**4.2.2 High diversity regional and local scale discussion**

The response to the temperature and *L .decidua* treatments in the high diversity regions highlights the importance of analysis of local climate and successional dynamics. The climate and *L. decidua* treatments do not have the strong effect on biomass in these high diversity regions of the RFE that is seen at the continental scale and in low diversity regions. Sites in the northern RFE (N RFE) region fall within larch's optimal growth ranges for both temperature and precipitation. Larch establishes strongly and competitively in N RFE in early succession followed by a transition to *Picea* spp. dominance under base climate conditions. The high diversity in the area indicates there is strong competition, and *L. decidua* does not have unique characteristics which allow it to establish in this region.

Under warmer conditions in the N RFE in late succession there is a transition, not to *Picea*  spp., but to *Pinus* spp. dominance. This highlights the importance of genus-level analysis. These genera are typically combined into a single evergreen conifer group, both for the purposes of global climate models and for the regional scale analysis presented here. Such a grouping prevents the detection of this shift between the two evergreen conifer species. This shift has the potential for further exploration in association with altered albedo values. Measured in the boreal forest of Canada, the summer albedo difference between *Pinus* spp. (0.086) and *Picea* spp. (0.081) is negligible, but the winter albedo for *Pinus* spp. (0.150) and *Picea* spp. (0.108) is more significant (Betts and Ball 1997). The albedo values in the RFE are likely similar to these Canadian species, and there is potential, with the winter albedo difference, to alter total annual absorption of surface radiation in response to the shift from *Picea* spp. to *Pinus* spp. with warming. Unlike the low diversity regions, the successional dynamics resulting from including *L. decidua* as a species in simulation are similar to those observed in response to warming, with late successional dominance by *Pinus* spp. This result suggests that introduction of *L. decidua* would not be a useful strategy for mitigation of vegetation shifts in this region.

The response to the climate and *L. decidua* treatments in southwestern (SW) RFE is also a product of local climate conditions. The SW RFE region has a climate with temperatures which place it in the upper limit of tolerance for larch, creating a climate in which the native larch species cannot compete and establish. Under base climate conditions, it is difficult for larch to compete, so it is not surprising that larch continues to be absent under warmer conditions. Even with increased tolerance for warmer conditions, *L. decidua* cannot effectively compete with other species in the SW RFE, and does not have a significant impact on biomass. Under base climate conditions, the SW RFE has mixed evergreen and deciduous species in the mature forest late in succession. With warming temperatures, there is an increase in biomass of mixed deciduous trees, and a decrease in evergreen conifer biomass. The decrease of evergreen conifers is balanced by the increase of mixed deciduous trees, so the effect on total biomass is not consistent and lasts only 60 years.

The high diversity areas in the RFE have high ecological resilience and stability, which allow them to maintain basic ecosystem function following climate change and avoid irreversible shifts to another vegetation state. Stability is the ability of a system to return to equilibrium following temporary disturbance, and resilience is the persistence of a system and its ability to absorb change and disturbance without changing state (Holling 1973). In other terms, stability is a persistence of the system state and a consequence of interactions within the

Resilience and Stability Associated with Conversion of Boreal Forest 209

The response of the low diversity regions to the altered climate treatment is similar to that observed at the continental scale. These low diversity regional scale responses, in conjunction with the low average diversity across the 372 sites considered at the continental scale, further emphasize the difference of the response in the high and low diversity areas. Similar to analysis at the continental scale, the response of the low diversity regions to treatments is connected to the transition from larch to evergreen dominance, which occurs across the southern and central portion of Siberia, and is accelerated by warming. Northwestern Siberia has the coldest temperatures compared to the other regions, and these cold temperatures naturally suppress evergreen conifer establishment during late succession under base climate. The patterns of successional dynamics in central Siberia under base conditions and in northwestern Siberia in response to warming climate suggest that the model is predicting consistent transitions. Under warming conditions, northwestern Siberia is exposed to temperatures more similar to those in the base condition in central Siberia. Thus, the forests in northwestern Siberia have similar dynamics under warming conditions to those seen in the central Siberia region for the base climate (Figure 7a,e). The temperature treatment results suggest that, with 4C of warming, the larch-dominated system across southern and central Siberia will be prematurely replaced with evergreen conifer and other deciduous trees, and the forests of northwestern Siberia will warm

The lack of response in northwestern and southern Siberia, and the late response in central Siberia, to the effect of precipitation change suggests a connection between evergreen conifer presence and seasonal precipitation. The precipitation treatment is significant in only one of the low diversity regions analyzed (i.e., central Siberia). Within the central Siberia region there is a short-lived response of total forest biomass to precipitation change at year 240, and a longer period of response of evergreen conifer biomass from year 300 to 500. Of the six regions analyzed, central Siberia has one of the lowest average seasonal precipitation curves; it is closest to that of northwestern Siberia. Forests within central Siberia naturally transition from a larch-dominated to an evergreen-conifer dominated system, with larch as a secondary species (Figure 7a). This successional transition is similar to local dynamics at sites in both East and West Irkutsk, regions in southern Siberia that both have higher precipitation than central Siberia. Northwestern Siberia, which has similar precipitation to central Siberia, does not experience a natural shift from larch- to evergreen conifer-dominance. The colder regional temperatures in northwestern Siberia suppress evergreen conifer growth, thereby helping larch to maintain dominance (Figure 7d). These observations suggest that the precipitation treatment response in central Siberia is a result of both the lower seasonal precipitation, and the presence of evergreen conifers late in succession in this region. In other words, the transition from larch- to evergreen conifer-dominant in central Siberia, combined with the low annual precipitation, creates a region which is responsive to precipitation change (Figure 6). The variability across the low diversity regions suggests that differences in response to climate treatments are a result of local conditions and species composition over time. The continental scale analysis shows a response to precipitation for both total forest and evergreen conifer biomass. This suggests that many of the sites considered in this continental scale analysis have low seasonal precipitation, and are dominated by moisture-sensitive evergreen conifers later in

There are marked differences between the results for the high and the low diversity regions, which demonstrate differences in the stability and resilience of these regions. The high diversity areas of the RFE have high ecological resilience and maintain basic ecosystem function as a result of similarly functioning species replacing one another following climate

**4.2.3 Low diversity regional and local scale discussion** 

enough that evergreen conifers will be able to effectively establish.

succession.

system where the next state of the system is predictable from within the system (Margalef 1968; Lewontin 1969; Child and Shugart 1972). Ecological resilience is therefore related to the magnitude of disturbance or change that can be absorbed before the system transitions to another stability domain (Folke et al., 2004; Gunderson 2000, Peterson et al., 1998). In the N RFE, the successional cycle from larch-dominance to evergreen-dominance is the expected cycle between vegetation states, because this is the response for base climate conditions. Both N RFE and SW RFE respond to an increase in temperature with a shift in the dominant species during late succession, but these late successional groups are functionally similar to the assemblage of species that exist under the base climate. This is an indication that this group of high diversity regions in the Amur region of the RFE does not experience a change in vegetation state with in response to climate change, and is resilient to the perturbations associated with this amount of climate change.

The concept of response diversity adds to the conclusion that the system in the Amur region of the RFE is resilient by defining the range of reactions to environmental change among species contributing to the functioning of a given ecosystem (Elmqvist et al., 2003; Folke et al., 2004). The RFE region has high response diversity, which means that it has functionally similar species sets which respond differently to environmental change and provide a buffer that protects the system against failure and increases tolerance to disturbance or climate change (Elmqvist et al., 2003; Folke et al., 2004). In other words, there are species in the system capable of maintaining the original state of ecosystem function under the new conditions following change; a concept also known as the insurance hypothesis (Folke et al., 1996, Naeem and Li 1997). High response diversity within an ecosystem increases the chances of reorganization or restart of the system into the desired state after disturbance (Chapin et al., 1997; Bengtsson et al., 2000; Elmqvist et al., 2003). The altered climate disturbed the ecosystem, but because of the high diversity of the SW and N RFE, the species can reorganize and maintain the same vegetation state and ecosystem function that is observed throughout succession when the system is not disturbed by climate change. Therefore the diversity of species in this region allows for replacement of one species with another functionally similar one under new climate conditions. It is also the adaptability of the system under altered climate which prevents a substantial contribution to biomass from the introduction of *L. decidua*. Neither a change in climate, nor the addition of a single species (*L. decidua)* leads to a change in ecosystem function in this high diversity system.

It is important to note that the RFE regions analysed display ecological resilience and high response diversity for the both temperature increase and precipitation change treatments evaluated. There is a response to temperature in both regions, but not at the same magnitude as that of the low diversity areas. These results suggest that increased amounts of climate change may have a stronger impact on the system. Further analysis is necessary to determine if the RFE system is equally resilient when the temperature is increased by more than 4C, or if it can restart under this altered climate condition following disturbance, such as fire or insect outbreak, that take the system back to bare ground. The results seen in this study indicate only a slight sensitivity of the mid- to late-successional stages, at and beyond year 200, which correspond to the time in the simulation when the temperature had increased to 4C. The early successional stages are comprised of a different set of species which may not show the same high resilience or response diversity displayed by the mid- to late-successional stages in the RFE.

system where the next state of the system is predictable from within the system (Margalef 1968; Lewontin 1969; Child and Shugart 1972). Ecological resilience is therefore related to the magnitude of disturbance or change that can be absorbed before the system transitions to another stability domain (Folke et al., 2004; Gunderson 2000, Peterson et al., 1998). In the N RFE, the successional cycle from larch-dominance to evergreen-dominance is the expected cycle between vegetation states, because this is the response for base climate conditions. Both N RFE and SW RFE respond to an increase in temperature with a shift in the dominant species during late succession, but these late successional groups are functionally similar to the assemblage of species that exist under the base climate. This is an indication that this group of high diversity regions in the Amur region of the RFE does not experience a change in vegetation state with in response to climate change, and is resilient to the perturbations

The concept of response diversity adds to the conclusion that the system in the Amur region of the RFE is resilient by defining the range of reactions to environmental change among species contributing to the functioning of a given ecosystem (Elmqvist et al., 2003; Folke et al., 2004). The RFE region has high response diversity, which means that it has functionally similar species sets which respond differently to environmental change and provide a buffer that protects the system against failure and increases tolerance to disturbance or climate change (Elmqvist et al., 2003; Folke et al., 2004). In other words, there are species in the system capable of maintaining the original state of ecosystem function under the new conditions following change; a concept also known as the insurance hypothesis (Folke et al., 1996, Naeem and Li 1997). High response diversity within an ecosystem increases the chances of reorganization or restart of the system into the desired state after disturbance (Chapin et al., 1997; Bengtsson et al., 2000; Elmqvist et al., 2003). The altered climate disturbed the ecosystem, but because of the high diversity of the SW and N RFE, the species can reorganize and maintain the same vegetation state and ecosystem function that is observed throughout succession when the system is not disturbed by climate change. Therefore the diversity of species in this region allows for replacement of one species with another functionally similar one under new climate conditions. It is also the adaptability of the system under altered climate which prevents a substantial contribution to biomass from the introduction of *L. decidua*. Neither a change in climate, nor the addition of a single species (*L. decidua)* leads to a change in ecosystem

It is important to note that the RFE regions analysed display ecological resilience and high response diversity for the both temperature increase and precipitation change treatments evaluated. There is a response to temperature in both regions, but not at the same magnitude as that of the low diversity areas. These results suggest that increased amounts of climate change may have a stronger impact on the system. Further analysis is necessary to determine if the RFE system is equally resilient when the temperature is increased by more than 4C, or if it can restart under this altered climate condition following disturbance, such as fire or insect outbreak, that take the system back to bare ground. The results seen in this study indicate only a slight sensitivity of the mid- to late-successional stages, at and beyond year 200, which correspond to the time in the simulation when the temperature had increased to 4C. The early successional stages are comprised of a different set of species which may not show the same high resilience or response diversity displayed by the mid- to

associated with this amount of climate change.

function in this high diversity system.

late-successional stages in the RFE.

#### **4.2.3 Low diversity regional and local scale discussion**

The response of the low diversity regions to the altered climate treatment is similar to that observed at the continental scale. These low diversity regional scale responses, in conjunction with the low average diversity across the 372 sites considered at the continental scale, further emphasize the difference of the response in the high and low diversity areas. Similar to analysis at the continental scale, the response of the low diversity regions to treatments is connected to the transition from larch to evergreen dominance, which occurs across the southern and central portion of Siberia, and is accelerated by warming. Northwestern Siberia has the coldest temperatures compared to the other regions, and these cold temperatures naturally suppress evergreen conifer establishment during late succession under base climate. The patterns of successional dynamics in central Siberia under base conditions and in northwestern Siberia in response to warming climate suggest that the model is predicting consistent transitions. Under warming conditions, northwestern Siberia is exposed to temperatures more similar to those in the base condition in central Siberia. Thus, the forests in northwestern Siberia have similar dynamics under warming conditions to those seen in the central Siberia region for the base climate (Figure 7a,e). The temperature treatment results suggest that, with 4C of warming, the larch-dominated system across southern and central Siberia will be prematurely replaced with evergreen conifer and other deciduous trees, and the forests of northwestern Siberia will warm enough that evergreen conifers will be able to effectively establish.

The lack of response in northwestern and southern Siberia, and the late response in central Siberia, to the effect of precipitation change suggests a connection between evergreen conifer presence and seasonal precipitation. The precipitation treatment is significant in only one of the low diversity regions analyzed (i.e., central Siberia). Within the central Siberia region there is a short-lived response of total forest biomass to precipitation change at year 240, and a longer period of response of evergreen conifer biomass from year 300 to 500. Of the six regions analyzed, central Siberia has one of the lowest average seasonal precipitation curves; it is closest to that of northwestern Siberia. Forests within central Siberia naturally transition from a larch-dominated to an evergreen-conifer dominated system, with larch as a secondary species (Figure 7a). This successional transition is similar to local dynamics at sites in both East and West Irkutsk, regions in southern Siberia that both have higher precipitation than central Siberia. Northwestern Siberia, which has similar precipitation to central Siberia, does not experience a natural shift from larch- to evergreen conifer-dominance. The colder regional temperatures in northwestern Siberia suppress evergreen conifer growth, thereby helping larch to maintain dominance (Figure 7d). These observations suggest that the precipitation treatment response in central Siberia is a result of both the lower seasonal precipitation, and the presence of evergreen conifers late in succession in this region. In other words, the transition from larch- to evergreen conifer-dominant in central Siberia, combined with the low annual precipitation, creates a region which is responsive to precipitation change (Figure 6). The variability across the low diversity regions suggests that differences in response to climate treatments are a result of local conditions and species composition over time. The continental scale analysis shows a response to precipitation for both total forest and evergreen conifer biomass. This suggests that many of the sites considered in this continental scale analysis have low seasonal precipitation, and are dominated by moisture-sensitive evergreen conifers later in succession.

There are marked differences between the results for the high and the low diversity regions, which demonstrate differences in the stability and resilience of these regions. The high diversity areas of the RFE have high ecological resilience and maintain basic ecosystem function as a result of similarly functioning species replacing one another following climate

Resilience and Stability Associated with Conversion of Boreal Forest 211

different scales in response to altered climate. The model simulated forest growth at high diversity sites in the Amur region of the RFE and low diversity sites across Siberia and the remainder of the RFE. The model successfully captures the natural successional dynamics

Results of the climate sensitivity analysis indicate that a 4C increase impacts the biomass of the total forest, *Larix* spp., and evergreen conifers at the continental scale, and in low diversity regions, early in succession. The effect of temperature is highly significant throughout most of the simulation in low diversity areas, whereas there is little effect of temperature in high diversity regions. Results at the continental scale suggest that the forests across much of Siberia and the RFE behave as a low diversity system. The early effect of temperature, across areas where larch is naturally dominant in early succession, suggests that larch is particularly vulnerable to temperature increase. In areas outside the cold northern portion of Siberia, larch is shown to abruptly collapse in response to warming. The effect of altered precipitation was significant at the continental scale, and in one low diversity region in central Siberia in mid- to late-succession following evergreen conifer establishment. Central Siberia, in addition to experiencing a late successional shift from larch- to evergreen conifer-dominance, has low seasonal precipitation. This suggests that sites that respond to the precipitation treatment have low annual precipitation, and are dominated by moisture-sensitive conifers. The precipitation effect was not significant in the high diversity regions, which have different dominant evergreen species in late succession. Concepts of ecological stability and resilience are used to explain the variable response of the high and low diversity areas to altered climate. The high diversity regions showed high stability and resilience for they maintained overall species and biomass dynamics in response to changing climate, and had replacement of one species by a functionally similar species under new climate conditions. Unlike the high diversity regions, the low diversity regions across Siberia show low ecological stability and resilience. The low diversity areas displayed a strong response of biomass to the climate treatments, and locally showed the collapse of the dominant larch species under increased temperatures. It is this the lack of diversity in the response of functionally similar species to environmental change, and thus an inability to maintain ecosystem function with altered conditions, which creates the low

*L. decidua*, the warmer adapted European larch, was added to the species list to gauge the potential of this species to prevent a premature shift to evergreen vegetation, and the associated albedo shift, in response to climate change. *L. decidua* established strongly in the low diversity system, but not in the high diversity areas of the RFE. Due to the increased diversity in the Amur region of the RFE, there are native species which can fill the same functional space under new climate conditions as the species which were dominant under base climate conditions, thereby creating high resilience and stability. It is this pool of locally available species which prevents *L. decidua* from significantly impacting biomass in the high diversity regions of the RFE. Within the low diversity regions, however, local scale results show that *L. decidua* becomes established and acts to prevent the collapse of larch in response to warming and delay the shift to an evergreen conifer-dominated forest. It is the low diversity which contributes to the lack of resilience and stability under altered climate, but also allows for strong establishment of *L. decidua*. Therefore *L. decidua* is uniquely adapted to establish in this low diversity system as well as to prevent the positive feedback

This study establishes that larch-dominated forests across Siberia will transition to a different vegetation state, and have an altered species composition, in response to climate

of forests for base climate conditions across the area.

ecological stability across the low diversity areas of Siberia.

associated with a premature shift to evergreen conifer-dominance.

change. The low diversity regions, however, have low resilience and cannot maintain basic ecosystem function following climate change, specifically temperature increase. This is shown by the collapse of dominant larch species following the 4C increase in sites in central and southern Siberia, and the fact that the northwestern Siberia sites shift to an entirely new stable state, which is not seen under base conditions in this region, in response to warming climate conditions. The collapse of larch in southern regions in response to a temperature increase suggests that larch is particularly vulnerable, and that the systems' response threshold may be exceeded with the 4C increase. With fewer species present, it is more likely that extinctions will alter ecosystem processes (Chapin et al., 1997). Furthermore, the diversity of the area is so low that there are no species capable of fulfilling the original ecosystem function under the new conditions following change. Additionally, differences in sensitivity among functionally different species, in this case larch and evergreen conifers, make the ecosystem vulnerable to change (Chapin et al., 1997).

Holling (1992) hypothesized that the vegetation of the boreal forest would buffer initial climate changes, but that there would be a limit to the buffering and an abrupt vegetation change would follow. These results follow Holling's hypothesis and suggest an abrupt shift in vegetation in response to temperature increase. This abrupt vegetation change is congruent with the identification of this system as having low resilience and stability when compared with the high diversity areas of the RFE. Chapin et al., (2004) suggested that vegetation within central portions of the boreal forest would remain stable for long periods followed by abrupt changes to a new state, which is what we see in the results for the low diversity areas in central Siberia. Low diversity areas do not have the appropriate pool of species to continue the cycle of succession and reorganization following change, thus the system in these areas is flipped into a different state (Bengtsson et al., 2000). The results presented here are consistent with field measurements documenting the shift of treelines northward or upslope of previous climate limits, and a reduction in cone and seed yield for *L. sibirica* (Kharuk et al., 2009; Soja et al., 2007). They are also consistent with bioclimatic model results predicting a replacement of taiga with forest-steppe or steppe environments across southern Siberia (Tchebakova et al., 2005; Vygodskaya et al., 2007; Tchebakova et al., 2009). These results also suggest that warming temperatures will lead to a shift in the ability of larch to establish and may signal a collapse of the species in this genus.

The introduction of *L. decidua* to the low diversity sites may help to buffer the perturbations associated with a warming climate. Unlike the high diversity regions in the RFE, the low diversity areas showed a strong response of biomass to the inclusion of *L. decidua*. Local scale analysis with warming conditions shows that the inclusion of *L. decidua* prevents the collapse of larch in central Siberia and delays transition to evergreen conifer dominance in northwestern Siberia. *L. decidua* is competitive in this low diversity area, and fills an important functional niche when temperatures are increased. Existing larch species cannot tolerate the warmer conditions, and their collapse opens functional space for the warm adapted *L. decidua*. These results, though theoretical, provide evidence that it is possible to address the issue of species replacement, and associated albedo shift, with techniques involving species management or introduction.
