Spatial analysis of cotton (Gossypium hirsutum L.) canopy responses to irrigation in a moderately humid area

Accurate irrigation scheduling is important to ensure maximum yield and optimal water use in irrigated cotton. This study hypothesizes that cotton water stress in relatively humid areas can be detected from crop stress indices derived from canopy reflectance or temperature. Field experiments were conducted in the 2003 and 2004 crop seasons with three irrigation treatments and multiple cultivars to study cotton response to water stress. The experiment plots were monitored for soil water potential (SWP), canopy reflectance and canopy temperature. Four crop stress indices namely normalized difference vegetative index (NDVI), green NDVI (GNDVI), stress time (ST) index and crop water stress index (CWSI) were evaluated for their ability to indicate water stress. These indices were analyzed with classic mixed regression models and spatial regression models for split-plot design. Rainfall was plentiful in both seasons, providing conditions representative of irrigated agriculture in relatively wet regions. Under such wet weather conditions, excessive irrigation decreased lint yield, indicating the necessity for accurate irrigation scheduling. The four crop stress indices showed significant responses to irrigation treatments and strong correlation to SWP at shallow (0.2 m) depth. Spatial regression models were able to accurately explain the effect of irrigation treatment, while classic split-plot ANOVA models were confounded by collinearity in data across space and time. The results also verified that extreme humidity can mask canopy temperature differences with respect to ambient temperature, adding errors to canopy temperature-based stress indicators.