Evaluation of phenotypic and photosynthetic indices to detect water stress in perennial grass species using hyperspectral, multispectral and chlorophyll fluorescence imaging

Knowledge of phenotypic and physiological traits associated with early responses to drought stress and the extent of stress damage is important for developing efficient irrigation programs and the selection of drought-tolerant cultivars. This study was conducted to identify major vegetation and photosynthetic indices from imaging technologies that are correlated to visual turf quality and leaf water status and responsive to drought stress by comparative analysis of different indices from multispectral, hyperspectral, and chlorophyll-fluorescence imaging for Kentucky bluegrass (Poa pratensis L.) exposed to drought stress. The progression of stress symptoms of plants was monitored using the three imaging technologies in controlled-environment chambers. Regression analysis demonstrated that the integrated vegetation indices from hyperspectral-imaging had better predictability for drought responses than those from multispectral or chlorophyll-fluorescence imaging. Among individual vegetation indices, SIPI and SRI from hyperspectral-imaging were more responsive to drought than other indices while PSRI and PRI from hyperspectral-and multispectral-imaging were highly correlated to leaf relative water content (RWC) or visual turf quality (TQ) under drought stress; NDVI or NDRE from hyperspectral and multispectral imaging were significantly correlated to TQ but were not as sensitive to drought stress as other indices. For chlorophyll-fluorescence photosynthetic indices, NPQ and Fv/Fm were significantly correlated to RWC or TQ while NPQ was most responsive to drought. Those vegetation or photosynthetic indices derived from the three imaging technologies that were responsive to drought stress and correlated to the extent of drought damages could be particularly useful traits for detecting and monitoring water stress in cool-season turfgrass.