As a new report shows greenhouse gas emissions are putting the globe on a track for 4-6ºC of warming by the end of the century I’ll continue my series on the Senate Recent trends in and preparedness for extreme weather events Inquiry. In the last instalment I looked at current and historical trends in extreme weather and attribution of them to climate change. In this post I look forward to the next ToR:
(b) based on global warming scenarios outlined by the Intergovernmental Panel on Climate Change and the Commonwealth Scientific and Industrial Research Organisation of 1 to 5 degrees by 2070:
(i) projections on the frequency of extreme weather events, including but not limited to drought, bushfires, heatwaves, floods and storm surges,
(ii) the costs of extreme weather events and impacts on natural ecosystems, social and economic infrastructure and human health, and
(iii) the availability and affordability of private insurance, impacts on availability and affordability under different global warming scenarios, and regional social and economic impacts;
Here we come across the first mistake of the terms of reference. Both the IPCC and the CSIRO projections were released in 2007 and based on models run prior to that. Although there has been no new report released by these bodies yet, there has been substantial work done by others in the meantime. Any projections based on these scenarios are likely to be conservative, and the inquiry should canvass the full range of projections.
I’ll address future projections of extreme weather, the current and future costs and issues of insurance.
A word on uncertainty
Firstly though, a word on uncertainty. There are massive uncertainties when it comes to climate change. We don’t know how successful emissions reduction efforts will be; we don’t have perfect historical data to calibrate computer models; computer models don’t simulate extreme weather very well; and they don’t produce regional projections with high confidence. There are also some fairly large uncertainties about the actual costs of natural disasters and extreme weather events.
However, risk is defined as the effect of uncertainty on objectives. The implication of this is simple: the higher the uncertainty, the higher the risk. Uncertainty should be a call to more urgent action to mitigate risk, not less.
Extreme Weather, future projections
The Australian Climate and Weather Extremes: Past, Present and Future, report also deals with projections of future climate impacts on extreme weather. Additionally the emergency services in Australia, concerned about climate change impacts on natural hazards, have been conducting and publishing reviews of the literature. At least NSW and QLD have commissioned specific reports on extreme weather, with NSW taking the additional step of regionalising those reports. Various states and territories are also undertaking climate modelling projects, with extreme weather being a key climate phenomena investigated (NSW/ACT, QLD, WA, TAS, SA these projects are at various stages and reports for some are available). Many local governments have also incorporated changes to the risk of extreme events in their climate change risk assessments.
As with my first post I’ll use these sources and try and update them with more current literature where available.
Various projects have forecast tropical cyclones in the Australian region to decrease in overall number and duration but increase in intensity and produce more rainfall. Some models have also shown tropical cyclone tracks to shift 100-200km southward. There is substantial variation in the magnitude of the changes shown by these models and the effects in different regions in Australia.
However due to the significant variability in the cyclone record, studies estimate that it would take 60-100 years for a trend to be observed in cyclone frequency, intensity and damage.
Obviously sea-level rise will also impact the storm surge from tropical cyclones having consequences for those areas already vulnerable to storm surge such as Cairns.
Increasing temperatures will drive an increase in the number of extreme fire weather days across much of the continent, except Tasmania. Many areas could experience an increase of 4-5 times the number/ of very extreme fire weather days and some areas could experience a catastrophic fire weather day on average every 3 years. Projected changes appear to mirror current trends, with the most significant increases being in inland NSW and northern Victoria.
Fire Danger in different areas will also be driven by changes to rainfall and relative humidity, e.g. both are expected to decrease in QLD where the number of Extreme Fire Danger days could increase by 2-4 times by 2050.
Climate change could also impact fuel loads, both by changing the vegetation patterns and the amount of fuel they produce and by possibly reducing the opportunity fire agencies have to undertake prescribed burning. One study has found that fuel loads in eucalyptus forests are expected to decrease, but not significantly. Changes could also occur to grass fire risk due to changes to cropping under a changing climate. Increases to the length and severity of the fire season are likely to limit the period for prescribed burning to winter in many areas, although there is likely to be regional variation. In some areas the number of days with suitable weather for prescribed burning could increase.
Both average and very hot temperatures are expected to increase with global warming. This will make heatwaves more intense. As with current trends most research has focussed on recurrence of very hot temperatures and not examined frequency of a heatwave ‘event. Some of the more recent modelling has examined days over 35ºC (the impact of which will vary considerably across the country). By 2100 the frequency of days with this maximum temperature will be 3 times higher in Melbourne, 4 times in Sydney, 2.5 times in Adelaide and Perth, 20 times in Brisbane, 3 times in Hobart and 6 times in Canberra. The average temperature will shift in Darwin such that the number of days over 35ºC will go from 9 days in 2008 to 10 months of the year by 2100.
In WA heatwave intensity is expected to increase, particularly over western inland and northern areas; frequency expected to decrease over the central region, but increase over the southwest; and duration to increase in central areas of south Western Australia.
Climate models are predicting a decline in average rainfall over much of Australia particularly during winter and spring. Much drought in Australia is driven by El Niño, the impact of climate change on the frequency and severity of which is unclear. The behaviour of periods of drought against this long term drying trend is likewise unclear. Nevertheless models are predicting more droughts over Australia and an increase in drought indices over eastern Australia.
There is considerable variation in modelled predictions for extreme rainfall with some models are predicting a general increase in rainfall intensity and others predicting a decrease although all have considerable spatial variation. There is a suggestion that increases in extreme rainfall at shorter durations will be larger which could potentially increase flash flooding in cities and other small catchments (which are mostly along Australia’s coasts).
In southwest WA, modelling is predicting a general decrease in extreme daily rainfall, although with some spatial variation. Similar effects are predicted in northwest WA.
In Tasmania modelling indicates increases in the number of very wet days, more intense extreme one day rainfall and a significant increase to the intensity of very short duration (6-minutes) rainfall, particularly in eastern Tasmania. One location is predicted to have the 200 year ARI (for rainfall in 24 hours) change to the 20 year ARI, although this is at the extreme end of projected changes. Seasonal changes to rainfall are also predicted. The TasFutures project has also been one of the few to directly feed these rainfall projections into flood models. They found that larger rivers will be unaffected by changes to rainfall intensity, though smaller catchments and the headwaters of larger catchments could see flood heights (for a given ARI) increase by 0.2-1.2m.
Storm Surge, Extreme waves and Coastal Erosion
As mean sea levels increase so will the frequency of extreme sea level events. The change will depend much on current variability in extreme sea levels. In Tasmania, for example, the current 100 year ARI extreme sea level could occur every year by the end of the century under rising sea levels.
Increased extreme sea levels will also have consequences for flooding in estuarine environments like coastal lakes. Changes to storms that cause extreme water levels are unclear but some studies have suggested an increase (above the changes caused by rising sea levels) to storm surge frequency and intensity.
Increased erosion of sandy beaches is expected with rising sea levels. The Bruun Rule (which has been the subject of some criticism) estimates that for every 3mm in sea level rise, beaches will erode by 0.15-0.3m. Thus for 50-100cm of sea level rise this century beaches could erode by 25-100m, although this will vary from location to location.
There is much uncertainty in projected changes to wave climate across the country. In NSW the maximum storm wave height may increase, but further research particularly on east coast lows is needed.
Increased mean sea levels will also increase tsunami risk around the country.
Thunderstorms and Hail
There has been considerable regional variation in predictions for future hailstorms. Modelling has forecast a decrease in hail frequency for Melbourne and Mt. Gambier, an increase in hail frequency for Queensland,and an increase in the frequency of large hailstones in Sydney.
Some models suggest that westerly gales, which are a common cause of wind damage in southern Australia could decline in some areas as wind belts shift southwards.
However extreme winds are expected to increase over much of Tasmania, increasing losses from damage to older buildings, although current building codes are considered to be adequate to 2100.
The costs and benefits of extreme weather events
In assessing the costs of extreme weather events and impacts on natural ecosystems, social and economic infrastructure and human health it is also important to recognise the real benefits of extreme weather. In some areas tropical cyclones and floods contribute significantly to seasonal rainfall (up to 50% in some parts of WA) recharging natural and man-made water storages. Periodic flooding is also important for many ecosystems and can have agricultural and pastoral benefits. Investigation of flooding in the Murray Darling Basin suggests that the benefits of flooding there are significantly underestimated.
Many systems, particularly natural systems, are well adapted to coping with extreme weather events. However climate changes to average conditions could stress systems to the point where they are no longer able to cope with extreme events. No study I know of has investigated this sort of interaction in any ecosystem or sector. In terms of ecosystem consequences this type of impact could very well be the most significant.
Analyses of the costs of natural disasters rarely cover all sectors (including the natural environment) and all events. Cost estimates will depend on the methodology used. There is significant discussion on how to estimate the loss in some sectors and how to value disaster benefits (e.g. rebuilding is an economic stimulus, the funding for which flows into Australia from the global reinsurance industry). Determining economic losses for disasters are very difficult and there are issues with comparing loss estimates derived for different events and hazards. Most analyses tend to focus on a specific event, disaster and/or sector. For example:
- This report on the economic impacts of the 2011 Queensland Floods
- Estimated costs of tropical cyclones in Western Australia
- This bulletin which discusses research on the CO2 released by bush fires and their impact on catchment hydrology, with forest regrowth absorbing more water.
- This appendix to the Final Report of the Victorian Bushfires Royal Commission which provides a comprehensive, though conservative, estimate of the total costs of the 2009 Black Saturday disaster.
The costs of preventing, preparing for and responding to extreme weather disasters also needs to be considered. A study investigating fire (primarily urban fire) found that the total cost of fire in Australia is over $12 billion (or 1.3% of GDP), however only 14% of this was from the consequences of fire with the remainder being spent on anticipation (e.g. compliance with building codes) and response (e.g. funding of the fire services).
The closest equivalent study on natural disasters only examined expenditure by federal, state and local governments directly associated with emergency services and disaster mitigation, response and recovery (ignoring the private sector and expenditure with co-benefits such as on land use planning and building code systems). The results showed that the majority of Federal expenditure was on relief and recovery, the majority of State and Territory expenditure was on preparedness and response and the majority of local government expenditure was on mitigation. Most disaster related expenditure is by the states and territories, followed by the Commonwealth, followed by local government.
A number of studies have given estimates of annualised costs of natural disasters in Australia ranging from 1.07 billion to 1.65 billion (adjusted for inflation to 2011 dollars) Some studies have confined themselves to insured losses whilst others have attempted to include economic losses, and the cost of deaths and injuries (including from lost productivity). Other methods of estimation focus on damage to the built environment, Risk Frontiers using its own propriety database estimates that the equivalent of 44,400 dwellings were destroyed between 1900 and 1999.
There are data issues for all these studies though as smaller disasters are likely to go unreported, there was little data collection for historical disasters, and detailed analyses tend to focus on larger more recent disasters. These difficulties mean that all estimates are likely to be conservative and the real cost of disasters much higher. These figures don’t include the costs of droughts or heatwaves.
Pricewaterhouse Coopers prepared a report on heatwaves which surveyed some of the impacts of heatwaves, but did not come to any annualised figure. Bambrick et. al. in a paper commissioned for the Garnaut review looked at current and projected loss of life, hospitalisations and lost workdays due to temperature-related morbidity and mortality. They found (relative to a baseline without climate change) deaths in the Northern Territory (1159%), Queensland (548%) and Western Australia (62%) would be higher, whilst they would fall in the southern states by 2100. They also found a similar pattern in loss of economic productivity. In pure productivity loss temperature related illness and death costs the Australian economy about $1.4 billion per year (assuming the value of a workday to be $400). This does not include the cost of hospitalisations (which would add about $50 million), temperature impacts on infrastructure and the social impact of temperature-related illness and death.
I have not been able to find any estimate of annualised costs of drought. A ‘back-of-the-envelope’ calculation (assuming 1 in 3 years are drought affected, and drought on average impacts about 0.6% of GDP) estimates that drought costs the Australian economy about $2-3 billion on average per year.
This would make drought the most costly extreme weather event. Adding estimates for temperature extremes, drought and other natural disasters suggest extreme weather costs Australia somewhere in the vicinity of $4-7 billion per year.
However to put things in context road accidents cost $27 billion per year, overweight and obesity $21 billion per year, influenza $7 billion per year, crime $36 billion per year, transport-related air pollution $2.7 billion per year, work related illness and injury $60.6 billion per year, illicit drugs $8 billion per year, alcohol $15 billion per year and smoking $31 billion per year.
Cost of other climate change impacts
The cost of climate impacts in Australia are expected to be 4.8% of GDP by the end of the century (about $65 billion per year in present dollar terms). There will be significant variation across the country with Queensland expected to lose 10% of gross state product by 2100.
The impact of climate change on natural hazards is unlikely to be its greatest cost for Australia. In the Murray Darling Basin alone 97% of agricultural production (about $18 billion per year) could be wiped out by 2100. These figures suggest that prolonged loss of rainfall, rather than drought per se, will be responsible for most of the economic damages to agriculture from climate change.
Damage to the tourism industry is also expected . The Great Barrier Reef is worth $5.4 billion annually to the national economy. Severe and irreversible damage is predicted to occur to the reef with climate change, which would wipe out this contribution.
In other sectors anticipating climate change impacts is more complex. For road infrastructure a generally drier climate could decrease maintenance costs. On the other hand more hot days would make roads more susceptible to intense rainfall which may increase in some areas. The magnitude of these effects could lead to an overall reduction, increase or no change in road infrastructure costs due to climate change.
Climate change migrants from Pacific islands could average 43,000 annually by 2050, many of whom would travel to Australia. Were they to be treated in a similar framework to current asylum seekers it could cost Australia upward of $4.7 billion per year.
It’s estimated that there is $226 billion of commercial, industrial residential and transport assets exposed to a sea level rise of 1.1 metres by 2100. Relocation of these assets will cost somewhere in the vicinity of $3 billion per year on average. However, coastal erosion and inundation will not happen gradually. If these assets are replaced when destroyed the peak annual cost can be expected to be much higher.
Availability and Affordability of Insurance
I’m going to mainly consider home insurance, which was examined extensively in the context of flooding in the Natural Disaster Insurance Review. I will draw on that inquiry’s final report and its submissions heavily here.
For the vast proportion of properties in Australia natural disaster risk doesn’t contribute significantly to household premiums. Availability and affordability of insurance is only currently an issue for hazards with high spatial variability, mainly flooding.
There are approximately 100,000 properties in Australia where flood insurance premiums are estimated to be significantly (3-14 times) above the base insurance price. These properties are located in areas with an ARI of flooding of 50 years or less. Using the risk bands suggested by the Insurance Council of Australia and assuming climate change would increase everyone to the next highest risk band (a conservative approximation) the total amount of flood premiums paid in Australia would double from about $400million to $800million. Most of this increase would be borne by those already facing unaffordable premiums although this group would grow by 50%.
Insurance for homes in North Queensland against tropical cyclone raises premiums, compared to similar homes in non-cyclone prone areas, by 3-4 times. However there is no evidence this has a great effect on rates of insurance in these areas. There have been some issues raised about affordability and availability of insurance in these areas, although it is unclear whether this is due to lack of competition in the local market or the underlying risk profile. As climate change is not expected to significantly increase tropical cyclone losses over the short-medium term it is unlikely to have a major impact on insurance in north Queensland.
One under-appreciated issue is coastal erosion. It is generally not covered by insurance and the future loss of the property and the land on which it sits is an almost certainty. This hazard will only increase with climate change. As the loss of a property will only occur once and be total, arrangements other than insurance will need to be investigated to transfer this risk.
The use of insurance as a climate change adaptation tool has been examined by Booth and Williams. They note that climate change could increase the unavailability and un-afforability of insurance, particularly for low income earners. Additionally Bell surveys the availability of coverage for different extreme weather events.
There remains considerable uncertainty about how climate change will impact many extreme weather events. Climate change will increase the incidence and severity of heatwaves, drought and bushfire weather over many parts of the country. Sea level rise will increase the frequency of storm surge and coastal erosion events. The number of weak tropical cyclones are expected to decrease but the number of intense tropical cyclones are expected to increase.
The trend for other hazards is less clear, although there is some evidence that the incidence of flash flooding will rise. There could be considerable regional variation with the incidence and severity of some extreme weather phenomena increasing in some areas and decreasing in others.
The cost of extreme weather events in Australia is approximately $4-7 billion, but there are difficulties in developing an accurate and comprehensive figure. Drought is likely the most costly hazard. With a limited understanding of present disaster costs it is difficult to develop projections for climate change, but it is likely that increases in the cost of extreme weather events will likely be dwarfed by other climate impacts.
With a limited understanding of the projected impacts of climate change on extreme weather it is difficult to predict how the insurance industry will respond. However current issues with the availability and affordability of flood insurance are likely to increase under climate change. Properties in the coastal zone that could be impacted by sea level rise will become an increasing problem that will need to be addressed by governments, perhaps through risk transfer arrangements such as land swaps.
- The scientific investigation of projections in extreme weather is rapidly evolving and involves many different weather variables and study techniques. As with past trends this makes it extremely difficult for non-experts to get a clear picture of the current science. In the report recommended above a survey of the literature on future projections of the incidence of extreme weather events and costs. This report should be regularly updated when new research is published.
- Federal, State and Local Governments should, consistent with their contexts, continue to undertake and improve on risk assessments on natural hazards:
- Federal State and Local Governments should, where possible, publish online copies of these risk assessments
- The Attorney-Generals Department should incorporate an easily searchable database on these risk assessments into the Emergency Management Knowledge Hub and improve linkages between the hub and other federal Government databases, for example the Australian Flood Studies Database.
- Further work is needed to better understand the costs of extreme weather events, including economy wide costs and the impact on human health. The costs expended by governments and the private sector on disaster preparedness and mitigation also need investigation.
- Federal and State and Territory Governments should continue programs of work to improve the affordability and availability of insurance in particular for flooding and tropical cyclone damage.
- Governments should also develop risk transfer arrangements for public and private assets in the coastal zone exposed to sea level rise, including land swap deals.