Eficiência de modelos de estimativa via sensoriamento remoto na evapotranspiração e coeficiente de cultura do algodoeiro
Data
2020-02-27
Autores
Moncada, Juan Vicente Liendro
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal de Mato Grosso
Resumo
Brazil is the fourth largest global producer of cotton and the second
largest exporter of this fiber, in addition to having the second place in terms of
productivity. In this world panorama, the State of Mato Grosso (MT) is the number one
cotton producer in the country with 66,61% of the total Brazilian production. Thus, the
problem of spatial and temporal estimation of water needs for cotton cultivation in
extensive agricultural production areas arises. Therefore, the objective of the study
was to determine the efficiency of estimation models by means of remote sensing in
evapotranspiration (ETc) and crop coefficient (Kc) of cotton (Gossypium sp. L.) during
the stages of the plant's phenological cycle. The research was carried out on eight
cotton fields located in the upper part of the Rio das Mortes (MT) hydrographic basin
using data and information accessible from the Campo Verde and Primavera do Leste
meteorological stations, associated with the National Institute of Meteorology (INMET)
from Brazil to determine the reference evapotranspiration (ETo) by the FAO PenmanMonteith method in the study area. The surface energy balance algorithms SEBAL
(Surface Energy Balance Algorithm for Land) and METRIC (Mapping
Evapotranspiration at High Resolution with Internalized Calibration) were implemented
using satellite images from the Landsat 8 program. The development of the research
took place in a Geographic Information Systems environment, using the capabilities of
the free software QGIS 3.6.2 and GRASS 7.6.1, and the EEFlux platform (Earth Engine
Evapotranspiration Flux) version 0.10.10 of the Google Earth Engine system. The
algorithm estimates were compared with determinations made by the FAO 56 method,
by simple differences in the case of Kc, and through statistical analysis of simple linear
regression for ETc. The results of the research show that in the set of fields analyzed,
the SEBAL model reached an average daily ETa close to 5,61 and 3,21 mm d-1, in the
intermediate and final stages of the cotton phenological cycle, with an overall efficiency
around 67%. The average Kc of the algorithm was close to 1,27 and 0,73 in the
intermediate and final stages, with global performances of approximately 92 and 96%
respectively. The average total water consumption was 775,43 mm. The model
showed an absence of data and information in the initial phase of the cycle, due to the
occurrence of the rainy season in the study area. Regarding the METRIC model, the
results indicate that it reached an average daily ETa close to 4,14 (initial stage); 3,68
(development stage); 3,28 (intermediate stage) and 2,86 mm d-1 (final stage), with
overall efficiency around 20%. The average Kc of the algorithm was close to 0,89 (initial
phase); 0,83 (intermediate phase) and 0,62 (final phase), with overall performances of
approximately 91; 73 and 100% respectively. The average of total water consumption
was 567,64 mm. In general, the SEBAL model surpassed the METRIC EEFlux model
in efficiency, in the ETc and Kc estimates of the cotton in the study area, when
compared with the FAO 56 method.
Descrição
Palavras-chave
Citação
MONCADA, Juan Vicente Liendro. Eficiência de modelos de estimativa via sensoriamento remoto na evapotranspiração e coeficiente de cultura do algodoeiro. 2020. 89 f. Dissertação (Mestrado em Engenharia Agrícola) - Universidade Federal de Mato Grosso, Instituto de Ciências Agrárias e Tecnológicas, Rondonópolis, 2020.