Soil Respiration after Bark Beetle Infestation along a Vertical Transect in Mountain Spruce Forest

It is estimated that global soil carbon stock represents 1700 Gt C [1] and the contribution of forest soil ecosystems, as stated by [2], equals about half of it (861 Gt C). Top-meter soil, live biomass, deadwood, and litter each contain a percentage of the total forest C stock: 44%, 42%, 8%, and 5%, respectively. It creates an enormous carbon pool with the potential to highly increase atmospheric CO₂ concentration [3] after large-scale disturbances. Carbon emission from global soil is considered the second largest carbon flux after photosynthesis, releasing 78–98 Gt C yr⁻¹ in total [4,5]. Global forest soil carbon flux is an important contributor to global soil respiration (SR) because forests cover 26% of the Earth‘s total land area [6]. Under non-disturbance conditions, C uptake in the forest exceeds C emissions from ecosystem respiration [7] unless forest disturbance changes the balance and the forest becomes a carbon source [8,9]. Then, SR as the main component of ecosystem respiration [10] changes its rate as well.

Under climate change primarily induced by rising CO₂ atmospheric concentrations [11], forests in Europe have experienced extreme heat and drought [12]. As a result, bark beetles as poikilothermic organisms have altered their behavior. They increase population size, and have extra generations per year [13]. Also, shifts of bark beetles to higher latitudes [14] and altitudes [15] have been observed. As a result, large-scale forest diebacks throughout the northern hemisphere have been caused by the European spruce bark beetle (Ips typographus L.) [16,17,18] recently.

Initially, after host tree death, the biogeochemical and biogeophysical processes (leaf area index, evapotranspiration, productivity, land surface temperature) of forest stands are altered [19,20]. Gross primary productivity (GPP) decreases as a result of tree mortality [21,22] and a reduction in the leaf area index corresponds to a contraction of GPP [19]. Fast nitrogen and carbon reduction take place in upper mineral soils, but an accumulation of soil inorganic N occurs as a result of the diminished contribution of roots, mycorrhizae, and rhizodeposition [23] which can be used by successional plant uptake [24], in the form of ammonium or nitrate [25]. The decline of fine root density increases with the level of tree mortality [26] and root biomass decomposition is accompanied by mycorrhizal fungi decomposition [27]. All of these factors contribute to changes in SR rate [28]. The decrease in SR after a disturbance event is mainly due to the reduced input of autotrophic SR [29] and nutrient losses [30]. After some time, infested trees shed their needles [20] and according to [31] approximately after 100 days spruce trees start to defoliate. Subsequently, increased solar radiation input [32,33,34] accelerates the decomposition rate of the litter as a consequence of higher temperatures [35].

Temperature is considered the most important factor influencing global SR [5,36,37]. SR positively correlates with ambient temperature [38]. With changes in temperature within different altitudinal zones, SR declines with increasing elevation [39,40]. Similarly, higher solar radiation intake on infested sites leads to rising soil and air temperatures on plots with bark beetle-infested trees [33]. SR follows seasonal dynamics of soil temperature with water surplus throughout the year [41], as it depends on both factors, with soil temperature having a more dominant effect [42,43].

Plant physiological–phenological relationships also influence SR rates, particularly the autotrophic component [44] and the successional stage after forest disturbance [45], as SR is linked to changes in phenology and photosynthesis [46]. After severe disturbance events, areas may shift to non-tree-dominated environments, with changes in plant species composition from shade-tolerant to shade-intolerant species due to altered solar conditions [32]. These plant communities, primarily grasses and perennial weeds, show increased biomass production and dispersive ability [47,48]. Over time after bark beetle attacks, grasses and perennials increase their surface coverage, leading to a greater influence on SR [49]. In the study site, four years after a windstorm, almost 100% of the area was covered by grasses and perennials [50]. The biomass production of successional grasses can exceed that of the original forest stand by a significant margin [51]. SR is also affected by Net Primary Productivity (NPP) [52,53,54]. Given that our study is on a vertical gradient on a homogenous slope, differences in physiological–phenological factors should be observed between infested and uninfested sites. Therefore, changes in SR are influenced not only by the soil temperature relationship but also by other factors [44], such as plant physiological–phenological dynamics and NPP.

There are studies that focus on SR within a vertical gradient [43,55,56] and studies that examine SR after windthrow or bark beetle attacks [22,42,49,57]. However, these studies do not address how different elevation zones of mountain spruce forests respond to disturbance events in terms of SR. Thus, in this study, we measured soil CO₂ efflux for two consecutive years and investigated if there is a significant difference in SR between infested sites and uninfested sites with living trees at different elevational zones during the vegetation period. Our first hypothesis is that SR under dead trees would be lower than under uninfested trees due to a decrease in autotrophic SR (tree roots). Secondly, we predict that SR will not significantly vary across different altitudes due to minimal microclimatic differences within the measured mountain slope. Additionally, we expect the highest SR rates during summer (June to August) due to increased heterotrophic and autotrophic SR stimulated by higher soil temperatures.

SR varied throughout the vegetation period at different altitudes and months in 2016 and 2017 (Figure 1). The mean SR during the vegetation period was highest at the elevation of 1200 m.a.s.l. at both infested and uninfested sites (Figure 1 and Figure 2). We observed that SR reached its peak in the summer (July or August) at each elevation zone. Almost every month and at each elevation zone, infested sites emitted more CO₂ than uninfested ones, but in many cases it was insignificantly higher. The mean SR at different elevations between infested and uninfested plots was found to be insignificant at the elevation zone of 1400 m and 1300 m in 2016 and 2017, respectively (Figure 1 and Figure 2). A faster increase in SR at infested sites towards the peak of the growing season (July–August) was observed. We also noticed a slower decrease in SR towards the end of the growing season at infested sites compared to undisturbed sites in both years. Rates of SR differed between forest statuses within a year (Figure 1). In both years, infested sites showed significantly higher yearly average values than uninfested ones. The mean annual SR (during the vegetation period) in 2016 reached 0.625 ± 0.335 g CO₂ m² h⁻¹ and 0.428 ± 0.248 g CO₂ m² h⁻¹ at infested and uninfested sites, respectively. In 2017, SR rates were 0.576 ± 0.275 CO₂ m² h⁻¹ and 0.438 ± 0.249 g CO₂ m² h⁻¹ at infested and undisturbed sites, respectively. We observed no statistical difference in SR between the two subsequent vegetation periods. The critical p-value was set at 0.05 (α = 0.05).

The monthly average fluxes from the soil surface under infested sites are higher in both years, although not always significantly (Figure 1). The average SR at undisturbed sites showed the same amount of carbon emitted throughout the vegetation period at each elevation zone, at a significant level (Figure 2). On the other hand, a higher variation of SR was observed under infested plots (Figure 2). Our results show a clear pattern of decrease or increase in soil efflux within the elevation gradient in 2017 at infested sites. However, no decline in SR with elevation was observed at undisturbed sites in both years, at a significance level of p Figure 2). In 2016 at infested sites, we observed a significant decrease in SR with elevation gain.

SR in both years and types of forests did not show a high dependence on soil temperature and there were significant variations in 2016 between infected and undisturbed spruce forests (Figure 3). The disturbed forest exhibited a higher correlation between SR and temperature compared to the undisturbed forest. Additionally, the year 2016 shows a more rapid change in SR with changes in soil temperature (Figure 3). In 2017, a minimal difference in the dependence between SR and temperature was observed between disturbed and undisturbed forests. The correlation slope showed almost identical values. We also noticed a decrease in soil temperature between the highest and the lowest elevation zone in both years and forest conditions. The highest soil temperatures were measured during the summer months (from June to August) in both years and no correlation between soil moisture and respiration was observed. As optimal values of soil moisture were reached, it did not significantly impact SR rates. Additionally, we found a higher variance in moisture content at infested sites. The difference in moisture content between infected and uninfected sites was statistically significant. Disturbed sites covered with grasses and perennials had higher soil moisture content in both years (Table 1).

We observed significantly higher mean annual soil moisture and temperature at infested sites than at uninfested sites. The increase in soil temperature at infested sites within two years of the experiment was also observed. Similarly, significantly higher values for soil moisture were observed at infested sites. Some differences in soil moisture between the years were measured, but they were at the limit of significance (Table 1).

In our experiment, we found that SR was significantly higher in the infested forest throughout both years. Changes in SR rate after disturbance events are not consistent throughout the published research papers [22,30,42,57,65,66,67]. In the girdling experiment in boreal Scots pine (Pinus sylvestris L.) forest, SR decreased by approximately 50% relative to ungirdled sites within one to two months [65], where forest mycorrhizae alone contribute to one-third of dissolved organic matter in forest soils, together with associated roots, contributing to 50% of dissolved organic matter [27]. However, a recent meta-analysis by [68] concluded that microbial, root, and mycorrhizal respiration contribute 57%, 28%, and 15%, respectively, to total SR. Soil fluxes decline as a consequence of altering key factors and nutrients [30] controlling SR rate.

Ref. [69] mentioned that up to 3 years after lodgepole pine (Pinus contorta Dougl. ex Loud.) infestation by mountain pine beetle, most needles remained on the trees. Therefore, no additional needle litter is added from dead pines to increase SR rate during this period of infestation. After a pulse of dead needles, SR almost fully recovered, lasting for up to 2 years and then followed by a decline again [22]. We suggest that needlefall and debris input from dying and dead spruce trees increased heterotrophic respiration, which compensates for the loss of autotrophic respiration. Since our study was conducted 5 to 6 years after the initial infestation by bark beetles, the results are quite similar to those of [22,49] with the difference that SR not only equaled undisturbed plots but exceeded them. This contradicts our initial hypothesis that SR will be higher at uninfested sites. Nonetheless, in our study, infested sites were covered by grasses and perennials with significant biomass productivity [51,70], which can increase autotrophic SR rates originating from the rhizosphere [71]. This phenomenon can also increase SR rates, in addition to hypothesized increased heterotrophic respiration due to higher mean soil temperatures [72]. The observed faster growth of SR towards the peak of the growing season (July and August) followed by a slower decrease in SR rates towards the end of the growing season at disturbed sites can be attributed to the gradual growth of perennials and grasses. This can be explained by the physiological–phenological development of a variety of plant species occupying research plots. Firstly, spring species start to grow (April–May), followed by summer species (May–June) and autumn species (June–August), creating a gradual and steady supply of autotrophic SR. Therefore, after disturbance, subsequent root biomass production of understory vegetation (grasses and perennials) [51,57,70] can account for increased SR within disturbed stands. Seven years after the disturbance, in terms of net primary productivity, successional understory vegetation acquires values that are only three times lower than forest stands before the disturbance event [73]. Nevertheless, developed understory vegetation combined with increased heterotrophic respiration caused by higher nutrient content [74,75] in the soil can compensate for severe tree death following a bark beetle attack.

Ref. [57] did not observe any significant changes in SR at a stand level from July to September over a period of 5 years between live lodgepole pine sites and bark beetle-infested sites. This phenomenon is attributed to surviving trees, understory vegetation, and the nutrient pulse from needlefall as noted by [74]. The decline in autotrophic respiration is offset by higher heterotrophic respiration induced by increased soil temperature, as discussed by [35,76]. Nonetheless, it is suggested that if mortality reaches 100%, total SR decreases to one-third of that in uninfested sites as stated by [57]. A similar pattern to that observed by [57] has been seen in fir-spruce forests by [42], attributed to root-respiring carbohydrates after tree death or a decrease in autotrophic respiration being replaced by heterotrophic respiration from dead roots and foliage. Dying roots and mycorrhizae release stored carbon for 2–3 years after disturbance, as reported by [77]. Additionally, an increase in soil temperature and soil moisture has been observed in infested ponderosa pine (Pinus ponderosa Laws.) forests by [33]. Ref. [49] confirmed that up to six years after disturbance, SR does not decrease at windthrow-disturbed sites. This is due to the substitution of decreased autotrophic SR with increasing heterotrophic SR supported by disturbance-induced alteration of soil temperature. In contrast, Ref. [30] concluded that the rapid decline of dissolved organic carbon, organic nitrogen, and phosphorus is accompanied by a decrease in SR after trees dieback, but after 4 years nutrients begin to recover due to litter mineralization. The post-disturbance chronosequence is an important factor influencing SR rate after bark beetle infestation. These varying results can be attributed to different mortality rates, gap size formation, and pre-existing understory vegetation.

As the temperature is considered the most important factor influencing global SR rates [5,36,37], with studies showing a positive correlation between SR and ambient temperature [71], we hypothesize that peak SR rates occur during the peak growing season in July and August, when temperatures are highest. This is likely due to the close relationship between SR and both air and soil temperature [78], as well as the increased contribution of heterotrophic respiration [79]. As temperatures decrease across different altitudinal zones [39,40], consequently, SR declines with increasing elevation [43,80]. Our results demonstrate a clear pattern of either decrease or increase in soil efflux along the elevation gradient in 2017 at infested sites, consistent with [56]. We observed a significant decrease in SR with elevation gain at infested sites in 2016, but no significant decline of SR with elevation had been observed at undisturbed sites in both years. This partially rejects our hypothesis that there will be no significant change in SR rates with increasing elevation.

The study of [75] conducted in the same vertical gradient shows that soil nutrient content increases with elevation and is statistically higher at infested sites. However, this finding contradicts the decreasing pattern of SR observed at infested sites in our study, as nutrient content plays a crucial role in supporting soil microorganisms that drive heterotrophic SR [30,81,82,83]. On the other hand, Ref. [75] only analyzed nutrient content for the year 2016. Nevertheless, we can suppose that increased nutrient availability of macroelements, increases autotrophic SR more than the heterotrophic component [84], but it can increase SR significantly because SR positively correlates with macroelement addition and increased temperature [81]. Also, like our study, Ref. [57] found no correlation between SR and moisture. We suppose that it can be supported by optimal moisture conditions during the studied period.