Water deficits at anthesis reduce CO2 assimilation and yield of lychee (Litchi chinensis Sonn.) trees

Ten-year-old 'Tai So' lychee (Litchi chinensis Sonn.) trees growing on a sandy loam soil in subtropical South Africa (latitude 25 degrees S) were watered weekly (well-watered treatment) or droughted from late July until January (drought treatment). After 16 weeks, at which time the trees obtained most of their water from below 150 cm, average soil water content at 0 to 150 cm depth was 14.5 +/- 0.1% in the well-watered treatment and reached a minimum of 7.6% in the drought treatment. At Week 7, minimum leaf water potential (psi L) in the morning and early afternoon declined to -2.6 and -2.8 MPa, respectively, in droughted trees compared with -1.5 and -2.2 MPa, respectively, in well-watered trees. From Week 9, stomatal conductance and net CO2 assimilation rate ranged from 70 to 300 mmol m-2 s-1 and 3 to 13 micromole CO2 m-2 s-1, respectively, in well-watered trees. The corresponding values for droughted trees were 50 to 180 mmol m-2 s-1 and 2 to 6 micromole CO2 m-2 s-1 Five weeks after rewatering the droughted trees, gas exchange had not recovered to the rate in well-watered trees, although tree water status recovered within a week of rewatering. In the well-watered trees, water use (Et) was 26 +/- 1 mm week-1 with evaporation (Ep) of 20 to 70 mm week-1 indicating a crop factor (kc = Et/Ep) of 0.4 to 1.2. Before anthesis, tree water status did not affect extension growth of floral panicles or leafy shoots. In contrast, no vegetative shoots were initiated after fruit set in the droughted trees when psi L in the morning declined to -2.5 MPa. Water deficits reduced initial fruit set by 30% and final fruit set by 70% as a result of fruit splitting (41.2 +/- 4.0% versus 10.0 +/- 1.3%). Water deficits did not alter the sigmoidal pattern of fruit growth, but reduced yield from 51.4 +/- 5.5 kg tree-1 in well-watered trees to 7.4 +/- 3.3 kg tree-1 in droughted trees.