Search

Orientador : Belmer Garcia Negrillo Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Matematica, Estatistica e Ciencia da Computação Resumo: Não informado Abstract: Not informed Mestrado Mestre em Estatística

The way the researcher groups his research data will influence the result of his work. In the literature, this phenomenon is treated as a Problem of the Modifiable Areal Unit. The objective of this article was to analyze the three spatial levels by Municipalities, Regional Centers and Mesoregions using the following data: gross domestic product, effective agricultural production, grain production and gross value of agricultural production for the state of Paraná-Brazil in the period since 2012 until 2015. The methodological procedure studied data from the Paranaense Institute for Economic and Social Development of the above-named variables collected on the website of the Paranaense Institute for Economic and Social Development of the 399 municipalities, 23 regional centers and 10 mesoregions. The results found show the presence of the Modifiable Areal Unit Problem, presenting different results for each level of grouping. The study revealed the problem of the modifiable areal unit is a relevant occurrence and it should be disregarded by researchers who work with clusters of spatial data in their studies. The results found allow a better understanding of the scale effect and demonstrate the efficiency of spatial analysis in socioeconomic data.

The strategic logistics of agricultural production and storage aggregates information related to production and storage. In this sense, time, location, and distance from producer and consumer markets are considered, emphasizing the importance of grain storage and production logistics. The Natural Neighbor and multiquadric equation are spatial interpolation methods used to predict these variables value at non-sampled locations, for asymmetric and categorical data, respectively. This study investigated the spatial prediction of grain production (tons) (soybean, first crop corn, second-crop corn, and wheat) in the 2016/2017 growing season and qualitative data on the static capacity of warehouses in the 2017/2018 growing season. The result obtained through the spatial interpolation using the natural neighbor method was coherent, as it showed the high variability of grain production relative to the different meso-regions. Therefore, the method was appropriate because it allowed predicting the behavior of grain production in the 2016/2017 growing season in the state of Paraná-Brazil, making it possible to identify regions of higher or lower production. The result of the spatial interpolation using the multiquadric equations allowed identifying a higher predominance of storage units with a low static capacity of warehouses, but also enabled the detection of regions with a static capacity of warehouses that varied from the medium to the high category in the state of Paraná, Brazil.

This paper is an analysis of the influence that the industrial units of the Agroindustrial Cooperative ‘LAR’ have on the industrial development of the cities in the area surrounding Medianeira in the State of Paraná. This paper was based on the information matrix elaborated to correlate employment in the sector as a base variable with the manpower distributed according to the branches of activity. This information matrix was then used to describe the behavior of this production sector and specify the localization and specialization measures to be taken. To calculate these measures, the information was organized into a matrix that relates the sector-space distribution of the manpower base variable distributed according to the branches of activity in the cities surrounding Medianeira, where the Agroindustrial Cooperative ‘LAR’ has set up its industrial plants. The results of the research demonstrated that the Agroindustrial Cooperative ‘LAR’ contributes to the industrial development of the cities where it has installed industrial plants. Este artigo apresenta uma análise da influência das unidades industriais da Cooperativa Agroindustrial LAR no desenvolvimento industrial dos municípios da região de Medianeira, estado do Paraná. Neste trabalho construiu-se uma matriz de informações que co-relacionam o emprego no referido setor, como variável-base, com mão-de-obra, distribuída por ramos de atividade, que permite descrever o comportamento desse setor produtivo. A partir de uma matriz de informações descrevem-se as medidas de localização e especialização. Para o cálculo dessas medidas organizaram-se as informações em uma matriz que relacionam a distribuição setorial-espacial da variável-base mão-de-obra, distribuída por ramos de atividade dos municípios da região de Medianeira que são: Medianeira, Itaipulândia, Missal, Matelândia, Céu Azul e Santa Helena, onde a Cooperativa Agroindustrial LAR tem implantado plantas industriais. Os resultados da pesquisa demonstraram que a Cooperativa Agroindustrial LAR contribui para o desenvolvimento industrial dos municípios onde mantém instaladas plantas industriais.

The study on spatial variability of soil properties performed through geostatistical techniques allow us to identify the spatial distribution of phenomena by means of a spatial model that considers degree of dependence among observed data, depending on distance and also the direction that separate them, if there is geometric anisotropy, in other words, a directional trend in spatial continuity. However, the main difficulty in decision making regarding the use of anisotropic spatial model focuses on its relevance to the parameters that express the geometric anisotropy in a spatial model exercise in relation to the estimation space. This study aims at identifying the degree of influence of geometric anisotropy on the accuracy of spatial estimation using simulated data sets with different sample sizes and soil chemical properties such as: Fe, potential acidity (H + Al), organic matter and Mn. Comparing the isotropic and anisotropic models, especially for smaller sample sizes (100 and 169) showed an increased sum of squares of differences between predictions anisotropy factor (Fα) equals 2. Furthermore, from Fα equals 2.5, over 50% of the simulations showed values of overall accuracy (OA) of less than 0.80 and values for the concordance index Kappa (K) and Tau (T) from 0.67 to 0.80, indicating differences between thematic maps. Similar conclusions were obtained for chemical properties of the soil, from Fα equals 2, showing that there are relevant differences regarding the inclusion or not of geometric anisotropy.

There is high spatial variability of nitrogen availability within the same area. Supplying the required N can increase income, and reduce unnecessary N supply and its leaching to underground water. Measuring spectral reflectance and calculating vegetation indexes is considered promising non-destructive and non-contact approach to instantly measure N deficiency in corn. This study was done to determine the influence of the direction of the sowing and of the hour of the day on determination of the vegetation indexes of corn crop. The sowing lines followed North-South and East-West direction. The sowing direction and sun trajectory influenced the indexes: the NIR/red ratio, the Normalized Difference Vegetation Index (NDVI), the Soil-Adjusted Vegetation Index (SAVI), the Optimized Soil-Adjusted Vegetation Index (OSAVI), the NIR/green ratio and the Green Normalized Difference Vegetation Index (GNDVI). GNDVI was more homogeneous with lower coefficient of variation, thus appears to be more appropriate for use in real time N application and in variable rates, when control of sowing direction is not possible. In other cases SAVI and OSAVI are the more suitable as their models had higher R2. A necessidade de nitrogênio (N) apresenta grande variabilidade espacial dentro de uma mesma área. A aplicação da real necessidade de N pode aumentar o rendimento, reduzir a aplicação desnecessária e o fluxo para os lençóis freáticos. A medida da reflectância espectral e conseqüente cálculo de índices de vegetação são considerados promissoras abordagens não destrutivas e sem contato para o sensoriamento instantâneo da deficiência de N em milho. Neste trabalho procurou-se estudar a influência da direção da semeadura e da hora do dia nos índices de vegetação no milho. As linhas de semeadura foram localizadas na direção Norte-Sul (NS) e Leste-Oeste (LO). O direcionamento da semeadura em relação à trajetória do sol influenciou os índices estudados: relação infravermelho próximo pelo vermelho - IVP/Verm, índice de vegetação da diferença normalizada - IVDN, índice de vegetação ajustado ao solo - SAVI , índice de vegetação ajustado ao solo aperfeiçoado – OSAVI, relação infravermelho próximo pelo verde - IVP/Verd, índice verde de vegetação da diferença normalizada - IVVDN. O IVVDN apresentou-se mais homogêneo e com menor coeficiente de variação, mostrando-se mais interessante na utilização da aplicação de N em tempo real e a taxas variáveis quando não for possível controlar a direção da semeadura. Caso contrário os índices SAVI e OSAVI são os mais indicados por apresentarem modelos com maior R2.