Correlations and Principal Components Analysis for Defining Management Zones in Cotton

Abstract

One approach for using variable rate fertilizer applications in precision agriculture is to divide an area into management zones. The objectives were: (i) to identify the chemical, physical and phenological properties that have the highest correlation with the yield; (ii) to use principal component analysis (PCA) to identify what physical, chemical, and phenological properties contribute to greater spatial variability; (iii) and to use these variables in the establishing management zones (MZ) for cotton through fuzzy k-means clustering analysis, associated with the geostatistics technique by the ordinary kriging method. The experiment was carried out in a cotton field in the Chapadões region in 2015. Phenological variables of cotton (plant height, number of bolls, number of capsules, opening percentage and Red Edge vegetation index) and chemical (pH, Ca, Mg, H+Al, V%, Ca/Mg, CEC, K, Al3+ and P) and physical (total soil porosity, soil density, soil moisture, soil mechanical resistance to penetration, clay content, and macro and micro-porosity) attributes of the soil were evaluated to define management zones. The variables that showed the highest correlation with cotton yield were pH, phosphorus, soil moisture measured at 39 and 70 days after cotton emergence (DAE), number of bolls at 107 DAE and red edge vegetation index at 53 DAE. The map with four MZ has a better

Identifying Key Indicators for Monitoring Water Environmental Services Payment Programs—A Case Study in Brazil

“…The water data were subjected to a descriptive statistical analysis (mean, standard deviation, and variation coefficient) to understand the variability of the limnological indicators used, and later a principal component analysis (PCA) was applied to the data set to assess the contribution of each indicator to the spatial-temporal variability of quality. This approach allows us to assess which variable has the greatest weight in statistical analysis and which are the most important variables from a statistical point of view [26].…”