Predicting thresholds of drought-induced mortality in woody plant species

Climate change is expected to result in an increased frequency of extreme events; while in some areas, heavier precipitation is predicted, other areas will suffer from more intense heat waves and droughts (Allison et al. 2009). Disturbingly, there is already evidence that increasing temperatures and reduced rainfall associated with climate change are responsible for an acceleration of large-scale forest dieback over the last century (Breshears et al. 2005; Allen et al. 2010). In the absence of mass mortality events, severe droughts can cause forest ecosystems to flip from net sinks to huge sources of CO2, with the response of the Amazon rainforest to recent droughts being the clearest example of this (Phillips et al. 2009). Increasingly, severe droughts are therefore expected to have profound effects on biodiversity, primary productivity and ecosystem function in many forested regions. Recent analyses have demonstrated that the majority of woody plant species across all forest biomes operate close to their safety margins for hydraulic failure, rendering them vulnerable to future shifts in precipitation and temperature (Choat et al. 2012). The safety margin referred to is the difference between the level of water stress experienced by a species in the field and the level of water stress that is likely to induce hydraulic failure in that species. In order to make accurate predictions of how different forest ecosystems and keystone species will be affected by drought, we require a detailed understanding of the physiology of plant death. The development of quantitative indices for mortality thresholds in woody plants is therefore essential for predictive models of vegetation response to drought. Despite this seemingly straight forward question (when is a plant dead?), reliable mortality indices are conspicuously lacking in the plant physiologist's tool kit (Anderegg et al. 2012). Hydraulic failure is now widely recognized as one of the key mechanisms of drought-induced mortality in woody plants (Tyree and Sperry 1989; McDowell et al. 2008). As the soil dries, greater tension develops in the xylem water column leading to an increased probability of embolism formation. Gas emboli block xylem vessels, reducing the capacity of the plant to move water to the canopy. In intense or prolonged droughts, xylem embolism becomes extensive, causing the death of the whole plant. In this issue of Tree Physiology, Urli et al. (2013) examine thresholds for hydraulic failure in a range of angiosperm species with a view to linking drought tolerance, survival and embolism resistance in five angiosperm tree species. Their results show that a physiological ‘point of no return’ can be predicted for these species based on the half time of their recovery after rewatering. This approach defines a critical water potential beyond which plants were unable to recover, even after 1 year of rewatering (Brodribb and Cochard 2009). The authors then compared the minimum recoverable water potential with vulnerability to embolism curves for each species. This represents an important step because we already possess vulnerability curve data from hundreds of species across the full range of forest biomes (Maherali et al. 2004; Choat et al. 2012). Vulnerability curves describe the relationship between xylem water potential and loss of hydraulic conductivity due to embolism for a given species. Species are most commonly compared by the water potential at which 50% of hydraulic conductivity is lost (P50). In a previous work, Brodribb and Cochard (2009) demonstrated that the minimum recoverable water potential for four gymnosperm species was very close to species P50 values. In contrast, Urli et al. (2013) show that the minimum recoverable water potential of five angiosperm species was closer to the point at which 88% loss of conductivity occurred (P88). This finding provides insight into how basic differences in xylem structure between gymnosperms and angiosperms translate to the physiological behaviour of these groups (Figure 1a). Gymnosperm xylem consists almost entirely of unicellular tracheids, which perform both transport and support functions for the plant. In angiosperm xylem, support and transport functions are performed by more specialized cell types than the tracheid: thick-walled fibres provide support, while multicellular vessels can attain much greater lengths and diameters (Choat et al. 2008). The wide vessels of angiosperm xylem allow them to attain much higher hydraulic conductivity than gymnosperms. How­ever, angiosperms also engage in a much riskier hydraulic strategy than gymnosperms. While angiosperms converge on their P50 safety margins, gymnosperms maintain much greater safety margins to P50, especially species with greater resistance to embolism (Figure 1b; Choat et al. 2012). This more conservative hydraulic strategy is logical if gymnosperm species tend to die closer to P50, while angiosperms die closer to P88. The fact that the stomatal behaviour of angiosperms leads them to more regularly experience higher levels of embolism may also be indicative of higher capacity for rapid embolism repair, although this remains to be confirmed (Johnson et al. 2012). (a) Transverse sections of xylem from a gymnosperm (top) and an angiosperm (bottom). The xylem of gymnosperms consists of 95% tracheids with a small proportion of ray parenchyma. Angiosperm vessels can attain greater diameters than gymnosperms and are surrounded by thick-walled fibres that provide biomechanical support. (b) Relationship between embolism resistance (P50) and the minimum water potenial observed in the field (Φmin) for angiosperm and gymnosperm species. The dashed line is the 1 : 1 relationship between P50 and Φmin. The safety margin is the difference bewteen the Φmin and P50 for each group. Modified from Choat et al. (2012). Readers may note that if gymnosperms die close to P50, then hydraulic failure in the stem is clearly not directly responsible for the death of the plant. Roots and leaves are usually less resistant to embolism within a plant and it is therefore possible that hydraulic failure in the root system and needles of gymnosperms may be a more direct driver of mortality than failure in the stem. Further research is required to establish whether this is the case and whether rundown of carbohydrate reserves plays a great role in some species (Mitchell et al. 2013). … and of course, some caveats. First, while the relationships between minimum recoverable water potential and P88 were strong in Urli et al. (2013), this relationship is based on a small number of species. Full confidence in this approach, and the observed differences between angiosperms and gymnosperms, will require data from a greater number of species. This approach may prove less precise in drought deciduous species, where leaf abscission occurs well before the minimum recoverable water potential has been reached. Plants capable of resprouting may present complications for the approach of defining a minimum survivable water potential. Finally, it is important to recognize that there is no magic line across which all species will consistently die. The data of Urli et al. (2013) and earlier studies (Kursar et al. 2009) show, while angiosperm species die close to P88, that there can be substantial between-species variation around this point. However, these issues are certainly not insurmountable and the work of Urli et al. (2013) brings us closer to an index that will allow for prediction of drought mortality thresholds for woody plant species across a wide range of environments.