Fusão espaço-temporal de imagens termais e avaliação da rede de monitoramento meteorológico da região oeste do estado do Paraná
| Ano de defesa: | 2023 |
|---|---|
| Autor(a) principal: |
Mendes, Isaque de Souza
 |
| Orientador(a): | |
| Banca de defesa: | , , , , |
| Tipo de documento: | Tese |
| Tipo de acesso: | Acesso aberto |
| Idioma: | por |
| Instituição de defesa: |
Universidade Estadual do Oeste do Paraná
Cascavel |
| Programa de Pós-Graduação: |
Programa de Pós-Graduação em Engenharia Agrícola
|
| Departamento: |
Centro de Ciências Exatas e Tecnológicas
|
| País: |
Brasil
|
| Palavras-chave em Português: | |
| Palavras-chave em Inglês: | |
| Área do conhecimento CNPq: | |
| Link de acesso: | https://tede.unioeste.br/handle/tede/6933 |
Resumo: | Global climate changes affect the spatial distribution of temperature at regional and local scales. Smart Farm tools assist in timely decision-making. However, data collection continuity is of paramount importance in understanding local-scale climatic dynamics, thus redundancy is fundamental for the continuity of data capture and filling in case of systemic failures. Orbital sensing data, as well as government stations distributed in the region of interest, can be used as redundancy tools for Smart Farming systems. Nonetheless, analyzing the spatial variability of temperature distribution can be limited in small and medium cultivation areas due to the low spatial or temporal resolution of orbital sensors. This study aimed to evaluate the correlation between daily average surface temperature resulting from 3 and 4 daily observations and the daily average air temperature collected by meteorological stations. It also aimed to merge, using the ESTARFM algorithm, and evaluate the use of synthetic surface temperature data images from MODIS – Terra and Aqua sensors and TIRS – Landsat 8 and 9 sensors. Additionally, the spatial distribution of publicly available stations in the Western Region of the State of Paraná was assessed. The images of daily average surface temperature resulting from 3 and 4 observations showed a strong correlation with the daily average air temperature, with a correlation coefficient rs of 0.92 for both observations. The model fit had an adjusted coefficient of determination R²adjusted of 0.85 for 3 observations and 0.86 for 4 observations, with RMSE values of 1.74 °C and 1.5 °C, respectively. The synthetic images of daily average surface temperature had a correlation coefficient rs of 0.69 and an adjusted coefficient of determination R²adjusted of 0.59. There was an overestimation of average surface temperature values in synthetic images, and cloud cover posed an obstacle to the generation of larger volumes of data. The evaluation of meteorological stations was carried out based on data interpolation, and when compared to the modeled spatial temperature distribution for air temperature, they showed errors mainly in urban clusters, with correlation coefficients rs of 0.42 in summer, 0.53 in autumn, 0.51 in winter, and 0.63 in spring. However, when compared to stations installed in agricultural areas with environmental characteristics similar to the station locations, no statistical differences in data were observed, and the correlation coefficient rs was 0.93. |