Exploring Spatial Variability and Yield-Determining Factors in Cotton Production: A Multivariate Analysis

Abstract

ABSTRACT The yield of cotton crops is severely affected by the soil’s chemical characteristics. This research probes the influence of soil properties on cotton yield, with a focus on assessing within-field spatial variability for advanced site-specific management in cotton production. Geostatistical analysis and semivariograms were conducted to examine the spatial variability of soil chemical properties. While multivariate analysis was employed to identify key yield-determining factors. Principal Component Analysis (PCA) was employed to explore leading soil indicators, as opposed to using a pre-defined set of indicators, and to analyze the spatial variability of soil properties. The soil characteristics encompassed saturation, soil pH, electrical conductivity (EC), organic matter (OM), phosphorous (P), potassium (K), calcium carbonate (CC), and micronutrients such as zinc (Zn), copper (Cu), iron (Fe), manganese (Mn), and boron (B). For this study, one thousand soil samples were meticulously collected from the Multan district, facilitated by Global Positioning System (GPS) devices. The PCA revealed that the first four PCs accounted for 57% of the total variation. The first PC, Cu, B, Mn, OM, and Fe emerged as the most influential variables, while the second PC exhibited a high correlation with CC, pH, and saturation. The third PC highlighted the dominant role of K and P in its construction, whereas the fourth PC was primarily shaped by Zn. The Principal Component Regression (PCR) analysis further revealed that EC, K, P, and Zn had positive impacts on cotton yield, while pH, Cu, Fe, Mn, and B had negative impacts.