Crop water stress index computation approaches and their sensitivity to soil water dynamics

There is a growing interest of using canopy temperature (Tc) based methods, including crop water stress index (CWSI), for irrigation management. However, different approaches exist to normalize Tc to microclimatic conditions, which can influence the accuracy and suitability of CWSI for irrigation scheduling. This study evaluated the performance of CWSI computation approaches and their sensitivity to changes in soil water depletion under different water stress levels. There were six different approaches – two empirical methods using developed lower baseline (i.e., CWSI-EB1, CWSI-EB2), two empirical methods using either artificial (CWSI-EA) or actual/natural (CWSI-EN) canopy reference surfaces, and two theoretical approaches which differ by how aerodynamic and canopy resistances are determined (CWSI-Th1, CWSI-Th2). Stationary infrared thermometers (IRTs) provided continuous Tc to calculate CWSI-EB, CWSI-Th, and CWSI-EN; whereas mobile IRTs and a thermal camera provided one-point-in-time Tc and temperatures of artificial canopy reference surfaces to calculate CWSI-EA. These measurements were all collected from full and deficit irrigated and rainfed maize plots in West Central Nebraska. Day-to-day variations within and across CWSI approaches were evident and their sensitivity to soil water depletion varied. Greater sensitivity and correlation strength to depletion (Dr,i) were observed with CWSI-Th and CWSI-EB under severe stress (i.e., Dr,i > 80%) at deeper soil depths of 1.8 and 2.1 m, producing r2 which ranged from 0.61 to 0.80 (slope: 0.03–0.05) and 0.69–0.79 (slope: 0.03–0.04), respectively. Observed differences in stress magnitudes among approaches and treatments, warrants a specific irrigation triggering threshold for each approach. Additionally, developing a robust index coupling both CWSI and soil water depletion is desirable to improve irrigation water management by accounting for both soil and plant water status.