1.3.2.2 Boreal forests & tundra

Boreal forests, also called ‘Taiga’, span around 1,135 million hectares, all located in the northern hemisphere (Pan et al., 2011) (Figure 1.3.4). They are vital for climate regulation, storing circa 272 (± 23) GtC, mostly below ground (Pan et al., 2011; Mayer et al., 2020). Management varies, but illegal logging constitutes a critical driver of boreal forest loss. Boreal forest growth is constrained by a short vegetation period, and their dynamics involve large-scale disturbances such as insect outbreaks and fire (with fire percolation dynamics important – see 1.3.2.4). While disturbance regimes differ in Eurasian and American forests, an overall increase in disturbances has been observed over past decades, fuelling worries about a wider loss of resilience.

Evidence for tipping dynamics

Boreal forest dieback has already been identified as a potential tipping element in the climate system in Lenton et al., (2008) and further assessed in the IPCC AR5 WG2 Report in 2014 and in the WG1 and WG2 reports of AR6 in 2021 and 2022. The IPCC SR1.5 (Hoegh-Guldberg et al., 2018) and also the most recent assessment by (Armstrong McKay et al., 2022) differentiate between southern boreal forest and northern tundra tipping points. The southern boreal forest tipping point refers to a dieback of southern boreal forests that lead to a state-shift to an almost treeless state (to steppe/prairie), while the northern tundra tipping point refers to an expansion of tree cover into currently treeless tundra ecosystems. 

There is little additional evidence for a boreal forest tipping point since the assessment of Armstrong McKay et al., (2022). Significant losses in tree cover driven by fires and logging were identified for the southern boreal forests of North America between 2000 and 2019 (Rotbarth et al., 2023). In contrast, interior boreal forests have become denser. There has been no clear sign of a northward expansion of the boreal forests of North America (Rotbarth et al., 2023). Similarly, Burrell et al. (2021) found that the forests of southern Siberia might have approached a tipping point as fire regimes have intensified, causing widespread regeneration failure. Moreover, Siberian larch foliage is sensitive to warming, with temperatures potentially exceeding a threshold by 2050 after which forest dieback can be expected (Rao et al., 2023).

A range of mechanisms contribute to the feedback processes associated with boreal tipping points (see Box 1.3.2 for more on forest feedbacks). For the southern boreal forest, the recent surge in forest disturbances, such as the extreme forest fires in Canada in summer 2023, is noteworthy because they constitute a substantial change in forest dynamics and resilience that, combined with failure to regenerate, could initiate regional tipping. In particular, the southern trailing edge of boreal forests has been identified as prone to compound and interacting disturbances, including droughts, windstorms, fires, large herbivores and insect outbreaks (Frehlich and Reich, 2010). For instance, increasing water stress reduces tree resistance against insects, and increases the size and severity of wildfires. 

Southern boreal tipping points are driven by forest dieback from disturbances (Lenton et al., 2008). Empirical evidence from satellite data suggests that disturbances are responsible for switches between states rather than causing gradual change (Scheffer et al., 2012; Abis and Brovkin, 2017). Rotbarth et al. (2023) confirm that processes dominating the dieback of southern boreal forests and the northward expansion of forests into tundra diverge and that a northward expansion is not compensating for declines in the southern boreal forests of North America.

Climate change will further intensify disturbance regimes (Seidl et al., 2017), with fire regimes expected to increase significantly in boreal forests (Velasco Hererra et al., 2022). In Canada, fire frequency could increase up to 50 per cent by the 21st century under climate change (Flannigan et al., 2013). A doubling of fire frequency and increased wind activity during the 21st century will likely cause a significant decrease in coniferous forests, potentially replaced by early successional broadleaved tree species (Anoszko et al., 2022; Liu et al., 2022). 

The increase in fire could potentially modify the forest microclimate, so that subsequent fires and droughts become more likely, causing a change in vegetation dynamics. For instance, Whitman et al. (2019) found that drought after fire exacerbates regeneration failure. Overall, drought-induced mortality will likely rise more in western than eastern North American regions (Peng et al., 2010). Moreover, insects, such as mountain pine beetles might expand into North American boreal forests, causing changes in ecosystem dynamics (Safranyik et al., 2010; Jarvis and Kulakowski, 2015). For instance, severe defoliation could impede birch forest recovery (Vindstad et al., 2018). 

If these changes in disturbances cause widespread mortality while, at the same time, forests fail to regenerate, the forest might tip into an almost treeless state. Stevens-Rumann et al., (2022) suggest that a combination of changing climate patterns and disturbance regimes could primarily cause regeneration failure in coniferous forests. Bailey et al., (2021) highlighted the importance of temperature-moisture interactions for successful seedling establishment at the upper treeline in the Southern Rocky Mountains. However, over the past decade, no seedling establishment occurred at any site, suggesting that a threshold for regeneration may have been passed. Regeneration failure of boreal forests might occur with warming alone (+1.6°C to +3.1°C increase in one local warming experiment), but temperature thresholds are reduced if an increase in temperature is combined with reduced precipitation (Reich et al., 2022). 

The sensitivity of coniferous tree recruits to climate change is overall higher than for broadleaved tree regeneration (Reich et al., 2022; Stevens-Rumann et al., 2022). In addition, natural disturbances might more likely cause state-shifts of coniferous than broadleaved-dominated boreal forest (Thom, 2023) as broadleaved tree species have an overall higher resprouting ability than conifers (Thom et al., 2021). Topographic complexity and peatlands may act as refugia from fire (Kuntzemann et al., 2023; Rogeau et al., 2018), thus reducing the likelihood of regeneration failure and state shifts. If widespread mortality becomes an increasing issue in northern forests, reduced microclimatic buffering of forests to increasing temperature might accelerate the thawing of permafrost in the boreal biome, causing additional releases of greenhouse gases – further interacting with the climate system [See 1.2.2.4 on Permafrost].

An increase in abundance of woody plants and advancing shrublines into the Arctic tundra is likely as climate changes (Mekkonen et al., 2021). This shrubification driven by warmer climate is also accompanied by northward treeline migration. A recent review of more than 400 treeline site locations suggested that at about two-thirds of treeline sites’ forest cover had increased in elevational or latitudinal extent (Hansson et al., 2021). Main drivers of treeline migration are an increase in the rate of seedling success through warmer summers and increased winter temperatures. The change from tundra and peatlands to boreal forests can be nonlinear. Experimental work in boreal peat bogs reveals positive interactions between shrub cover and tree recruitment in which shrub cover favours tree seedlings and, in turn, higher tree basal area fosters shrub biomass, potentially triggering tipping towards high tree cover (Holmgren et al., 2015). As with southern dieback, interaction with permafrost thaw is also likely, but is complex and currently uncertain. 

There are no clear thresholds for boreal forest dieback beyond the initial estimates already presented in Armstrong McKay et al., [2022]. With low confidence, they estimate a southern dieback tipping point at a global warming threshold of ~4°C (1.4-5°C) and a tipping timescale of ~100 (50-?) years, and a northern expansion into tundra tipping point at an estimated global warming threshold of ~4°C (1.5-7.2°C) and a tipping timescale of ~100 (40-?) years. Regeneration failure of southern boreal forests might occur with warming alone, while those thresholds are even lower if precipitation amounts also decrease (Reich et al., 2022).

Assessment and knowledge gaps

We assess with medium confidence that larger parts of boreal forests will approach a southern dieback tipping point and with low confidence that they will expand northwards as global temperatures increase by 3-4°C, if precipitation amounts and patterns remain similar. Yet, this threshold depends on multiple factors such as human and natural disturbances. 

The capacity for adaptation and resilience is among the key uncertainties. Biodiversity, among other factors, might influence tipping dynamics as a diverse ecosystem may be more resistant to reaching tipping points, yet the effects of compositional and structural diversity require further investigation. Furthermore, although there is strong evidence and confidence in the increase of natural disturbances in boreal forests it remains uncertain whether they will truly lead to the transgression of a tipping point, pushing the southern range of boreal forests into an alternative, treeless state. 

While in the southern boreal region the main mechanisms causing tipping points are relatively clear, for the northern tundra expansion tipping point the mechanisms sustaining large-scale abrupt state-shifts are not as evident. Disturbances in this region may be weaker and more localised, and the replacement of tundra by forest might occur more gradually. Yet, it is unclear if regeneration failure drives a self-sustaining feedback loop hindering recovery due to soil dryness or extreme conditions, causing a tipping point.

Further uncertainties linked to tipping points requiring further investigation include testing: 

1. interactions between climate, atmospheric forcing and disturbances; 
2. cascading and compounding disturbances; 
3. the existence of a ‘fast-in, fast-out’ behaviour of release and recovery in boreal forests; 
4. whether changes are self-reinforcing and perpetuating forest loss (or gain in the case of the northern tipping point); 
5. the extent of southern forest loss vs. northern forest expansion; and 
6. the role of human interventions, such as forest management on tipping dynamics.