Forage Mass Estimation in Silvopastoral and Full Sun Systems: Evaluation through Proximal Remote Sensing Applied to the SAFER Model

• Autor(a) principal: Luns Hatum de Almeida, Samira [UNESP]
• Data de Publicação: 2023
• Outros Autores: Brunno Costa Souza, Jarlyson [UNESP], Furlan Nogueira, Sandra, Ricardo Macedo Pezzopane, José, Heriberto de Castro Teixeira, Antônio, Bosi, Cristiam, Adami, Marcos, Zerbato, Cristiano [UNESP], Carlos de Campos Bernardi, Alberto, Bayma, Gustavo, Pereira da Silva, Rouverson [UNESP]
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://dx.doi.org/10.3390/rs15030815; http://hdl.handle.net/11449/249651
• topic: digital agriculture; meteorological data; proximal sensor; Urochloa brizantha

The operational slowness in the execution of direct methods for estimating forage mass, an important variable for defining the animal stocking rate, gave rise to the need for methods with faster responses and greater territorial coverage. In this context, the aim of this study was to evaluate a method to estimate the mass of Urochloa brizantha cv. BRS Piatã in shaded and full sun systems, through proximal sensing applied to the Simple Algorithm for Evapotranspiration Retrieving (SAFER) model, applied with the Monteith Radiation Use Efficiency (RUE) model. The study was carried out in the experimental area of Fazenda Canchim, a research center of Embrapa Pecuária Sudeste, São Carlos, SP, Brazil (21°57′S, 47°50′W, 860 m), with collections of forage mass and reflectance in the silvopastoral systems animal production and full sun. Reflectance data, as well as meteorological data obtained by a weather station installed in the study area, were used as input for the SAFER model and, later, for the radiation use efficiency model to calculate the fresh mass of forage. The forage collected in the field was sent to the laboratory, separated, weighed and dried, generating the variables of pasture total dry mass), total leaf dry mass, leaf and stalk dry mass and leaf area index. With the variables of pasture, in situ, and fresh mass, obtained from SAFER, the training regression model, in which 80% were used for training and 20% for testing the models. The SAFER was able to promisingly express the behavior of forage variables, with a significant correlation with all of them. The variables that obtained the best estimation performance model were the dry mass of leaves and stems and the dry mass of leaves in silvopastoral and full sun systems, respectively. It was concluded that the association of the SAFER model with the proximal sensor allowed us to obtain a fast, precise and accurate forage estimation method.