Crop area estimate from original and simulated spatial resolution data and landscape metrics

Abstract

Images acquired at the same day by the ETM+/Landsat-7 (30 m of spatial resolution) and MODIS/Terra (250 m) sensors were used to estimate areas of three major crops (soybean, sugarcane, and corn) with different landscape patterns in Southeastern Brazil. Majority filtering of ETM + classification results was applied to describe the behavior of 15 selected landscape metrics at distinct simulated spatial resolutions (90, 150, 210 and 270 m). By using regression models, the performance of MODIS and derived metrics to predict adequately the crop area, considering ETM+ data as reference, were analyzed. Results showed that the MODIS instrument overestimated the areas of soybean (15%) and sugarcane (1%), and underestimated the area of corn (12%). Multiple regression results indicated that coarse spatial resolution sensors can be used to predict adequately the area viewed by the 30 m spatial resolution instruments only for crops with low fragmentation pattern such as soybean. These sensors cannot be used to predict the area of corn due to aggregation pixel effects of the less fragmented crops (soybean and sugarcane) over the most fragmented one (corn), as demonstrated by the spatial resolution simulation using majority filtering of the ETM+ image. Landscape metrics improved MODIS area estimates only for sugarcane, as indicated by higher values of R 2 for multiple than for simple regression. Only a small set of metrics was select to compose the multiple regression models because most of them were not preserved across different spatial resolutions (30 m and 250 m). Keywords: MODIS, remote sensing, regression, soybean, sugarcane ESTIMATIVA DE ÁREA AGRÍCOLA COM DADOS DE RESOLUÇÃO ESPACIAL ORIGINAL E SIMULADA E MÉTRICAS DE PAISAGEMRESUMO: Imagens coletadas no mesmo dia pelos sensores ETM+/Landsat-7 (30 m de resolução espacial) e MODIS/Terra (250 m) foram utilizadas para estimar a área de três importantes culturas agrícolas (soja, cana-de-açúcar e milho) com diferentes padrões de paisagem no Sudeste Brasileiro. Filtragem de Maioria dos resultados da classificação da imagem ETM+ foi aplicada para descrever o comportamento de 15 métricas em diferentes simulações de resolução espacial (90, 150, 210 e 270 m). Utilizando modelos de regressão, o desempenho do MODIS e de suas métricas para predizer a área das culturas, considerando os dados ETM+ como referência, foi analisado. Os resultados mostraram que o sensor MODIS superestimou as áreas de soja (15%) e cana-de-açúcar (1%) e subestimou a área de milho (12%). A regressão múltipla indicou que sensores de resolução espacial grosseira podem ser usados para predizer adequadamente a área vista por instrumentos com 30 m de resolução espacial apenas para culturas com baixo padrão de fragmentação como soja. Estes sensores não podem predizer adequadamente a área de milho devido aos efeitos de agregação de pixels das culturas menos fragmentadas (soja e cana-de-açúcar) sobre a mais fragmentada (milho), conforme demonstrado pela simulação da resolução espacial por filtragem de maioria...

Estimating soybean crop areas using spectral-temporal surfaces derived from MODIS images in Mato Grosso, Brazil

“…This confusion can be explained by the pixel aggregation effect in spatial resolutions of 250 m, as described by Soares et al (2008). In fact, these authors, when working with sensors of different spatial resolutions, found that pixels from more fragmented classes, characterized by small polygons, were aggregated by less fragmented classes.…”

“…Previous works, carried out with the TM-Landsat and the CCD-CBERS sensors, with a spatial resolution of approximately 30 and 20 m respectively, showed that these sensors are appropriate for the quantifi cation of crop areas from the point of view of its spatial resolution (Chen, 1990;Sanches, 2004;Soares et al, 2008). However, the temporal resolution of such satellites, which ranges around 16 days or more, is still a problem in the operational use of this kind of data for crop forecasting, given the high probability of cloud cover along the summer season (Ippoliti-Ramilo, 1999), especially in non-irrigated crops at tropical regions.…”

Simulated multipolarized MAPSAR images to distinguish agricultural crops

“…Due to the continental dimension and the intense agricultural involvement of Brazil, information acquired from remote sensing images are important to various agricultural applications, such as crop monitoring, mapping and estimation of crop production (Galvão et al, 2009;Rizzi and Rudorff, 2007;Rudorff et al, 2010;Soares et al, 2008;Xavier et al, 2006). However, the availability of cloud free images during the primary growth period of summer crops is limited, particularly in tropical regions and, therefore, restricting the use of optical remote sensing imagery to monitor agricultural activities.…”

Transição Florestal no estado de São Paulo, Brasil: fatores associados ao desmatamento e recuperação das matas nativas

“…As São Paulo's forest cover is constituted mainly by very small patches(Ribeiro et al 2009), larger municipalities possibly have more isolated patches which may not be quantified by SOS, decreasing the difference between SOS and IBGE.EFFECTS OF LANDSCAPE CONFIGURATION ON DIFFERENCE BETWEEN IF'S AND SOS'S FOREST COVER ESTIMATESOur results indicate that discrepancies between forest cover estimates from different spatial resolution images are influenced by different landscape structure characteristics.High total forest area (CA) in the municipality is related to higher differences between IF and SOS, indicating that SOS may be underestimating the forest cover in municipalities with more remnants. Forest cover in most São Paulo's municipalities are highly fragmented and composed of very small patches(Ribeiro et al 2009) which can be underestimated by coarser spatial resolution maps(Soares et al 2008), such SOS maps. For the same reason, large Permanent Protection Areas (PPA), resulting in more linear patches along rivers (high SHAPE values) and longer edges (high total edge length) increase the differences between IF and SOS.…”