Determinação das unidades diferenciadas de gestão por meio dos atributos físico-hídricos e componentes espectrais, aplicado à cafeicultura

Abstract

Precision coffee farming requires an understanding of the spatial and temporal variability of fields in order to optimize management and ensure productive sustainability. This study aimed to delineate Management Zones (MZs) in a coffee plantation by integrating soil physical–hydric attributes and spectral components. The study was conducted at the IFMG Campus in São João Evangelista, evaluating the 2022 and 2024 growing seasons. The methodology consisted of georeferenced sampling of Soil Penetration Resistance (SPR) in the 0–25 cm and 25–50 cm layers, Field Capacity (FC), topographic survey (elevation), and monitoring of leaf miner incidence (Leucoptera coffeella) and leaf cover. Geostatistical techniques were applied to model spatial dependence, and clustering algorithms were used to define management zones.The results indicated that leaf miner distribution varied according to leaf tissue availability, showing a continuous pattern (drift) in 2022 (range of 5,688 m) and a fragmented pattern in 2024 (range of 75 m). Cluster analysis defined four distinct MZs: MZ 1 was characterized by hydrological limitation (lowest FC); MZ 2, located at higher elevations and with lower compaction levels, stood out for greater vegetative vigor, but also acted as an ecological refuge for pests in stress years; MZ 3 showed the highest water retention, but with spectral noise due to weed presence; and MZ 4 exhibited higher SPR levels at greater depths. It is concluded that the multivariate approach allowed the plantation to be segmented into units with distinct productive potential and constraints, validating the use of remote sensing and soil physical attributes as complementary tools for agronomic decision- making.