Ponderosa pine (PP; Pinus ponderosa
Lawson and C. Lawson) is a geographically diverse, ecologically significant, and economically important tree species in western North America. In the Northern Rockies, U.S.A., PP occurs from approximately 700–2300 m elevation, making it not only a widespread species but one that also has an extensive altitudinal gradient [1
]. Along the elevational gradient, significant climatic differences exist, with the warmest and driest conditions occurring at the lowest elevations and progressively cooler and wetter conditions with increasing elevation. Global circulation models for the Northern Rockies predict substantially warmer (3.1–4.5 °C) and slightly drier (~0.3 cm decrease per month) conditions in the next several decades [3
] and these future conditions create an impetus to examine the historical responses of this dominant coniferous tree to changing climate and atmospheric composition.
If climatic change in the Northern Rockies [5
] continues, regional PP forests likely will experience more frequent summer drought conditions. Prior work with PP in this region has demonstrated that the growth of PP is positively related to wetter and cooler conditions during the spring and summer months [10
]. Thus, warmer conditions and more frequent droughts should negatively impact overall forest productivity. But the question remains as to what degree and, more importantly, which individuals will be most impacted? Are PP growing at the lowest elevations the most prone to growth declines or mortality because they grow along the climatic margin for survivability?
In a broad synthesis of research examining relationships between changing climate and tree mortality, Allen et al. [14
] reported that the evidence linking recent increases in temperature and the frequency/severity of drought is associated with increasing tree senescence in multiple ecosystems worldwide. More specifically, McDowell et al. [15
] (p. 399) reported that PP growing at lower elevations are more likely to exhibit “chronic water stress” and thus are more likely to die during periods of drought, with long-term implications for PP ecosystem modifications if the vectors for increasing aridity are manifest. Ganey et al. [16
], hypothesized similar outcomes to McDowell et al. [15
] in that tree senescence would be greatest at the low elevation sites as they are more frequently stressed by a combination of high temperature and limited moisture availability. Ganey et al. [16
] examined mortality rates of multiple species within mixed PP forests that had experienced high rates of senescence in recent decades yet found that senescence was not closely linked with elevation. Tague et al. [17
] examined rates of PP mortality and productivity along elevational transects to determine if models incorporating carbon allocation and hydroclimatic parameters can be used to understand spatial patterns of drought-induced tree senescence along elevational transects. Their elevation-specific results are mixed. They noted that net primary productivity (NPP) at high sites is greater largely due to higher precipitation rates and greater moisture storage at high elevations. However, the relationship between NPP and precipitation was similarly strong at both high and low elevation sites. They concluded that spatiotemporal patterns of temperature and moisture are strongly controlled by elevation and equally important in modeling the probability of tree senescence. While Lloret et al. [18
] noted that climate change, which is manifest through warming and increasing drought frequency, is related to tree senescence globally. They found that rates of radial growth recovery for PP following drought are largely unchanged through time. They concluded that, long-term, climate change-induced senescence is not closely related to decreasing resilience but is rather controlled by tree reactions to specific events (e.g., an individual drought).
Previous work leaves the question: Could it be that many of these low-elevation PP populations already possess distinct characteristics that would confer climatically marginal populations an advantage over those PPs growing in the cooler and wetter conditions that exist with increasing elevation? Elevation has been identified as a key component regulating the growth of PP [19
] with physiologic plasticity suggested as a means by which individuals adapt to elevational heterogeneity [20
]. The degree to which the persistence of a species is influenced by local adaptation and/or phenotypic plasticity across populations could be critical to understanding the impact of climate change on overall populations [21
Here, we address these climate/growth-response questions and discrepancies in the literature based on a dendrochronological sampling of PP along vertical transects in western Montana. We hypothesize there are differences in the climate-growth response of PP across elevational classes. Specifically, we posit that trees growing at lower elevations will exhibit a greater sensitivity to drought and an overall decline in radial growth. To investigate this hypothesis we: (1) assess drivers of PP radial growth at three study sites in western Montana; (2) compare standardized ring width of PP to climate along study-site specific elevational transects (i.e., relative elevation); and (3) investigate PP growth in response to varying extreme climatic conditions along an integrated elevational transect (i.e., absolute elevation).
Minimal differences exist in macroscale climatic conditions among the three study sites (Figure 2
). As expected, the highest site (TGF) is the coldest, and the lowest site (FLF) is the warmest. Small differences in monthly precipitation totals are present with FCF receiving the greatest amount of annual precipitation (~63 cm), while TGF receives the least amount of annual precipitation (~51 cm).
Our investigation of the relationships between standardized growth and monthly measures of climate shows that moisture availability in mid-summer (i.e., July PDSI) is the principal driving force of PP growth. PP respond negatively to mid-summer temperatures and positively to precipitation (Table 2
; Figure 3
). For PDSI, the strongest relationships occur in the month of July. Since PDSI values in each month are partially dependent on moisture supply and demand in the preceding months [31
], there is a cumulative component to the climate response for PP.
The overall (i.e., all trees in the site level chronology) climate response is comparable across the three study sites (Table 2
). While small differences exist between sites when the analyses are divided into TOP and BOTTOM groupings of trees (Table 2
), the Fisher z-test results showed there were no significant differences (p
< 0.05) in the climate/growth relationships between TOP and BOTTOM trees for the strongest monthly relationship for each climate variable.
The ALL chronology (i.e., the chronology based on absolute elevation) reveals a significant and similar association with the same climatic variables discussed at the specific site level (Table 3
). The Fisher z-test resulted in no significant difference in climate response between ALL_TOP and ALL_BOTTOM groupings along the combined elevational transect. When correlating (Spearman) raw radial growth of trees along the combined transect to elevation during wet and dry (PDSI > 2; PDSI < −2) conditions, we found no significant association. However, for standardized radial growth values, we found that low elevation trees have significantly (p
< 0.001) higher radial growth than higher elevation trees during dry periods, and high elevation trees have significantly (p
< 0.003) greater radial growth than low elevation trees during wet periods (Figure 4
). Lastly, our Mann–Whitney U Test used to compare the difference in means of high and low groupings resulted in significant differences between both dry (p
= 0.000) and wet periods (p
While slight differences exist in site-specific climatic conditions (Figure 2
), the temperature patterns are logical (i.e., temperature decreased with elevational increase). We expected the highest annual precipitation at TGF due to the combination of higher elevation and orographic lifting processes, but the PRISM data suggest this location receives slightly less precipitation than FCF and FLF. This is likely a function of the location of TGF, east of the spine of the Bitterroot Mountain range, resulting in a rain shadow effect. (Figure 1
). Although there are small site differences in precipitation and temperature, the overall climate response of trees is congruent (Figure 3
; Table 2
and Table 3
). For change detection, a Mann–Kendall trend test showed that temperatures in Montana Climate Division 1 have a significant positive trend over the last 50-years (1862–2011; p
= 0.025), but no trends were evident over both longer (1895–2011; p
= 0.439) or shorter periods (1982–2011; p
= 0.080). We also found no trends at 30-, 50- or the 117-year periods in precipitation or PDSI in Montana Climate Division 1, so the increasing temperatures have not translated into increasing aridity as measured via a water balance-based metric (i.e., the PDSI). However, summer temperature in the region is projected to continue to increase, and precipitation is projected to decrease [3
]. If this occurs, PP in the region will experience increasing aridity during the growing season in future decades.
Raw radial growth patterns during dry and wet years in Montana Climate Division 1 have no directional relationship based on elevation (Figure 4
a). For example, there is no propensity for lower (higher) elevation trees to grow faster than higher (lower) elevation trees during dry (wet) periods. The only discernible pattern is that trees with the highest rates of radial growth are growing faster than trees with the lowest rates during both wet and dry periods. Although all of our sampled trees were mature, this could be a function of declining radial growth with tree age. However, when assessing standardized radial growth based on absolute elevation during wet and dry years, distinct trends exist within the data (Figure 4
b,c). The pattern reveals that, collectively, higher elevation trees outperform lower elevation trees during wet periods, and lower elevation trees outperform higher elevation trees during dry periods (Figure 4
Both prior work (e.g., [13
]) and our results (Figure 3
) demonstrate that PP is positively related to wetter and cooler conditions during the spring and summer months. Thus, warmer, drier conditions and more frequent droughts should negatively impact overall forest productivity, and PP growing at the lowest elevations may be the most prone to growth declines or mortality because they grow along the climatic margin for survivability [15
]. Conversely, these low-elevation populations may already possess distinct characteristics, such as exceptional water-use efficiency, that would confer climatically marginal populations an advantage over those PPs growing in the cooler and wetter conditions that exist with increasing elevation. Specifically, the ability to withstand xylem cavitation is often a characteristic of individuals growing in the more xeric portion of a species range [32
], and individuals may experience decreasing growth response to drought with increasing elevation [33
]. Elevation has been identified as a key component regulating the growth of PP, with physiologic plasticity suggested as a means by which individuals adapt to elevational heterogeneity [21
]. Our findings suggest this interpretation is correct, but only across wider ranges of elevation than are typically encountered by PP trees growing on site-specific elevational transects.
Previous research has found that increased water stress and aridity are projected to lead to more mortality and senescence of PP growing at low elevations in the western United States [14
]. The results of our study, however, suggest that this generalization may not be operational in the northern Rocky Mountains of western Montana. One plausible explanation for our differing results is the location of our study sites. Lascoux et al. [36
] documented a divergence of PP approximately 250,000 years ago into two varieties: eastern and western. These varieties moved northward over time and reconvened in what Latta and Mitton [37
] (p. 769) described as “west-central Montana”. Latta and Mitton [37
] (p. 769) further noted that at this transition zone, “gene flow between the two varieties will introduce genes to potentially different adaptive regimes. The varieties are interfertile…”. This finding suggests that PP growing in the transition zone may possess certain genetic adaptations that allow for greater resiliency in a changing climate, and this might explain why trees from our study sites reveal differing broader climate responses to changing climatic conditions.
Detecting and/or modeling PP health/mortality, specifically at lower elevations, in response to warmer and drier climate conditions in the western United States, has been successful [15
]. For example, Van Mantgem et al. [34
] modeled tree mortality rates across 76 forest plots in western North America. They found significant increases in mortality rates through time in all regions and species, across all diameter classes of trees, and at all elevations. For elevation, they found the greatest increases in mortality in the mid-elevation ranges (1000–2000 m). However, other investigations of PP health have been unable to either detect or confidently make linkages to large-scale climate change. For example, McCullough et al. [38
] investigated the climate response of 161 PP chronologies by grouping sites with similar climate response. They generally found that trees growing in the more western populations were overall less sensitive to climate than eastern populations. More importantly, McCullough et al. [38
] suggested that making generalizations about a species that occupies large elevation and spatial gradients comes at the risk of oversimplification. They concluded by saying they believe that most PP ecosystems will experience changing climate regimes, yet the responses of trees will be based on local conditions. Our findings, which show no within-site difference in climate-growth or growth-elevation responses, but do show elevational differences when trees across a wider range of elevations are grouped, suggest that local (i.e., site-specific) environmental differences are less important to PP radial growth rates than those experienced over a wider range.