Data assimilation has been widely used for improvement of crop models’ estimates, for example to incorporate the effects of external events or compensate calibration errors in large areas. The term describes multiple approaches for those who want to take advantage of satellite imagery to reduce uncertainty or improve accuracy of model estimates. Kalman Filters are among the most used methods for achieving these goals. But their use in new contexts, i.e., from open field to protected environments, requires untangling aspects of the pipeline that are often performed in many different ways without guidelines, such as which variables to assimilate or how to ascribe uncertainty to observations or model estimates. This study is then divided in two parts. In the first, we review details on how uncertainty is ascribed on crop model estimates and in observations for applications of the Kalman Filter and three variations of the method, i.e., the Extended, Unscented and Ensemble, as well as which state variables are often updated and the frequency with which assimilation may occur, as well as how these aspects are connected to each other. In the second part, we apply different approaches from the reviewed literature in a greenhouse tomato crop model. We use artificial data with controlled noise levels as well as artificial data generated by simulation using other tomato crop model. We assess the impacts of using different methods and different approaches for ascribing uncertainty in model estimates and in observations, by assimilating artificial observations of fruit and of mature fruit biomass. We note that covariances should not be fixed values, that there are trade-offs between ascribing model uncertainty to the state itself and to other elements of the process, that observation covariance may have been considered disproportionality higher when using some ensemble generation approaches in the EnKF, and that bias in model estimates may lead to worse outcomes even when observations are high-quality ones. While we discussed aspects that should be considered in a new environment, many of them are also important for field crops, and we concluded assimilation should follow an assessment of which variables could be useful for assimilation.

ABSTRACT: Sugarcane mills in Brazil collect a vast amount of data relating to production on an annual basis. The analysis of this type of database is complex, especially when factors relating to varieties, climate, detailed management techniques, and edaphic conditions are taken into account. The aim of this paper was to perform a decision tree analysis of a detailed database from a production unit and to evaluate the actionable patterns found in terms of their usefulness for increasing production. The decision tree revealed interpretable patterns relating to sugarcane yield (R2 = 0.617), certain of which were actionable and had been previously studied and reported in the literature. Based on two actionable patterns relating to soil chemistry, intervention which will increase production by almost 2 % were suitable for recommendation. The method was successful in reproducing the knowledge of experts of the factors which influence sugarcane yield, and the decision trees can support the decision-making process in the context of production and the formulation of hypotheses for specific experiments.

ABSTRACT: The present research hypothesized that the thermal, lighting and acoustic environments in commercial swine farrowing rooms vary over time and from crate to crate. This study was conducted on 27 replicates in two commercial farrowing rooms in North Central Indiana, each equipped with 60 farrowing crates. Temperature, relative humidity, light intensity, sound intensity, and air velocity were continuously monitored and estimated for each crate at the sow level, for 48 h post-farrowing, which is usually a critical period for piglet survivability. Average daily temperature for all the crates in Room 1 was 24.1 ± 2.0 °C, 1.0 °C lower (p < 0.05) than in Room 2. Although the overall mean temperature was similar between rooms and seasons, frequency distribution diagrams revealed that the proportion of time spent within distinct limits of mean daily temperature ranged from 15.0 °C to 28.0 °C and varied substantially between rooms and seasons. Similar results were found for all variables measured in this study. Differences in temperature, relative humidity, light intensity, air velocity, and sound intensity in crates were as high as 9.6 °C, 57 %, 3,847.3 Lx, 0.87 m s–1, and 38.7 dBC, respectively, in the same farrowing room when measured at the same instant. The results of the present research indicate that aspects that go beyond the physical environment of the sows, such as thermal, lighting, and acoustic environment can vary substantially over time and between crates of automatically climate controlled farrowing rooms. These differences should be taken into consideration in production setting and research.

Resumo O objetivo deste trabalho foi avaliar o cultivo da cana-de-açúcar, em análise harmônica aplicada a uma série temporal de índices de vegetação Modis, com a representação dos termos harmônicos. Os dados diários de precipitação pluvial para o Estado de São Paulo foram obtidos do Agritempo, acumulados para o período de 16 dias das composições Modis, ao longo das safras de 2004/2005 a 2011/2012. Utilizaram-se o índice de vegetação por diferença normalizada (NDVI) e o índice de realce da vegetação (EVI) em séries temporais decompostas em termos harmônicos pela análise harmônica. Para visualizar a condição de crescimento da vegetação de áreas agrícolas, em especial a informação de fase, aplicou-se a transformação HLS aos termos harmônicos obtidos com o algoritmo Hants, por meio do programa Envi. O cultivo da cana-de-açúcar no Estado de São Paulo apresenta padrões espaciais coerentes com o ciclo de desenvolvimento da cultura e consistentes com a variabilidade das precipitações sazonais que afetam diretamente o momento de máximo dos índices de vegetação. O pico vegetativo da cana-de-açúcar ocorre durante a fase de crescimento, em anos com chuvas normais; porém, em anos de pouca chuva, a fase de maturação da cultura é antecipada e, em anos de muita chuva, a fase de crescimento é antecipada, o que causa o atraso da maturação.

Abstract The objective of this work was to evaluate sugarcane cultivation, in a harmonic analysis applied to a time series of Modis vegetation indices, with the representation of harmonic terms. Daily rainfall data were obtained from Agritempo for the state of São Paulo, Brazil, and accumulated for a period of 16 days of Modis compositions, from the 2004/2005 to 2011/2012 crop seasons. The normalized difference vegetation index (NDVI) and the enhanced vegetation index (EVI) were used in time-series decomposed in harmonic terms by the harmonic analysis. In order to visualize the growing conditions of vegetation in agricultural areas, specially the phase information, the HLS transformation was applied to the harmonic terms obtained by the Hants algorithm, using Envi software. Sugarcane cultivation in the state of São Paulo shows spatial patterns that are coherent with the sugarcane development cycle and consistent with the variability of seasonal rainfall that directly affect the maximum period of vegetation indices. The peak growth stage of sugarcane occurs in years of normal rainfall; however, in years with below normal rainfall, sugarcane maturation phase is anticipated, and, in years with above normal rainfall, the growth phase is anticipated, which causes maturation delay.

Incubatório de ovos é um setor de grande importância na Avicultura de postura. Com a redução dos custos dos equipamentos de informática cresce o armazenamento de dados para gerenciamento do processo produtivo. A Mineração de Dados surge como uma técnica para identificar conhecimentos novos e úteis nos bancos de dados. Objetivou-se, neste trabalho, explorar a técnica Arvore de Decisão em banco de dados de incubatórios de matrizes de postura, visando a elaboração de padrões de incubação. Foram disponibilizados, pela empresa Hy-Line do Brasil Ltda, dados de incubação entre os anos de 2002 e 2006 da linhagem Hy-Line W-36. Dois experimentos foram realizados. Em um deles, valores acima dos estabelecidos pela empresa como desejado para o índice "fêmeas nascidas vendáveis" foram identificados como relevantes para a geração das regras. No outro, valores abaixo dos estabelecidos pela empresa foram identificados como relevantes para a geração das regras. Foi utilizado o algoritmo Entropia C 4.5 e o software SAS-Enterprise Miner como ferramenta de análise . Como conclusão deste estudo, foi possível observar que com a técnica estudada, os dados utilizados no gerenciamento de produção são suficientes para identificar conhecimentos novos, úteis e aplicáveis a fim de melhorar a produtividade das empresas incubadoras, atendendo à demanda com diminuição do desperdício.

Hatchery is a very important sector in egg production. As computers become cheaper, there is an increase in data storage for the production management process. Data Mining has appeared as a technique to identify new and useful knowledge in databases. The objective of this work was to explore the Decision Tree technique in hatchery databases to identify the best standards of the incubation process. The data set used in this research was supplied by Hy-Line do Brasil Ltda., corresponding to the incubation period of 2002-2006, from the strain Hy-line W-36. Two experiments were carried out. In the first experiment, values higher than the company's standards for saleable females were identified as relevant to generate the rules. In the second experiment, values below those established by the company were identified as relevant for the generation of rules. Entropy C 4.5 algorithm and the software SAS-Enterprise Miner were used for data analysis. The conclusion is that, with the technique studied, the data used for production management are sufficient to identify new, useful and applicable knowledge in order to increase productivity of hatcheries, catering for the demand with less waste.

Um dos desafios da agricultura de precisão é oferecer subsídios para a definição de unidades de manejo para posteriores intervenções. Portanto, o objetivo deste trabalho foi avaliar os atributos químicos do solo e a produtividade da cultura de cana-de-açúcar por meio da geoestatística e mineração de dados pela indução da árvore de decisão. A produtividade da cana-de-açúcar foi mapeada em uma área de aproximadamente 23ha, utilizando-se o critério de célula, por meio de um monitor de produtividade que permitiu a elaboração de um mapa digital que representa a superfície de produção para a área em estudo. Para determinar os atributos de um Argissolo Vermelho-Amarelo, foram coletadas as amostras no início da safra 2006/2007, utilizando-se uma grade regular de 50 x 50m, nas profundidades de 0,0-0,2m e 0,2-0,4m. Os dados dos atributos do solo e da produtividade foram analisados por meio da técnica de goestatística e classificados em três níveis de produção para indução de árvore de decisão. A árvore de decisão foi induzida no programa SAS Enterprise Miner, sendo utilizado algoritmo baseado na redução de entropia. As variáveis altitude e potássio apresentaram os maiores valores de correlação com a produtividade de cana-de-açúcar. A indução de árvores de decisão permitiu verificar que a altitude é a variável com maior potencial para interpretar os mapas de produtividade de cana-de-açúcar, auxiliando na agricultura de precisão e mostrando-se uma ferramenta adequada para o estudo de definição de zonas de manejo em área cultivada com essa cultura.

One of the challenges of precision agriculture is to offer subsidies for the definition of management units for posterior interventions. Therefore, the objective of this work was to evaluate soil chemical attributes and sugarcane yield with the use of geostatistics and data mining by decision tree induction. Sugarcane yield was mapped in a 23ha field, applying the cell criterion, by using a yield monitor that allowed the elaboration of a digital map representing the surface of production of the studied area. To determine the soil attributes, soil samples were collected at the beginning of the harvest in 2006/2007 using a regular grid of 50 x 50m, in the depths of 0.0-0.2m and 0.2-0.4m. Soil attributes and sugarcane yield data were analyzed by using geostatistics techniques and were classified into three yield levels for the elaboration of the decision tree. The decision tree was induced in the software SAS Enterprise Miner, using an algorithm based on entropy reduction. Altitude and potassium presented the highest values of correlation with sugarcane yield. The induction of decision trees showed that the altitude is the variable with the greatest potential to interpret the sugarcane yield maps, then assisting in precision agriculture and, revealing an adjusted tool for the study of management definition zones in area cropped with sugarcane.

O objetivo deste trabalho foi desenvolver árvores de decisão como modelos de alerta da ferrugem-do-cafeeiro em lavouras de café (Coffea arabica L.) com alta carga pendente de frutos. Dados de incidência mensal da doença no campo coletados durante oito anos foram transformados em valores binários considerando limites de 5 e 10 pontos percentuais na taxa de infecção. Foi gerado um modelo para cada taxa de infecção binária a partir de dados meteorológicos e do espaçamento entre plantas. O alerta é indicado quando a taxa de infecção, prevista para o prazo de um mês, atingir ou ultrapassar o respectivo limite. A acurácia do modelo para o limite de 5 pontos percentuais foi de 81%, por validação cruzada, chegando até 89% segundo estimativa otimista. Esse modelo apresentou bons resultados para outras medidas de avaliação importantes, como sensitividade (80%), especificidade (83%) e confiabilidades positiva (79%) e negativa (84%). O modelo para o limite de 10 pontos percentuais teve acurácia de 79%, e não apresentou o mesmo equilíbrio entre as demais medidas. Em conjunto, esses modelos podem auxiliar na tomada de decisão referente ao controle da ferrugem-do-cafeeiro no campo. A indução de árvores de decisão é alternativa viável às técnicas convencionais de modelagem e facilita a compreensão dos modelos.

The objective of this work was to develop decision trees as warning models of coffee (Coffea arabica L.) rust in growing areas with large fruit load. Monthly data of disease incidence in the field collected during eight years were transformed into binary values considering limits of 5 and 10 percentage points in the infection rate. Models were generated from meteorological data and space between plants for each binary infection rate. The warning is indicated when the infection rate is expected to reach or exceed the respective limit in a month. The accuracy obtained by cross-validating the model to the limit of 5 percentage points was 81%, reaching up to 89% according to an optimistic estimate. This model showed good results for other important evaluation measures, such as sensitivity (80%), specificity (83%), positive reliability (79%), and negative reliability (84%). The model for the limit of 10 percentage points had a 79% accuracy and did not show the same balance among the other evaluation measures. Together, these two models may support the decisions about coffee rust control in the field. The decision tree induction is a viable alternative to conventional modeling techniques, thus facilitating the comprehension of the models.

As ondas de calor provocam perdas na produção animal pela sua exposição ao estresse por calor aumentando a mortalidade, e consequentemente, perdas econômicas. Bancos de dados zootécnicos e meteorológicos históricos podem conter informações que permitem modelar a mortalidade de frangos devido à incidência de ondas de calor. O objetivo foi analisar bancos de dados de frangos de corte associados a dados meteorológicos utilizando técnicas de mineração de dados, seleção de atributos e classificação (árvore de decisão) para modelar o impacto da incidência de onda de calor na mortalidade de frangos de corte. O Índice de Temperatura e Umidade (ITU) foi utilizado para descrever parte dos dados ambientais. A técnica de Mineração de Dados permitiu a construção de três modelos compreensíveis para estimar a alta mortalidade em frangos de corte. Os modelos gerados pela abordagem de seleção de atributos por Análise dos Componentes Principais e Wrapper apresentaram igual desempenho com uma precisão total de 89,3% e a classificação para alta mortalidade foi de 83,3%. Quando a seleção foi feita por especialistas do domínio, a precisão do modelo foi de 85,7%, e a da classificação para alta mortalidade foi de 76,9%. Resultados meteorológicos e o ITU calculada a partir de estações meteorológicas permitiram identificar condições ambientais prejudiciais para frangos entre 29 e 42 dias de vida. A técnica de Mineração de Dados é aplicável para construir modelos preditivos para a produção animal.

Heat waves usually result in losses of animal production since they are exposed to thermal stress inducing an increase in mortality and consequent economical losses. Animal science and meteorological databases from the last years contain enough data in the poultry production business to allow the modeling of mortality losses due to heat wave incidence. This research analyzes a database of broiler production associated to climatic data, using data mining techniques such as attribute selection and data classification (decision tree) to model the impact of heat wave incidence on broiler mortality. The temperature and humidity index (THI) was used for screening environmental data. The data mining techniques allowed the development of three comprehensible models for estimating specifically high mortality during broiler production. Two models yielded a classification accuracy of 89.3% by using Principal Component Analysis (PCA) and Wrapper feature selection approaches. Both models obtained a class precision of 0.83 for classifying high mortality. When the feature selection was made by the domain experts, the model accuracy reached 85.7%, while the class precision of high mortality was 0.76. Meteorological data and the calculated THI from meteorological stations were helpful to select the range of harmful environmental conditions for broilers 29 and 42 days old. The data mining techniques were useful for building animal production models.