A site-specific and dynamic modeling system for zoning and optimizing variable rate irrigation in cotton

Cotton irrigation has been rapidly expanding in west Tennessee during the past decade. Variable rate irrigation is expected to enhance water use efficiency and crop yield in this region due to the significant field-scale soil spatial heterogeneity. A detailed understanding of the soil available water content within the effective root zone is needed to optimally schedule irrigation. In addition, site-specific crop-yield mathematical relationships should be established to identify optimum irrigation management. This study aimed to design and evaluate a site-specific modeling system for zoning and optimizing variable rate irrigation in cotton. The specific objectives of this study were to investigate (i) the spatial variability of soil attributes at the field-scale, (ii) site-specific cotton lint yieldwater relationships across all soil types, and (iii) multiple zoning strategies for variable rate irrigation scenarios. The field (73 ha) was sampled and apparent soil electrical conductivity (ECa) was measured. Landsat 8 satellite data was acquired, processed, and transformed to compare indicators of vegetation and soil response to cotton lint yields, variable irrigation rates, and the spatial variability of soil attributes. Multiple modeling scenarios were developed and examined. Although experiments were performed during two wet years, supplemental irrigation enhanced cotton yield across all soil types in comparison with rain-fed conditions. However, length of cropping season and rainfall distribution remarkably affected cotton response to supplemental irrigation. Geostatistical analysis showed spatial variability in soil textural components and water content was significant and correlated to yield patterns. There was as high as four-fold difference between available water content between coarse-textured and fine-textured soils on the study site. A good agreement was observed (RMSE = 0.052 cm³ cm^-3 [cubic centimeter per cubic centimeter] and r = 0.88) between predicted and observed water contents. ECa and space images were useful proximal data to investigate soil spatial variability. The site-specific water production functions performed well at predicting cotton lint yield with RMSE equal to 0.131 Mg ha^-1 [megagram per hectare] and 0.194 Mg ha-1 in 2013 and 2014, respectively. The findings revealed that variable rate irrigation with pie shape zones could enhance cotton lint yield under supplemental irrigation in west Tennessee.