Continuous measurement of plant and soil water status for irrigation scheduling in plum

The usefulness of continuous measurement of soil and plant water status for automated irrigation scheduling was studied in a drip-irrigation experiment on plum (Prunus salicina Black Gold). Two levels of water restriction were imposed at different phenological periods (from pit-hardening to harvest, post-harvest) and compared with a well irrigated control treatment. Soil matrix water potential (psi(soil)) was measured with granular matrix sensors (Watermark); and short-period trunk diameter variation (TDV) was measured with linear variable displacement transformers. The Watermark sensor readings were in reasonable agreement with the irrigation regime and showed a good indication of plant water status across the season (r(2)=0.62), although they were a better predictor of stem water potential (psi(stem)) in the dry range of psi(soil). Nonetheless, the most important drawback in their use was the high variability of readings (typical CV of 35-50%). From TDV measurements, maximum daily shrinkage (MDS) and trunk growth rate (TGR) were calculated. Their performance was also compared with psi(stem), which had the lowest variability (CV of 7%). During most of the fruit growth period, when TGR was minimum, MDS was higher in the less-irrigated treatment than in the control and correlated well (r(2)=0.89) with psi(stem). However, after harvest, when TGR was higher, this correlation decreased as the season progressed (r(2)=0.73-0.52), as did the slope between MDS and psi(stem), suggesting tissue elasticity changes. Later in the season, TGR was better related to plant water status. These observations indicate some of the difficulties in obtaining reference values useful for irrigation scheduling based exclusively on plant water status measurements.