Drought-related tree mortality occurs worldwide, including recent episodes in piñon-juniper woodlands of the American west. Although the physiological mechanisms of mortality are poorly understood, carbon starvation may occur in trees that limit transpiration (E) to avoid hydraulic failure. Results from a hydraulic model suggest that isohydric species, those that tightly regulate gas exchange to maintain constant leaf water potential (Ψ_l) as soil water potential (Ψ_s) fluctuates through time, may be more susceptible to carbon limitation. Anisohydric species, those that allow Ψ_l to decline along with Ψ_s during drought, are more susceptible to catastrophic hydraulic failure. 

We examined these contrasting mechanisms of mortality in isohydric piñon pine (Pinus edulis) and anisohydric oneseed juniper (Juniperus monosperma) at an ecosystem scale rainfall manipulation experiment in central New Mexico which was implemented in late summer 2007. We measured Ψ_s and Ψ_l and used sap flow estimates of E in control and droughted plots to parameterize a plant and soil hydraulic model. The model was used to estimate the safety margin between E and E_crit, the value of E leading to transport failure. We predicted that water use during drought would reflect predisposition to one of these two mortality mechanisms such that:

1) piñon avoids hydraulic failure by restricting Ψ_l  and maintaining a wider margin between E and E_crit, and

2) juniper Ψ_l  declines with Ψ_s, maintaining gas exchange further into drought but increasing the likelihood of hydraulic failure as E approaches E_crit. 

Results/Conclusions

Piñon restricted Ψ_l to -2.5 MPa in 2008 but not in 2007. Juniper did not regulate Ψ_l to the extent that piñon did; in 2008 juniper Ψ_l was ~2 MPa lower in the drought treatment than in the control. Under wet conditions, piñon maintained a larger margin between E and E_crit than juniper. As soils dried, E_crit decreased more rapidly in piñon than in juniper. Both species experienced E ≈ 0 at the same time so our prediction about juniper maintaining gas exchange longer into drought does not seem to hold. 

The observed patterns of E and Ψ_l suggest that while juniper may approach its hydraulic limits by allowing Ψ_l to decline, both species may experience carbon limitation under the imposed drought treatment because both spent approximately 6 consecutive months with E close to zero. Droughted piñon suffered approximately 35% mortality after less than one year of treatment, supporting the paradigm of carbon limitation in isohydric species during drought.