Differences between rootstocks, ‘Cleopatra’ mandarin and ‘Carrizo’ citrange, in soil–plant water relations and the influence of these factors on vigor, crop yield, fruit quality and mineral nutrition were evaluated in field-grown Clemenules mandarin trees irrigated at 100% of potential seasonal evaporation (ET_c) (control treatment), or irrigated at 100% ET_c, except during Phases I and III of fruit growth and post-harvest when no irrigation was applied (deficit irrigation (DI) treatment), for 3 years. Differences between rootstocks in plant–soil water relations were the primary cause of differences among trees in vegetative development and fruit yield. After 3 years of DI treatment, trees on ‘Cleopatra’ showed more efficient soil water extraction than trees on ‘Carrizo’, and maintained a higher plant water status, a higher gas exchange rate during periods of water stress and achieved faster recovery in gas exchange following irrigation after water stress. The DI treatment reduced vegetative development more in trees on ‘Carrizo’ than in trees on ‘Cleopatra’. Cumulative fruit yield decreased more in DI trees on ‘Carrizo’ (40%) than on ‘Cleopatra’ (27%). The yield component most affected by DI in ‘Cleopatra’ was the number of fruit, whereas in ‘Carrizo’ it depended on the severity of water stress reached in each phase (severe water stress in Phase I affected mainly the number of fruit, whereas it affected fruit size the most in Phase III). In the third year of DI treatment, water-use efficiency decreased sharply in trees on ‘Carrizo’ (70%) compared to trees on ‘Cleopatra’ (30%). Thus, trees on ‘Cleopatra’ were able to tolerate moderate water stress, whereas trees on ‘Carrizo’ were more sensitive to changes in soil water content.

We investigated the effects of regulated deficit irrigation (RDI) during the pre-harvest period (kernel-filling stage) on water relations, leaf development and crop yield in mature almond (Prunus dulcis (Mill.) D.A. Webb cv. Cartagenera) trees during a 2-year field experiment. Trees were either irrigated at full-crop evapotranspiration (ET_c = 100%) (well-irrigated control treatment) or subjected to an RDI treatment that consisted of full irrigation for the full season, except from early June to early August (kernel-filling stage), when 20% ET_c was applied. The severity of water stress was characterized by measurements of soil water content, predawn leaf water potential (Ψ_pd) and relative water content (RWC). Stomatal conductance (g_s), net CO₂ assimilation rate (A), transpiration rate (E), leaf abscission, leaf expansion rate and crop yield were also measured. In both years, Ψ_pd and RWC of well-irrigated trees were maintained above –1.0 MPa and 92%, respectively, whereas the corresponding values for trees in the RDI treatment were –2.37 MPa and 82%. Long-term water stress led to a progressive decline in g_s, A and E, with significant reductions after 21 days in the RDI treatment. At the time of maximum stress (48 days after commencement of RDI), A, g_s and E were 64, 67 and 56% lower than control values, respectively. High correlations between A, E and g_s were observed. Plant water status recovered within 15 days after the resumption of irrigation and was associated with recovery of soil water content. A relatively rapid and complete recovery of A and g_s was also observed, although the recovery was slower than for Ψ_pd and RWC. Severe water stress during the kernel-filling stage resulted in premature defoliation (caused by increased leaf abscission) and a reduction in leaf growth rate, which decreased tree leaf area. Although kernel yield was correlated with leaf water potential, RDI caused a nonsignificant 7% reduction in kernel yield and had no effect on kernel size. The RDI treatment also improved water-use efficiency because about 30% less irrigation water was applied in the RDI treatment than in the control treatment. We conclude that high-cropping almonds can be successfully grown in semiarid regions in an RDI regime provided that Ψ_pd is maintained above a threshold value of –2 MPa.