Results
Overall, eight of the 48 total trees died during the two-year experiment. Bark beetle attacks occurred over 10 weeks, starting early June and ending in middle of August. Attacks peaked in late June and declined through August. All eight of the dead trees were attacked by bark beetles, with six of them being in the trenched-attacked treatment and two in the untrenched-attacked treatment. Trees in the combined treatment died sooner, with all of the mortality to trenched-attacked trees occurring in 2014 and all of the mortality to untrenched-attacked trees occurring in 2015.
As reported by Kolb et al . (2019), predawn water potential over the summer of 2013 averaged -0.97 MPa for trenched trees and -0.83 MPa for untrenched trees. In the second year of the experiment (2014), predawn water potential was approximately 0.5 MPa more negative for trenched trees than untrenched trees between late May and early July during the typical spring drought period of the region (Fig. 1 in Kolbet al . 2019). Predawn water potential of trees in the trenched-attacked trees became much more negative than trees in all other treatments for the remainder of summer in 2014, reaching an average value of -2.5 MPa by middle of September compared with -0.5 MPa in other treatments. Differences in light saturated net photosynthetic rate among treatments in 2014 paralleled differences in predawn water potential, with values of trenched trees 14% lower than untrenched trees in May (mean of 6.1 and 5.3 umol m−2s−1, respectively) and 90% lower in late June (mean of 0.4 and 0.04 umolm−2 s−1, respectively). Trenching continued to cause water stress during dry periods in the third year of the experiment (2015), with values of trenched trees 0.2 MPa more negative than untrenched trees in late June.
In general, chemical profiles of ponderosa pine trees changed with treatment (Fig. 1, Fig. S1, Tables S1 & S2), year (PerManova: F(1.313)=4.45, P=0.009) and month (PerManova: F(8.313)=7.11, P<0.001). Therefore, we tested whether the NCSs differed among treatments for each month and each year. The results showed that with the exception of September 2015, chemical profiles varied among treatments in the remaining sampling times (Fig. 1, Figs. S1 & S2, Table S3). Note that terpenes of trees challenged with either bark beetle attacks or inoculated with crushed beetles represent “induced” concentrations, whereas terpenes of trees without any biological challenge treatments represent “constitutive” concentrations. Thus, the results should not be applied to define the question of whether terpene concentrations predict which trees live or die.
Do treatments influence NSCs of ponderosa pine trees?Overall, results show the importance of cumulative stresses in tree NSCs reduction in the xylem and mortality during drought as the effects of trenching on NSCs were negligible in trees without any additional biological treatment especially in the trenched-attacked treatment (Fig. 1, Table S2, Fig. S1). For instance, bark beetle attacks with trenching significantly lowered the total NSCs of trees relative to the trenching without such attacks (Fig. 1).
Effects of trenching on NSCs. Overall, chemical profiles did not differ between trenched-control and untrenched-control trees (Fig 2a, Table S4). We further tested if the NSCs were affected by trenching in each month. The difference was only significant in October 2014 (Fig. 3, Table S5) when starch, total sugars and total NSCs were higher in the trenched-control than the untrenched control.
Effects of bark beetle attacks on NSCs. For sugars, starch and total NSCs, treatments and month were significant, but treatment-month interaction was only significant for starch (Fig. 2b, Table S4). We then tested if NSCs were a function of beetle attacks in each month and found that bark beetle attacks significantly lowered sugars, starch, and total NSCs in July, August, September, October, November in 2014 and in May, June, August in 2015 (Fig. 3, Table S5).
Effects of crushed beetle inoculations on NSCs . Although treatment and month were significant for starch, total sugars, and total NSCs, the interaction was not (Fig. 2c, Table S4). We then tested whether NSCs were affected by inoculation in each month and found that inoculation significantly reduced total sugars and total NSCs in July, August, September, October in 2014 and in May, June, July, and August in 2015 and starch in July, August, and September in 2014 and in May, June, July, and August in 2015 (Fig. 3, Table S5).
Trench-biotic stress interactions. We did not find significant interaction between trenching and biotic (beetle attacks or inoculations) treatments for NSCs and diterpenes (p>0.05). However, trenching did impact monoterpene concentrations in response to inoculation (F(1,,21)=4.69, P=0.042). Trenched-inoculated trees had significantly greater monoterpene concentration (8.85±1.89 µg mg-1) than trenched-control (2.80±0.60 µg mg-1) or untrenched-control (3.82±0.82 µg mg-1) trees. Means of trenched-inoculated and untrenched-inoculated (5.42±1.16 µg mg-1) were statistically similar. Please note that monoterpenes
NSCs and defense metabolite interactions. Relationship between NSCs and diterpenes. There was a significant negative relationship between diterpenes and NSCs (Total NSCs: F (1,267)=36.22, P< 0.001, Fig. 4a; Total Sugars: F(1,270) =34.41, P<0.001, Fig. 4b; Starch: F (1,271) =25.49, P<0.001, Fig. 4c). However, treatments did not affect these relationships (Fig. 4a).
Relationship between NSCs and monoterpenes . We did not find any relationship between monoterpenes and NSCs (Total NSCs: F(1,291)=0.082, P=0.774, Fig. 4d; Total Sugars: F(1,293) =0.080, P=0.783, Fig. 4e; Starch: F(1,265)=0.001, P=0.969, Fig. 4f). However, the relationship between starch and monoterpenes varied by treatments (Fig. 4f). Trees in the untrenched-attacked treatment had a significant negative relationship between monoterpenes and starch; in contrast, we found a significant positive relationship in the untrenched-inoculated and trenched-inoculated trees (t=-4.54, P<0.001, and t=-3.75, P=0.003, respectively). Moreover, although there was a marginal effect of treatment on the relationship between total NSCs and monoterpenes, we did not find a significant interaction for total sugars (significance of interactions are provided in Fig. 4 caption).
Concentrations of NSCs and terpenes between dying and live trees (2014 data only). Treatment, month, and their interaction significantly affected sugars, starch, or total NSCs (Suppl. Table S6). We then compared the dying and live trees in the same trenched-attacked category for each month and found that live trees had significantly higher total sugars, starch, and total NSCs than dying trees in August, September, October, and November in 2014 (Fig. 5).
We investigated similar relationships for diterpenes and monoterpenes of dying and live trees in the same treatment category. Treatment, time and their interaction were not significant effects for either terpenes (Table S6), showing that the amount of diterpenes or monoterpenes produced in the dying versus live trees were statistically similar (Fig. 5).
Comparisons of carbohydrates between the six dying trees in the trenched-attacked treatment and live trees in the untrenched-control treatment showed a significant effect of month (Table S7). In a pairwise comparison between treatments for each month, we found that live trees had significantly higher starch, total sugars, and total NSCs than dying trees starting from July until November in 2014 (Fig. 6). We only used the data in 2014 because all dying trees in the attacked-trenched category died in 2014.
We investigated relationships between the six dying trees in the trenched-attacked treatment and live trees in the untrenched-control treatment for diterpenes and monoterpenes. Although the two treatments did not vary significantly, the monthly comparisons showed that constitutive tissue of live trees had significantly lower diterpenes in August, September, October, and November than tissues of attacked trees (Fig. 6, Table S7). Likewise, for monoterpenes, the overall interaction was not significant but the monthly comparisons showed that live trees had significantly lower monoterpenes in July, August, September, and October than dying trees (Fig. 6).
Explanation of tree mortality by NSCs-terpene interactions. We investigated the relationship between NSCs and terpenes from 2014 for dying trees in the trenched-attacked treatment and all treatments of live trees averaged over months. We found significant negative relationships of NSCs with either with diterpenes or monoterpenes (Fig. 7). For live trees, diterpenes or monoterpenes had a significant negative relationship with starch, total sugars, or total NSCs. For dying trees, only monoterpenes had a significant relationship with starch (P < 0.05). Other relationships between carbohydrates and terpenes were not significant. The low sample size for dying trees may have impacted the lack of significance in correlations compared to live trees.