Modelación del rendimiento de cultivo de palma de aceite a partir de imágenes satelitales multiespectrales, en el Pacífico Central de Costa Rica

Oil palm cultivation is one of the three largest crops in Costa Rica. It has grown from 16,830.20 ha in 1984 to 92,456 ha in 2017. A success factor is that 63.7% of the planted areas are in the hands of small, independent producers and cooperatives. Pre-harvest forecasting of fresh fruit bunches is an important means of assessing total production and therefore provides useful information for management decisions related to storage, distribution, and marketing budgets. However, due to climate variability, oil palm production yields fluctuate. Despite existing crop planning and harvest estimation methods, forecasts have wide margins of error due to random sampling by field personnel.

On the other hand, oil palms, being perennial trees, have a canopy structure similar to a forest, compared to other agricultural crops; therefore, remote sensing of oil palms can be based on aerial or satellite images. When using images to monitor plantations, the different conditions in different regions of the world have been a challenge for researchers. Specifically in the tropics, mapping vegetation cover is challenging due to the predominant cloud cover and dense biomass that can be mistaken for forests or other crops. Locally, freely accessible multispectral satellite images are available, providing information of up to 11 wavelengths (bands), with acceptable resolutions because the oil palm cultivation management areas in Costa Rica exceed 4 hectares

Automating estimates and reducing margins of error in order to improve management throughout the oil palm cultivation chain is an urgent requirement for producers. Due to the climatic variability of the Central Pacific of Costa Rica, it is urgent to find relationships between the multispectral nature of freely available images, yield, climatic and soil conditions, and crop age. This is to explain and predict production during safe lag times for production management, using satellite images that allow producers to manage more efficiently with available technology that does not represent onerous investments

Generar una herramienta que permita estimar la producción futura de palma aceitera a partir de imágenes multiespectrales satelitales de acceso libre, en plantaciones del Pacífico Central de Costa Rica