Experimental evidence for heat plume-induced cavitation and xylem deformation as a mechanism of rapid post-fire tree mortality

Abstract

Recent work suggests that hydraulic mechanisms, rather than cambium necrosis, may account for rapid post-fire tree mortality. We experimentally tested for xylem cavitation, as a result of exposure to high-vapour-deficit (D) heat plumes, and permanent xylem deformation, as a result of thermal softening of lignin, in two tree species differing in fire tolerance. We measured percentage loss of conductance (PLC) in distal branches that had been exposed to high-D heat plumes or immersed in hot water baths (high temperature, but not D). Results were compared with predictions from a parameterized hydraulic model. Physical damage to the xylem was examined microscopically. Both species suffered c. 80% PLC when exposed to a 100°C plume. However, at 70°C, the fire-sensitive Kiggelaria africana suffered lower PLC (49%) than the fire-resistant Eucalytpus cladocalyx (80%). Model simulations suggested that differences in PLC between species were a result of greater hydraulic segmentation in E. cladocalyx. Kiggelaria africana suffered considerable PLC (59%), as a result of heat-induced xylem deformation, in the water bath treatments, but E. cladocalyx did not. We suggest that a suite of 'pyrohydraulic' traits, including hydraulic segmentation and heat sensitivity of the xylem, may help to explain why some tree species experience rapid post-fire mortality after low-intensity fires and others do not.

Keywords: Eucalyptus cladocalyx; Kiggelaria africana; cavitation; fire; hydraulic failure; pyrohydraulics; tree mortality; xylem deformation.

Figure 1

Leaf specific canopy conductance (K _leaf, ± 1SE) measured for Eucalyptus cladocalyx shoots following the control (CONT), simulated heat plume (PLUME) and water bath (WB) treatments at 70 and 100°C. Following initial measurement, stems were flushed (F) and re‐measured to determine whether any loss of K _leaf was recoverable. Means were tested by ANOVA (Table 1). Significantly different means (post‐hoc Tukey's honestly significant difference (HSD) test) are indicated with unique letters.

Figure 2

Leaf specific canopy conductance (K _leaf, ± 1SE) measured for Kiggelaria africana shoots following the control (CONT), simulated heat plume (PLUME) and water bath (WB) treatments at 70 and 100°C. Following initial measurement, stems were flushed (F) and re‐measured to determine whether any loss of K _leaf was recoverable. Means were tested by ANOVA (Table 1). Significantly different means (post‐hoc Tukey's honestly significant difference (HSD) test) are indicated with unique letters.

Figure 3

Percentage loss of conductance (PLC), calculated as the leaf specific canopy conductance (K _leaf) of the unflushed treatment relative to K _leaf of the unflushed control (± 1SE), in Eucalyptus cladocalyx and Kiggelaria africana.

Figure 4

Light (× 40 and × 100 magnification) and scanning electron microscopy (SEM) of stem cross‐sections, showing xylem damage to Kiggelaria africana in 70 and 100°C water bath treatments, but not to Eucalyptus cladocalyx. Bars, 20 μm.

Figure 5

Temperature–time curves for Kiggelaria africana (solid lines) and Eucalyptus cladocalyx (dashed lines) for thin (4–5 mm, thin lines) and thick (8–9 mm, thick lines) stems in the 70 and 100°C simulated plume (black lines) and water bath (blue lines) treatments. Note the water bath treatments heat up far more rapidly than the plume treatments, and remain above 60°C for longer than the plume treatments. In the plume treatments, only the thin stems heated to above 60°C (the minimum temperature for the lignin glass transition) during the 6‐min experimental treatment.

Figure 6

Model simulations of the percentage loss of conductance (PLC) vs temperature compared with experimental data (solid symbols) from the simulated plume treatments in Eucalyptus cladocalyx and Kiggelaria africana. For a full description of the model, see the Materials and Methods section and model diagram (Supporting Information Fig. S1). For each panel, the solid black line is the initial model result with no surface tension decrease (Fig. S1, box iii), the dashed black line includes surface tension effects (Fig. S1, box v), the shaded area represents the range of possible solutions encompassing assumptions of constant canopy stomatal conductance (g _s) (Fig. S1, Loop 1) and rapid decline of g _s to cuticular conductance (g _c) (Fig. S1, Loop 2), and the dotted black line is the effect of decreased surface tension only (holding D constant). (a, b) Model results with no parameter adjustments. (c, d) Model results after reducing P ₅₀ to simulate PLC in petioles. (e, f) Model results after reducing initial g _s.