Determination of soil depth and its water threshold for diagnosing water deficit of winter wheat based on grain yield

Abstract: Clarifying response of crop yield to water deficit is important for deficit irrigation. Field experiment of winter wheat with seven treatments, irrigated at different growth stages and supplied with varying irrigation schedules to maintain their variation in soil water, was conducted at a site in Hengshui, Hebei Province in North China Plain. The well-watered control was given four irrigations, and the most stressed treatment was rainfed. Each treatment had three replicates with plot size of 12 m × 8 m and was arranged in a completely randomized block design. Soil water was measured by TDR within 1.6 m soil depth at 0.2 m interval. Grain yield was measured on three 8-m2 area for each treatment. Meteorological data were collected through a weather station at the test site. By examining soil water dynamics, its average over growing season and its variability over soil profile, and by exploring the relations of grain yield of winter wheat with soil moisture of different depth at various growth stages, we determined suitable soil depth for detecting water deficit and its corresponding soil water threshold. We found that the largest difference in soil water between the deficit treated and well-watered control occurred at shallow soil layer, and the magnitude in difference decreased as soil depth increased, which was 19.7%-36.5%, 9.3%-21.7% and 2.9%-9.7% at the depth of 0-0.4, 0-0.8 and 0-1.2 m, respectively. Variability in soil water also reduced with increase of soil depth, shown by the average variation coefficient of all the treatments as 0.149, 0.129, 0.116, 0.108, 0.100 and 0.090 respectively for 0-0.4, 0-0.6, 0-0.8, 0-1.0, 0-1.2 and 0-1.6 m soil layer. Relationship between grain yield of winter wheat and soil water of various depths at different growth stages was described by downward-opening parabolic function, but the correlation significance changed with both growth stage and soil depth, being most significant during grain filling (coefficient of determination was 0.953-0.981, P< 0.01) and at the soil layer of 0-0.4 m. Soil water threshold that affected grain yield decreased with progression of growing season. From jointing-booting to milk maturing, they reduced from 95.0%, 91.1%, 83.1%, 86.9% and 94.1% to 75.7%, 73.7%, 72.3%, 74.3% and 90.0% of the field water capacity for the soil layer of 0-0.4, 0-0.6, 0-1.0, 0-1.2 and 0-1.6 m, respectively, indicating a diminished sensitivity of winter wheat growth to water supply. These results, the largest difference and variability in soil water at shallow layer, as well as the most close correlation between yield and shallow-layer soil water, all pointed to the fact that 0-0.4 m was the most suitable soil depth for detecting water deficit since large variability can give a stronger signal to be detected. The corresponding water threshold at jointing-booting, heading, flowering, beginning-, middle-, late-grain filling and milk maturing was respectively 95.0%, 98.4%, 79.9%, 73.7%, 88.6%, 79.6% and 75.7%. Since the detection depth of soil water deficit has never been studied as an important variable for deficit irrigation and that most previous studies concern threshold affecting plant physiological process, our results are of practical value for guiding deficit irrigation practice in wheat production and for ease monitoring of soil water in Northern China agricultural production areas.