1.3.2.6 Freshwater ecosystems

The scientific content of this chapter is closely based on the following scientific manuscript: Hessen et al., (in review) Lake ecosystem tipping points and climate feedbacks, Earth System Dynamics Discussion.

Freshwater bodies such as lakes are common across most biomes, forming unique and sometimes isolated ecosystems (Figure 1.3.12). In natural sciences, the hysteretic behaviour of lakes (Scheffer et al., 2007) has informed the concept of tipping points at the ecosystem level, leading to the development of the alternative stable states theory in shallow lakes (Scheffer et al., 1993; Carpenter et al., 1999; Carpenter 2005). They represent archetypal case studies for how tipping points relate to theories of ecological stability and resilience that can underpin preventative management approaches (Andersen et al., 2009; Spears et al., 2017). Despite this, significant uncertainty remains on the geographical extent of tipping points in lakes and the wider relevance for the Earth’s climate system. 

Lakes are also good examples of social-ecological systems, with their ecological dynamics closely intertwined with the socio-economic dynamics of surrounding populations who often depend on them for key ecosystem services and adaptively respond to changes in lake condition (Martin et al., 2020). Given the global vulnerability of freshwaters and the pervasive nature of major pressures acting upon them (e.g. nutrient pollution and climate change), tipping points in these systems could have significant societal impacts, including on human and environmental health, food production and climate regulation. The capacity to detect discontinuous ecosystem responses to pressure changes in natural systems has been challenged (e.g. Hillebrand et al., 2020). Nevertheless, there are several studies that have reported real tipping points, i.e. shifts from one stable state to another in small shallow lakes (the most common lake type globally, (Messager et al., 2016).

Empirical analyses, process modelling and experimental studies are advanced for shallow lakes providing a good understanding of ecosystem behaviours around tipping points, typically starting with positive/amplifying feedback loops, then entering a runaway phase before finally the tipping point brings the system into a different stable state (Nes et al., 2016). For example, the well documented increase of phosphorus (P) loading across European lakes in the last century (e.g. from agricultural and waste water pollution) has uncovered critical loading thresholds beyond which lakes can shift rapidly from a clear water, macrophyte rich state to a turbid, phytoplankton rich state (Scheffer et al., 2001; Jeppesen et al., 2005; Tátrai et al., 2009), and vice versa when nutrient loading decreases. 

Adding to such well-described and mechanistically well-understood changes, there is a wide range of local or single lake shifts that may be categorised as tipping points. The question remains as to whether tipping points are merely isolated phenomena in single lakes, or specific types of lakes, or whether they manifest, or will in the future, across geographically distinct populations of lakes experiencing similar environmental change, with the potential for regional or global extent (Figures 1.3.12 and 1.3.13).

It is well established that lakes are sensitive to the effects of climate change, including warming and changes in precipitation and storminess (Meerhoff et al., 2022). Emerging evidence suggests that they may also play an important role in climate regulation, through both the emission of greenhouse gases (predominantly methane – Downing et al., 2021) and carbon burial (Anderson et al., 2020). It is therefore relevant to consider the extent to which potential tipping points may drive, or be driven by, climate change, leading to higher-level feedbacks to the Earth’s climate system. In this context we will constrain the discussion to potential tipping points that are more generic, at least with some regional or biome-wise impact, and that could feedback to the climate, while not necessarily being driven or triggered by climate change per se.

Here, we adhere to tipping points as defined in this report (and matching Nes et al., 2016). Based on this we discuss candidate tipping points in freshwaters (Table 1.3.2), focusing on lakes and ponds, with the potential for global or at least regional or biome-scale relevance.

Evidence for tipping dynamics

Eutrophication-driven anoxia and internal P-loading

The mobilisation of P from sediments, a process known as internal loading (Sondergaard et al., 2001), is well described and plays a key role in hysteresis in preventing lakes recovering from human-driven eutrophication (Boström et al., 1982; Jeppesen et al., 1991; Spears and Steinman 2020). The process may be enhanced by lake warming, and there are feedbacks to climate since water anoxia and internal P-loading (which features the actual tipping point) could offset CO₂-fixation by increased release of GHGs. Consequent changes in biota also strengthen hysteresis (Brabrand et al., 1990), not least when cyanobacterial blooms develop. The phenomenon is local but widespread, and likely to increase as a result of global warming (Meerhoff et al., 2022). Increases in precipitation, and high-intensity rainfall events, are also expected to significantly increase runoff of P from agricultural catchments to surface freshwaters (Ockenden et al., 2017), further promoting eutrophication and its manifestations. Warming increases stratification and thermal stability promoting anoxia (Maberly et al., 2020; Woolway et al., 2020), internal fertilisation and increased GHG emissions. In addition to anoxia, there are other feedback mechanisms for lake eutrophication tipping points, such as the macrophyte-nutrient-algae-turbidity and macrophyte-zooplankton/fish-algae-turbidity loops (Wang et al., 2022). Shifts in trophic cascades, i.e. a top-down control of zooplankton and reduced grazing on phytoplankton, could also help drive eutrophication (Carpenter et al., 1985; Carpenter and Kitchell 1988). However, feedback to the climate is primarily related to anoxia.

Increased loading of DOM and anoxia

Increased export of terrestrially derived dissolved organic matter (DOM) to lakes and rivers in boreal regions (“browning”) is a widespread phenomenon partly linked to reduced acidification, but also driven by land use changes (notably afforestation) and climate change (CO₂-fertilisation of forests, warming and hydrology) (de Wit et al., 2016; Creed et al., 2018; Monteith et al., 2023). Wide-scale regime shifts in boreal lakes caused by increased loadings of DOM can promote a prolonged and more intensified stratification period (implications summarised above, described for DOM by Spears et al., 2017), amplified by warming. Increased terrestrial DOM loadings intensify net heterotrophy in the systems (i.e. through increased light attenuation and increased access to organic carbon) (Karlsson et al., 2009; Thrane et al., 2014; Horppila et al., 2023). While at present the thresholds around these effects have not been well constrained, the impacts may be significant at the global scale for GHG emissions (Tranvik et al., 2009) and regionally for coastal productivity (Opdal et al., 2019)

Both eutrophication and browning are to some extent driven by climate change, and warming of lakes will promote the effects by increasing thermal stratification, promoting anoxia which again promotes internal loadings of phosphorus, leading in some cases to self-sustaining change (i.e. tipping). Increased release of GHGs will serve as another feedback to the climate (Fig. 1.3.14).

Disappearance/appearance of waterbodies

A global reduction in lake water storage (Yao et al., 2023) and climate-related creation or, more frequently, disappearance, of water bodies is a large-scale concern (Woolway et al., 2022). For example, current and future permafrost thaw and glacier melting can both create new and drain old waterbodies, providing a strong link to the fate of the cryosphere (Smith et al., 2005; Olefeldt et al., 2021). Such small but numerous waterbodies over vast areas in the high Arctic may also serve as major conduits of greenhouse gases and historical soil carbon stocks to the atmosphere (Laurion et al., 2010) and play an important role in mediating nutrient delivery to the polar oceans (Emmerton et al., 2008), potentially affecting global productivity (Terhaar et al., 2021). 

Despite the scale considered here, the extent of open water globally is relatively easy to quantify using remote sensing, and loss of waterbodies can be predicted from water balance and thresholds for permafrost thaw with high confidence. However, while representing a binary shift between two states, driven by climate, this should not be classified as tipping events as in most cases no self-sustaining feedback is involved. Lake appearance or disappearance can be driven by cryosphere tipping points though – for example, thermokarst lake formation or abrupt drainage due to permafrost thaw (Turetsky et al., 2020; Teufel and Sushama, 2019) (see Chapter 1.2) – and in such cases the lake forms part of a coupled thermokarst system capable of tipping.

Switch from N to P-limitation

Regions receiving increased nitrogen (N) deposition may shift from prevailing P- to N-limitation (Elser et al., 2009). Conversely, increased N-loss by denitrification, eventually associated with increased internal P-loading, may shift systems from P to N-limitation (Weyhenmeyer et al., 2007). Changes in N- versus P-limitation of productivity are associated with changes in community structure, both for the phytoplankton and macrophyte communities, which could involve ecological tipping points. However, while the switch between N and P-limitation represents a binary switch with ecological consequences, it is not itself classified as a tipping point according to our criteria, as self-sustaining feedbacks have not been identified. There is currently weak evidence for this shift’s impact on climate feedbacks.

Salinisation

Salinisation is a prevalent threat to freshwater rivers, lakes and wetlands and is caused by a range of anthropogenic actions including water extraction, pollution and climate change (Herbert et al., 2015). It has severe consequences for aquatic communities (Short et al., 2016; Cunillera-Montcusí et al., 2022) with salinity thresholds likely strongly impacted by other stressors – including eutrophication (Kaijser et al., 2019). Salinisation has a strong societal impact, particularly related to domestic and agricultural water supply in arid and semi-arid regions (Williams et al., 1999). Salinisation tends to decrease CH₄ emissions (Herbert et al., 2015) and, in that sense, is a negative/damping feedback with respect to climate change. Salinisation may induce ecological regime shifts, for example leading to microbial mat dominance (Sim et al., 2006), and results in some hysteresis, with salinised sediments remaining salty also after the system is flushed with fresh water (Van Dijk et al., 2019), but is not in itself driven by self-sustaining feedbacks.

Spread of invasive species

Freshwaters are especially vulnerable to species loss and population declines as well as species invasions due to their isolation. Substantial ecosystem changes by reinforcing interactions between invasive species and alternative stable states (i.e. macrophyte – aquatic plant – versus phytoplankton dominance, as described above) may occur (Reynolds and Aldridge, 2021). The spread of several invasive species can be facilitated by climate change (Rahel and Olden, 2008) and may have some self-sustaining properties. Such changes could thus drive a regime shift for a given system, but in most cases are hypothetically reversible if the original driver (the invasive species) were removed. Species invasion is hard to predict and difficult to quantify, despite the risk of species ingress as ranges expand with climate change.

Assessment and knowledge gaps

Table 1.3.2: Candidate tipping events from the literature with potential to occur at local to regional scales, their association with climate change, and whether tipping points and hysteresis have been associated with them. Brackets indicate higher uncertainty. Bold entries represent categories that qualify as tipping points in this context, while the others are either simply binary shifts between states, threshold effects, or similar.

Type of event	Local	Regional	Climate driver	Climate feedback	Tipping event	Hysteresis
Eutrophication-driven anoxia and internal P-loading	x		x	x	x	x
Increased loadings of DOM		x	x	x	x	(x)
Disappearance/appearance of waterbodies		x	x	x	(x) (linked to cryosphere tipping)	(x)
Switch between N and P limitation		x	x	(x)
Salinisation		x	x	x		(x)
Spread of invasive species	x	(x)	(x)			(x)

Abrupt changes driven by warming, eutrophication or increased loadings of organic matter, leading to changes in the production to respiration ratio (i.e. systems shifting from net autotrophic to net heterotrophic), and/or onset of bottom-water anoxia have clear tipping dynamics (high confidence) and strong feedback to the climate via GHG emissions (Meerhoff et al., 2022) (Table 1.3.2). Whether the widespread effect of increased loading of organic matter (browning) in boreal lakes can drive tipping points is more of a knowledge gap, yet the feedback of lake browning to climate through increased GHG emissions is evident. Loss of waterbodies residing on permafrost or suffering negative water balance and eventually complete disappearance represents a binary shift, which has major ecological consequences (Woolway et al., 2022) but is not considered a tipping event sensu stricto. The same holds for other types of binary shifts, threshold effects or local changes. The role of warming as a catalyst on the changes driven by eutrophication and browning is a critical knowledge gap. Quantification of GHG release from lakes represents major feedbacks to climate, and to quantify the impact of eutrophication, browning and warming in this context should have high priority.