contrarily to the prevailing opinion, the intensity functions of three-dimensional convex objects are not necessarily convex, which led us to developing models for describing such profiles. The simplest suitable model includes four parameters that can be easily estimated by a standard least square constrained optimization procedure. After merging the selected edges that fall on circles, a second analysis is performed to remove false positive detections and eliminate multiple detections of apples. The procedure was demonstrated on 51 grey-level images that were recorded in a Golden Delicious apple variety orchard under natural light conditions. On average, together with preliminary pre-processing operations, the convexity test removed 99.8% of the edges initially identified by Canny filter. Analysis based on the remaining edges led to correct detection of 94% of the apples visible in the images. Fourteen percent of the identified objects were "false positive" detections, mainly due to leaves or parts of leaves that generated convex surfaces very similar to apples, or by leaves that lay on apples and created misleading edges.

Determination and monitoring of tunneling induced ground displacement is an important component in tunneling design and construction. In recent years several technologies for distributed strain measurement along fiber optics have been developed, namely the Brillouin Optical Time Domain Reflectometry (or Analysis) - BOTDR/A and the Rayleigh backscatter wavelength interferometry (OBR). This paper presents how these technologies could be used to monitor and define ground displacement models through an appropriate 2D and 3D optimization and signal analysis of information derived from a horizontally laid fiber above the tunnel. The suggested approach is evaluated in two field investigations, one involving excavation of a 3. m diameter tunnel by TBM at depth of 18. m, and the other installation of a 1. m diameter water main by pipe-jacking at depth of 6. m. Comparison between the results obtained by the different technologies shows that they are equally suitable for the suggest approach. The suggests approach allows reliable determination of the parameters involved in empirical ground displacement models, and allows field validation that the tunneling process lies within the design bounds. An interesting observation, supported by the analytical models, is that non-perpendicular alignment of the fiber, relatively to the tunnel line, results in a shift in the peak strain location as the tunnel advances. It was demonstrated that the rate of change in peak strain location, with tunnel advancement, can be used to obtain the settlement trough length parameter, without the need for complete evaluation of all other model parameters.

Understanding and quantifying N transformations in soil is critical for sustainable use of this important plant nutrient and for understanding the mechanisms through which polluting N species are discharged to the environment. Advanced methods such as the "isotope dilution technique", which uses stable N-isotopes to estimate gross mineralization and nitrification rates, answer this need. In this study the use of Fourier transform infrared-attenuated total reflectance (FTIR-ATR) spectroscopy for measuring isotopic N species concentrations directly in soil pastes was tested as a complementary technique to the commonly used isotope ratio mass spectrometry (IRMS). It is shown that, with proper chemometric tools (e.g., partial least squares [PLS]), FTIR-ATR enables simple tracking of changes in the concentrations of the isotopic species of nitrate and ammonium and allows estimation of the gross reaction rates of N transformations in soil. Soil incubations were performed by adding either ¹⁵NO₃^- or ¹⁵NH₄ ⁺ to the soils. The incubations with added ¹⁵NH ₄⁺ yielded a gross mineralization rate of 6.1 mg N kg ^-1 dry soil d^-1 compared with a net mineralization rate of 4.1 mg N kg^-1 dry soil d^-1 and a gross nitrification rate of 40.9 mg N kg^-1 dry soil d^-1 compared with a net nitrification rate of 29.5 to 25.3 mg N kg^-1 dry soil d^-1. The incubations with added ¹⁵NO₃^- yielded a gross nitrification rate of 18.6 mg N kg^-1 dry soil d^-1 compared with a net nitrification rate of 11.9 to 18.3 mg N kg^-1 dry soil d^-1. The combined use of FTIR-ATR and ¹⁵NO ₃^- or ¹⁵NH₄⁺ enrichment appears to provide an effective tool for almost real-time quantification of N-dynamics in soils with minimal interference.

A 3DOF mobile Cartesian robotic harvester for two-dimensionally distributed crops such as melons is being developed. A two-step procedure to calculate the trajectory of its manipulator that will result in the maximum number of melons harvested is described in this article. The goal of the first step is to calculate the minimum-time trajectory required to traverse between any two melons while adhering to velocity, acceleration, location, and endpoint constraints. This step is accomplished in a hierarchal manner by solving several subproblems involving optimal control and nonconvex optimization, enabling optimal (maximum) melon harvesting to be formulated as an orienteering problem with time windows. In the second step, the orienteering problem is solved using the moving branch and prune method, based on dynamic programming. This enables suboptimal sequences of melons (out of all options) to be eliminated on the fly without the need to solve the entire problem at once. An example is shown to demonstrate the efficacy of the algorithm.

Emissions of N₂O from agricultural soils are an important source for this greenhouse gas. The present work examines the potential of Fourier transformed infrared (FTIR) spectroscopy coupled with a long path (LP) infrared (IR) gas cell for on-line measurement of concentration and isotopic signature of N₂O emitted from soils. Nitrous Oxide was spectrally monitored during incubations of soil samples in a closed system under different conditions. Its emission from a Grumosol (Vertisol) was measured in presence and absence of acetylene, with various additions of nitrate and glucose, under aerobic and anaerobic conditions, and for two soil thickness layers. For comparison N₂O emissions from a Terra Rossa (Cambisol) and a Hamra (Luvisol) were measured in the presence and absence of acetylene. In an additional experimental set the isotopic signature of emitted N₂O was quantified after enrichment with K¹⁵NO₃. Acetylene addition led to an increase in N₂O emissions in all three soils but at various extents. Under aerobic conditions, N₂O emission from the Grumosol became detectable only when running the experiments with a thicker soil layer (10 mm), suggesting the existence of coupled nitrification- denitrification. Nitrate addition to soils enhanced N₂O emissions especially when coupled with glucose addition. Addition of ¹⁵NO ₃^- to the Grumosol resulted in the emission of all four N₂O isotopologues: ¹⁴N₂O, ¹⁵N ₂O, ¹⁴N¹⁵NO, and ¹⁵N¹⁴NO. The observed slight delay in appearance of the species containing ¹⁵N and the relatively lower ¹⁵N enrichment of the N ₂O compared with the soil nitrate, indicate isotopic fractionation during denitrification. Yet, within the accuracy of our isotopic analysis, temporal emission patterns of ¹⁴N¹⁵NO and ¹⁵N¹⁴NO were similar indicating low possibility for "site preference" under the specific experimental conditions.

AimsStudies of species distribution patterns traditionally have been conducted at a single scale, often overlooking species-environment relationships operating at finer or coarser scales. Testing diversity-related hypotheses at multiple scales requires a robust sampling design that is nested across scales. Our chief motivation in this study was to quantify the contributions of different predictors of herbaceous species richness at a range of local scales.MethodsHere, we develop a hierarchically nested sampling design that is balanced across scales, in order to study the role of several environmental factors in determining herbaceous species distribution at various scales simultaneously. We focus on the impact of woody vegetation, a relatively unexplored factor, as well as that of soil and topography. Light detection and ranging (LiDAR) imaging enabled precise characterization of the 3D structure of the woody vegetation, while acoustic spectrophotometry allowed a particularly high-resolution mapping of soil CaCO₃ and organic matter contents.Important FindingsWe found that woody vegetation was the dominant explanatory variable at all three scales (10, 100 and 1000 m²), accounting for more than 60% of the total explained variance. In addition, we found that the species richness-environment relationship was scale dependent. Many studies that explicitly address the issue of scale do so by comparing local and regional scales. Our results show that efforts to conserve plant communities should take into account scale dependence when analyzing species richness-environment relationships, even at much finer resolutions than local vs. regional. In addition, conserving heterogeneity in woody vegetation structure at multiple scales is a key to conserving diverse herbaceous communities.

Non-invasive methods that enable continuous monitoring of spatial and temporal variations of physical and chemical parameters in the roots and the rhizosphere in situ are required in order to study the complex soil-roots interactions. In this work we present the use of laser-induced fluorescence imaging for visualizing these processes. The system consists of a Nd:YAG Q-switched laser that excites the root and rhizosphere autofluorescence, and an intensified gated camera (intensified charged coupled device (ICCD)) that is synchronized with the laser so that it captures the very short auto-fluorescence signal even in broad daylight. By using very short gating times (1 ns) and varying the delay between the laser pulse and the gating operation, time-resolved fluorescence profiles are obtained for each pixel in the image. The potential of the system is illustrated with several examples that show that both fluorescence intensity and temporal evolution profiles provide information about root activity and rootesoil interactions.

The advantages of object-oriented programming for crop modeling are discussed. Contrarily to procedural languages such as Fortran77, object oriented programming yields a code that is intrinsically modular and whose architecture is close to that of the actual crop. This simplifies model merging and code re-use, which in turn tends to extend the model life-span and increases its chances of being adopted outside the research community. The tomato model TOMGRO is used as a case study and a new version of this model in C# is outlined. One of the advantages of C# is that it is fully integrated in the .Net framework and offers developers a powerful class library. Moreover, the inclusion of such a model in a web-based decision support system using Web services and ASP.NET would be straightforward.

A control strategy for greenhouse cooling with natural ventilation and variable high-pressure fog was evaluated via computer simulation and verified experimentally. Two set points were used, one based on air specific enthalpy (56 kJ kg^-1) for determining the vent openings, and the other based on vapor pressure deficit (VPD) of the air (1.0 kPa) for controlling the fogging rate. These set points would maintain the greenhouse air at a temperature of 24°C and a relative humidity of 67%. Achieving the VPD set point was the priority, and excessive air exchange was avoided through adjustments of the vent openings when fogging demands were beyond the operational capacity of the fogging system. Results of simulations and experiments from four different days (18, 19, 29 June and 3 July 2011) were in good agreement. The performance of the control strategy developed was satisfactory to maintain the greenhouse indoor climate close to the set points with 1.1 ±0.4 kPa and 26°C ±1.7°C for inside air VPD and temperature, respectively (with relative humidity of 67% ±8%). The implementation results demonstrated that the control strategy was capable of reducing the air VPD by an average of 4.2 kPa when the average outside air VPD was 5.4 kPa and reducing the air temperature by an average of 10.5°C when the average outdoor air temperature was 37°C, and the greenhouse relative humidity was increased by an average of 52% compared to outside during the four experiment days. Deviations between the simulated and measured variables inside the greenhouse were attributed to a prevailing high and variable-magnitude outside wind speed during the experiments as well as to the time delay on the retrieval of climate data needed for computations of the strategy. Real-time climate parameters with no delay are preferable for effective computations of ventilation rates and desired fogging rates. The strategy that was developed in this study maintained the VPD close to the selected set point for all the experimental periods evaluated. Finally, the control strategy developed effectively maintained desirable climate conditions inside the greenhouse, and the simulation results were experimentally validated.

Cross-border smuggling tunnels enable unmonitored movement of people and goods, and pose a severe threat to homeland security. In recent years, we have been working on the development of a system based on fiber- optic Brillouin time domain reflectometry (BOTDR) for detecting tunnel excavation. In two previous SPIE publications we have reported the initial development of the system as well as its validation using small-scale experiments. This paper reports, for the first time, results of full-scale experiments and discusses the system performance. The results confirm that distributed measurement of strain profiles in fiber cables buried at shallow depth enable detection of tunnel excavation, and by proper data processing, these measurements enable precise localization of the tunnel, as well as reasonable estimation of its depth.

The expansion of an existing crack in chicken eggs exposed to an abrupt pressure drop was investigated. The initial crack was induced in the eggshells by two weak mechanical impacts close enough to induce a macro-crack oriented mainly along the direction of the meridian. The cracked eggs were placed in a custom-made pressure chamber in which the pressure was abruptly decreased to -25 kPa. During this pressure drop, crack expansion was measured using an imaging system and three phases of expansion (evolution of equivalent crack width vs. time) were observed: a linear mode, an exponential mode and a transition zone between them. A model that associates these phases of expansion with the stretching and tearing of the outer membrane is proposed. According to this model, the linear mode corresponds to reversible membrane stretching, which ends with membrane tearing during the transition zone. This model was supported by visual observations that revealed that membrane fragments became visible in the transition zone, and by the crack expansion curves obtained during weaker pressure drops (-10 kPa), which showed that only linear (reversible) crack expansion occurred. It was also shown that the air cell that is initially enclosed between the inner and outer membranes at the egg blunt-end plays in important role in the crack expansion. A simple model of the airflow that escapes from the egg through the crack shows that the airflow from the blunt-end to the crack, rather than the airflow through the crack itself, is the limiting factor.

Passive ventilation in greenhouse production systems is predominant worldwide, limiting its usability and profitability to specific regions or for short production cycles. Evaporative fogging systems have increasingly been implemented in Arid and Semi-Arid regions to extend the production cycle during the warmest season, and also to achieve near-optimum environments for year-round production. However, appropriate control strategies for evaporative fogging systems are still lacking or limited despite its reported benefits in terms of environmental uniformity and potential savings in water and energy usage, when compared to fan and pad systems. The present research proposes a neural network predictive control approach for optimizing water and energy usage in a naturally ventilated and fog cooled greenhouse while providing a near-optimum and uniform environment for plant growth. As a first step the dynamic behavior of the greenhouse environment, defined by air temperature and relative humidity, was characterized by means of system identification using a recurrent dynamic network (NARMX). The multi-step ahead prediction capability of NARMX allows for the optimization of the control actions (vent configuration and fogging rate) for its implementation in the NN predictive control scheme. Greenhouse environmental data from a set of experiments consisting of several vent configurations (0/50, 0/100, 50/50, 50/100 and 100/100, percent opening of the side/roof vents) and three fogging rates (17.5, 22.3 and 27.0 g m^-2min^-1) during several days throughout the year were used in the system identification process. The resulting NN model accurately predicted the dynamic behavior of the greenhouse environment, having coefficients of determination (R²) of 0.99 for each parameter (air temperature and relative humidity). These NN model will be incorporated into the NN predictive control scheme and its feasibility is in a naturally ventilated greenhouse equipped with a variable-rate fogging system is discussed, while achieving a greenhouse environment within defined permissible ranges of air temperature and relative humidity.

Cooling must be supplied for greenhouses located in semiarid climates most of the year to provide desired climate conditions for year-round crop production. High-pressure fogging systems have shown promising results for cooling, however the lack of effective control strategies, especially under passive ventilation, have limited their use. In this study, a new proposed climate control strategy, which considers the contribution on cooling and humidification from plants, is tested through simulation. The developed strategy using variable pressure fogging (VPF) and variable vent configurations was compared to a constant pressure fogging (CPF), fixed vents cooling strategy. In both cases, the control of fog was based on vapor pressure deficit (VPD) set points. Results showed that on average, VPF based system was able to save 15.2% of water and consumed 10.1% less energy. Pump cycling was reduced by 78.5% and lower temperature and relative humidity fluctuations were achieved by adjusting fog rates through manipulating the system working pressure. Finally, simulations also showed that by reducing the number of nozzles, a smaller fogging rate was achieved and the system performance and savings on water and energy were enhanced during morning hours of operation.

Even though several models to predict evapotranspiration (ET) of greenhouse crops have been developed, previous studies have evaluated them under fixed greenhouse conditions. It is still not clear which model is more appropriate, accurate, and best suited for applications such as inclusion in greenhouse cooling strategies for different crops, climatic conditions and greenhouse cooling settings. This study evaluated three theoretical models (Stanghellini, Penman-Monteith and Takakura) to simulate the ET of two crops (bell pepper and tomato), under two greenhouse cooling settings (natural ventilation with fog cooling and mechanical ventilation with pad and fan), and for three growing seasons (spring, summer, fall). Predictions of ET from the models were compared to measured values obtained from sap flow gauges. Inputs of internal and external crop resistances for Stanghellini and Penman-Monteith models were calibrated separately by crop and by model. Even though Stanghellini model produced the smallest deviations of the predicted ET from the measured ET, having the best overall performance under all conditions evaluated, an analysis of variance of the daily mean square errors did not show significant differences (α= 0.05) between the three models. This suggested that any of the three models could be used for inclusion in a greenhouse cooling climate control strategy. However, parameter adjustments such as stomatal and aerodynamic resistances, and the need of leaf area index (LAI) in the models of Penman-Monteith and Stanghellini represent a limitation for this application. The Takakura model was found to be easier to implement; however as the crop grows, careful adjustments on the height of the solarimeter used for this approach are required. Such adjustments determine the field of view of the solarimeter and play a significant role on the determination of radiation balances and the average apparent temperature of the evaporative surface.

This work details the development and validation of an algorithm for estimating the number of apples in color images acquired in orchards under natural illumination. Ultimately, this algorithm is intended to enable estimation of the orchard yield and be part of a management decision support system. The algorithm includes four main steps: detection of pixels that have a high probability of belonging to apples, using color and smoothness; formation and extension of "seed areas", which are connected sets of pixels that have a high probability of belonging to apples; segmentation of the contours of these seed areas into arcs and amorphous segments; and combination of these arcs and comparison of the resulting circle with a simple model of an apple. The performance of the algorithm is investigated using two datasets. The first dataset consists of images recorded in full automatic mode of the camera and under various lighting conditions. Although the algorithm detects correctly more than 85% of the apples visible in the images, direct illumination and color saturation cause a large number of false positive detections. The second dataset consists of images that were manually underexposed and recorded under mostly diffusive light (close to sunset). For such images the correct detection rate is close to 95% while the false positive detection rate is less than 5%.

In addition to ventilation, daily cooling must be provided for greenhouses located in semiarid climates to maintain the desired climate conditions for year-round crop production. High-pressure fogging systems have been successfully developed for greenhouse cooling. However the lack of control strategies, in combination with ventilation systems, especially passive ventilation, has limited their capabilities. A new cooling control strategy, which considered the contribution of humidification and cooling from the crop, was evaluated by computer simulations. The strategy controlled the amount of fog introduced into the greenhouse, as well as the percentage of vent openings to maintain desired values of greenhouse atmospheric vapour pressure deficit (VPD) and enthalpy, respectively, which would consequently affect air temperature. The performance was compared to constant fogging rate strategy, which was based on VPD. On average, the new strategy saved 36% water and consumed 30% less electric energy. Smaller air temperature and relative humidity fluctuations, and more consistent control, were achieved by varying the fog system operating pressure to provide a more optimum amount of fog for evaporative cooling. It was demonstrated by simulations that dynamically varying the fog rate and properly selecting the number of nozzles, savings of water and electric energy were increased, while still maintaining acceptable VPD and temperature. The improvements in the greenhouse climate achieved by the new strategy were due to its ability to dynamically manipulate fog rates, as well as, the vent configurations.

A climate control system for a small greenhouse equipped with a variable-pressure fogging system and variable-speed extracting fans was developed and validated. The controllers were designed using the robust control method quantitative feedback theory (QFT) that guarantees adequate performance of the controlled system despite large modelling uncertainties and disturbances. In order to simplify the design of the controllers and achieve better performances, partial decoupling between the two control loops was achieved by describing the greenhouse climate in terms of air enthalpy and humidity ratio. This led to using ventilation for achieving the desired air enthalpy, and fogging for achieving the desired humidity ratio, assuming that the ventilation rate was approximately known. The implementation results demonstrated the good performance of the controllers, with mean tracking errors of ~100 J kg^-1 [dry air] and ~0.1 g [water] kg^-1 [dry air] for enthalpy and humidity ratio, respectively. For practical applications, the desired climate was expressed in terms of air temperature and relative humidity, which were converted into enthalpy and humidity ratio using the psychrometric relationships. In this case, the mean tracking errors were ~0.2 °C air temperature and less than 1% relative humidity, and the maximum mean deviations over a 10-min period with constant setpoints were 2.5 °C air temperature and 5% relative humidity.

This study focused on quantitative chemical analysis of organic fouling of ultrafiltration membranes by protein (bovine serum albumin-BSA) in the presence of a polysaccharide (alginic acid-AA). Filtration experiments were performed in a dead-end cell with two microporous membranes. Prior to the filtration experiments, the ultraviolet-visible absorbance spectra of 140 solutions with various BSA and AA concentrations were used to calibrate a partial least square (PLS) model for estimating BSA concentration. This model, which had a validation error of less than 0.7. mg BSA/l, was used in the filtration experiments to estimate the concentration of BSA in the feed and permeate solutions, as well as in a solution containing the BSA fouling removed by simple water rinsing. These measurements, together with a mass balance of the different fractions of the filtration process, were used to estimate the reversible and irreversible fouling. Direct analysis of the fouled membrane was done by photoacoustic spectroscopy (FTIR-PAS), and a PLS model was used to estimate the amount of BSA present on the membrane. The fouling estimated from the FTIR-PAS spectra by PLS correlated well with the estimates from the mass balance approach, with typical errors of less than 10% of the range investigated.

This work presents the development and validation of robust controllers for greenhouses equipped with a climate control system that consists of variable-pressure fogging and variable-speed extracting fans. The controllers were designed using the robust control method "Quantitative Feedback Theory" that guarantees adequate performance of the controlled system despite large modeling uncertainties and disturbances. The design of the controllers and their initial validation were performed in a small experimental greenhouse, and the same controllers were later implemented in a much larger greenhouse located in a warmer and drier region. Good tracking performances were observed in both greenhouses.

An algorithm for estimating the number of apples in color images acquired in orchards is presented. The algorithm includes four main steps: (1) Detection of pixels that have a high probability of belonging to apples; (2) Formation and extension of "seed areas", which are connected sets of pixels that have a high probability of belonging to apples; (3) Segmentation of the contours of these seed areas into arcs and linear segments; and (4) Combination of these arcs and comparison of the resulting circle with a simplistic "apple" model. The performance of the algorithm is investigated using two datasets. For the first datasets, which consists of images recorded in full automatic mode of the camera and under various lighting conditions, more than 85% of the apples are correctly detected but direct illumination and color saturation cause a large number of false positive detections. For the second dataset, which consists of images that were manually underexposed and recorded under mostly diffusive light, close to 90% of the apples are detected while the false positive detection rate is less than 5%.

This work presents an algorithm for estimating the number of apples on trees using images acquired with a standard color CCD camera. The proposed system is capable of correctly identifying and localizing more than 85% of the apples in the images. To achieve this high detection rate, color and texture analyses are combined together with shape analysis. In the first step, pixels with a high probability of belonging to an "apple object" are detected according to their color and texture. In the second step, "seed areas" consisting of connected sets of pixels with a high probability of belonging to an apple object are detected. Each seed area is then extended to cover the entire visible area of the apple to which it belongs. Finally, each blob is segmented into simple components that can either be combined into circles or are discarded, so that each of the resulting circles corresponds to an apple.

This paper details a procedure for identifying edges that may belong to three-dimensional convex objects of an approximate size, such as apples, from a much larger set of edges. The identification is achieved by analyzing a number of intensity profiles that originate at each edge and determining whether they have a shape that is suitable with a 3D convex object of the correct size. We show that contrarily to the prevailing opinion, the intensity or luminance functions of three-dimensional convex objects are not necessarily convex, which led us to developing models for describing such profiles. The simplest suitable model includes four parameters that can be easily estimated by a standard constrained least squares optimization procedure. The proposed procedure is applied to images of apple trees recorded in an orchard. The procedure correctly removes 77-100% of the edges not belonging to apples, with minimal loss of edges belonging to such objects.

Evaluation of pesticides' fate in the atmosphere is important in terms of environmental effects on non-target areas and risk assessments analysis. This evaluation is usually done in the laboratory using analytical grade materials and is then extrapolated to more realistic conditions. To assess the effect of the pesticide purity level (i.e. analytical vs. technical) and state (i.e. sorbed film vs. airborne particles), we have investigated the oxidation rates and products of technical grade cypermethrin as thin film and in its airborne form, and compared it with our former results for analytical grade material. Technical grade thin film kinetics for both ozone and OH radicals revealed reaction rates similar to the analytical material, implying that for these processes, the analytical grade can be used as a good proxy. Oxidation products, however, were slightly different with two additional condensed phase products: formanilide, N-phenyl and 2-biphenyl carboxylic acid, which were seen with the technical grade material only. OH experiments revealed spectral changes that suggest the immediate formation of surface products containing OH functionalities. For the ozonolysis studies of airborne material, a novel set-up was used, which included a long-path FTIR cell in conjugation with a Scanning Mobility Particle Sizer (SMPS) system. This set-up allowed monitoring of real-time reaction kinetics and product formation (gas and condensed phases) together with aerosol size distribution measurements. Similar condensed phase products were observed for airborne and thin film technical grade cypermethrin after ozonolysis. Additionally, CO, CO₂ and possibly acetaldehyde were identified as gaseous oxidation products in the aerosols experiments only. A kinetic model fitted to our experimental system enabled the identification of both primary and secondary products as well as extraction of a formation rate constant. Kinetic calculations (based on gaseous products formation rate) have revealed values similar to that of the thin film experiments. Interestingly, heterogeneous oxidation of cypermethrin was also found to generate ultra fine secondary organic aerosols. Again, no significant difference was observed between analytical and technical grade materials. However, particle size distribution was much broader when films were exposed to OH and ozone than to ozone alone.

Cross-borders smuggling tunnels enable unmonitored movement of people, drugs and weapons and pose a very serious threat to homeland security. Recent advances in strain measurements using optical fibers allow the development of smart underground security fences that could detect the excavation of smuggling tunnels. This paper presents the first stages in the development of such a fence using Brillouin optical time domain reflectometry (BOTDR). Two fiber optic layouts are considered and evaluated in a feasibility study that includes evaluation of false detection and sensitivity: (1) horizontally laid fiber buried at a shallow depth, and (2) fibers embedded in vertical mini-piles. In the simulation study, two different ground displacement models are used in order to evaluate the robustness of the system against imperfect modeling. In both cases, soil-fiber and soil-structure interactions are considered. Measurement errors, and surface disturbances (obtained from a field test) are also included in the calibration and validation stages of the system. The proposed detection system is based on wavelet decomposition of the BOTDR signal, followed by a neural network that is trained to recognize the tunnel signature in the wavelet coefficients. The results indicate that the proposed system is capable of detecting even small tunnel (0.5. m diameter) as deep as 20. m (under the horizontal fiber) or as far as 10. m aside from the mini-pile (vertical fiber), if the volume loss is greater than 0.5%.

This paper details the design and implementation of a robust controller that stabilizes an unmanned motorcycle. A linearized model and Quantitative Feedback Theory (QFT) are used to derive a low order cascaded controller that stabilizes the motorcycle for velocities ranging from 2.5 to 6.5. m/s. Stabilization is achieved by measuring roll angle and roll rate and controlling the steering torque. The approach is validated through simulations and experiments with a 50cc scooter whose throttle, brakes, and reference roll angle are radio-controlled.

Previous studies on high pressure fogging have shown their capability for maintaining temperature and humidity in acceptable ranges most of the year in greenhouses located in semiarid regions. The heat load, and therefore cooling demand, inside the greenhouse vary during the day and throughout the seasons. Thus, it may be advantageous to use a variable pressure fogging (VPF) system, where specific fog rates can be supplied based on the cooling demand. However, the absence of effective cooling strategies is one of the drawbacks limiting the extensive use of these systems. A well defined control strategy should account for plant's contribution on cooling and humidification in the control algorithm. This study compared the accuracy of three evapotranspiration models using measured values from greenhouse grown pepper plants. The results showed that Stanghellini model (R²=0.93) predicted measured evapotranspiration rates slightly better than Penman-Monteith (R²=0.84) and Takakura models (R²=0.79). Furthermore, a computer simulation was developed to compare a proposed control algorithm for VPF to a typical on/off fixed pressure fogging system based on vapor pressure deficit (VPD). Results showed that VPD based fixed pressure fogging strategy consumed more water and energy compared to the VPF system. Cycling of the pump was smaller and higher stability of temperature and relative humidity were achieved by the operation of the VPF system.

Greenhouse crop production systems have been established throughout the world, including arid and semi-arid regions, to fulfill a market demand of locally grown produce consistently through the year. In these particular regions while they have the advantage of sunshine year-round, production during the summer is a challenge due to elevated air temperatures. Fog systems have proven to be a good economical alternative for evaporative cooling while potentially providing a more uniform environment when compared to fan and pad systems. High-pressure fogging systems equipped with variable frequency drives can be operated at different pressures to meet the varying cooling demands during the day. This feature adds the flexibility of varying the fog flow rate by operating at lower pressures or by changing the number of working fog lines accordingly to the cooling demands. These systems may offer the potential advantage of energy and water saving by operating at a low frequency while providing the proper amount of fog accordingly to the cooling loads. A variable pressure fogging systems operating in the range of 4.8 to 10.3 MPa (700 to 1500 psi) was recently installed in a greenhouse at the University of Arizona Controlled Environment Agriculture Center (UA-CEAC) for the purpose of developing advanced control strategies for optimum greenhouse environments. This study experimentally evaluated the dynamics of air and canopy temperatures, crop evapotranspiration rates, and climate uniformity in the greenhouses working under various fogging system operational pressures and greenhouse side/roof vent opening configurations.

Cross-border smuggling tunnels enable unmonitored movement of people, drugs and weapons and pose a very serious threat to homeland security. Recently, Klar and Linker (2009) [SPIE paper No. 731603] presented an analytical study of the feasibility of a Brillouin Optical Time Domain Reflectometry (BOTDR) based system for the detection of small sized smuggling tunnels. The current study extends this work by validating the analytical models against real strain measurements in soil obtained from small scale experiments in a geotechnical centrifuge. The soil strains were obtained using an image analysis method that tracked the displacement of discrete patches of soil through a sequence of digital images of the soil around the tunnel during the centrifuge test. The results of the present study are in agreement with those of a previous study which was based on synthetic signals generated using empirical and analytical models from the literature.

In recent years open-path FTIR systems (active and passive) have demonstrated great potential and success for monitoring air pollution, industrial stack emissions, and trace gas constituents in the atmosphere. However, most of the studies were focused mainly on monitoring gaseous species and very few studies have investigated the feasibility of detecting bio-aerosols and dust by passive open-path FTIR measurements. The goal of the present study was to test the feasibility of detecting a cloud of toxic aerosols by a passive mode open-path FTIR. More specifically, we are focusing on the detection of toxic organophosphorous nerve agents for which we use Tri-2-ethyl-hexyl- phosphate as a model compound. We have determined the compounds' optical properties, which were needed for the radiative calculations, using a procedure developed in our laboratory. In addition, measurements of the aerosol size distribution in an airborne cloud were performed, which provided the additional input required for the radiative transfer model. This allowed simulation of the radiance signal that would be measured by the FTIR instrument and hence estimation of the detection limit of such a cloud. Preliminary outdoor measurements have demonstrated the possibility of detecting such a cloud using two detection methods. However, even in a simple case consisting of the detection of a pure airborne cloud, detection is not straightforward and reliable identification of the compound would require more advanced methods than simple correlation with spectral library.

A quantitative analysis to asses the influence of a non-measured imaginary index spectrum on the extracted real refractive index is presented. The investigation was done on the Mid-IR spectral range, where the ''measured'' imaginary spectrum is defined between 800 and 4500 cm^-1. The influence of bands of various locations and shapes in the non-measured IR spectral region (0-800 cm^-1) on the n values obtained by the computational procedure of the Kramers-Kronig transform was investigated. Additional analysis was conducted to estimate the relevance of different assumptions that are commonly made with regard to the non-measured range (e.g. linear extrapolation or the effect of uncertainty in the precise band location). The results show that the contribution of an unmeasured band at any wavenumber over(ν, ̃)₀ is well described by a simple function of the band location and over(ν, ̃)₀, regardless of the band shape. Furthermore, the error caused by incorrect band location can also be described by a simple function of the band location, the band location error and over(ν, ̃)₀. The simple functions can be used to estimate the impact that ignoring or misplacing a band will have on the extracted n spectrum, without performing the whole KK integration. These relationships were validated on two data sets of optical constants of crystalline ammonium sulfate and water in the Mid-IR range.

The paper describes a simplified dynamic model of a greenhouse tomato crop, and the optimal control problem related to the seasonal benefit of the grower. A HJB formalism is used and the explicit form of the Krotov-Bellman function is obtained for different growth stages. Simulation results are shown.

A direct method for extracting optical constants in the mid-infrared (IR), using small particle's spectra is presented. The method is based on the direct extraction of the optical constants from the measured spectra using the Rayleigh approach for absorbance cross section of small particles. This was achieved by using an experimental system combining a scanning mobility particle sizing system attached to a long-path IR cell, allowing simultaneous measurements of aerosol size distribution and their IR spectra. The inversion procedure was tested on crystalline ammonium sulfate aerosols, for which high resolution set of optical constants was obtained and were found to be in good agreement with recently published data. Since the extraction of the k and n spectra is deduced from the refractive index dependent complex function, the exact band features can be obtained, unlike the commonly used iterative methods that modify simultaneously both band features and scale of k and n during the calculation procedure. The suggested procedure is simple to apply; nevertheless, it is sensitive to scaling errors of the final constants resulting from uncertainties in total particle volume measurements.

Real-time information about milk composition would be very useful for managing the milking process. Mid-infrared spectroscopy, which relies on fundamental modes of molecular vibrations, is routinely used for off-line analysis of milk and the purpose of the present study was to investigate the potential of attenuated total reflectance mid-infrared spectroscopy for real-time analysis of milk in milking lines. The study was conducted with 189 samples from over 70 cows that were collected during an 18 months period. Principal component analysis, wavelets and neural networks were used to develop various models for predicting protein and fat concentration. Although reasonable protein models were obtained for some seasonal sub-datasets (determination errors <∼015% protein), the models lacked robustness and it was not possible to develop a model suitable for all the data. Determination of fat concentration proved even more problematic and the determination errors remained unacceptably large regardless of the sub-dataset analyzed or of the spectral intervals used. These poor results can be explained by the limited penetration depth of the mid-infrared radiation that causes the spectra to be very sensitive to the presence of fat globules or fat biofilms in the boundary layer that forms at the interface between the milk and the crystal that serves both as radiation waveguide and sensing element. Since manipulations such as homogenisation are not permissible for in-line analysis, these results show that the potential of mid-infrared attenuated total reflectance spectroscopy for in-line milk analysis is indeed quite limited.

This work reports a first step towards the development of an artificial vision system for real - time estimation of cow breathing rate. A simple video camera placed at four meters above the barn floor was used to record 30-second sequences at a frame rate of 30 frames/second. Each sequence was decomposed into consecutive individual frames, which were analyzed to determine the width of the cow close to the regio abdominis lateralis. The contour of the "cow" object within each image was extracted using the "active contour" approach after rotating and translating the image to bring the cow to a standard location within the frame. After extracting the cow's contour, the width of the cow was calculated at a predetermined location. Finally, the breathing rate of the cow was estimated by performing fast Fourier transform (FFT) of this signal and identifying the FFT coefficient corresponding to the breathing rate. The results showed that in most cases this system yielded good estimate of the cow breathing rate. However, most of the cows used in this study were mostly white and much poorer results were observed with mostly black cows due to poor image contrast.

Nitrification and mineralization of organic nitrogen (N) are important N transformation processes in soil, and mass spectrometry is a suitable technique for tracing changes of¹⁵N isotopic species of mineral N and estimating the rates of these processes. However, mass spectrometric methods for tracing N dynamics are costly, time consuming, and require long and laborious preparation procedures. This study investigates midinfrared attenuated total reflection (ATR) spectroscopy as an alternative method for detecting changes in¹⁴NO₃-N and¹⁵NO₃-N concentrations. There is a significant shift of the v₃ absorption band of nitrate according to N species, namely from the 1275 to 1460 cm^-1 region for¹⁴NO₃^- to the 1240-1425 cm^-1 region for¹⁵NO₃^-. This shift makes it possible to quantify the N isotopes using multivariate calibration methods. Partial least squares regression (PLSR) models with five factors yielded a determination error of 6.7-9.2 mg N L^-1 for aqueous solutions and 5.9-7.8 mg N kg^-1 (dry soil) for pastes of a Terra rossa soil. These PLSR models were used to monitor the changes of¹⁵NO₃-N and ¹⁴NO₃-N content in the same Terra rossa soil during an incubation experiment in which [¹⁵NH₄]₂SO ₄ was applied to the soil, allowing the estimation of the contributions of applied N and mineralized N to the net nitrification rate, the potential losses of the applied¹⁵NH₄-N, and the net mineralization of soil organic N.

Despite the advantages of fogging systems over more traditional cooling methods, this relatively new technique is not widely applied in practice. The major reason for this is the shortage of advanced control algorithm for controlling the fogging system together with ventilation. This work presents a MIMO control formulation of the temperature and humidity tracking problem. The control algorithm is tested inside a small experimental greenhouse. The actuators of the greenhouse consist of variable-speed fans together with two high pressure water lines. By defining the above two actuators to be the control inputs and the outside solar radiation, dry-bulb temperature and relative humidity as the disturbances, a MIMO control problem is formulated and solved under the framework of the H^∞ loop shaping optimization technique. This yields a robust multivariable (2×2) sub-optimal controller which succeeds to maintain the dry-bulb temperature and relative humidity inside a small experimental greenhouse within ±2°C and ±10% of their respective set points.

This work presents the design and validation of a robust controller for regulating greenhouse temperature and humidity via forced ventilation and fogging. The control problem is formulated under the H^∞ framework, which takes into account the strong temperature-humidity coupling as well as large parametric uncertainties. Validation experiments conducted in a small experimental greenhouse show that once properly tuned, the controller maintains the temperature and relative humidity within ±2°C and ±10% in their respective setpoints during most of the day.

Cross-borders smuggling tunnels enable unmonitored movement of people, drugs and weapons and pose a very serious threat to homeland security. Recent advances in strain measurements using optical fibers allow the development of smart underground security fences that could detect the excavation of smuggling tunnels. This paper presents the first stages in the development of such a fence using Brillouin Optical Time Domain Reflectometry (BOTDR). In the simulation study, two different ground displacement models are used in order to evaluate the robustness of the system against imperfect modeling. In both cases, soil-fiber interaction is considered. Measurement errors, and surface disturbances (obtained from a field test) are also included in the calibration and validation stages of the system. The proposed detection system is based on wavelet decomposition of the BOTDR signal, followed by a neural network that is trained to recognize the tunnel signature in the wavelet coefficients. The results indicate that the proposed system is capable of detecting even small tunnel (0.5m diameter) as deep as 20 meter.

This paper presents a relatively simple approach for determining optimal paths for car-like vehicles operating in orchards. The method relies on the vehicle's configuration space and uses the well-accepted A* algorithm to determine the optimal path, taking into account constraints that are specific to the type of vehicle and environment considered: limited steering angle, limited range of pitch and roll angles permissible, forward motion preferable, and frequent turning not desirable. All these constraints are expressed as penalty terms weighted via user-defined parameters that reflect the type of path implicitly sought by the user. A number of examples are presented, which show the ability of the procedure to find the optimal path in various situations and illustrate how the weighting parameters can be used to shape the solution.

Idowu et al., in this edition of Plant and Soil describe a welcome new tool that allows farmers to optimize the management of their fields. It is based on a holistic approach that integrates assays of soil physical, chemical and biological parameters into an easily understandable, low-cost and quantitative estimate of soil quality or "health". It can facilitate tracking and understanding of the short and long term effects of different management approaches on soil (and crop) health. Most of the assays used are based on traditional, labour-intensive techniques. The tool could be improved through employment of assay technologies that are under development. For example, Idowu et al., in this edition investigated the use of near infra-red spectroscopy as a replacement for some of the traditional soil assays. Similarly, future developments in remote spectroscopic assays, using satellite, airborne or land-based platforms may facilitate further improvement.

Fourier transform infrared (FT-IR) attenuated total reflection (ATR) spectroscopy was used to discriminate five commonly encountered soil-borne fungi that cause severe economic damage to agriculture: Colletotrichum, Fusarium, Pythium, Rhizoctonia, and Verticillium. Contrary to previous studies related to microorganism discrimination using FT-IR-ATR spectroscopy, the pathogen samples were not dried on the ATR crystal, which is a time-consuming operation. Rather, after removing some pathogen filaments from the solution using tweezers, these were placed directly on a flat ATR crystal and pressure was applied using a pressure clamp. Following water subtraction, baseline correction, and normalization of the spectra, principal component analysis was used as a data-reduction step and canonical variate analysis was used for discrimination. Discrimination was performed at the genus level and at the strain level for Colletotrichum. For discrimination between the five fungi at the genus level, the success rate for the validation samples ranged from 75% to 89%. For discrimination between the two Colletotrichum strains, the success rate was 78%. Comparison with spectra of similar fungi dried on the ATR crystal showed that both types of spectra were very similar, indicating that drying the samples on the ATR crystal is not required and can be replaced by mathematical post-processing of the spectra. For routine analyses that involve rapid screening of very large amounts of samples, this approach allows for increasing significantly the number of samples that can be analyzed daily.

Numerous works have demonstrated that ion exchange membranes can be used for accurate determination of the availability of nitrate and other key nutrients in soils. After letting the ion exchange membrane interact with the soil for a known period of time, the membrane is typically immersed in a strong reagent to desorb the ions and the resulting solution is analyzed by standard chemical methods. The present study shows that mid-infrared photoacoustic spectroscopy can be used to estimate directly the amount of nitrate sorbed onto such membranes and could replace advantageously the standard chemical analysis. The study was conducted with two commercially available membranes, and in both cases the average determination error achieved with a simple partial least squares model was approximately 1.6-1.8 μeq, which under the specific experimental conditions corresponds to approximately 4.5-5.0 mg[N]/kg[dry soil]. Such errors are about 30% larger than those reported in a previous study in which the membrane was analyzed by transmittance spectroscopy. However, the present method is suitable for a much wider range of membranes, such as those commonly used for water treatment and that are too thick for transmittance measurements. For such membranes, photoacoustic spectroscopy is a very cheap and rapid alternative to standard chemical analysis.

This study investigated the use of Fourier transform infrared photoacoustic spectroscopy (FTIR-PAS) for rapid identification of agricultural soil samples. The PAS spectra of 166 air-dried samples belonging to five Mediterranean soil types most common in Israeli agriculture were recorded. The various soil types exhibited distinctive mid-IR bands, especially around the 2900-3700 cm^{- 1}, 2500-2550 cm^{- 1}, 1800-2050 cm^{- 1} and 900-1600 cm^{- 1} regions. Following smoothing and normalization of the spectra, principal component analysis (PCA) was used to reduce the dimensionality of the data and the PCA scores were used in classifiers based either on linear discriminant analysis or on probabilistic neural networks. The two classifiers based on four PCA scores yielded very similar results and correctly identified over 96% of the 77 validation samples. Comparison with the attenuated total reflectance (ATR) spectra of similar soils used in a previous study showed that the PAS spectra contained more information than the ATR ones, both in terms of the number of soil-specific bands and in terms of the bands' distinctiveness. The results clearly show that FTIR-PAS can be used for rapid soil identification and the abundance of information in the PAS spectra indicates that this technique could be further developed to assess important soil features.

A model-free statistical method for detecting failures in dynamic systems controlled via output feedback is described. In such systems, the impact of the failure on the output is minimized but the failure causes a change in the control signal, which can be detected. Assuming that historical data of the process with and without failure is available, optimization theory can be used to determine detection thresholds that ensure any desired level of false alarms or detection rate. The proposed method is iDustrated on a hypothetical bio-reactor and the results show that the proposed method allows for both rapid detection of the faliure and continued operation despite the failure.

A simple algorithm for planning the optimal path for a car-like vehicle operating in outdoor environments is presented. The algorithm is based on the well-known A approach, which is applied in the vehicle's "configuration space" and is adapted to take into account constraints that are specific to the type of vehicle and environment considered: limited steering angle, limited range of pitch and roll angles permissible, forward motion preferable, frequent turning not desirable. The performance of the algorithm is illustrated on several examples based on randomly generated maps as well as on more realistic orchard maps.

Linearized equations of motion for a motorcycle with small roll angles are derived and used to design a robust cascade control scheme that stabilizes the motorcycle over a range of speeds. Stabilization is achieved by measuring the roll angle and its rate of change, and controlling the steering torque. The approach is validated via simulations and experiments performed with a radio-controlled scooter.

The need for rapid and inexpensive techniques for soil characterization has led to the investigation of modern technologies, and in particular those based on reflectance spectroscopy. While near-infrared has been traditionally used, midinfrared in the 400-4000 cm-1 range is becoming increasingly common due to the specificity of the absorbance bands in this spectral range. The present work discusses two methods based on mid-infrared spectroscopy for soil classification: attenuated total reflectance (ATR) and photoacoustic spectroscopy. The ATR method requires a soil sample close to water saturation, and as a result only the 800-1600 cm-1 interval of the spectrum yields a useful signal. Typical ATR soil spectra consist mostly of several broad bands in the 800-1200 cm-1 region and a calcium carbonate band around 1450 cm-1. By comparison, photoacoustic measurements are conducted with air-dried samples, and the photoacoustic spectra exhibit a larger number of clearly-defined bands. Both methods were tested on data sets containing over 100 samples of various soils commonly used in Israeli agriculture. Data analysis was conducted by wavelet decomposition and neural network classifiers. Very good classification performances were achieved, with correct classification rates of the validation samples typically above 95%.

This study investigates the use of photoacoustic spectroscopy (PAS) for rapid soil analysis. Photoacoustic spectroscopy requires very minimal sample preparation (air-drying), which is a major advantage compared to the more traditional transmittance technique, which requires timeconsuming preparation of pellets. The amount of information contained in the PAS spectra appears to be similar to that contained in transmittance spectra, and the PAS spectra exhibit a large number of bands that can be associated with various soil constituents such as quartz, calcium carbonate, and various types of clay. Comparison with attenuated total reflection (ATR) spectra of saturated soil pastes shows that the PAS spectra provide much more information than the ATR spectra due to the strong water bands present in the latter. PAS quantitative analysis of clay, calcium carbonate, and organic matter is presented, with respective determination errors of ∼12% clay, ∼5% CaCO₃, and ∼0.2% organic matter.

Direct determination of nitrate and soil moisture can significantly improve N-application management and thus reduce N-derived environmental pollution related to agriculture. Several studies have shown that Fourier transform infrared attenuated total reflectance (FT-IR/ATR) spectroscopy could be used to estimate the nitrate content of standardized soil pastes. Paste standardization appeared to be the main obstacle to in situ application of this approach, and the present study shows how FT-IR/ATR can be used to estimate both water content and nitrate concentration of field soil samples. Water content and nitrate concentration are determined sequentially using two subsamples of the initial soil sample. An a priori determined amount of highly concentrated nitrate solution is added to the first subsample and the ATR spectrum of this paste is used to estimate the sample water content. It is then possible to calculate the amount of water that should be added to the second subsample so that the resulting paste is very close to the ideal standard paste. Nitrate concentration, mg [N]/kg [dry soil], is estimated using the FT-IR/ATR spectrum of this second paste. Results are presented for a laboratory experiment with four agricultural soils, as well as for a field trial with a calcareous soil. For water content, the determination errors range from 0.01 to 0.02 g [water]/g [dry soil]. For nitrate concentration, the errors for three of the soils range from 5.9 to 8.4 mg [N]/kg [dry soil], while for the fourth, calcareous clay soil, the determination error is 13.6 mg [N]/kg [dry soil]. The determination errors obtained for the field trial are similar to the ones obtained for a similar soil under laboratory conditions, which shows the potential usefulness of the approach for improving N-application management and reducing environmental pollution.

Accurate simulation models for short-term (∼hours) changes in hydroponic crop growth and nitrate uptake are needed for rapid fault detection in hydroponic systems. Comparison between model-predicted and measured values for crop growth and nitrate uptake is proposed as the basis for such a fault detection system. To this end, the Nitrate Control in Lettuce (NiCoLet) model was used to evaluate both short- and long- term changes in growth and nitrate accumulation. Three replicated experiments were conducted with lettuce (Lactuca sativa L. cv. Flandria), including: (1) plants subjected to either low or high nitrate concentration treatments for model calibration; (2) collection of growth data every 2 days for model validation; and (3) frequent (every 4 mol m^-2 of accumulated light) collection of growth and shoot nitrate concentration data to validate short-term predictions. After a minor modification (maximum nitrogen uptake rate restricted) and calibration, the NiCoLet model accurately simulated lettuce crop growth and nitrate uptake on a long-term basis and provided evidence of short-term behaviour, including statistically significant predictions of diurnal patterns. This is a first step in realising fault detection systems based on mechanistic simulation models.

This study investigates the combined use of an anion exchange membrane and transmittance mid-infrared spectroscopy for determining nitrate concentration in aqueous solutions and soil pastes. The method is based on immersing a small piece (2 cm²) of anion exchange membrane into 5 mL of solution or soil paste for 30 minutes, after which the membrane is removed, rinsed, and wiped dry. The absorbance spectrum of the charged membrane is then used to determine the amount of nitrate sorbed on the membrane. At the levels tested, the presence of carbonate or phosphate does not affect the nitrate sorption or the spectrum of the charged membrane in the vicinity of the nitrate band. Sulfate affects the spectrum of the charged membrane but does not prevent nitrate determination. For soil pastes, nitrate sorption is remarkably independent of the soil composition and is not affected by the level of soil constituents such as organic matter, clay, and calcium carbonate. Partial least squares analysis of the membrane spectra shows that there exists a strong correlation between the nitrate charge and the absorbance in the 1000-1070 cm^-1 interval, which includes the vi nitrate band located around 1040 cm^-1. The prediction errors range from 0.8 to 2.1 μeq, which, under the specific experimental conditions, corresponds to approximately 2 to 6 ppm N-NO₃^- on a solution basis or 2 to 5 mg [N]/kg [dry soil] on a dry soil basis.

Mid-infrared (mid-IR) spectroscopy experiments were conducted to detect added nitrate in various soil types both in the laboratory and field. Soil pastes from ten different soils, including sandy loam, clay, and peat soils, were analysed for soil nitrate contents using the Fourier transform infrared (FTIR) attenuated total reflectance (ATR) technique. Nitrate concentrations for the laboratory experiments varied from approximately 0-1000 ppm. NO₃-N while concentrations for the field experiments varied from approximately 0-140 ppm. NO₃-N. Three-dimensional plots were created by graphing the wavelet deconvoluted values at 32 scales for each sample. From each plot, the volume of the nitrate peak was determined and correlated to nitrate concentrations. Results of the laboratory experiments indicated values for the coefficient of determination R² as high as 0·99 and standard errors as low as 24 ppm. NO₃-N for soil-specific calibrations. Results of the field experiments gave values for R² as high as 0·98 and standard errors as low as 5 ppm NO₃-N for soil-specific calibrations. An alternative technique to determine nitrate content was developed in which wavelet analysis was used to identify a few wavenumbers at which interferences from other ions were minimal. This method produced calibration equations that were soil independent and gave superior results to those obtained based on correlating wavelet deconvoluted volumes to nitrate concentrations.

Direct determination of nitrate in soil is required for improving N-application management, which would help reduce soil and water pollution. Several works have demonstrated that mid-infrared Fourier transform infrared attenuated total reflectance (FTIR-ATR) spectroscopy could be used to determine nitrate concentration in soil pastes. The present work further investigates this approach, and proposes to combine nitrate determination with soil identification in order to improve the determination accuracy. The study focuses on soils commonly used for agriculture, which are classified according to soil taxonomy and their carbonate and clay contents. Soil identification is investigated using the 800-1200 cm ^-1 and 1250-1550 cm ^-1 intervals of the spectrum, using either cross-correlation with a reference library or principal component analysis (PCA) decomposition followed by neural network (NN) classifier. When applied to the 1250-1550 cm ^-1 interval, the PCA-NN method leads to correct identification of all the samples, while the other approaches lead to poorer results. Nitrate determination is achieved using several partial least-squares regression models, each model being associated with a soil type. Determination errors range from 6·2 to 13·5 mg[N]/kg[dry soil], depending on the soil type, with the lowest errors for light sandy soils. These determination errors are appreciably smaller than those obtained using a single model calibrated using all the data (19·1 mg[N]/kg[dry soil]).

Water deficit caused by addition of polyethylene glycol 6000 at -0.5 MPa water potential to well-aerated nutrient solution for 48 h inhibited the elongation of maize (Zea mays) seedling primary roots. Segmental growth rates in the root elongation zone were maintained 0 to 3 mm behind the tip, but in comparison with well-watered control roots, progressive growth inhibition was initiated by water deficit as expanding cells crossed the region 3 to 9 mm behind the tip. The mechanical extensibility of the cell walls was also progressively inhibited. We investigated the possible involvement in root growth inhibition by water deficit of alterations in metabolism and accumulation of wall-linked phenolic substances. Water deficit increased expression in the root elongation zone of transcripts of two genes involved in lignin biosynthesis, cinnamoyl-CoA reductase 1 and 2, after only 1 h, i.e. before decreases in wall extensibility. Further increases in transcript expression and increased lignin staining were detected after 48 h. Progressive stress-induced increases in wall-linked phenolics at 3 to 6 and 6 to 9 mm behind the root tip were detected by comparing Fourier transform infrared spectra and UV-fluorescence images of isolated cell walls from water deficit and control roots. Increased UV fluorescence and lignin staining colocated to vascular tissues in the stele. Longitudinal bisection of the elongation zone resulted in inward curvature, suggesting that inner, stelar tissues were also rate limiting for root growth. We suggest that spatially localized changes in wall-phenolic metabolism are involved in the progressive inhibition of wall extensibility and root growth and may facilitate root acclimation to drying environments.

A four-compartment lettuce model designed to predict the effects of abrupt changes in nitrogen availability is described. Sensitivity analysis is used to determine the parameters to which the predictions are most sensitive. These parameters are estimated using data from experiments in which nitrogen availability was changed abruptly. The model predicts quite well the main observations associated with interruption of nitrogen supply: reduced shoot growth, increased root-to-shoot ratio (RSR), fast depletion of nitrate, and increased dry matter content. The shoot of plants deprived of nitrogen for 25 days is seven to ten times smaller than the shoot of unstressed plants, while their root-to-shoot ratio is about 12 times larger. Restoring nitrogen supply to previously N-starved plants has opposite effects. A simplified model, obtained by imposing a fixed RSR, gives predictions of shoot fresh mass, dry matter content, nitrate concentration and reduced-N content similar to the predictions of the model with variable RSR, except for extreme cases in which the plants are N-deprived throughout the whole growing period. This indicates that proper modeling of root-shoot partitioning is not essential for predicting the effect of abrupt changes of nitrogen availability on shoot growth and internal plant composition.

The use of mid-infrared attenuated total reflectance (ATR) spectroscopy enables direct measurement of nitrate concentration in soil pastes, but strong interfering absorbance bands due to water and soil constituents limit the accuracy of straightforward determination. Accurate subtraction of the water spectrum improves the correlation between nitrate concentration and its ν₃ vibration band around 1350 cm^-1. However, this correlation is soil-dependent, due mostly to varying contents of carbonate, whose absorbance band overlaps the nitrate band. In the present work, a two-stage method is developed: First, the soil type is identified by comparing the "fingerprint" region of the spectrum (800-1200 cm^-1) to a reference spectral library. In the second stage, nitrate concentration is estimated using the spectrum interval that includes the nitrate band, together with the soil type previously identified. Three methods are compared for estimating nitrate concentration: integration of the nitrate absorbance band, cross-correlation with a reference spectrum, and principal component analysis (PCA) followed by a neural network. When using simple band integration, the use of soil specific calibration curves leads to determination errors ranging from 5.5 to 24 mg[N]/kg[dry soil] for the mineral soils tested. The cross-correlation technique leads to similar results. The combination of soil identification with PCA and neural network modeling improves the predictions, especially for soils containing calcium carbonate. Typical prediction errors for light non-calcareous soils are about 4 mg[N]/kg[dry soil], whereas for soils containing calcium carbonate they range from 6 to 20 mg[N]/kg[dry soil], which is less than four percent of the concentration range investigated.

A method for automatic analysis of high-dimensional spectra is presented. The method is based on a so-termed extended auto-associative neural network, which is an auto-associative (bottleneck) neural network with an additional output that corresponds to the response variable of interest. The input of the neural network consists of the absorbance at selected wavelengths. Once the model has been calibrated, the contribution of each input is estimated, the least contributing input and the corresponding output are removed (pruned), and the training procedure is repeated. This procedure leads to a compact non-linear mapping between the absorbance at a few wavelengths and the response variable of interest. The auto-associative architecture of the model prevents the overparametrization and overfitting that occur in straightforward mapping between spectra and variable of interest. Overparametrization and overfitting are further reduced by retaining only the inputs required for predicting the variable of interest. The method is illustrated on two case studies involving mid-infrared absorbance spectra. The results are compared to those obtained with partial least squares (PLS) and principal component analysis (PCA) followed by a neural network. In both case studies, the proposed approach leads to models with fewer parameters and smaller prediction errors.

Experiments were conducted involving adding nitrate to various soil types both in the laboratory and field. Soil pastes from ten different soils, including sandy loam and clay soils, were analyzed for soil nitrate content using the Fourier Transform Infrared (FTIR) Attenuated Total Reflection (ATR) technique. Three of the soil types were known to be calcareous, thus containing large carbonate amounts. Nitrate concentrations for the laboratory experiments varied from approximately 0 to 1000 ppm NO₃-N while concentrations for the field experiments varied from approximately 0 to 100 ppm NO ₃-N. Wavelet analysis was applied to the spectra obtained from the soil pastes in order to allow for calibration equations to be developed to predict nitrate concentrations. Three-dimensional plots were created by graphing the wavelet deconvoluted values for each sample. From each plot, the volume of the nitrate peak was determined. Calibrations equations were developed by correlating the volume of these peaks to nitrate concentrations. High correlation values were found for all the soil types. Results of the laboratory experiments indicated R²-values as high as 0.99 and standard errors as low as 24 ppm NO₃-N. Results of the field experiments gave R ²-values as high as 0.98 and standard errors as low as 5 ppm NO ₃-N. In both cases, the slopes of the calibration equations depended on the soil types, indicating site specific calibration may be needed. Another technique that involved correlating absorbances at a fixed number of wavenumbers to nitrate concentrations was used to develop calibration equations. This method led to calibration equations that were soil independent and gave superior results to those obtained based on correlating wavelet deconvoluted volumes to nitrate concentrations. Also, these calibration equations allowed for the calcareous soils to be pooled with the noncalcareous soils for predicting nitrate concentrations.

This work presents a user-friendly, Internet-based, version of the NiCoLet lettuce model that includes most of the versions currently available. The potential use of this simulation tool is as a decision-support aid for growers and extension agents, as well as for use in teaching students crop simulation modelling. The model, which was previously limited to a handful of researchers, is made available to all Internet users using the Matlab Webserver toolbox. Matlab Webserver is a promising method for disseminating powerful decision-support tools.

A three-compartment lettuce model designed to predict growth and nitrate concentration under severe nitrogen stress conditions is described and calibrated using data of N-limitation experiments. Since the limited amount of data does not allow calibration of all the model parameters, sensitivity analysis was used to determine which parameters should be estimated using this data. The calibrated model predicts quite well the main observations associated with severe nitrogen stress: reduced growth, fast depletion of nitrate, gradual reduction of reduced-N content, and considerable increase of dry matter content. After being deprived of nitrate for 21 days, stressed plants are five to ten times smaller than unstressed plants, and their dry matter content is two to three times higher than normal. Nitrate concentration of N-stressed plants drops to an insignificant level within three days, while reduced-N content decreases steadily throughout the experiment. All these observations are correctly predicted by the model.

This paper investigates the possibility of determining nitrate concentration in soil pastes using spectral absorbance at several fixed wavebands. A three-step procedure for determining the most appropriate wavebands, as well as their width, is described. This procedure is applied to a dataset that includes eleven soils with various nitrate concentrations ranging from 0 to approximately 150 mg [N]/kg [dry soil]. The results show that nitrate concentration can be determined quite acceptably using only four 12 cm ^-1 wide wavebands, centered at 1280, 1330, 1379, and 1430 cm ^-1. The prediction errors range from approximately 1.5 to 14.0 mg [N]/kg [dry soil], depending on soil composition and moisture content, with the lighter and more vulnerable (pollution-wise) soils having errors inferior to 10 mg [N]/kg [dry soil]. These results are similar to results obtained by applying partial least square to the 'continuous' spectrum, and indicate that the development of a soil nitrate attenuated total reflectance (ATR) sensor based on a few fixed mid-infrared (MIR) wavebands could be considered.

This study investigates the potential use of attenuated total reflectance spectroscopy in the mid-infrared range for determining protein concentration in raw cow milk. The determination of protein concentration is based on the characteristic absorbance of milk proteins, which includes 2 absorbance bands in the 1500 to 1700 cm^-1 range, known as the amide I and amide II bands, and absorbance in the 1060 to 1100 cm^-1 range, which is associated with phosphate groups covalently bound to casein proteins. To minimize the influence of the strong water band (centered around 1640 cm ^-1) that overlaps with the amide I and amide II bands, an optimized automatic procedure for accurate water subtraction was applied. Following water subtraction, the spectra were analyzed by 3 methods, namely simple band integration, partial least squares (PLS) and neural networks. For the neural network models, the spectra were first decomposed by principal component analysis (PCA), and the neural network inputs were the spectra principal components scores. In addition, the concentrations of 2 constituents expected to interact with the protein (i.e., fat and lactose) were also used as inputs. These approaches were tested with 235 spectra of standardized raw milk samples, corresponding to 26 protein concentrations in the 2.47 to 3.90% (weight per volume) range. The simple integration method led to very poor results, whereas PLS resulted in prediction errors of about 0.22% protein. The neural network approach led to prediction errors of 0.20% protein when based on PCA scores only, and 0.08% protein when lactose and fat concentrations were also included in the model. These results indicate the potential usefulness of Fourier transform infrared/attenuated total reflectance spectroscopy for rapid, possibly online, determination of protein concentration in raw milk.

The NICOLET model has been developed to predict the growth and nitrate content of greenhouse lettuce. Four single-organ versions have been developed: [1] abundant supply of nitrogen (1998), [2] mild N-stress (1999), [3] severe N-stress (2003), and [4] ontogenetic changes of organic-N and water content (in preparation). The 'abundant-N' model [1] and the 'mild-stress' model [2] have two compartments: 'structure' and 'vacuole', while the 'severe-stress' model [3] requires a third compartment: 'excess-carbon', and the 'ontogenetic' model [4] has separate 'metabolic' and 'support' structural sub-compartments. The main special features of the NICOLET model are [1] the osmotica balance of the 'vacuole', where nitrate and hydrocarbons play a complementary role in maintaining a constant osmotic potential, [2] the excess-carbon compartment, where 'dry' carbon compounds are stored, and [3] the sub-division of the 'structure' into sub-compartments of different compositions. Loosely speaking, the first feature controls the nitrate concentration, the second controls the organic-N and water contents, and the third controls the ontogenetic changes. The NICOLET model has been able to mimic 'normal' seasonal variations of nitrate content, as well as the effects of drastic N-stress treatments. These results are illustrated by comparing measured data with model-simulations. Accurate prediction of nitrate concentration is difficult, due to its sensitivity to changes in the environment. Exact control of nitrate under commercial conditions may require transient corrective measures, such as N-interruption, in conjunction with a good plant-nitrate monitoring system.

This paper investigates the use of Fourier transform infrared (FT-IR) attenuated total reflectance (ATR) spectroscopy as a fast and simple way for direct determination of nitrate concentration in soil pastes, which would assist precision fertilizer placement and reduce nitrate pollution. Eight types of soils are investigated, with nitrate concentrations ranging from 0 to 1000 ppm-N. The spectral region around the nitrate band (1300-1550 cm^-1) is analyzed by (1) principal component regression (PCR), (2) partial least squares (PLS), and (3) cross-correlation with reference libraries that include spectra of pure ions and/or soils. The main obstacle to accurate nitrate measurement appears to be an interfering band present in calcareous soils. This band, which may be due to carbonate, is located around 1450 cm^-1 and overlaps with the nitrate band centered around 1370 cm^-1. For non-calcareous soils, and in particular for light sandy agricultural soils, PLS and cross-correlation with a reference library containing only spectra of ions in water give similar results (about 8 ppm-N on dry soil basis), while PCR leads to slightly poorer results. When calcareous soils are included in the analysis, the prediction errors are about twice as large. In this case, the best results are obtained using PLS, followed by PCR, while cross-correlation with reference libraries leads to poorer results.

Greenhouse operation and inside climate strongly depend on the outside weather. This implies that at least a year of data collection is required to cover the whole operational domain. Greenhouse-climate models calibrated with data limited to only a small region of the operating domain (weather and control), may therefore, produce erroneous predictions when applied to unfamiliar conditions. A comparison is made between the performance of three types of models trained with several seasonal sub-sets of data: (1) black-box (BB) sigmoid neural network (NN) trained only with in situ data, (2) hybrid physical-RBF (radial basis function) model, and (3) sigmoid neural network trained with a combination of in situ data and synthetic data generated with a physical model (termed 'prior-K sigmoid model'). The BB sigmoid model gives the best predictions within the training domain, but performs very badly outside it. On the other hand, the hybrid and prior-K sigmoid models produce useful predictions over the whole operating domain, although they are slightly less accurate within the training domain.

Root disease, particularly disease caused by Pythium aphanidermatum, is a major problem in commercial production of hydroponic spinach (Spinacia oleracea L.). Cooling the nutrient solution has been shown to suppress root disease, but it has the disadvantage of not eliminating completely the pathogen from the hydroponic system, and of reducing the shoot growth rate. Another strategy for disease management involves early disease detection, which, ideally, would allow for corrective measures to be taken even before plants exhibit severe disease symptoms, such as a reduction in shoot dry mass. This study attempted to detect plant stress caused by root disease through continuous monitoring of dissolved oxygen in temperature-controlled hydroponic ponds. The hypothesis is that root disease affects the exchange of oxygen between the plants and the nutrient solution, and this change is large enough to make early disease detection feasible, The results indicate that plants inoculated with P. aphanidermatum in 18°C, 24°C and 30°C ponds have noticeably lower levels of oxygen consumption than their non-inoculated counterparts. While the timing of this reduction depends on pond temperature, reduction of the oxygen consumption rate appears to occur before a significant reduction in shoot dry mass is observed, indicating the potential usefulness of the approach.

Water stress of a greenhouse crop results in a lower-than-normal transpiration and photosynthesis rates, which affects the humidity and CO₂ concentration of the greenhouse air. This change may be detected and used for stress diagnosis if the signal-to-noise ratio is sufficiently large. Under midday conditions, when water stress is likely to occur, the greenhouse is often ventilated, resulting in a rather small signal-to-noise ratio. It is suggested here that the signal level may be increased by exciting the system, namely by suspending the ventilation for a short period of time. The transient response of temperature, humidity and CO₂ concentration to the abrupt change of ventilation may be monitored, analysed and compared to a modelled normal response for the same greenhouse with an unstressed crop. A significant discrepancy between the two would indicate that the crop is stressed. Lawn, serving as a crop substitute, has been submitted to several irrigation-drought sequences in a small experimental greenhouse. Each day, in the early afternoon, ventilation was suspended for three periods of 30 min each. Differences in the response of CO₂ concentration enabled the detection of all stress periods as soon as the rate of transpiration decreased to below normal. The humidity response was slightly less sensitive, and the temperature response was the least sensitive. Water-stress detection via the excitation method was considerably better than detection based on observations at full ventilation. 2003 Silsoe Research Institute. All rights reserved. Published by Elsevier Science Ltd.

This work presents an adaptation of the unknown input observer (UIO) failure diagnosis method to non-linear input output systems, and its application to an experimental greenhouse. By comparison with the UIO results for linear systems, the present developments lead to time-varying observers, which depend on the partial derivatives of the model. Conditions for perfect failure/disturbance decoupling similar to the ones obtained in the linear case, are derived. The failures investigated in the greenhouse include sensor failures, actuator failures, and changes in the greenhouse dynamics. Using the method developed, all the failures investigated are detected and isolated, with typical detection time ranging from less than 30 min for ventilation failure, to several hours for slowly developing failures. 2002 Elsevier Science Ltd. All rights reserved.

A model-based method for the detection and identification of single failures in greenhouses is presented. The method relies solely on climatic measurements currently available in commercial greenhouses, and combines hybrid physical/neural-network models with robust failure detection and identification theory for non-linear systems. Both sensor and actuator failures are considered, and the detection of crop water stress is also addressed. The first part of the paper is devoted to a simulation study to estimate the economical cost of the failures. In the second part, a method for robust failure detection and identification is described and tested on experimental data. The application of the method to experimental data shows that, under most circumstances, the failures are correctly detected and identified, leading to a significant reduction of the losses caused by the failures. (C) 2000 Elsevier Science B.V.

The first part of this paper presents an analysis of the effect of different failures on greenhouse operation. A simulation study shows that failures prevent the greenhouse climate from being controlled optimally, and may result in significant financial losses to the grower. Subsequently, two methods for failure detection and isolation are presented. The first method is based on a comparison between the measured greenhouse climate and the predictions of a reference model. This comparison is performed according to the Unknown Input Observer approach, which ensures robustness of the diagnosis with respect to noise and modeling errors. The second method is aimed specifically at detecting crop water stress during noontime hours. At such times, the high heat load requires strong ventilation, which reduces the detection capability of the first method. Crop water stress detection is achieved by suspending the ventilation for short periods of time (30 minutes), and comparing the greenhouse climate response to some reference response.

The present work focuses on the control of greenhouse air temperature and CO₂ concentration by means of simultaneous ventilation and enrichment. Such an operation, which may a priori seem contradictory, was shown by several authors to be required to maintain optimal temperature and CO₂ setpoints. The control process is divided into two distinct control loops, the first maintaining the temperature by adjusting the ventilation, and the second maintaining the CO₂ concentration by adjusting the enrichment. The CO₂ concentration controller assumes the ventilation rate to be constant and approximately known over 2-min intervals. Implementation in an experimental greenhouse shows the ability of the controllers to meet the requirements.

CO₂ enrichment in warm climates requires a delicate balance between the need to ventilate and the desire to enrich. Model-based optimization can achieve this balance, but requires reliable models of the greenhouse environment and of the crop response. This study assumes that the crop response is known, and focuses on the greenhouse model. Neural network greenhouse models were trained using data collected over two summer months in a small greenhouse. The models were reduced to minimum size, by predicting separately the temperature and CO₂ concentration, and by eliminating any unessential input. The resulting models not only fit the data well, they also seem qualitatively correct, and produce reasonable optimization results. Using these models, the effect of evaporative cooling on extending the enrichment duration is demonstrated.

The formulation of a simple optimization criterion led to two approaches for the enhancement of CO₂enrichment. Both result in a reduced ventilation rate and hence in less 'waste' of CO₂. One approach is to reduce the energy load on the greenhouse and the other, for which some examples are shown, is to increase the humidity of the exhaust air. Optimal control trajectories showed the existence of three daytime operating regimes, which switch from one to the next as the energy load increases: A: Enrichment without ventilation. B: Simultaneous ventilation and enrichment. C: Ventilation without enrichment. Evaporative cooling increases the duration of regime A at the expense of regime C. Reference to canopy temperature, rather than to air temperature, reduces the apparent advantage of evaporative cooling, except when it is applied at or close to the canopy, or when it prevents water stress.