Climate change: uncertainty is inevitable but risk is certain

There is a strong consensus amongst experts on climate that the Earth is warming at a rate that can only be explained by rapid increases in the concentration of greenhouse gases in the atmosphere. These result largely from the burning of fossil fuels and destruction of forests. The Earth's climate in the past has changed due to changes in the Earth's orbit around the Sun and for short periods by major volcanic eruptions, but the present rate of change cannot be explained without taking account of increases in greenhouse gases.

This global warming is fastest in high northern latitudes due to large land areas which warm faster than the oceans, and the melting of ice and snow cover which allows more sunshine to be absorbed rather than reflected back into space. High latitudes are also warming faster than the tropics because warming increases evaporation, especially in the tropical oceans. This takes latent heat of evaporation into the air, where the winds transport much of it to higher latitudes where the water vapour condenses, releasing the latent heat and increasing rainfall at higher latitudes.

These processes are complicated so it needs suitably qualified scientists to research and understand what is going on. Uncertainties remain about the rates of change and the local and regional consequences but it is certain that the water cycle of evaporation and condensation is accelerating. This causes more severe and more frequent extreme events such as droughts and flooding rains, modified locally by regional weather patterns which also change with global warming.

The oceans are warming slower than the continents due to their large heat capacity, but the warming causes the water to expand, making the sea level rise. Added to that is more water from melting mountain glaciers and increasingly from more rapid outflow of ice from the Greenland and Antarctic ice sheets. These effects have been measured from ground-based and satellite observations, so there is no doubt they are happening. What is uncertain is how rapidly these processes may accelerate in future.

Observations show that average global warming since the industrial revolution is already nearing 1ºC, and global average sea level has risen in the same time by some 20 cm. Best estimates suggest that by 2100 global average temperature could increase by about 2 to maybe 4 or 5ºC, and global average sea level may rise anywhere from about 60 cm to well over a metre.

All this implies that so-called natural disasters such as severe storms, floods, droughts, wildfires, and coastal flooding will increase in frequency and intensity. Take the 2012 storm Sandy effects on the highly populated New Jersey/New York coastal area. It was made worse by the roughly 20 cm sea level rise in the last century, but an additional 1 meter rise in the 21^st century would make it far far worse. The same is true, but with much more drastic consequences, in poorer countries such as the Philippines.

This raises the question of what we should do about global warming. The science of climate change is thus policy-relevant and scientists have a duty to explain its relevance by referring to possible courses of action that might make the best of the situation. This necessarily involves managing risk. The future impact of climate change is not certain, but those of us who understand the problem are duty bound to discuss the possibility of reducing global warming by reducing greenhouse gas emissions, and of adapting to any changes we cannot avoid.

This essay will focus on the latter, by discussing the question as to how can we best adapt to the effects of global warming. This is a question of risk management, given that, while we know there will be increased impacts, especially of extreme climatic events, their exact magnitude and location remain uncertain.

This is best expressed in terms of probability. A 50% probability means there is one chance in two of a given event happening, whereas a 95% probability means that the chance of it happening is 19 out of 20. Traditionally scientists have regarded something as "proven", or at least well-established, if the probability of it being true is at least 95% and preferably better than 99% (i.e., 1 in 100). However, if the consequences of some given event happening is serious or disastrous a lower probability of it occurring should also be taken seriously despite the uncertainty.

For example, we generally pay our annual insurance premium on our house against fire, not because we are certain that it will burn down this coming year, but because it just might. Similarly engineers when they design culverts, bridges or large dams have to design these to withstand floods of various magnitudes. For a culvert, where the consequences of it overflowing would be minor, they may design it to cope with a 1 in 10 year flood, but a bridge to withstand a 1 in 100 or 1000 year flood, and a large dam to withstand a 1 in 1,000 or 10,000 year flood. This is because in the latter cases the consequences of failure would be much greater.

Thus risk management must take account of both probability and consequences. In the case of climate change the consequences of increases in the frequency and intensity of extreme events may be very serious, causing large damage bills, disrupting society and costing many lives. Examples include the disastrous floods in Pakistan a few years ago, the huge storm disasters caused by tropical cyclone Katrina in New Orleans in 2005 and storm Sandy that hit New York and New Jersey in 2012, and Typhoon Haiyan in the Philippines this year . Australia has had its share of disastrous floods and droughts in recent years, as well as disastrous bushfires. These have cost hundreds of lives and billions of dollars.

Both investors and governments need to take the risk of such events into account in planning, especially for new developments in risky locations. Governments have to work out how best to adapt existing settlements and infrastructure from increasing risks and investors to properly anticipate risks.

Business and investors refer to a quasi-legal concept of "sovereign risk" by which they mean the risk of additional costs due to a government changing the rules, as if it were the government's fault. However, if the risk is reasonably well known due to scientific evidence and warnings, then governments have a duty to take account of that risk. It is not the government's fault that there is a risk which requires action – that is due to the laws of Nature, and it is the laws of Nature that scientists try to understand. If people add greenhouse gases to the atmosphere and it has consequences, the consequences are not due to government action or of scientists' research results, but due to the laws of Nature.

Given warnings and risk, if the government fails to take preventive action, who is responsible for future losses? Is it the individual, the local council, the state or federal government, the developer or investor, the scientist or engineer, or the insurance company (and by implication others who pay insurance premiums)? Who will lose out, and who will be sued in decades to come?

There is a communication problem here. Research on climate change is becoming increasingly policy-relevant, but there are often large ranges of uncertainty. Scientists often apply the traditional standards of proof for "proving" something to be true (i.e., the 95% or 99% probability) rather than be "alarmist" by highlighting the risk-relevant probability. Thus scientists often offer a central or "best estimate", rather than the policy-relevant risk probability. They emphasise "scientific rigour" and stress caveats, rather than stressing the small but non-negligible likelihood of really dangerous outcomes that must be avoided. Thus an outcome with only a 10% probability of occurring should not be labelled as "unlikely" (and thus often dismissed), but rather as "possible" (and thus worthy of examination).

So-called "sceptics", who want no action on climate change, focus on the low impact end of the uncertainty range, "pure" scientists focus on the middle "best guess" result, while applied scientists such as engineers and managers focus on the high impact end of the range of possibilities as these are the ones to be avoided or coped with. It is not only engineers that focus on the high consequence risks – so too do military planners and insurance companies, because it is their business to do so.

So what are the risks from climate change? A short list would include:

• temperature rises → heat waves, crop failures
• heavy rains → floods, damages, loss of life, erosion, siltation
• droughts → crop losses, water shortages, soil erosion
• sea-level rise → coastal erosion, estuarine flooding, salinisation of coastal aquifers, damage to coastal buildings and infrastructure
• wild fires → loss of crops, wildlife, forests; loss of life and property; increased soil erosion and runoff
• food shortages, displaced people, "climate refugees"
• failed investments.

And how do we best adapt to climate change, given that we will not stop it entirely in the foreseeable future?

Prof. Gilbert White, in his 1945 Ph. D. thesis, provided a rigorous examination of how to live on a flood plain. He identified eight forms of adjustment, and stated that these need to be pursued in appropriate combinations:

1. Abating floods by land treatments: e.g., decrease upstream erosion and runoff by tree plantings and fire control.
2. Protection via dams, diversions and levee banks: these are 'obvious' but costly in dollar terms and to the environment, increase exposure to larger extremes, and involve management conflicts (e.g., flood control requires keeping dams as empty as possible, but water supply requires keeping them as full as possible).
3. Elevating land or buildings: a permanent solution (maybe), but costly.
4. Emergency warnings and evacuation: good if timely and efficient but mainly saves lives rather than investments.
5. Structural changes to buildings and transport: reduces losses, safeguards services, but expensive.
6. Change land use planning: avoids loss of vulnerable investments.
7. Distributing relief: essential if other adjustments inadequate, but can be repetitive.
8. Compulsory un-subsidised insurance: this identifies risks and dissuades unwise investments.

Following Gilbert's example, but with much less expertise, here are my suggestions for forms of adjustment to sea level rise:

1. Estimate effects: of sea-level rise, changing wave climates and storm surges, and their effects on local aspects such as cliffs, dunes, currents and estuaries.
2. Invest in "hard" defences such as sea walls, groynes and imported sand: but these are costly and have side effects.
3. Plan "soft" defences: escape routes and evacuation.
4. Planning: retreat, zoning, engineering standards, re-routing of transport and drains.
5. Invest at owner's risk: written in to permits or titles.
6. Mandate unsubsidised insurance to dissuade development in unsafe areas.

In summary, managing risk from climate change involves taking notice of small chances of large impacts and looking at a range of options to deal with these to minimise losses. Given the likelihood of ongoing climate change, despite efforts to reduce greenhouse gas emissions, these adjustments will be expensive but cheaper than suffering the consequences of inaction. In some cases such measures will be unpopular, particularly to those who do not understand the real problem. Serious efforts must therefore be made to educate the people and communicate the real situation and options. Communicating the need for risk management is vital. This is especially so for investors in business enterprises that require major infrastructure such as open-cut mines, railways and ports. Full returns on such investments will take decades, and on this time-scale the risks may be large indeed.