Use of kriging techniques to estimate Conilon coffee productivity

ABSTRACT

Knowing the spatial variability of a coffee-crop helps to understand some phenomena which can be correlated with its productivity. This study aimed estimating the productivity (kg ha-1) of conilon coffee-tree in three consecutive harvests using the cokriging method, considering the number of plagiotropic branches per plant as a covariable. In the center of a coffee-crop we settled a sampling grid with 109 georeferenced points, each point consisted of five plants. Subsequently the vegetative branches were counted for three consecutive growing seasons before performing manual harvesting of coffee beans. The harvested coffee was separated into cherry, yellowish and green and dried to a standard moisture of 12%. The dry cherry productivity (kg ha-1) was then evaluated. The three harvests evaluated showed linear correlation and spatial dependence among productivity and number of vegetative branches. Estimation of productivity by the number of vegetative branches showed similarities in cokriging, indicating that vegetative branches is an efficient covariable to estimate conilon coffee productivity.

Key words:
geostatistics; spatial variability; cross variogram