Dissertations / Theses on the topic 'Crop forecasting'

Inter-annual variations in crop yield are intricately linked to fluctuations in the weather. Accurate yield forecasts prior to harvest are possible if crop-weather relationships are integrated into models that are responsive to the major yield determining factors. A network of meteorological stations was selected across the Australian wheat belt and monthly rainfall regressed with wheat yields from the surrounding shires. Autumn rains that permit an early sowing and finishing rains after July are important for higher yields. As the rainfall distribution becomes more winter dominant in nature, both crop yield variability and the usefulness of early winter rainfall decreases. Waterlogging has a large negative effect in the south-west of Western Australia, such that the rainfall distribution is more important than the amount in this region. A national sowing date survey determined that regional sowing dates have become earlier during the 1980’s and that these vary considerably, especially to the north-east. In Western Australia, earlier sowing combined with higher nitrogen inputs from fertilizers and legumes caused a significant upward trend in recent yields. Trends have been smaller in other states. Yields were also regressed with broad scale atmospheric indicators. Up to a year in advance of harvest, changes in the amplitude of the trough in the upper level westerlies (South Pacific) precede major anomalies in yields. Trends in the Southern Oscillation Index (SOI) around sowing time account for half the variance in the national yield, due to a persistence in following rainfall anomalies. Agrometeorological index models that combine the features of simulation and regression are shown to be the most appropriate models for yield forecasting. At a shire level they account for an average 55% of the yield variance in Western Australia, but 60 to 80% of the variation in eastern states yields. Satellite spectral data also resolved similar amounts of yield variance when sensor calibration bias was removed. With a mean regional index determined by station weighting, crop-weather models account for 87 to 92% of the variance of state and national yields. Tests with operational model forecasts equalled, or were more accurate than, official forecasts in 4 out of 5 years. Seasonal outlooks incorporated into model calculations brought further gains in accuracy in extreme years. Overall, the broad scale extent of yield anomalies across the Australian wheat belt is highlighted. Extreme yields, which are of most interest to the grain industry, are inseparably coupled to the ENSO phenomenon and the broad scale atmospheric circulation. Crop-weather models adjust rapidly to these anomalies in the weather and should be applied in an operational environment to provide early indications of crop prospects.

Thesis (Ph.D.)--Industrial and Systems Engineering, Georgia Institute of Technology, 2007.<br>Committee Co-Chair: Griffin, Paul; Committee Co-Chair: Serban, Nicoleta; Committee Member: Liang, Steven; Committee Member: Sharp, Gunter; Committee Member: Tsui, Kwok-Leung

While the impacts of the El Nino Southern Oscillation (ENSO) are well documented on topics ranging from agricultural production to socio-economic factors, a closer consideration of key interaction terms in this complex relationship is pivotal for better understanding of future production impacts and as well as relevant policy implications. In this thesis, the ENSO link to staple crop production in the US is derived through an econometric approach, in particular taking advantage of recent advances in the nonlinear parameterization of climate variables such as temperature. Via the comparison of competing model specifications, across all major Corn and Soybean producing regions in the United States, the findings of the present study suggest the ENSO link with crop yields manifests itself primarily via extreme degree days. Following this conclusion, this study further extends previous literature by examining the effect of ENSO anomalies on agricultural production in an out-of-sample setting. Optimal producer strategies can be a powerful adaptive measure to anticipated/forecasted ENSO outcomes, predominantly planting date and crop mix. Key results prove valuable to such strategies, particularly in those regions where the channel of ENSO influence for production is obvious, and statistically significant in a pseudo-forecasting environment.

Agricultural policy changes directly affect the economic viability of Texas crop producers because government payments make up a significant portion of their net farm income (NFI). NFI projections benefit producers, agribusinesses and policy makers, but an economic model making these projections for Texas did not previously exist. The objective of this study was to develop a model to project annual NFI for producers of major crops in Texas. The Texas crop model was developed to achieve this objective, estimating state prices, yields and production costs as a function of their national counterparts. Five hundred iterations of national price and yield projections from the Food and Agricultural Policy Research Institute (FAPRI), along with FAPRI’s average production cost projections, were used as input to the Texas crop model. The stochastic FAPRI Baseline and residuals for Ordinary Least Squares (OLS) equations relating Texas variables to national variables were used to incorporate the risk left unexplained by OLS equations between Texas and U.S. variables. Deterministic and probabilistic NFI projections for Texas crops were compared under the January 2005 and January 2006 FAPRI Baseline projections. With production costs increasing considerably and prices rising moderately in the January 2006 Baseline, deterministic projections of 2006-2014 Texas NFI decreased by an average of 26 percent for corn, 3 percent for cotton, 15 percent for peanuts, and 12 percent for rice, and were negative for sorghum and wheat. Probability distributions of projected NFI fell for all program crops, especially sorghum and wheat. Higher hay price projections caused deterministic projections of NFI for hay to rise roughly 13 percent, and increased the probability distributions of projected hay NFI. Deterministic and probabilistic projections of total NFI decreased for each year, especially for 2006-2008 when fuel price projections were the highest. The Texas crop model can be used to simulate NFI for Texas crop producers under alternative FAPRI baselines. The model shows the impact of baseline changes on probability distributions of NFI for each crop and for Texas as a whole. It can also be useful as a policy analysis tool to compare impacts of alternative farm and macroeconomic policies on NFI.

Crop yield forecasting models are needed to help farmers and decision makers cheaply detect crop condition early enough to assess and mitigate its impacts on grain production. A precise estimate of crop production requires an accurate measure of the total cultivated area and well-established knowledge of crop yield. The first requirement is no longer a problem as is technically solved through various techniques such as area frame sampling. With respect to the second, great efforts have been made to find an accurate definition of the crop yield with respect to the actual factors that shape its growth through out the season. Agrometeorological models have found a wide range of applications in agricultural research and technology and are playing an increasing role in translating information about climate variability into assessments, predictions and recommendations tailored to the needs of agricultural decision makers. However these models have generally been developed and tested for application at the scale of a homogeneous plot. They are criticized for their inability to address large-scale yield estimates at regional or even national levels in addition to their high cost of application. This is because field conditions during the period of crop establishment at the regional scale may be quite variable and poorly represented by standard parameter values of the crop model.

Crop yield forecasting models are needed to help farmers and decision makers cheaply detect crop condition early enough to assess and mitigate its impacts on grain production. A precise estimate of crop production requires an accurate measure of the total cultivated area and well-established knowledge of crop yield. The first requirement is no longer a problem as is technically solved through various techniques such as area frame sampling. With respect to the second, great efforts have been made to find an accurate definition of the crop yield with respect to the actual factors that shape its growth through out the season. Agrometeorological models have found a wide range of applications in agricultural research and technology and are playing an increasing role in translating information about climate variability into assessments, predictions and recommendations tailored to the needs of agricultural decision makers. However these models have generally been developed and tested for application at the scale of a homogeneous plot. They are criticized for their inability to address large-scale yield estimates at regional or even national levels in addition to their high cost of application. This is because field conditions during the period of crop establishment at the regional scale may be quite variable and poorly represented by standard parameter values of the crop model.

The objective of this study is to examine land use change for cropping systems in North Dakota. Using Seemingly Unrelated Regression with full information maximum likelihood estimation method, acreage forecasting models for barley, corn, oats, soybean, and wheat were developed to examine the extent to which farmers? expectations of prices and costs affect their crop choices. The results of the study show that farmers? decision for acreage allocation is varied across the crops depending on how responsive they are to price, cost and yield of its own and competing crops. Substitutability and complementarity relationship of crops in the production have positive effect on crops selection when facing price, cost, and yield changes. In addition, the results revealed that expected prices have little effect on acreage response compared to expected costs and yield variables in most of the crop models.<br>IIE team Fulbright sponsorship<br>North Dakota State University. Department of Agribusiness and Applied Economics<br>National Science Foundation (NSF). Grant Number IIA-1355466

The production of grain crops plays an important role in the economy of the Canadian Prairies and early reliable crop yield forecasts over large areas would help policy makers and grain marketing agencies in planning for exports. Forecast models developed from satellite data have the potential to provide quantitative and timely information on agricultural crops over large areas. The use of nonlinear modeling techniques from the field of machine learning could improve crop forecasting from the linear models most commonly used today. The Canadian Prairies consist of the provinces of Alberta, Saskatchewan and Manitoba and three of the major crops in this region are barley, canola and spring wheat. The agricultural land on the Canadian Prairies has been divided into Census Agricultural Regions (CAR) by Statistics Canada. A clustering model was applied to the crop yield data to group the CARs for the development of forecast models. The normalized difference vegetation index (NDVI) and enhanced vegetation index (EVI) derived from the Moderate Resolution Imaging Spectro-radiometer (MODIS), NDVI derived from the Advanced Very High Resolution Radiometer (AVHRR) and several climate indices were considered as predictors for crop yields. A correlation analysis between crop yield and the time series of each potential predictor was performed to determine which variables showed the most forecasting potential and at what time during the growing season their values were most correlated to crop yield. Various combinations of MODIS-NDVI, MODIS-EVI and NOAA-NDVI were used to forecast the yield of barley, canola and spring wheat. Multiple linear regression as well as nonlinear Bayesian neural networks and model-based recursive partitioning forecast models were developed using the various sets of predictors. The models were trained using a cross-validation method and the forecast results of each model were evaluated by calculating the skill score from the mean absolute error, with 95% confidence intervals for the skill scores calculated using a bootstrap method. The results were compared in an effort to determine the optimal set of predictors and type of forecast model for each crop.

This Thesis examines the main issues surrounding crop modelling by detailed studies of (i) multi-model ensemble forecasting using a simple dynamical system as a proxy for seasonal weather forecasting, (ii) probabilistic forecasts for crop models and (iii) an analysis of changes in US yield. The ability to forecast crop yield accurately on a seasonal time frame would be hugely beneficial to society in particular farmers, governments and the insurance industry. In addition, advance warning of severe weather patterns that could devastate large areas of crops would allow contingency plans to be put in place before the onset of a widespread famine, potentially averting a humanitarian disaster. There is little experience in the experimental design of ensembles for seasonal weather forecasting. Exploring the stability of the results varying, for example, the sample size aids understanding. For this a series of numerical experiments are conducted in an idealised world based around the Moran Ricker Map. The idealised world is designed to replicate the multi-model ensemble forecasting methods used in seasonal weather forecasting. Given the complexity of the physical weather systems experiments are instead conducted on the Moran Ricker Map [56,70]. Additionally, experiments examine whether including climatology as a separate model or blending with climatology can increase the skill. A method to create probabilistic forecasts from a crop model, the Crop Environment Resource Synthesis Maize model (CERES-Maize) [19, 37] is proposed. New empirical models are created using historical US maize yield. The skill from equally weighting the crop model with a simple empirical model is investigated. Background reviews of weather and yield data is presented in new ways for the largest maize growing state Iowa. A new method separating the impacts of favourable weather from technology increases in a crop yield time series is explored.

The impacts of a changing climate on the social and economic development of humanity have been increasingly studied in the last decades. According to the Intergovernmental Panel on Climate Change (IPCC), the lack of implementation of effective and adequate measures for contrasting green house gases emissions will lead to increasingly severe and partially irreversible impacts on the environment, and consequently on the society. The estimate of possible impacts on food production, starting from agriculture, is essential to develop strategies to alleviate the consequences of climate change. In this context, the evaluation of the future dynamics of plant diseases plays a key role because they determine actual production levels for many crops in many areas, therefore deeply influencing food availability and security. In order to perform such analyses, process-based simulation modelling offers the capability to capture the high non-linearity characterizing the responses of biophysical processes to boundary conditions. However, such models have been marginally used to estimate scenarios of plant diseases impact on crop production, because of the limited availability of modelling approaches and tools. This work constitutes an attempt to respond to the need of developing a software framework for the simulation of a generic fungal plant airborne disease which can be easiliy coupled with a crop simulator in order to improve the estimation of the levels of crop productions under climate change scenarios. The first section of the work deals with the evaluation of models for the estimation of meteorological data and for the simulation of leaf wetness, driving variable of the infection process of fungal plant pathogens. These assessments were justified by the need of feeding the disease models with high quality data, and by the scarce availability of hourly data in large area databases. The second section presents the implementation and the calibration of the generic fungal plant epidemic framework, and its test via an extensive use of sensitivity analysis techniques. The third section deals with the application of the developed modelling solutions, coupled with crop simulators, for the forecasting of the impact of climate change on crop production in Latin America. In the last section, new criteria and metrics for biophysical model evaluation and analysis are presented, aimed at considering the models performance under heterogeneous climatic conditions such as those explored in climate change and large area application studies.

La météo et le climat ont un impact important sur les rendements agricoles. De nombreuses études basées sur différentes approches ont été réalisées pour mesurer cet impact. Cette thèse se concentre sur les modèles statistiques pour mesurer la sensibilité des cultures aux conditions météorologiques sur la base des enregistrements historiques. Lors de l'utilisation d'un modèle statistique, une difficulté critique survient lorsque les données sont rares, ce qui est souvent le cas pour la modélisation des cultures. Il y a un risque élevé de sur-apprentissage si le modèle n'est pas développé avec certaine précautions. Ainsi, la validation et le choix du modèle sont deux préoccupations majeures de cette thèse. Deux approches statistiques sont développées. La première utilise la régression linéaire avec régularisation et validation croisée (c.-à.-d. leave-one-out ou LOO), appliquée au café robusta dans la principale région productrice de café du Vietnam. Le café est une culture rémunératrice, sensible aux intempéries, et qui a une phénologie très complexe en raison de sa nature pérenne. Les résultats suggèrent que les informations sur les précipitations et la température peuvent être utilisées pour prévoir l'anomalie de rendement avec une anticipation de 3 à 6 mois selon la région. Les estimations du rendement du robusta à la fin de la saison montrent que les conditions météorologiques expliquent jusqu'à 36 % des anomalies de rendement historiques. Cette première approche de validation par LOO est largement utilisée dans la littérature, mais elle peut être mal utilisé pour de nombreuses raisons : elle est technique, mal interprétée et nécessite de l'expérience. Une alternative, l'approche “leave-two-out nested cross-validation” (ou LTO), est proposée pour choisir le modèle approprié, évaluer sa véritable capacité de généralisation et choisir la complexité du modèle optimale. Cette méthode est sophistiquée mais simple. Nous démontrons son applicabilité pour le café robusta au Vietnam et le maïs en France. Dans les deux cas, un modèle plus simple avec moins de prédicteurs potentiels et d'entrées est plus approprié. Utiliser uniquement la méthode LOO peut être très trompeur car cela encourage à choisir un modèle qui sur-apprend les données de manière indirecte. L'approche LTO est également utile dans les applications de prévision saisonnière. Les estimations de rendement du maïs en fin de saison suggèrent que les conditions météorologiques peuvent expliquer plus de 40 % de la variabilité de l'anomalie de rendement en France. Les impacts du changement climatique sur la production de café au Brésil et au Vietnam sont également étudiés à l'aide de simulations climatiques et de modèles d'impact ou “suitability models”. Les données climatiques sont cependant biaisées par rapport au climat réel. De nombreuses méthodes de “correction de biais” (appelées ici “calibration”) ont été introduites pour corriger ces biais. Une présentation critique et détaillée de ces calibrations dans la littérature est fournie pour mieux comprendre les hypothèses, les propriétés et les objectifs d'application de chaque méthode. Les simulations climatiques sont ensuite calibrées par une méthode basée sur les quantiles avant d'être utilisées sur nos modèles d'impact. Ces modèles sont développés sur la base des données de recensement des zones caféières, et les variables climatiques potentielles sont basées sur un examen des études précédentes utilisant des modèles d'impact pour le café et des recommandations d'experts. Les résultats montrent que les zones propices à l'arabica au Brésil pourraient diminuer d'environ 26 % d'ici le milieu du siècle dans le scénario à fortes émissions, les régions compatibles avec la culture du robusta vietnamien pourraient quant à elle diminué d'environ 60 %. Les impacts sont significatifs à basse altitude pour les deux types de café, suggérant des déplacements potentiels de la production vers des endroits plus élevés<br>Weather and climate strongly impact crop yields. Many studies based on different techniques have been done to measure this impact. This thesis focuses on statistical models to measure the sensitivity of crops to weather conditions based on historical records. When using a statistical model, a critical difficulty arises when data is scarce, which is often the case with statistical crop modelling. There is a high risk of overfitting if the model development is not done carefully. Thus, careful validation and selection of statistical models are major concerns of this thesis. Two statistical approaches are developed. The first one uses linear regression with regularization and leave-one-out cross-validation (or LOO), applied to Robusta coffee in the main coffee-producing area of Vietnam (i.e. the Central Highlands). Coffee is a valuable commodity crop, sensitive to weather, and has a very complex phenology due to its perennial nature. Results suggest that precipitation and temperature information can be used to forecast the yield anomaly with 3–6 months' anticipation depending on the location. Estimates of Robusta yield at the end of the season show that weather explains up to 36 % of historical yield anomalies. The first approach using LOO is widely used in the literature; however, it can be misused for many reasons: it is technical, misinterpreted, and requires experience. As an alternative, the “leave-two-out nested cross-validation” (or LTO) approach, is proposed to choose the suitable model and assess its true generalization ability. This method is sophisticated but straightforward; its benefits are demonstrated for Robusta coffee in Vietnam and grain maize in France. In both cases, a simpler model with fewer potential predictors and inputs is more appropriate. Using only the LOO method, without any regularization, can be highly misleading as it encourages choosing a model that overfits the data in an indirect way. The LTO approach is also useful in seasonal forecasting applications. The end-of-season grain maize yield estimates suggest that weather can account for more than 40 % of the variability in yield anomaly. Climate change's impacts on coffee production in Brazil and Vietnam are also studied using climate simulations and suitability models. Climate data are, however, biased compared to the real-world climate. Therefore, many “bias correction” methods (called here instead “calibration”) have been introduced to correct these biases. An up-to-date review of the available methods is provided to better understand each method's assumptions, properties, and applicative purposes. The climate simulations are then calibrated by a quantile-based method before being used in the suitability models. The suitability models are developed based on census data of coffee areas, and potential climate variables are based on a review of previous studies using impact models for coffee and expert recommendations. Results show that suitable arabica areas in Brazil could decrease by about 26 % by the mid-century in the high-emissions scenario, while the decrease is surprisingly high for Vietnamese Robusta coffee (≈ 60 %). Impacts are significant at low elevations for both coffee types, suggesting potential shifts in production to higher locations. The used statistical approaches, especially the LTO technique, can contribute to the development of crop modelling. They can be applied to a complex perennial crop like coffee or more industrialized annual crops like grain maize. They can be used in seasonal forecasts or end-of-season estimations, which are helpful in crop management and monitoring. Estimating the future crop suitability helps to anticipate the consequences of climate change on the agricultural system and to define adaptation or mitigation strategies. Methodologies used in this thesis can be easily generalized to other cultures and regions worldwide

Sugar is an important global agricultural commodity and a significant input to the advanced industrialised world. Annual average global sugar production is around 120 million tonnes, with consumption around 118 million tonnes. Sugar is produced under a broad range of climatic conditions in some 120 countries and is one of the most heavily traded agricultural commodities (FAO, 2001). Plants produce sugar as a storehouse of energy that is used as required. Approximately 70% of sugar is produced from sugar cane while the remaining 30% is produced from sugar beet (Sugar Knowledge International, 2001). Thailand's cane and sugar industry is now one of the major sources of foreign income for the country. The value of sugar exports (around 35 billion baht or AUD $1.5 billion per annum) ranks among the top ten exported commodities of the Thai economy. Approximately 9.2% of annual global sugar production is exported from Thailand (WTO, 2001). The sugar industry is extremely complex and comprises individual links and components in the supply and demand chain that are more delicately in balance than with most other commodity based industries. Thailand's sugar production has been characterized by greater extremes of variability than in most other sugar producing countries. A unique combination of pests, disease, climate, soils, problems with plant available moisture and the low technology basis of crop management has increased production risk and uncertainty for the crop. Total tonnage of cane and sugar is notoriously difficult to predict during the growing season and for a mature crop before the harvest. Accordingly, the focus of this research is on the development and testing of methods, algorithms, procedures and output products for Sugar Cane Crop Forecasting and Yield Mapping. The resulting spatial and temporal information tools have the potential to provide the basis of a commercially deployable decision support system for Thailand's sugar industry. The scope of this thesis encompasses several levels within a geographical hierarchy of scales; from regional, district, farm, and plot within a study area in northeastern Thailand. Crop forecasting at regional level will reduce production risk uncertainty while yield mapping and yield estimation at local, farm and plot scales will enable productivity to be improved by identifying, diagnosing the cause of and reducing yield variability. The research has three main objectives. These are to: Develop statistical analysis procedures and empirical algorithms expressing the relationship between yield potential and spectral response of sugar cane yield as a basis for mapping, monitoring, modeling, forecasting and management of sugar production in Thailand. Evaluate the validity of a technology based versus conventional approach to crop forecasting and yield mapping, commencing with a series of testable null-hypotheses and culminating in procedures to calibrate and validate empirical models against verifiable production records. Outcomes are used to review and evaluate existing and potential future approaches to regional crop forecasting, localised yield mapping and yield estimation tools for operational use within Thailand's sugar industry. Identify, evaluate and establish performance benchmarks in relation to the practicality, accuracy, timeliness, cost effectiveness and value proposition of a satellite based versus conventional approach to crop forecasting and yield mapping. The methodology involved time series analysis of recorded sugar cane yields and production outcomes paired with spectral response statistics of crops derived from satellite imagery and seasonal rainfall records over a three year period within four provinces, forty five component districts and 120 representative farms. Spectral statistics were derived fiom raw multi-spectral satellite imagery (multitemporal SPOT- VI at regional scale and Landsat 7 ETM+ imagery at local scale) acquired during the 1999 to 2001 sugar cane seasons. Crop area and production statistics at regional scale were compiled and furnished by the provincial sugar mill and verified through government agencies within Thailand. Selective cutting at sample sites within nominated fields owned by collaborating growers was undertaken to validate localised differences in productivity and to facilitate yield variance mapping. Acquisition, processing, analysis and statistical modeling of remotely sensed satellite spectral data, rainfall records and production outcomes were accomplished using an empirical approach. Resulting crop production forecasting algorithms were systematically evaluated for reliability by assessing accuracy, spatial and temporal variability. Long term rainfall and district sugar cane yield and production records were used to account for district and season specific differences between estimated and recorded yields, to generate error probability functions and to improve the accuracy and applicability of empirical models under more extreme conditions. Limitations on finding and length of records constrained the number of seasons and the area for which satellite imagery with contrasting levels of spatial and spectral resolution could be acquired. The absence of verifiable long term production records combined with limitations on the duration and area able to be covered by field trips meant that time series analysis of paired data was necessarily constrained to a three year period of record coinciding with the author's period of candidature. Accordingly, although a comprehensive set of well correlated district and month specific yield forecasting algorithms was able to be developed, temporal restrictions on data availability constrained the extent to which they could be subjected to thorough accuracy and reliability analysis and extended with confidence down to farm and field scale. A variety of approaches, using different parameter combinations and threshold values, was used to combine individual districts and component farms into coherent groups to overcome temporal data constraints and to generate more robust production forecasting algorithms, albeit with slightly lower levels of apparent accuracy and reliability. The procedures adopted to optimise these district groupings are systematically explained. Component differences in terrain, biophysical conditions and management approaches between district groupings are used to explain differences in production outcomes and to account for apparent differences between forecast versus actual yields between districts both within and between different groups. The outcomes of this research - particularly the data acquisition and analysis procedures, empirical modeling, error assessment and adjustment techniques, and the optimisation procedures used to facilitate grouping of districts - provide a practical basis for the deployment of an operational sugar cane production forecasting and yield mapping information system to facilitate planning and logistical management of production, harvesting, transportation, processing, domestic marketing and export of sugar from northeastern Thailand. At the local and farm level, yield maps and plot based yield estimates will assist users to improve productivity by recognising, identiwing and responding to potential causes of within and between field spatial variability. However, before such an information system can be confidently deployed, additional resources will be required to obtain paired production records, spectral data fiom satellite imagery and biophysical input data over a longer period to ensure that the empirical models are operationally robust and to validate their accuracy under a wider range of conditions by comparing forecasts with actual outcomes over larger areas during the next few seasons.

Thesis (Ph. D.)--Ohio State University, 2005.<br>Title from first page of PDF file. Document formatted into pages; contains xiii, 155 p.; also includes map, graphics (some col.) Includes bibliographical references (p. 149-155). Available online via OhioLINK's ETD Center

Uma das atividades mais relevantes para a economia brasileira é a agricultura. Entre os produtos de maior importância no cenário agrícola nacional, estão a soja e o arroz, os quais representam uma grande parcela da produção. Somente o Estado do Rio Grande do Sul é responsável por aproximadamente 67% da produção nacional de arroz e 10% de soja (IBGE, 2012). Portanto, informações confiáveis sobre a produção agrícola são relevantes para o desenvolvimento do setor e o desenvolvimento de metodologias capazes de auxiliar no monitoramento das áreas agrícolas torna-se peça importante na geração de dados confiáveis e com maior rapidez de obtenção. Desta forma, o objetivo deste trabalho foi desenvolver uma metodologia de baixo custo para a execução do mapeamento da área cultivada de arroz irrigado e soja, em escala municipal e estadual, baseado na análise do comportamento espectro-temporal de índices de vegetação de imagens de satélite de alta resolução temporal. O estudo foi realizado no Estado do Rio Grande do Sul, abrangendo os 497 municípios no ano safra 2011/2012. Para realizar o estudo, foram utilizadas imagens multitemporais do sensor MODIS, índices de vegetação EVI e NDVI. Foi aplicado o modelo HAND para gerar as áreas de inundação, as quais foram utilizadas para discriminar a cultura do arroz irrigado de outras culturas, especialmente a soja. Para avaliar os resultados foram utilizados como dados de referência, os dados coletados a campo, dados de área cultivada do IBGE e dados do mapeamento gerados a partir de imagens do satélite RapidEye. Os resultados mostraram que a metodologia proposta foi satisfatória, com valores médios do índice Kappa de 0,90 para a cultura de arroz irrigado e de 0,84 para a soja. Não houve diferença significativa entre as estimativas de área cultivada utilizando os dados EVI e NDVI para ambas as culturas. A utilização do Modelo HAND para discriminar o arroz irrigado de outros cultivos, mostrou-se muito eficiente, separando as áreas de várzea, que são mais aptas para o cultivo de arroz irrigado. Apesar dos resultados terem sido considerados como satisfatórios alguns municípios apresentaram problemas de subestimação ou superestimação quando foram comparados com os dados oficiais do IBGE. Esses problemas podem estar relacionados ao caráter subjetivo de aquisição de dados por parte do IBGE e também o fato de ter sido utilizada para a validação dos dados da safra 2011/2012 a média das últimas três safras, podendo desta maneira ter fragilizado ou comprometido os resultados para alguns municípios. Portanto, técnicas de sensoriamento remoto e geoprocessamento podem ser úteis no auxilio dos atuais métodos de monitoramento e mapeamento de culturas agrícolas, melhorando as estatísticas oficiais do arroz irrigado e soja.<br>One of the most relevant activities for the Brazilian economy is agriculture. Among the products of greatest importance in the national agricultural, are soybeans and rice, which represent a large portion of the production. Only the State of Rio Grande do Sul is responsible for approximately 67% of the national rice production and 10% of soybean (IBGE, 2012). Therefore, reliable information on agricultural production are relevant to the development of the sector and the development of methodologies capable of assist in the monitoring of agricultural areas becomes important part in the generation of reliable data and faster of obtaining. Thus, the objective of this work was to develop a methodology of low cost to implement the mapping of acreage irrigated rice and soybeans, at the municipal and state levels, based on the analysis of the spectral-temporal behavior of vegetation indices from satellite images high temporal resolution. The study was conducted in the state of Rio Grande do Sul, covering 497 municipalities in crop year 2011/2012. To conduct the study, images were used multitemporal MODIS vegetation indices EVI and NDVI. HAND model was applied to generate the inundation areas, which were used to discriminate the rice culture of other crops, especially soybeans. To evaluate the results were used as reference data, data collected in the field, the cultivated area data from the IBGE and mapping data generated from satellite images RapidEye. The results show that the proposed method was satisfactory, with mean values of Kappa 0.90 for irrigated rice and 0.84 for soybeans. There was no significant difference between the estimates of acreage using EVI and NDVI data for both crops. The use of the HAND model to discriminate irrigated rice from other crops, was very efficient, separating the lowland areas, which are more suitable for the cultivation of irrigated rice. Although the results were considered satisfactory as some municipalities had problems underestimation or overestimation when they were compared with the official data. These problems may be related to the subjective nature of data acquisition by the IBGE and the fact of having been used for the validation of data from 2011/2012 season the average of the last three years, and may in this way be weakened or compromised results for some municipalities. Therefore, techniques of remote sensing and GIS can be useful in the aid of the current methods of monitoring and mapping of agricultural crops, improving the official statistics of irrigated rice and soybeans.

The raising global demand for agricultural products and the exacerbated inter-annual fluctuations of food productions due to climate change are increasing world food price volatility and threatening food security in developing countries. In this context, the availability of reliable operational yield forecasting systems would allow policy makers to regulate agricultural markets. However, the reliability of the current approaches (the most sophisticated being based on crop models) is undermined by different sources of uncertainty. In particular, large area simulations can be affected by errors deriving from the uncertainty in input data (e.g., sowing dates, information on cultivar/hybrid grown, management practices) and upscaling assumptions, as well as from the incomplete adequacy of crop models to reproduce the effects of key factors affecting inter-annual yield fluctuations (e.g., extreme weather events, pests, diseases). The aim of this Ph.D. project was to reduce the uncertainty affecting the existing model-based forecasting systems through: (i) the implementation of approaches for the estimation of the impact of biotic and abiotic stressors on crop yields (based on dynamic models and on dedicated agro-climatic indicators), and (ii) the integration of remote sensing information within crop models. Concerning the first objective, the approaches for the simulation of transplanting shock and cold-induced spikelet sterility in rice included in Oryza2000 and WARM models, respectively, were improved, by increasing the model adherence to the underlying systems. Moreover, generic approaches for the simulation of the impacts of extreme weather events on crop yields were developed and evaluated, as well as approaches specific for sugarcane. For the second objective, remote sensing information was used to derive rice-cropped areas and sowing dates varying with time and space, as well as for the assimilation of exogenous leaf area index data using both recalibration and updating techniques (to account for factors not explicitly reproduced by the model within large-area applications). The application of the improved forecasting systems to different crops and agro-climatic contexts worldwide led to marked improvements compared to existing approaches. This was achieved through an increase in the percentage of inter-annual yield variability explained. On the one hand, the simulation of the impact of weather extremes (cold shocks, heat waves, water stress and frost) allowed to reduce the tendency of CGMS (the monitoring and forecasting system of the European Commission) to overestimate cereal yields in case of unfavorable seasons. Moreover, the integration of dynamic crop models and of agro-climatic indicators led to enhance the predicting capacity of available approaches. On the other hand, the integration of remote-sensing information within high resolution simulation chains allowed to decidedly reduce the uncertainty of the standard CGMS-WARM system when applied to the main European rice districts.

Sendo o cereal mais produzido no mundo e em larga expansão, os sistemas de produção de milho são altamente complexos e sua produção é diretamente dependente de fatores ligados tanto ao clima local quanto ao manejo da cultura. Para auxiliar na determinação tanto dos patamares produtivos de milho quanto quantificar o impacto causado por condições adversas tanto de clima quanto de manejo, pode-se lançar mão do uso de modelos de simulação de culturas. Para que os modelos possam ser devidamente aplicados, uma base solida de dados meteorológicos deve ser consistida, a fim de alimentar esses modelos. Nesse sentido, o presente estudo teve como objetivos: i) avaliar dois sistemas de obtenção de dados meteorológicos, o NASA-POWER e o DailyGridded, comparando-os com dados medidos em estações de solo; ii) calibrar, testar e combinar os modelos de simulação MZA-FAO, CSM DSSAT Ceres-Maize e APSIM-Maize, a fim de estimar as produtividades potenciais e atingíveis do milho no Brasil; iii) avaliar o impacto na produtividade causado pelo posicionamento da semeadura em diferentes tipos de solo; iv) desenvolver e avaliar um sistema de previsão de safra baseado em modelos de simulação; v) mapear as produtividades potencial, atingível e real do milho no Brasil, identificando regiões mais aptas ao cultivo e vi) determinar e mapear as quebras de produtividade, ou yield gaps (YG) da cultura do milho no Brasil. Comparando os dados climáticos dos sistemas em ponto de grade com os dados de estações meteorológicas de superfície, na escala diária, encontrou-se boa correlação entre as variáveis meteorológicas, inclusive para a chuva, com R2 da ordem de 0,58 e índice d = 0,85. O desempenho da combinação dos modelos ao final da calibração e ajuste se mostrou superior ao desempenho dos modelos individuais, com erros absolutos médios relativamente baixos (EAM = 627 kg ha-1) e com boa precisão (R2 = 0,62) e ótima acurácia (d = 1,00). Durante a avaliação da influência das épocas de semeadura e do tipo de solo no patamar produtivo do milho, observou-se que esse varia de acordo com a região estudada e apresenta seus valores máximos e com menores riscos à produção quando a semeaduras coincidem com o início do período de chuvas do local. O sistema de previsão de safra, baseado em modelos de simulação de cultura teve seu melhor desempenho simulando produtividades de milho semeados no início da safra e no final da safrinha, sendo capaz de prever de forma satisfatória a produtividade com até 25 dias antes da colheita. Para o estudo dos YGs, 152 locais foram avaliados e suas produtividades potenciais e atingíveis foram comparadas às produtividades reais, obtidas junto ao IBGE. Os maiores YGs referentes ao déficit hídrico se deram em solos arenosos e durante os meses de outono e inverno, usualmente mais secos na maioria das regiões brasileiras, atingindo valores de quebra superiores a 12000 kg ha-1. Quanto ao YG causado pelo manejo, esse foi maior nas regiões menos tecnificadas, como na região Norte e na Nordeste, apresentando valores superiores a 6000 kg ha-1. Já as regiões mais tecnificadas e tradicionais na produção de milho, como a região Sul e a Centro-Oeste, os YGs referentes ao manejo foram inferiores a 3500 kg ha-1 na maioria dos casos.<br>Maize is the most important cereal cultivated in the world, being its production system very complex and its productivity directly affected by climatic and crop management factors. In order to quantify the impacts caused by water and crop management deficits on maize yield, the use of crop simulation models is very useful. For properly apply these models, a solid basis of meteorological data is required. In this sense, the present study had as objectives: i) to evaluate two meteorological gridded data, NASA-POWER and DailyGridded, by comparing them with measured data from surface stations; (ii) to calibrate, evaluate and combine the MZA-FAO, CSM DSSAT Ceres-Maize and APSIM-Maize simulation models to estimate the maize potential and attainable yields in Brazil; iii) to evaluate the impact caused by the different sowing dates and soil types on maize yield; iv) to develop and evaluate a crop forecasting system based on crop simulation models and climatological data; v) to map the potential and the attainable maize yields in Brazil, identifying the most suitable regions for cultivation, and vi) to determine and map maize yields and yield gaps (YG) in Brazil. Comparing the gridded climatic data with observed ones, on a daily basis, a good agreement was found for all weather variables, including rainfall, with R2 = 0.58 and d = 0,85. The performances of the combination of the models at the end of the calibration and evaluation phases were better than those obtained with the individual models, with relatively low mean absolute error (EAM = 627 kg ha-1) and with good precision (R2 = 0.62) and accuracy (d = 1.00). During the evaluation of different sowing dates and soil types on maize yield, it was observed that this variable depends on the region and presents the maximum values and, consequently, the minimum risk during the sowings in the beginning of the rainy season of each site. The crop forecasting system, based on crop simulation models, had its best performance for simulating maize yields when the sowings were performed at the beginning of the main season and at the end of the second season, when it was able to predict yield satisfactorily 25 days before harvest. For the YG analysis, 152 sites were assessed and their potential and attainable yields were compared to the actual yields reported by IBGE. The highest YGs caused by water deficit occurred for sandy soils and during the autumn and winter months, usually dry in most of Brazilian regions, reaching values above 12000 kg ha-1. For YG caused by crop management, the values were higher in the less technified regions, such as in the North and Northeast regions, with values above 6000 kg ha-1. In contrast, more traditional maize production regions, such as the South and Center-West, presented YG caused by crop management, lower than 3500 kg ha-1 in most cases.

Orientadores: Rubens Augusto Camargo Lamparelli, Jansle Vieira Rocha<br>Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Agricola<br>Made available in DSpace on 2018-08-12T21:23:17Z (GMT). No. of bitstreams: 1 Ramme_FernandoLuizProchnow_D.pdf: 9393591 bytes, checksum: a6d5183861f5be0ab0ed23ba7a8838da (MD5) Previous issue date: 2008<br>Resumo: O objetivo do trabalho foi avaliar a relação entre as fases do crescimento da cana-de-açúcar com as formas de curvas do perfil temporal do Índice de Vegetação por Diferença Normalizada - NDVI, obtidas a partir do sensor remoto orbital MODerate-resolution Imaging Spectroradiometer - MODIS, na região de estudo. A avaliação desta relação é realizada utilizando-se técnicas de sensoriamento remoto para a geração do perfil temporal do NDVI, ao longo do ciclo de desenvolvimento fenológico da cana-soca, nas maturações Precoce, Média e Tardia. Os talhões de cana-soca analisados foram agrupados de acordo com a variedade, solo, data de plantio e corte, e contigüidade. A visualização gráfica das formas de curvas analisadas é realizada através de aplicativo, desenvolvido neste trabalho na linguagem de programação Java, e do sistema gerenciador de banco de dados PostgreSQL. O aplicativo realiza a filtragem de ruídos presentes nas imagens, composição na resolução temporal de 8 dias, através dos dados da banda de controle de qualidade do produto MOD09Q1, realiza a eliminação de valores discrepantes ao longo do perfil temporal do NDVI para a safra analisada, corrige as influências dos períodos de corte e rebrota da cana-soca, e propicia a suavização da forma de curva através do filtro Savitzky-Golay. Três janelas temporais de monitoramento da cultura são apresentadas neste trabalho. Cada janela temporal é determinada em função do tipo de maturação da cultura, do coeficiente de cultura (Kc) ao longo do ciclo fenológico da cana-soca e do comportamento na evolução do perfil temporal do NDVI. Concluiu-se que na região de estudo, diferentes maturações são caracterizadas por diferentes formas de curvas do perfil temporal do NDVI<br>Abstract: The objective of the work was to evaluate the relationship among the phases of the growth of the sugarcane with the forms of curves of the Normalized Difference Vegetation Index - NDVI temporal profile, obtained from remote sensor orbital MODerate-resolution Imaging Spectroradiometer - MODIS, in the study area. The evaluation of this relationship is accomplished by using of the techniques of remote sensing to generate the NDVI profile, along the phenological development phase of stubble-cane, in the Carly, Medium and Late maturations. The fields of stubble-cane analyzed were contained in agreement with the variety, soil, planting date and cut, and proximity. The graphic visualization of curves shape analyzed is accomplished through application, developed in this work in the Java programming language, and of the PostgreSQL system database manager. The application accomplishes the filtering of present noises in the images, composition in the temporal resolution of 8 days, through the data of the band of quality control of the MOD09Q1 product, accomplishes the elimination of outliers along the NDVI temporal profile for the culture analyzed, corrects the influences of the cut periods and regrowth of the stubble-cane, and propitiates the smoothing in the curve shape through the filter Savitzky-Golay. Three temporal windows of culture monitoring are presented in this work. Each temporal window is determined in function of the type of crop maturation, of the culture coefficient (Kc) along the phenological development phase of stubble-cane and of the behavior in the evolution of the NDVI profile. It concluded that in the study area, different maturations are characterized by different forms of NDVI profile curves<br>Doutorado<br>Planejamento e Desenvolvimento Rural Sustentável<br>Doutor em Engenharia Agrícola

Climate change and the reduction of available agricultural land are two of the most important factors that affect global food production especially in terms of wheat stores. An ever increasing world population places a huge demand on these resources. Consequently, there is a dire need to optimise food production. Estimations of crop yield for the South West agricultural region of Western Australia have usually been based on statistical analyses by the Department of Agriculture and Food in Western Australia. Their estimations involve a system of crop planting recommendations and yield prediction tools based on crop variety trials. However, many crop failures arise from adherence to these crop recommendations by farmers that were contrary to the reported estimations. Consequently, the Department has sought to investigate new avenues for analyses that improve their estimations and recommendations. This thesis explores a new approach in the way analyses are carried out. This is done through the introduction of new methods of analyses such as data mining and online analytical processing in the strategy. Additionally, this research attempts to provide a better understanding of the effects of both gradual variation parameters such as soil type, and continuous variation parameters such as rainfall and temperature, on the wheat yields. The ultimate aim of the research is to enhance the prediction efficiency of wheat yields. The task was formidable due to the complex and dichotomous mixture of gradual and continuous variability data that required successive information transformations. It necessitated the progressive moulding of the data into useful information, practical knowledge and effective industry practices. Ultimately, this new direction is to improve the crop predictions and to thereby reduce crop failures. The research journey involved data exploration, grappling with the complexity of Geographic Information System (GIS), discovering and learning data compatible software tools, and forging an effective processing method through an iterative cycle of action research experimentation. A series of trials was conducted to determine the combined effects of rainfall and temperature variations on wheat crop yields. These experiments specifically related to the South Western Agricultural region of Western Australia. The study focused on wheat producing shires within the study area. The investigations involved a combination of macro and micro analyses techniques for visual data mining and data mining classification techniques, respectively. The research activities revealed that wheat yield was most dependent upon rainfall and temperature. In addition, it showed that rainfall cyclically affected the temperature and soil type due to the moisture retention of crop growing locations. Results from the regression analyses, showed that the statistical prediction of wheat yields from historical data, may be enhanced by data mining techniques including classification. The main contribution to knowledge as a consequence of this research was the provision of an alternate and supplementary method of wheat crop prediction within the study area. Another contribution was the division of the study area into a GIS surface grid of 100 hectare cells upon which the interpolated data was projected. Furthermore, the proposed framework within this thesis offers other researchers, with similarly structured complex data, the benefits of a general processing pathway to enable them to navigate their own investigations through variegated analytical exploration spaces. In addition, it offers insights and suggestions for future directions in other contextual research explorations.

L'action des facteurs climatiques sur la croissance d'un peuplement de fetuque elevee (festuca arundinacea schreb). A ete etudiee au moyen de deux approches complementaires: une approche agronomique qui a permis de proposer un modele de prevision de la croissance printaniere du peuplement vegetal. Ce modele permet de relier la precocite de croissance et la vitesse de croissance et la vitesse de croissance aux deux facteurs climatiques essentiels (temperature et rayonnement), une approche ecophysiologique qui a cherche a analyser les mecanismes d'action des facteurs climatiques sur les composantes de la croissance: photosynthese nette et morphogenese (apparition et elongation des feuilles, tallage)

Precision management of the mango crop can be aided by quantitative measures of indices relevant to fruit maturation and measurement of the fruit load per tree. Such measures address fruit quantity and quality, respectively. Mango fruit DM at harvest is an index of ripened fruit eating quality of ripened fruit, and DM is also useful in assessment of the stage of maturation of the fruit. This thesis considered the robustness of NIR spectroscopy based DM models for use with fruit from different harvest events and of five cultivars (Calypso™, Honey Gold, Keitt, Kensington Pride and R2E2), from two distinct growing regions (Northern Territory and Queensland, over two seasons). Individual cultivar calibrations achieved cross validation statistics of Rcv2 = 0.85-0.93; RMSECV = 0.61-0.96% DM, while a combined cultivar (global) model had satisfactory statistics of Rcv2 = 0.84; RMSECV = 0.99% DM.

Seasonal climate forecasts may be coupled with crop models to provide quantitative forecasts of crop yield, assess sensitivity to farm management decisions and manage risk associated with seasonal climate variability. Today, seasonal climate forecasts are produced by computationally expensive, physically-based global climate models, which capture large-scale climate patterns well. However, their coarse spatial resolution (typically >50km) means they do not reliably depict daily weather at sub-grid locations, limiting their direct use in crop models. Consequently, operational crop forecasting systems in Australia typically use alternative meteorological forcings such as historical climate analogues based on El Niño - Southern Oscillation phases, which may be less skillful than global climate model forecasts. An emerging tactic for coupling global climate model forecasts and crop models is to apply quantile mapping (otherwise known as cumulative distribution function matching) to adjust forecast ensemble members according to the historical distribution of observations. However, quantile mapping assumes the global climate model forecasts are highly skilful and well-behaved (which they are often not). The overly simplistic formulation of quantile-mapping propagates an assortment of model errors. Additionally, quantile-mapping cannot be used for downscaling to multiple sub-grid locations owing to its deterministic nature. Accordingly, an increasing number of studies are reporting negative results arising from coupling global climate model forecasts and crop models using quantile mapping. Hence, the overarching objective of this thesis is to develop more robust, spatially and temporally relevant post-processing methods to harness global climate model forecasts for use in crop models. To this end, I develop a new multivariate forecast post-processing workflow that combines Bayesian parametric methods and non-parametric methods to calibrate and downscale global climate model forecasts for use in crop models. Forecast calibration means to (1) minimise systematic error such as forecast bias, (2) ensure forecast uncertainty is reliably conveyed by ensemble spread, and (3) ensure forecasts are at least as skilful as climatology. Downscaling means, depending on the context, either: (1) producing a revised forecast with the correct local weather variability at a spatial scale smaller than the GCM grid (2) producing a local forecast based on large-scale climate drivers (e.g. sea surface temperature patterns) (this approach is also referred to as bridging), or (3) spatial or temporal disaggregation of a forecast. Crop forecasting models require physically-coherent inputs of rainfall, temperature and solar radiation. Previous research has established the suitability of the Bayesian joint probability modelling approach for calibrating monthly and three-monthly rainfall forecasts from global climate models. The Bayesian joint probability modelling approach has not previously been applied to post-process temperature or solar radiation forecasts or to post-process multivariate forecasts. However, it is formed on the general assumption that the joint distribution of two or more variables can be modelled as a multivariate normal distribution in transformed space. It can theoretically be extended for multivariate forecast post-processing with a relevant transformation for each variable. Thus the first objective of this thesis is to develop and evaluate several strategies for calibrating multivariate global climate model forecasts using the Bayesian joint probability modelling approach. Three strategies are compared: (1) simultaneous calibration of multiple climate variables in a single statistical model, which explicitly models inter-variable dependence via the covariance matrix; (2) univariate calibration coupled with an empirical ensemble reordering method (the Schaake Shuffle) that injects inter-variable dependence from historical data; and (3) quantile-mapping, which borrows inter-variable dependence from the raw forecasts. Applied to Australian seasonal (three-month) forecasts from the European Centre for Medium-range Weather Forecasts System4 model, univariate calibration paired with the Schaake Shuffle performs best in terms of univariate and multivariate forecast verification metrics. Direct multivariate calibration is the second-best method, with its far superior performance in in-sample testing vanishing in cross-validation, likely because of insufficient data to reliably infer the sizeable covariance matrix. Bayesian joint probability post-processing is confirmed to outperform quantile-mapping. Hence the Bayesian joint probability modelling approach and the Schaake Shuffle should, therefore, be preferred to quantile-mapping as a basis for calibrating GCM forecasts for crop forecasting applications. Global climate model forecast skill is best captured by post-processing on seasonal time scales. However, crop models require daily forecast sequences. Also, it is observed that some operational crop forecasting systems run separate crop models for multiple locations within a region and then aggregate the results into a regional forecast. Therefore, spatial forecasts are also needed. Accordingly, the second objective of this thesis is to develop and evaluate downscaling and disaggregation methods for post-processing global climate model forecasts to higher spatial and temporal resolutions. To this end, I develop an empirical multivariate downscaling method that imparts observed spatial, temporal and inter-variable relationships into disaggregated forecasts whilst completely preserving the joint distribution of forecasts post-processed at coarser spatial and/or temporal scales. Specifically, a Euclidean distance metric is devised to identify a nearest-neighbour in historical observations for each forecast ensemble member. The method of fragments is subsequently applied to simultaneously disaggregate the forecast spatial and temporally. The new method is demonstrated to perform well for downscaling skilful forecasts of rainfall, temperature and solar radiation for six locations in northeast Australia. The climatological distributions of the downscaled forecasts mirror observations and the observed frequency of wet days is also reproduced in forecasts. The new downscaling method is a step towards full integration of calibrated seasonal climate forecasts into crop models and has a significant advantage over quantile-mapping in that it can be applied for multiple sub-grid locations. The final objective of this thesis is to feed global climate model forecasts, post-processed using the new methods, to a crop decision support system to demonstrate an end-to-end solution for linking global climate model forecasts with a crop model to produce yield forecasts. The first crop forecasting application of the new methods is for sugarcane yield forecasting in Tully. The region is selected because it is a non-irrigated region, and it is thus suitable for assessing the value of climate forecasts. Two sets of post-processed forecasts are produced for the Tully Mill weather station in North-east Queensland. The first set is obtained by applying the Bayesian joint probability modelling approach to calibrate monthly rainfall, temperature and solar radiation forecasts for the grid cell containing Tully. The second set is obtained by using global climate model forecasts of the Niño 3.4 climate index (commonly associated with the El Niño Southern Oscillation), also using the Bayesian joint probability modelling approach, to produce local forecasts of monthly rainfall, temperature and solar radiation. In both cases, the monthly forecasts are subjected to the Schaake Shuffle and subsequently downscaled to daily sequences using identical methods. The calibration and bridging forecasts are used to drive a sugarcane crop model to generate long-lead forecasts of biomass in north-eastern Australia from 1982-2016. A rigorous probabilistic assessment of forecast attributes suggests that the calibration forecasts provide the most skilful forecasts overall although the bridging forecasts give more skilful yield forecasts at certain times. The biomass forecasts are unbiased and reliable for short to long lead times, suggesting that the new downscaling methods are effective. My end-to-end solution for linking global climate model forecasts and crop models enables quantitative modelling and risk management at the farm level. It has the potential to improve farm productivity and profitability through better decisions. Future research should investigate the value of the post-processing methods for a wide range of crops.

碩士<br>輔仁大學<br>統計資訊學系應用統計碩士班<br>100<br>As the increase of the economy, wages and consumer price, the pressure of Inflation causes production and marketing cost to increase. In addition, after joined the WTO, Taiwan faced the trend of trade liberalization, and had to compete against the whole world, this situation lead Taiwan to be confronted with a significant challenges. Therefore, to get correct and useful information, and grasp the changes of market supply and demand are able to react the changes in market. This study took cabbage for example, under the situation that full of uncertainty in the process of agricultural produce, regarded import and export trade, origin price, trading volume, and climate information as influence variables, and used data mining techniques to establish CRISP-DM process included regression analysis, time series, neural network, SVR and Random Forests and MARS prediction methods to find out the best agricultural forecasting model of crop price and yield. The results showed that MARS is the best model in Yield and SVR is the best model in price. This study expect the results can assist the related governmental units to obtain detailed price and yield early warning system quickly, and make countermeasure in advance, to improve the ability of agricultural information and production stability.

The current study focused on forecasting and monitoring of durum wheat productions in Val d’Orcia (Tuscany region), with the following objectives: Objective 1. Evaluate the impact of temperature and water conditions on grain yield and GPC. Objective 2. Assess the performance of the complex crop model CERES-Wheat in the simulation of yield and GPC and in determining of key growth stages and of weather variables with greatest effect on the harvest. Objective 3. Revisit the algorithms adopted by CERES-Wheat for GPC simulation and carry out a diagnosis to trace the model deficiencies. Objective 4. Set up forecasting indices suitable for operational applications at farm level, and able to provide information about the quantity and quality of the harvest in order to assist with the application of late fertilization. Objective 5. Assess the improvement in the yield and GPC simulations by crop model due to the integration with RS data, based on a relatively simple procedure for the model output spatialization. Objective 6. Compare the performance of the satellite imagery related to RS indices in the monitoring of yield and GPC variability and trace the deficiencies in GPC description.

D.Litt. et Phil. (Geography)<br>The aim of this study is to integrate LANDSAT multispectral digital data with agricultural statistics, to analyse, explain and forecast the spatial variation of crop production in the Vermaas development area (south of Lichtenburg, Western Transvaal). This aim answers the urgent need for a reliable agricultural data base that can be quickly and cheaply obtained and used for the timely planning of an environment's limited agricultural resources. With such a data base available, early decisions about imports and exports can be taken in connection with the expected agricultural commodities of an area: the year-to-year fluctuation in crop yields is still the main problem in relation to the overall planning of agricultural food production. The study has been conducted according to two main analytical phases, i.e. (i) the interpretation of the data, which in turn was subdivided into: - the cartographic-analytical evaluation of the agricultural information, and - the recognition of rural land-use patterns from LANDSAT digital data. (i i) the integration process. The LANDSAT land-use information was integrated with the observed agricultural statistics with the aid of two integration models: an empirical and an operational model. The data for the research consisted of the multispectral digital data of LANDSAT-l and available agricultural statistics. The LANDSAT data was acquired from the Satellite Remote Sensing Centre at Hartbeeshoek, while the agricultural data was obtained from the Department of Agriculture (Highveld Region) and other official soures. These analytical phases were conducted at the computer centres of the CSIR and RAU. Existing computer programme packages were used - the VICAR system for pattern recognition, and the BMD and SYMAP systems for the analytical evaluation of the agricultural information and for the implementation of the integration models. The following results were obtained: 3.1 The integration of the LANDSAT information with the agricultural statistics was reasonably successful. The success of any study of this nature can be ascertained from the accuracy with which the necessary information is derived from the LANDSAT multispectral digital data. 3.2 This analysis highl ighted the cultivated area as a major factor for consideration. The type of crop and the area covered by it are the two most important sets of information that can be obtained from the LANDSAT data and used in an integration model. 3.3 The results (predicted crop yields) that were obtained from the integration process could probably be improved, if the detrimental influence of collinearity, which existed between some of the agricultural variables, was el iminated. 3.4 The identification of different crops from the LANDSAT digital data was not possible - a fact which can be attributed to the lack of a crop calendar for this farming area. Besides the above-mentioned results, the following can also be listed: 4.1 The spatial variation In maize production was well analysed in terms of the integration results, In spite of the fact that the accuracy of the agricultural statistics was, in certain cases, questionable. 4.2 The important influence of time upon the spatial variation in crop production could not be implicated, because of the one point in time consideration of this study. 4.3 Only the agricultural variables that were directly related to farm area could be used as input data for this study. 4.4 The potential usefulness of the LANDSAT digital data as geographical information is mainly determined by its quality (cloudcover, resolution, etc.). 4.5 The application of multispectral digital data depends on certain specific techniques, with which the researcher must acquaint himself for a successful and useful interpretation of the digital data.

Questo studio concerne l’analisi dei mercati delle materie prime da una prospettiva sia micro che macroeconomica e si compone di tre parti. Nella prima parte vengono analizzate le strategie di gestione ex-ante del rischio, in particolare la selezione di un portafoglio diversificato di colture, messe in atto in paesi poveri da famiglie produttrici fortemente dipendenti per la loro sussistenza dalla produzione di materie prime agricole. A tal fine, viene specificato un modello di portafoglio da cui si ottengono stime strutturali dei parametri relativi alla tecnologia, al consumo e alle preferenze verso il rischio dei produttori. Il modello è stimato utilizzando dati longitudinali da un campione di produttori di caffè in Etiopia. La seconda parte concerne la modellizzazione dei mercati delle materie prime in un contesto di aspettative razionali. L’attenzione è posta in particolare sulle colture pluriennali e viene specificato un modello globale per il mercato del cacao che incorpora la domanda di stoccaggio per fini speculativi. Dal modello strutturale si deriva una forma ridotta nelle variabili prezzo e stock, utilizzando due nuove variabili costruite per rappresentare eccessi di offerta di breve e lungo periodo. Viene quindi illustrata la derivazione analitica di restrizioni che emergono dall’utilizzo dell’ipotesi di aspettative razionali. Inoltre, viene proposta in un contesto deterministico un’analisi di equilibrio del mercato, utilizzando la soluzione in aspettative razionali del modello rispetto al variabile prezzo, al fine di studiare l’evoluzione del mercato in risposta a shock esogeni. Nella terza parte, la forma ridotta nelle variabili prezzo e stock, derivata in precedenza da una versione di breve periodo del modello, viene stimata utilizzando dati annuali relativi al mercato del cacao e le restrizioni derivanti dall’ipotesi di aspettative razionali vengono testate. Le stime ottenute attraverso il metodo dei momenti generalizzato (GMM) vengono quindi confrontate con quelle ottenute da un modello vettoriale autoregressivo (VAR) che presenta una specificazione equivalente.<br>This work concerns the analysis of primary commodity markets from both a micro and macro perspective and is composed of three parts. In the first part, we investigate how rural households in poor countries, depending for their livelihood on crop production, cope ex-ante with risk through a strategy of a diversified portfolio of crops. To this end, a portfolio model of production is set up from which structural estimates of risk preference and technology parameters are derived. The model is fit to longitudinal data from a sample of coffee producers in Ethiopia. The second part is concerned with the modelling of commodity markets within a rational expectations approach. A special emphasis is placed on perennial crops and a world model for the cocoa market is specified, accounting for speculative stockholding. From the structural model a solved form in price and stocks is derived, using two constructed variables capturing excess supply in the short and long term. The derivation of restrictions stemming from the hypothesized rational expectations by stockholders is then illustrated. Furthermore, an equilibrium analysis of the model is carried out using the price rational expectations solution, in order to investigate the qualitative response of the system to shocks. In the third part, the reduced form in price and stocks previously derived from a short-run version of the model is estimated using annual data on the cocoa market and the rational expectations restrictions are tested. The estimates obtained using the generalized method of moments (GMM) are then compared with those obtained from a restricted vector autoregressive (VAR) model presenting a matching specification.

The agriculture sectors in Canada are highly vulnerable to a wide range of inter-related weather risks linked to seasonal climate variability (e.g., El Ni ̃no Southern Oscillation(ENSO)), short-term extreme weather events (e.g., heatwaves), and emergent disease(e.g., grape powdery mildew). All of these weather-related risks can cause severe crop losses to agricultural crop yield and crop quality as Canada grows a wide range of farm products, and the changing weather conditions mainly drive farming practices. This dissertation presents three machine learning-based statistical models to assess the weather risks on the Canadian agriculture regions and to provide reliable risk forecasting to improve the decision-making of Canadian agricultural producers in farming practices. The first study presents a multi-scale, cluster-based Principal Component Analysis(PCA) approach to assess the potential seasonal impacts of ENSO to spring wheat and barley on agricultural census regions across the Canada prairies areas. Model prediction skills for annual wheat and barley yield have examined in multi-scale from spatial cluster approaches. The ’best’ spatial models were used to define spatial patterns of ENSO forcing on wheat and barley yields. The model comparison of our spatial model to non-spatial models shows spatial clustering and ENSO forcing have increase model performance of prediction skills in forecasting future cereal crop production. The second study presents a copula-Bayesian network approach to assess the impact of extreme high-temperature events (heatwave events) on the developments of regional crops across the Canada agricultural regions at the eco-district-scale. Relevantweather variables and heatwave variables during heatwave periods have identified and used as input variables for model learning. Both a copula-Bayesian network and Gaussian-based network modeling approach is evaluated and inter-compared. The copula approach based on ’vine copulas’ generated the most accurate predictions of heatwave occurrence as a driver of crop heat stress. The last study presents a stochastic, hybrid-Bayesian machine-learning approach to explore the complex causal relationships between weather, pathogen, and host for grape powdery mildew in an experimental farm in Quebec, Canada. This study explores a high-performance network model for daily disease risk forecast by using estimated development factors of pathogen and host from recorded daily weather variables. A fungicide strategy for disease control has presented by using the model outputs and forecasted future weather variability. The dissertation findings are beneficial to Canada’s agricultural sector. The inter-related weather risks explored by the three separate studies in multi-scales provide a better understanding of the interactions between changing weather conditions, extreme weather, and crop production. The research showcases new insights, methods, and tools for minimizing risk in agricultural decision-making<br>Graduate<br>2021-08-19

South Africa is the largest producer of sugar in Africa and one of the ten largest sugarcane producers in the world. Sugarcane in South Africa is grown under a wide range of agro-climatic conditions. Climate has been identified as the single most important factor influencing sugarcane production in South Africa. Traditionally, sugarcane mill committees have issued forecasts of anticipated production for a region. However, owing to several limitations of such committee forecasts, more advanced technologies have had to be considered. The aim of this study has been to develop, evaluate and implement a pertinent and technologically advanced operational sugarcane yield forecasting system for South Africa. Specific objectives have included literature and technology reviews, surveys of stakeholder requirements, the development and evaluation of a forecasting system and the assessment of information transfer and user adoption. A crop yield model-based system has been developed to simulate representative crops for derived Homogeneous Climate Zones (HCZ). The system has integrated climate data and crop management, soil, irrigation and seasonal rainfall outlook information. Simulations of yields were aggregated from HCZs to mill supply area and industry scales and were compared with actual production. The value of climate information (including climate station networks) and seasonal rainfall outlook information were quantified independently. It was concluded that the system was capable of forecasting yields with acceptable accuracy over a wide range of agro-climatic conditions in South Africa. At an industry scale, the system captured up to 58% of the climatically driven variability in mean annual sugarcane yields. Forecast accuracies differed widely between different mill supply areas, and several factors were identified that may explain some inconsistencies. Seasonal rainfall outlook information generally enhanced forecasts of sugarcane production. Rainfall outlooks issued during the summer months seemed more valuable than those issued in early spring. Operationally, model-based forecasts can be expected to be valuable prior to the commencement of the milling season in April. Current limitations of forecasts include system calibration, the expression of production relative to that of the previous season and the omission of incorporating near real-time production and climate information. Several refinements to the forecast system are proposed and a strong collaborative approach between modellers, climatologists, mill committees and other decision makers is encouraged.<br>Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2005.

У роботі розглядаються теоретичні аспекти дослідження та прогнозування кон’юнктури ринку, узагальнене поняття ринкової кон’юнктури, визначені особливості товарного ринку, розглянуті основні методики дослідження та прогнозування кон’юнктури ринку. Виявлені різновиди та структура ринку зерна. Проведено аналіз кон’юнктури ринку зерна. Визначені основні суб’єкти на ринку зерна. Проаналізовано кон’юнктуру ринку зерна в Україні. Приведена загальна характеристика кон’юнктурних змін на світовому ринку зерна, позиції та перспективи України. Визначені проблеми розвитку зернового ринку України. Спрогнозовані кон’юнктурні показники ринку зерна України за методом балансу змінних та методом ВФ Кобба-Дугласа-Тімбергена.<br>The theoretical aspects of research and prognostication of the state of affairs of market, generalized concept of the market state of affairs, certain features of commodity market, are in process examined, basic methodologies of research and prognostication of the state of affairs of market consider revising. Varieties and grain market structure are revealed. An analysis of the grain market conjuncture is carried out. The main subjects in the grain market are identified. The conjuncture of the grain market in Ukraine is analyzed. The general characteristic of conjunctural changes on the world grain market, the position and prospects of Ukraine are presented. The problems of development of grain market of Ukraine are determined. Predicted conjuncture indicators of the Ukrainian grain market.

博士<br>國立中興大學<br>農藝學系所<br>103<br>The brown planthopper (Nilaparvata lugens, Stal) is an important pest insect which affects rice production in Taiwan. Incorrect control strategies will reduce the quality and yield of rice crops. The most commonly used traps to monitor the migration of brown planthoppers in Taiwan are light traps and air borne net traps placed in the rice field. Other sampling methods include visual counting and sweep net counts in monitored fields after transplanting. The population abundance is an important issue because it affects the rice yielding loss. As usual, data on the brown planthoppers are monitored sequentially by time. Thus, using the statistical method of time series to estimate the population abundance is a feasible approach. In this study, population abundance data based on four survey methods collected from monitored paddy fields and traps from 1988 to 2012 in Chiayi County were investigated. In our approach, the rice development was stratified into three phases and the change rate of the brown planthopper population was estimated according to three rice phases. Population abundance was estimated by using an exponential growth observation error model (EGOE). Relationships between population abundances and each of the survey methods were presented. The results showed that population change rates during reproductive phase were larger than those of the other phases. The 40th day after transplanting is a critical time point for the outbreak of the brown planthopper population. According to our data, during reproductive phase, the population change rate of the brown planthopper population increased drastically. Without prompt and appropriate action, this could lead to serious reduction in quality and yield for rice crops. The brown planthoppers can generally immigrate into Taiwan every year from neighboring areas. Moreover, the immigration time and population abundance are often changeable. By a system of long-term monitoring of the insect, a time-series population fluctuation of brown planthoppers can be recorded. The forecasting system for the outbreak time of brown planthoppers can provide early warning and information on chemical application for safety rice production. In this study, population fluctuations based on daily data collected from light traps in Chiayi County from 1988 to 2012 were used. Owing to the autocorrelation of the data, the three-regime threshold autoregressive (TAR) statistical model of time series was utilized. Firstly, the data from 1988 to 2003 was used to establish the predicted model. Secondly, the data from 2004 to 2012 were employed to test the validity of the predicted model. A long-term forecast provided 110-160 days of prediction after the first prediction date and was used to estimate daily forecasting data in the second crop season. Results showed that most of the forecasting trends are near the trends of the observed data. For short-term forecasting, we used the results of one day forecasting to those of fourteen day forecasting to describe the precision of the forecasting model. The results indicated that the trend of seven day forecasting is recommended. That is, our forecasting model could effectively estimate population fluctuations seven days in advance. In short, the results of this study are helpful in characterizing fluctuations in brown planthopper population and in providing trends for a forecasting model for these fluctuations, and their effects on rice production in Taiwan. Thus, I believe that our finding can be helpful in the field of plant protection.

У роботі розглядаються сутність конкурентоспроможності підприємства, методи оцінки конкурентоспроможності підприємства, аналіз основних чинників та напрямів підвищення конкурентоспроможності підприємства, надається загальна характеристика торгівельного підприємства ДП «Адідас-Україна»; проаналізовано фінансовий стан підприємства; пропонуються шляхи підвищення конкурентоспроможності; надано оцінку економічної ефективності запропонованих заходів.<br>Thesis consists of three chapters. Object of study is the process of taking place in the grain market in Ukraine, its modern state and prospects for its further development. The theoretical aspects of research and prognostication of the state of affairs of market, generalized concept of the market state of affairs, certain features of commodity market, are in process examined, basic methodologies of research and prognostication of the state of affairs of market consider revising. Varieties and grain market structure are revealed. An analysis of the grain market conjuncture is carried out. The main subjects in the grain market are identified. The conjuncture of the grain market in Ukraine is analyzed. The general characteristic of conjunctural changes on the world grain market, the position and prospects of Ukraine are presented. The problems of development of grain market of Ukraine are determined. Predicted conjuncture indicators of the Ukrainian grain market.

The El Niño – Southern Oscillation (ENSO) has far-reaching impacts across the globe and provides the most reliable source of seasonal to interannual climate prediction over North America. Though numerous studies have discussed the impacts of ENSO teleconnections on North America during boreal winter, it is becoming more and more apparent that the regional impacts of ENSO teleconnections are highly sensitive to the seasonal evolution of ENSO events. Also, the significant impacts of ENSO are not limited to the boreal winter seasons. To address these knowledge gaps, this thesis examines the seasonal dependence of ENSO teleconnections and impacts on North American surface climate, focusing on two examples. Chapter 1 examines the relationship between El Niño – California winter precipitation. Results show that the probability of the anomalous statewide-wetness increases as El Niño intensity increases. Also, the influences of El Niño on California winter precipitation are statistically significant in late winter (Feb-Apr), but not in early winter even though that is when El Niño usually reaches its peak intensity. Chapter 2 further investigates why the strong 2015/16 El Niño failed to bring above normal winter precipitation to California, focusing on the role of westward shifted equatorial Pacific sea surface temperature anomalies (SSTAs) based on two reasons: the maximum equatorial Pacific SSTAs was located westward during the 2015/16 winter compared to those during the 1982/83 and 1997/98 winters, both of which brought extremely wet late winters to California. Also, the North American Multi-Model Ensemble (NMME) forecasts overestimated the eastern tropical Pacific SSTAs and California precipitation in the 2015/16 late winter, compared to observations. The Atmospheric General Circulation Model (AGCM) experiments suggested that the SST forecast error in NMME contributed partially to the wet bias in California precipitation forecast in the 2015/16 late winter. However, the atmospheric internal variability could have also played a large role in the dry California winter during the event. ENSO also exerts significant impacts on agricultural production over the Midwest during boreal summer. Chapter 3 examines the physical processes of the ENSO summer teleconnection, focusing on the summer when a La Niña is either transitioning from an earlier El Niño winter or persisting from an existing La Niña winter. The results demonstrate that the impacts are most significant during the summer when El Niño is transitioning to La Niña compared to that when La Niña is persisting, even though both can loosely be defined as developing La Niña summer. During the transitioning summer, both the decaying El Niño and the developing La Niña induce suppressed deep convection over the tropical Pacific and thereby the corresponding Rossby wave propagations toward North America, resulting in a statistically significant anomalous anticyclone over northeastern North America and, therefore, a robust warming signal over the Midwest. These features are unique to the developing La Niña transitioning from El Niño, but not the persistent La Niña. In Chapter 4, we further evaluate the performance of NCAR CAM5 forced with historical SSTA in terms of the La Niña summer teleconnections. Though the model ensemble mean well reproduces the features in the preceding El Niño/La Niña winters, the model ensemble mean has very limited skill in simulating the tropical convection and extratropical teleconnections during both the transitioning and persisting summers. The weak responses in the model ensemble mean are attributed to large variability in both the tropical precipitation, especially over the western Pacific, and atmospheric circulation during summer season. This thesis synthesizes the physical processes and assessments of climate models in different seasons to establish the sensitivity of regional climate to the seasonal dependence of ENSO teleconnections. We demonstrate that the strongest impacts of ENSO on North American regional climate might not be necessarily simultaneous with maximum tropical Pacific SST anomalies. We also emphasize the importance of the multi-year ENSO evolutions when addressing the seasonal impacts on North American summertime climate. The findings in this thesis could benefit the improvement of seasonal hydroclimate forecasting skills in the future.