**1. Introduction**

A clear understanding of boreal forest dynamics is critical to developing an accurate representation of the Earth's response to climate change. The Russian boreal forest is the largest continuous forest region on Earth and a tremendous repository of terrestrial organic carbon. The boreal forest has experienced significant warming over the past several decades and is expected to be impacted by global climate change (Chapin et al., 2000; McGuire et al., 2002; Soja et al., 2007). Siberian summers in the past century were warmer than any century in the past millennium, and future climate scenarios indicate that the region will continue warming, by some accounts between 2° and 10°C by 2100 (IPCC 2007; Soja et al., 2007). Warming climate will likely exert influence on species distributions and land cover types in the boreal forest regions (Ustin and Xiao 2001; Tchebakova et al., 2005; Tchebakova et al., 2009). In particular, these temperature increases have led to the shift of treelines northward or upslope of previous climate limits, and a reduction in cone and seed yield for *Larix sibirica* and *Pinus slyvestris* which changes forest composition and structure (Kharuk et al., 2009; Soja et al., 2007). These changes are important indicators of how Eurasian boreal forests may respond to, and ultimately amplify, increases in average global temperature.

These land cover changes can force alterations in regional climate through modifications in surface albedo and land/atmosphere energy fluxes (Bonan et al., 1992; Chapin et al., 2000; Baldocchi 2000; Amiro 2001; Beringer et al., 2005; Soja et al., 2007), as well as in global climate through changes in carbon sequestration and release patterns (Bonan 2008; Snyder et al., 2004). Global climate model (GCM) results have shown that clearing boreal forest alters surface albedo, and substantially cools the Earth, not only in the boreal region but across the Northern Hemisphere (Bonan et al., 1992), and has the greatest effect on global mean temperature when compared to the removal of other biomes (Snyder et al., 2004). Betts (2000) found surface albedo changes associated with the growth of coniferous evergreen trees led to significant increases in average global temperature large enough to overshadow the effect of carbon storage by growing evergreen forest in that region. Bioclimatic modeling predicts that by 2090 vegetation change across Siberia will create an albedo shift and increase overall net radiation, thereby producing enhanced warming above that already predicted for the high latitudes (Vygodskaya et al., 2007). Larch (*Larix* spp.) forest, dominated by both *L. sibirica* and *L. gmelinii,* covers extensive regions in Siberia. Field observations have documented shifts from larch to evergreen conifer forests, dominated by trees such as spruce (*Picea* spp.) or fir (*Abies* spp.) that are tolerant of higher temperatures

Resilience and Stability Associated with Conversion of Boreal Forest 197

model included a simulation of forest composition and basal area at different elevations on Changbai Mountain in China, with statistical comparison to inventory data and then qualitative comparisons to observed forest type at 31 sites in the Russian Far East and Siberia (Yan and Shugart 2005). Further validation of the model using linear regression of model generated and independent forest inventory data indicated that FAREAST successfully captures the natural biomass dynamics of mixed-species forests across the vast geographic area and varied climatic conditions of Russia (Shuman and Shugart 2009; Shuman 2010). The use of gap models allows for the evaluation of novel conditions or the addition of a new species for the purpose of evaluating the impact on existing vegetation. The impact that changing climate has on forests at local and regional scales has been explored with several different forest gap models (Shugart 1984; Solomon 1986; Pastor and Post 1988; Dale and Franklin 1989; Urban et al., 1993; Lasch and Lindner 1995; Bugmann 1996; Yan and Zhao 1996; Bugmann and Solomon 2000; Zhang et al., 2009). IBMs can be used to develop a vegetation "signature" for the response of ecosystems to change, especially climate change. Using a boreal forest gap model to assess climate change effects, Bonan (1989a,b) investigated the responses to several climate change predictions from global climate models along the north-facing and south-facing slopes of boreal forests near Fairbanks, Alaska. The black spruce forests growing on cold north-facing slopes were largely unaffected by the climatic warming, but white spruce forests on the relatively warmer south-facing slopes were strongly affected by the change in climate. Conditions predicted in the climate change scenarios for south-facing slopes were outside the ecological conditions under which the common tree species near Fairbanks are known to be able to persist. For white spruce, the limiting condition identified by the model results appeared to be moisture stress and not the direct effects of temperature change. A decade later, Barber et al., (2000) investigated tree ring data to determine the effect that several decades of warmer than usual temperatures in the Fairbanks area had had on white spruce stands and confirmed Bonan's model predictions with evidence for moisture-stress

In this study, the FAREAST model is used to simulate forest composition and biomass at 372 sites across Siberia and the Russian Far East for the purpose of evaluating forest response to climate change. Climate sensitivity analysis is performed in order to assess the resilience and stability of forest structure and composition to altered climate at multiple spatial scales. The model was used to simulate the impact of changes in temperature and precipitation on both total and genus-specific biomass at sites across Siberia and the Russian Far East, and for six different regions representing areas of high and low diversity. Comparisons of regions within areas of high and low diversity provide a tool to evaluate the relationship between diversity and the response of the system to changing climate. Model runs with and without European Larch (*Larix decidua*) are compared in order to assess the potential for the introduction of this species to mitigate the effects of climate change, especially the positive

FAREAST was run at a total of 372 sites across Siberia and the Russian Far East (RFE) from the eastern coast to the western border of the range limits of *L. sibirica*. FAREAST uses monthly climate parameters derived from historical station data to compute daily

feedback among temperature, forest type and surface albedo.

**2.1 Model simulation across Siberia and Russian Far East** 

effects in the tree ring dataset.

**2. Methods** 

(Kharuk et al., 2007). Because larch is a deciduous conifer, this shift to evergreen dominance would lead to an albedo decrease, particularly in winter, when evergreen trees tend to mask laying snow relative to deciduous species (Betts and Ball 1997). The difference in summer albedo is smaller but also significant, with larch albedo measured at approximately 0.13 and evergreen species around 0.09 (Hollinger et al., 2010) This reduction of albedo associated with the shift in forest type indicates that increased temperatures may lead to a positive feedback response: a warmer climate accelerates the natural succession from larch to evergreen conifer forest and the resultant albedo promotes additional warming. Areas of southern Siberia identified as vulnerable to premature replacement of larch by evergreen conifers would undergo a local significant albedo shift of approximately 5.1 W m-2 following conversion from dominant larch to evergreen conifer stands (Shuman et al., 2011).

Identification of areas prone to vegetation change is crucial in efforts to mitigate the effects of potential forest type conversion. Remote sensing technology has advanced to a point which allows for estimation of biomass and detailed evaluation of land cover and land use change. Estimation of Russian forest biomass directly from Moderate Resolution Imaging Spectroradiometer (MODIS) data provided estimates of a distribution of biomass classes that correlated well to ground measured forest biomass with signatures from a minimum of training sites (Houghton et al., 2007). Detailed characterization of vegetation by remote sensing technology provides land cover maps for areas within Russia that are not easily accessible, and at a more frequent temporal resolution than is possible to obtain using fieldbased mapping techniques. A NOAA/AVHRR-derived vegetation map for a remote section of northern Siberia provides detailed information regarding latitudinal transition zones, vegetation differences inside each zone and variability along vertical transects for mountainous areas (Kharuk et al., 2003). This type of high quality map for a remote area provides land cover information essential to evaluating vegetation changes in response to climate change. Remotely sensed data can be used to identify areas undergoing a change in land cover type and assess the direction and magnitude of any albedo shift associated with such a change. Vegetation models can be used to provide information regarding the specific type of vegetation change and the location where this change is most likely to occur, and can thus inform vegetation monitoring efforts based on remotely sensed data.

In the past 20 years, individual-based models (IBMs) have been used to provide increasingly accurate predictions and simulations of forests (Mladenoff 2004; Scheller and Mlandenoff 2007). The model used in this study, FAREAST, is in a class of IBMs called "gap models" (Shugart and West 1980) that simulate individual trees, specifically their growth, mortality, and decomposition into litter in a relatively small area, typically the size of a forest gap. Forest gap models established according to the approach of Botkin et al., (1972) and Shugart and West (1977) are based upon the concept of "gap phase" replacement (Watt 1947). Gap models account for competition among individuals of multiple tree species for light and other resources, with the outcome determining the composition and structure of the forest through aggregation of homogenous mosaic patches through time (Shugart 1984). Testing of gap models is divided into verification and validation (Mankin et al., 1977; Cale et al., 1983; Rykiel 1996; Sargent 1984), and involves evaluating the ability of the model to predict species successional dynamics and biomass accumulation for the region of model development. Gap models have been verified and validated for a variety of forests world-wide (Botkin et al., 1972; Shugart and West 1977; Shugart, 1984 and its reprinting 1998, 2003; Kienast 1987; Leemans and Prentice 1989; Kienast and Krauchi 1991; Bugmann 2001). The initial tests of the FAREAST

(Kharuk et al., 2007). Because larch is a deciduous conifer, this shift to evergreen dominance would lead to an albedo decrease, particularly in winter, when evergreen trees tend to mask laying snow relative to deciduous species (Betts and Ball 1997). The difference in summer albedo is smaller but also significant, with larch albedo measured at approximately 0.13 and evergreen species around 0.09 (Hollinger et al., 2010) This reduction of albedo associated with the shift in forest type indicates that increased temperatures may lead to a positive feedback response: a warmer climate accelerates the natural succession from larch to evergreen conifer forest and the resultant albedo promotes additional warming. Areas of southern Siberia identified as vulnerable to premature replacement of larch by evergreen conifers would undergo a local significant albedo shift of approximately 5.1 W m-2 following

conversion from dominant larch to evergreen conifer stands (Shuman et al., 2011).

thus inform vegetation monitoring efforts based on remotely sensed data.

In the past 20 years, individual-based models (IBMs) have been used to provide increasingly accurate predictions and simulations of forests (Mladenoff 2004; Scheller and Mlandenoff 2007). The model used in this study, FAREAST, is in a class of IBMs called "gap models" (Shugart and West 1980) that simulate individual trees, specifically their growth, mortality, and decomposition into litter in a relatively small area, typically the size of a forest gap. Forest gap models established according to the approach of Botkin et al., (1972) and Shugart and West (1977) are based upon the concept of "gap phase" replacement (Watt 1947). Gap models account for competition among individuals of multiple tree species for light and other resources, with the outcome determining the composition and structure of the forest through aggregation of homogenous mosaic patches through time (Shugart 1984). Testing of gap models is divided into verification and validation (Mankin et al., 1977; Cale et al., 1983; Rykiel 1996; Sargent 1984), and involves evaluating the ability of the model to predict species successional dynamics and biomass accumulation for the region of model development. Gap models have been verified and validated for a variety of forests world-wide (Botkin et al., 1972; Shugart and West 1977; Shugart, 1984 and its reprinting 1998, 2003; Kienast 1987; Leemans and Prentice 1989; Kienast and Krauchi 1991; Bugmann 2001). The initial tests of the FAREAST

Identification of areas prone to vegetation change is crucial in efforts to mitigate the effects of potential forest type conversion. Remote sensing technology has advanced to a point which allows for estimation of biomass and detailed evaluation of land cover and land use change. Estimation of Russian forest biomass directly from Moderate Resolution Imaging Spectroradiometer (MODIS) data provided estimates of a distribution of biomass classes that correlated well to ground measured forest biomass with signatures from a minimum of training sites (Houghton et al., 2007). Detailed characterization of vegetation by remote sensing technology provides land cover maps for areas within Russia that are not easily accessible, and at a more frequent temporal resolution than is possible to obtain using fieldbased mapping techniques. A NOAA/AVHRR-derived vegetation map for a remote section of northern Siberia provides detailed information regarding latitudinal transition zones, vegetation differences inside each zone and variability along vertical transects for mountainous areas (Kharuk et al., 2003). This type of high quality map for a remote area provides land cover information essential to evaluating vegetation changes in response to climate change. Remotely sensed data can be used to identify areas undergoing a change in land cover type and assess the direction and magnitude of any albedo shift associated with such a change. Vegetation models can be used to provide information regarding the specific type of vegetation change and the location where this change is most likely to occur, and can model included a simulation of forest composition and basal area at different elevations on Changbai Mountain in China, with statistical comparison to inventory data and then qualitative comparisons to observed forest type at 31 sites in the Russian Far East and Siberia (Yan and Shugart 2005). Further validation of the model using linear regression of model generated and independent forest inventory data indicated that FAREAST successfully captures the natural biomass dynamics of mixed-species forests across the vast geographic area and varied climatic conditions of Russia (Shuman and Shugart 2009; Shuman 2010).

The use of gap models allows for the evaluation of novel conditions or the addition of a new species for the purpose of evaluating the impact on existing vegetation. The impact that changing climate has on forests at local and regional scales has been explored with several different forest gap models (Shugart 1984; Solomon 1986; Pastor and Post 1988; Dale and Franklin 1989; Urban et al., 1993; Lasch and Lindner 1995; Bugmann 1996; Yan and Zhao 1996; Bugmann and Solomon 2000; Zhang et al., 2009). IBMs can be used to develop a vegetation "signature" for the response of ecosystems to change, especially climate change. Using a boreal forest gap model to assess climate change effects, Bonan (1989a,b) investigated the responses to several climate change predictions from global climate models along the north-facing and south-facing slopes of boreal forests near Fairbanks, Alaska. The black spruce forests growing on cold north-facing slopes were largely unaffected by the climatic warming, but white spruce forests on the relatively warmer south-facing slopes were strongly affected by the change in climate. Conditions predicted in the climate change scenarios for south-facing slopes were outside the ecological conditions under which the common tree species near Fairbanks are known to be able to persist. For white spruce, the limiting condition identified by the model results appeared to be moisture stress and not the direct effects of temperature change. A decade later, Barber et al., (2000) investigated tree ring data to determine the effect that several decades of warmer than usual temperatures in the Fairbanks area had had on white spruce stands and confirmed Bonan's model predictions with evidence for moisture-stress effects in the tree ring dataset.

In this study, the FAREAST model is used to simulate forest composition and biomass at 372 sites across Siberia and the Russian Far East for the purpose of evaluating forest response to climate change. Climate sensitivity analysis is performed in order to assess the resilience and stability of forest structure and composition to altered climate at multiple spatial scales. The model was used to simulate the impact of changes in temperature and precipitation on both total and genus-specific biomass at sites across Siberia and the Russian Far East, and for six different regions representing areas of high and low diversity. Comparisons of regions within areas of high and low diversity provide a tool to evaluate the relationship between diversity and the response of the system to changing climate. Model runs with and without European Larch (*Larix decidua*) are compared in order to assess the potential for the introduction of this species to mitigate the effects of climate change, especially the positive feedback among temperature, forest type and surface albedo.
