Busca avançada
Ano de início
Entree


EFFECT OF INTERPOLATION METHODS ON THE GENERATION OF FERTILIZER PRESCRIPTION MAPS UNDER LIMITED SAMPLING SCENARIOS

Texto completo
Autor(es):
Bejarano, Laura D. ; Oliveira, Agda L. G. ; Oldoni, Henrique ; Sanchez, Dario C. ; Amaral, Lucas R. Do
Número total de Autores: 5
Tipo de documento: Artigo Científico
Fonte: Engenharia Agrícola; v. 45, p. 12-pg., 2025-01-01.
Resumo

In precision agriculture (PA), the evaluation of soil spatial variability to optimize crop management requires dense sampling. This costly activity often results in sparser sampling grids and may compromise both map quality and the return on variable rate fertilizer applications. This study evaluated whether the choice of interpolation method influences the quality of fertilizer prescription maps under sampling limitations. In two areas with different amounts of samples and degrees of spatial dependence, the interpolation of phosphorus (P) and potassium (K) contents was evaluated using deterministic methods (TPS, IDW) that consider mathematical functions based on distance, stochastic methods (OK, KED) that consider the spatial autocorrelation of the data, and machine learning methods (SVM, RFSI) that learn complex relationships. TPS demonstrated superior performance in predicting content, and the agreement of fertilization classes (Lin's correlation coefficient and Kappa's agreement index were higher). However, considering the recommendation classes, the differences between the methods were reduced. Thus, if this approach to creating recommendation maps is adopted, the interpolator can be selected based on the simplicity of the method, with TPS being a promising alternative. (AU)

Processo FAPESP: 24/01557-3 - Desenvolvimento de uma abordagem (framework) para otimização de interpolação em cenários com diferentes densidades de amostras de solo
Beneficiário:Laura Delgado Bejarano
Modalidade de apoio: Bolsas no Brasil - Mestrado
Processo FAPESP: 22/03160-8 - Mapeamento da variabilidade espacial dos solos e amostragem otimizada com o apoio de técnicas de sensoriamento: bases para uma agricultura de precisão mais eficiente e sustentável
Beneficiário:Lucas Rios do Amaral
Modalidade de apoio: Auxílio à Pesquisa - Projeto Inicial