Proximity to Tipping Points

Proximity to tipping points

COP21 successfully negotiated a target to limit temperature rise to 1.5°C of warming in order to prevent 'dangerous climate change' limit of 2°C. Lenton and Schellnhuber (2007) attempted to relate IPCC climate change scenarios to multiple tipping elements. The figure shows policy relevant tipping elements which are elements which may exhibit tipping behaviour in the future for a 1.1-6.4°C warming. It is interesting to learn that the three major tipping points covered (Arctic, Amazon and THC) can possibly be triggered this century under three different warming trajectory (1-2°C, 3°C and >4°C). It is evident that even the widely used 2°C limit may not prevent the tipping of some elements and that societal tipping points (this post) - drastic technological and societal shift to a carbon-free economy is needed.

Source

While the number of 2°C is widely debated and deserves its own blog, it is essential that one looks beyond this number as the 2°C number is more of a political construct than a binary threshold/tipping point as defined in this blog. Obsessing over a symbolic 'end goal' for climate change not only detracts from public understanding but also fails to promote mitigation and fails to understand the reality by which anthropogenic activities are edging components of the earth system towards tipping behaviour (Knutti et.al. 2015). Although the few tipping points featured in detail in this blog mostly derived from increasing GHG emissions, they are not the only forcing to induce tipping behaviour (eg. aerosol emissions forcing re-organization of the Indian Monsoon and landscape fragmentation for regional terrestrial tipping points). Furthermore, it must be recognized that the tipping of one element may promote/induce rapid tipping behaviour of a different but positively connected element. Additionally, due to the inherent inertia and the non-linearity of different climate and biosphere systems as explored in this blog, it is probable that some tipping points may already have been surpassed but its impacts not yet fully realised. 

This feels like a good place to end my blogging journey. In the next post, I will reflect on major environmental news in 2016 and provide some reflections on the lessons learnt in the past 3 months.

Tipping Point III - Amazon Dieback

Welcome back! In this post, I will look at a probable ecological (terrestrial biosphere) tipping point with potentially subcontinental to global impacts. The most notable tipping element under continual anthropogenic greenhouse gas induced warming would be a significant and relatively rapid dieback of the Amazonian rainforest. There is no better case study than a possible Amazon dieback to illustrate the widespread impacts of humans to the biosphere and the global cascading effects a collapse of local ecosystem may be capable of inducing.

Characterizing planetary tipping points within the biosphere, Barnosky et.al. 2012 argued that based on the characteristics of historical planetary tipping points, present day anthropogenic forcings are capable of causing an ever-increasing number of local-scale tipping points which will trigger transitions across a critical threshold over a larger area than the originally affected region (such as an Amazonian dieback) and eventually contribute to a global scale regime shift. 

Amazonian Dieback

The general health and ecosystem functioning (provision and delivery of ecosystem services) of tropical rainforests  are increasingly affected and dominated by anthropogenic activities and anthropogenically induced climate change (Lewis et.al. 2015). The Amazonian rainforest is the richest and most important region on Earth in terms of biodiversity and biogeochemical flows. It is currently increasingly subjected to human activities which has fundamentally altered the Amazonian landscape. It is with this alarming notion in mind when scientists proposed the possibility of a transition to a less biodiverse 'savannah' state if continually subjected to landscape stress and anthropogenic induced climate change (Blaustein 2011). When a certain threshold of biomass loss is passed, a much larger area of the Amazonian rainforest would be affected and would subsequently irreversibly transform into an impoverished state. Impacts of this may be subcontinental - the loss of the world's most biodiverse rainforest, or global - through the release of stored carbon (90-120 billion metric tonnes - 50% of tropical forest carbon!), thus capable of regulating global climate with potentially catastrophic effects on other tipping points (ice sheets stability, precipitation and evapotranspiration). 

Trajectories of change in rainfall regime using multi-model GCM approach.
Greyscale background indicates relationship between precipitation, water deficit and vegetation state

The Amazon had experienced previous episodes of extreme droughts in the past, most notably being the 2005 drought (the worst drought on record) and more recently, a dry episode in 2010. These droughts are increasingly seen as analogues for future events to assess potential response of the rainforest to extreme weather (caused by anthropogenic climate change) coupled with increasing landscape stress. Using multiple GCMs, Malhi et.al. 2009 modelled changes in the rainfall regime and concluded that a significant dry-season water stress is prevalent across the 21st century. Increased length of dry season and increased annual water demand, critical threshold needed to sustain seasonal forest instead of savannas are the two main causes of a transition to savannah states. Recent research also suggested a graded, spatially heterogenous transition from accumulation of water stress conditions of individual plants (Levine et.al. 2016). Interactions between deforestation and removal of forest cover, modification of local climate and presence of fire may also contribute significantly to a possible abrupt shift in vegetative state. 

Biosphere Tipping Points

In this post, I aim to distinguish between arguments and propositions for global planetary scale ecological tipping points and local, sub-continental/regional tipping points and whether one causes the other.

Local vs Global

In true geography fashion, no issue is resolved without discussing scale. The majority of recent scientific studies have pointed to the presence of global, planetary scale climatic state shifts with catastrophic consequences, as epitomised by the planetary boundaries concept. However, questions have been raised on whether genuinely global tipping points are scientifically probable and whether non-climatic elements (eg. terrestrial and ecological systems) can have global scale tipping points. Listed below are main characteristics for planetary  tipping points:

1) The magnitude, extent and rates of present global forcings (human population growth, energy consumption, climate change) initiated by anthropogenic activities have surpassed global forcings which caused past global state shifts (Barnosky et.al.2012)

2) Planetary scale tipping points originate from the accumulation of local system behaviour where local scale forcings propagate through scales to cause global state shifts (Steffen et.al. 2011)

3) Planetary tipping points may not contain early warning signs or trajectory may be smoothed out despite impending critical thresholds (Scheffer et.al. 2009)

4) Internal evolutionary events causing changes on a global scale (Lenton and Williams 2013)

5) Global tipping points may not be synchronous or sudden. Internal inertia may cause incremental and gradual change relative to human timescales. Speed and abruptness should not be criterion for global tipping points (Hughes et.al. 2013).

While it is established that local ecological and biological systems have had tipping points, there have been substantial arguments against the presence of planetary scale thresholds:

1) Unlikely due to high spatial heterogeneity and low connectivity within regions of the biosphere. Global forcings unlikely to cause synchronous tipping due to spatial heterogeneity of and differing environmental impacts between local regions (Brook et.al. 2013)

• Eg. landscape fragmentation from the building of roads in which current roadless areas are fragmented into >600,000 patches at the expense of terrestrial biodiversity (Ibish et.al. 2016)

2) Most proposed planetary tipping points does not have genuinely global biophysical boundaries Global limits may limit local/regional action (Blomqvist et.al. 2012) 

3) Dichotomy thinking of 'safe' and catastrophic in planetary tipping points may encourage inaction and may distract from fundamental local regime shifts and biological change (Brook et.al. 2013)

In the next post, I will look at the possibility of an Amazon dieback. After that, I will conclude the blog by looking at present-day anthropogenic forcing.

Learning from the past...

The presence of tipping points are not new and examples of crossing critical thresholds are plentiful in Earth's long history. It is highly conceivable that comparable high impact regime shifts are increasingly probable in the future due to human forcing. In this week's post, I will look back in Earth's long history and discuss some of the tipping points and abrupt climate and biosphere changes.

Ancient tipping points

Abrupt Climate Change

Widespread, global in scale and abrupt climate change occurred repeatedly in the past when tipping point dynamics were evident in the Earth system through geological and paleoclimatic records.

Glaciation

Glacial-interglacial transitions between multiple stable states of the climate system are an example of past abrupt climate shifts with global impacts. Triggers for these abrupt climatic oscillations were attributed mainly to orbital forcing (Milankovitch cycles; the eccentricity, obliquity and precession of the Earth's orbit). Warming prior to transition into an interglacial period is often abrupt and exceeds magnitude of solar radiation variations. This is due to positive feedback loops of ice-albedo relationships and atmospheric CO2 when stored CO2 are released and reinforces warming. Similarly, abrupt shifts within glacial periods are also evident. D-O and Heinrich events characterizes these shifts and represents a critical transition between warm-cold states. Dansgaard-Oescheger (D-O) events are rapid warming episodes in the last glacial period (happened 25 times!) while Heinrich events are intensely cold periods in between D-O cycles (Ahn and Brook 2008).

Younger Dryas

Possibly the most well-known example of abrupt climate change, the Younger Dryas event 14,500 years ago saw global climate abruptly shift to near-glacial conditions in a period when the world is gradually shifting to an interglacial warm state. It is largely recognized that increased freshwater discharge in the North Atlantic from Lake Aggaiz and subsequent impacts on the thermohaline circulation was a major cause of the abrupt shift. However, the forcing causing additional freshwater input is still debated (Carlson 2010) as the outlet of Lake Aggaiz may have remained closed until after the onset of the Younger Dryas with other sources such as meltwater or bollide impacts are notable contenders.

Desertification of N.Africa

Widespread areas of North Africa and the current Sahara Desert were 'green' and were covered with vegetation and lakes until abrupt desertification marking the end of the green and wet phase around 4.5 kyr ago. Green Sahara was mainly attributed to changes in the Earth's orbit and tilt resulting in the solar irradiation and intensification of the African summer monsoon. However, positive preciptiation-vegetation feedback loops also amplifies monsoon, causes radiative cooling and suppresses convective precipitation, encouraging the spread of Sahelian vegetation and desert conditions (Claussen 2008). It is debatable whether the desertification occurred abruptly with singular aridification events or on gradual timescales from accumulation of local changes across inhomogeneous land surface (Bathiany et.al. 2016). 

Source

Ecological Change

Tipping points in ecological systems are less clear as it is often hard and controversial to discern between ecological change as a consequence of critical threshold or as part of the tipping mechanism (Lenton and Williams 2013). There are also widespread debate over whether tipping points in the biosphere can be truly planetary in scale. Nonetheless, there has been well documented abrupt changes in the terrestrial biosphere.  

Mass Extinctions

There has been 5 mass extinctions in Earth's history with widespread extinctions occurring abruptly over short timescales. Major biotic changes can be self-reinforcing through trophic cascades after initial forcings from climate change or volcanism. There has been repeated episodes of loss in >75% of species on Earth and fundamental reorgnizations of biota (Barnosky et.al. 2012). Having said that, tipping points in ecological systems need not require climate feedbacks to occur and can occur through intrinsic thresholds and internal feedbacks (eg. changes to predator-prey dynamics) and self-propagation (Seekell 2016).

Are we causing the 6th mass extinction?

Cambrian Explosion

Global scale re-organization of biota ~540 million years ago with the explosion in and dominance of complex multicellular organisms over single celled microbes. This was attributed to crossing a critical oxygen level threshold (de-oxygenation) which encouraged genetic complexity and sustained metabolic processes.

Evolutionary tipping points

Another possible category of ancient tipping points lies with evolution and is based on the fact that evolutionary adaptations and changes can result in abrupt changes global in scale. Evolution of new traits, speciation and adaptive radiation can all contribute to internal dynamics which tips ecosystem to an alternative state (Williams and Lenton 2010). Examples of this may include evolution of traits in oxygenic photosynthesis which may have caused the Great Oxidation event (~2.3 billion years ago) which caused mass extinction and a transition to 'snowball Earth' glaciation period (Lenton and Williams 2013).

This also raises the question on whether the Anthropocene can be characterized as an evolutionary tipping point on a global scale. Global forcings in the present day include intense population growth, resource consumption, land fragmentation and climate change. All of which are already causing observable response from biota in the biosphere and the respective magnitude of change far exceeds forcings seen in past abrupt ecological shifts (Barnosky et.al. 2012).

It is very interesting to see strikingly similar parallels in system responses (eg. AMOC freshening, Arctic ice melting, marine dead zones etc.) in play in the present day. Past abrupt shifts provide an analogy for causes of shifts and possible system response in the future. 

Hot off the press...

A quick post to highlight a new report published by the Arctic Council highlight resilience and potential tipping points in the Arctic. This report is particularly relevant to this blog and the past few posts on tipping points in and around the Arctic region. 

CLICK HERE TO WATCH A VIDEO FROM THE GUARDIAN

The authors identified 19 explicit tipping points which consists of both climatic and non-climatic (socio-ecological) elements with local to global drivers and impacts. The tipping points identified fully embraced the very definition of tipping points I intended to define in the first post. Interestingly, this report also uses an ecosystem services approach when classifying and analyzing various impacts of crossing certain tipping points. Popularized in the Millennium Ecosystem Assessment 2005, the ecosystem services framework aims to link anthropogenic change to biophysical and economic values of ecosystem functions. [Self promotion: I have been blogging about ecosystem services in another blog for my other university module (check it out here!).]

Source

This report efficiently sums up the wide variety (climatic and non-climatic, local or global) of impending tipping points with a multitude of different drivers (socio-ecological, climate change) with a wide range of impacts (local to global, ecosystem services).  An example of such impacts would be a loss of Arctic summer sea ice after surpassing a critical threshold and a subsequent loss of cultural ecosystem services of subsistence hunting and transportation of indigenous Alaskans. Although I focused on climatic tipping points until now, I will also be blogging about ecological and societal tipping points in future posts. 

Detecting Critical Thresholds

I have yet to touch upon how scientists identify or detect tipping points and how they determine where a critical threshold is located in time. In this post, I will outline the major characteristics of a system approaching critical thresholds and how scientists determine them.

General properties

Three general properties characterizes the point at which a critical threshold is surpassed:

1. Rate of change increase sharply then what prevailed over previous stable periods
2. System state exceeds range of historical variations 
3. Rate of change increased at a pace which exceeds the abilities of nations to respond

The timing, type of transition and magnitude of change depends on the nature of interactions and heterogeneity within the system. Recent increases in attempts to predict the type and temporal onset of tipping points can mainly be characterized in two ways (Thompson and Seiber 2010):

1. Models

Climate models have been widely used to predict the presence, timing and magnitude/extent of changes in identified tipping elements. However, the nature of climate models means that the magnitude of threshold effects and when feedback induced thresholds are reached are highly uncertain and varies among model specifications and parameterization (Maslin and Austin 2012). Models also have varying interpretations of mechanisms behind the earth system and varying sophistication in processes represented. Climate models are often used to predict future changes (eg. Arctic summer ice loss seen in Holland et.al. 2006) and validated by its ability to adequately simulate past abrupt changes (eg. 'dangerous climate change' prediction model in Hansen et.al. 2006).

2. Time Series - statistical analysis from observational climatological or ecological time series data to identify statistical characteristics that precedes tipping points/bifurcation

A prime example of this is early warning systems for natural hazards risk management. As Earth have underwent a long history of abrupt climatic changes, scientists often make use of natural archives (eg. ice cores; diatoms; pollen etc.) to infer past climatic change and fluctuations (Thomas 2016). Paleo records of Earth surpassing ancient critical thresholds provides scientists with an observational records from which early warning signals can be identified. Scientists using time series statistical analysis to identify early warning signals showed universal signals across multiple ancient abrupt climate shifts (icehouse -> greenhouse; Younger Dryas; N.Africa climatic shift) (Davos et.al. 2008).

Early warning signals

Tipping points could affect decision making if adequate knowledge on timing, occurrence, and impacts were available. There is therefore a whole field of research aimed at finding preceding signals and characteristics of tipping points from historical abrupt transitions. Early warning identification can take the path of qualitative assessment or quantitative prediction of timing of impending thresholds (Lenton 2011). Listed below are early warning signals identified in past abrupt tipping points (Scheffer et.al. (2012). These are true for a range of complex systems, ranging from climatic systems to financial and social systems.

1. Critical slowing down - Decreased rates of change as system approaches critical thresholds; increasingly sluggish system response; increase in amplitude and reduction in fluctuations

2.  Skewness and kurtosis - Asymmetric fluctuations; presence of extreme values measured through high skewness and kurtosis as system approaches bifurcation

3.  Increased autocorrelation - Decreased rates of change lead to state of system being more and more like its past state prior to critical threshold, thus increase in correlation

4. Spatial patterns - often ecological; signals derived from spatial/temporal persistence and presence of species 

Thank you for reading! In the next post, I will continue this discussion by looking at identified tipping points in Earth's history.

Tipping Point II - Circulation Change

This week, I will be outlining research and modelling exercises on the potential shutdown of the Atlantic Meridional Overturning Circulation (AMOC), also known as the Thermohaline circulation. 

Thermohaline Circulation (THC)

The thermohaline circulation, popularly called the global ocean conveyer belt, is an integral feature of the present day climate system. It sustains the current climate and is a major contributor to the global heat budget. As illustrated in Fig 1,  global oceanic circulation is driven by density gradients related to the formation of deep water. The density of seawater is governed by temperature and salinity. The THC process is briefly outlined below (Broecker 1997):

1. Warmer, saltier water brought into NE Atlantic, warms the European continent
2. Warm water cools and mixes with cold Arctic water, becomes dense and sinks, forming North Atlantic Deep Water 
3. Further sinking of dense water occurs near Antarctica with cool water sinking from the effect of the circumpolar currents, forming Antarctic Bottom Water 
4. Cold dense water returns to the surface throughout the world's oceans and forms a closed loop of exchanges between warm, surface water and cool, dense deep water

Modes of THC

Driven by density differences, the THC is particularly sensitive to the freshwater budget which will disrupt overturning of deepwater by reducing salinity. Looking in past abrupt changes to the THC, scientists have identified three possible modes (stable states) with fundamentally different climates (Rahmstorf 2000):

1. Warm - interglacial (current) mode where deep water forms in Nordic Seas 
2. Cold - glacial mode where deep water forms south near Greenland, Iceland and Scotland
3. Off - shutdown of THC, no formation of deep water in North Atlantic

Transitions between modes would cause abrupt climate changes. Previous transitions between modes have occurred in the form of Dansgaard-Oeschger cycles and Henrich events where large scale breakdown of N Atlantic icebergs dramatically increases freshwater input (Alley 2000). 

Tipping Point

As the thermohaline circulation is driven by density differences which are particularly sensitive to temperature and salinity, both sufficient heat from continued increase in GHG concentrations and alteration of the freshwater budget from melting ice can lead to fundamental re-organization of ocean circulation and transition to a alternative state (Clark 2002). Classified as being 'low probability with high impacts' by the IPCC, a critical threshold may be observed with hysteresis characteristics from non-linear behaviour. Modelling studies using coupled atmosphere-ocean General Circulation Models have suggested that the THC is particularly sensitive to freshwater infiltration on the order of 0.1 Sv with a transition across critical threshold within range between 0-0.15 Sv (Clark 2002; Rahmstorf 2000).

Modelling responses of the THC to rising CO2 concentrations with warming and melting ice effects, Wood et.al. 1999 proposed that a shutdown of convection in the Labrador Sea would result in a 20-25% reduction in deep water formation by the time CO2 quadruples pre-industrial levels. Salinity in the Nordic seas declined since the 1960s, suggesting a possibility of critical threshold of freshening in this century (Curry and Mauritzen 2005). A 20th century slowing down of the AMOC was witnessed by a region of cooling in Northern Atlantic after 1970 due to increased freshwater input with further uncertainty from melting of the Greenland ice sheet (Rahmstorf et.al. 2015). 

Impacts

The impacts of a THC tipping point has been widely studied with models and experiments. A shutdown of AMOC would lead to cooling effects which may outweigh and reverse current CO2 induced temperature trends whereas a AMOC weakening are dominated by increased CO2 domination (Yin et.al. 2006). Using the Met Office HadCM3 GCM in a modelling experiment, an artificial collapse of the AMOC in 2049 would cause reduction in precipitation in Western Europe and a regional cooling of the Northern Hemisphere by -1.7 degrees with up to -9 degrees cooling locally. Global primary production from vegetation will also decrease by 5% due to temperature and moisture changes. Drying trends will also be noticed in Central America and SE Asia with impacts extending globally within 30 years (Vellinga and Wood 2003; Vellinga and Wood 2008). 

Causing the shutdown of the THC would no doubt constitute as 'dangerous level of interference' (Hansen et.al. 2006), possibly characterising the Anthropocene epoch. The uncertainty in coupled GCMs on the full range of feedback responses and the 'low probability' claim of the IPCC is certainty not an excuse for inaction and a shutdown in the current century should not be ruled out. This also highlights the need for better oceanic monitoring equipment and research to ensure development of early warning systems and proper anticipation of global impacts.

Thank you for reading this post! In the next post, we will look at the possibilities of detecting critical thresholds.