RELATIONSHIP BETWEEN COFFEE CROP PRODUCTIVITY AND VEGETATION INDEXES DERIVED FROM OLI / LANDSAT-8 SENSOR DATA WITH AND WITHOUT TOPOGRAPHIC CORRECTION

Sulimar M C Nogueira,Maurício A Moreira,Margarete M L Volpato

doi:10.1590/1809-4430-eng.agric.v38n3p387-394/2018

Sulimar M C Nogueira, Maurício A Moreira + Show 1 more

Open Access

https://doi.org/10.1590/1809-4430-eng.agric.v38n3p387-394/2018

Copy DOI

Abstract

ABSTRACT The reflectance values of a coffee crop are influenced by several factors such as planting direction, crop spacing, time of the year, plant age and topography which reduces the accuracy of the estimates derived from remote sensing data. In this context were evaluated the relationships between coffee productivity and values of NDVI, SAVI and NDWI vegetation indexes with and without topographic reflectance correction for different coffee phenological phases for the crop years 2013/2014 (low productivity) and 2014/2015 (high productivity). The evaluations were made through the standard deviation of vegetation indices (VIs), linear relationship between the cosine factor and the VIs and between VIs and coffee productivity. The best phenological phases of coffee to determine productivity from spectral indexes were the stages of dormancy and flowering. The results indicated that the NDVI was the best index to estimate the productivity of coffee trees with coefficient of determination (R2) that ranged from 0.58 to 0.90. There was an increase in R2 between productivity and NDVI with topographic correction in the dormancy phase in the year of low productivity; between productivity and NDVI with topographic correction in the flowering phase in the year of high productivity; and between productivity and SAVI and NDWI with topographic corrections in the flowering phase in the year of high productivity.

Highlights

Phenological information supports crop productivity and crop management (Sakamoto et al, 2005; Couto Junior et al, 2013)
On July 31, 2013 SoilAdjusted Vegetation Index (SAVI) presented the highest values of R2 for plots 1, 4, 5 and 8, while Normalized Difference Vegetation Index (NDVI) showed higher correlations with cosine factor for plots 2, 3 and 7
In the year of high productivity (2014/2015), SAVI presented lower correlation with cosine factor, with NDVI being more sensitive to topographic effects

Summary

Introduction

Phenological information supports crop productivity and crop management (Sakamoto et al, 2005; Couto Junior et al, 2013). Vegetation indices (VIs) are sensitive to phenological changes and have been used to correlate with agricultural productivity (Bolton & Friedl, 2013; Kogan et al, 2013; Fu et al, 2014), to estimate attributes such as Foliar Area Index that can be related to agricultural yield (Rembold et al, 2013; Taugourdeau et al, 2014; Jiang et al, 2014; Li et al, 2017; Liaqat et al, 2017) or to be incorporated into modeling (Padilla et al, 2012; Meroni et al, 2013; Kowalik et al, 2014). According to the same authors, the Engenharia Agrícola, Jaboticabal, v.38, n.3, p.387-394, may/jun. 2018

Objectives

Methods

Results

Conclusion