Bushfires in Southeast Australia

Over the Christmas break this year, I hopped onto the plane and headed down under, towards Australia! During the trip, I visited the Blue Mountains, just west of Sydney, some parts of which were recently ravaged by major bushfires in October. While bushfires may not be as widely researched and discussed as the other types of natural disasters that I have brought up in the previous posts, it has nonetheless been a recurring phenomenon in Australia for millions of years. As such, I will briefly touch on the impacts of climate change on bushfires in Australia, particular the southeastern region, in this post.

Bushfires on Blue Mountain in October 2013. Credits: AFP

Southeastern Australia is one the top 3 fire prone regions in the world, together with southern California and southern France. In the past century, bushfires have destroyed thousands of homes and claimed hundreds of lives including the infamous Black Friday fires in 1939 and Ash Wednesday fires in 1983. One of the driving factors that have caused southeastern Australia to be particularly vulnerable to bushfires is its climate – hot, dry summers and mild, wet winters. The precipitation received during winter and spring allows fuel (the vegetation) to grow, while the dry summers favour the development of bushfires (Lucas et al. 2007). Moreover, periods of drought have exacerbated the dry conditions and fire risks in the region. Over the past decade, it has been observed that temperatures of the region have shifted towards higher temperatures, while rainfall has declined below the 1961-1990 mean (Murphy and Timbal 2008). The extended period of dry conditions have contributed to the large fires that burned with little control in 2006/07.  

Eucaplyptus trees that are commonly found in Australia. Credits: Joon Ting

Another factor that has been blamed for the bushfires is the predominant type of vegetation that lines the landscape of Australia – eucalypts. There are more than 800 endemic species found in Australia and forms the main diet for koalas. These trees are highly flammable as they contain oil, which gives them their distinct spicy fragrance. During periods of dry and windy conditions, their flammable oil can cause small fires to develop into huge firestorms very rapidly. Yet, these trees are extremely fire resistant themselves and tend to survive the bushfires, allowing for the regeneration of the eucalyptus forest after the fire. Hence, they tend to be naturally selected in regions prone to bushfires including Australia and California, where they continue to dominate the landscape. As such, it seems inevitable that southeast Australia experiences such frequent bushfires.

Eucaplytus trees that cover the Blue Mountains. Credits: Joon Ting

It has already been projected that southeast Australia will become hotter and drier under climate change (Suppiah et al. 2004). Modelling studies have been further carried out to determine how this projected change in climate will affect the fire risks of the region. The Forest Fire Danger Index (FFDI) has been used to quantify the fire risks and is calculated based on observations of temperature, relative humidity and wind speed. Modelling studies carried out by Hennessy et al. (2005) on 17 sites in southeast Australia have suggested that the combined frequencies of days with very high and extreme FFDI rating are likely to increase by 4-25% by 2020 and 15-70% by 2050. This corresponds with more recent studies by Lucas et al. (2007) that show that the increase in annual cumulative FFDI is generally 0-4% in the low scenarios and 0-10% in the high scenarios by 2020, and 0-8% in the low scenarios and 10-30% in the high scenarios by 2050.

Nonetheless, there is still large uncertainty in these studies given that much of the climate of southeast Australia is dominated by interannual and interdecadal variability that is influenced by complex systems including ENSO and Southern Hemisphere Annular Mode (SAM). The evolution of these systems under future climate change is still not fully understood thus it is difficult to ascertain how fire weather and risks will change in future when such variability is taken into account (Lucas et al. 2007).

Droughts and climate change

With the past few posts looking at floods, let’s now move on to the next category of natural disasters – droughts. Droughts are generally defined as ‘a period of abnormally dry weather long enough to cause a serious hydrological imbalance’ (IPCC 2012). Droughts have been a relatively common phenomenon in several regions including Australia, Sahel, North America, and even the United Kingdom (though they are generally much less severe than those experienced in the other countries mentioned). In fact, since 2012, North America has been hit with one of its worst droughts, with 81% of the contiguous United States being covered with at least abnormally dry conditions (DO) at its peak on July 17 2012 (Fig.1). Currently, the situation has improved significantly although the western parts of the US are still affected by the droughts (Fig.2).

Figure 1 US drought monitor report for July 17 2012. 81% of the contiguous US is affected by at least abnormally dry conditions. Source: http://droughtmonitor.unl.edu/

Figure 2 US drought monitor report for December 10 2013. Most of the western part of the US is still affected by the drought. Source: http://droughtmonitor.unl.edu/

When discussing about droughts, it is important to note that there are different types of droughts – meteorological, hydrological and agricultural (Fig.3). The definitions for each category of droughts are taken from Mishra and Singh (2010):

• Meteorological – lack of precipitation over a region for a period of time.
• Hydrological – a period with inadequate surface and subsurface water resources for established water uses of a given water resources management system.
• Agricultural – a period with declining soil moisture and consequent crop failure. 

Figure 3 Links between the different types of droughts and the main drivers for the droughts. Source: IPCC SREX report (2012)

However, out of all the atmospheric hazards, droughts are at the moment the least well-understood and predictable (Mishra and Singh 2010). There are still large uncertainties involved regarding observed global-scale trends in droughts as well as how anthropogenic climate change will affect droughts in future (IPCC 2012).

Firstly, there has been a lack of consensus with regards to the observed trends in drought over the recent past. This is mainly due to the fact that there are few direct observations of drought-related variables available for global analysis and hence drought indices, which attempt to integrate precipitation, temperature and other variables, are used instead to infer the changes in drought conditions. One of the most prominent index used is the Palmer Drought Severity Index (PDSI) (Palmer 1965) that measures the cumulative departure in surface water balance. Based on this index, Dai et al. (2004) have suggested that globally, very dry areas have doubled in extent since 1970. However, such studies based on the PSDI have been criticised greatly due to the fact that the PSDI relies on a temperature-based method of calculation of potential evaporation (PE), known as the Thornthwaite equation. Sheffield et al. (2012) argued that temperature-based PE tend to overestimate the extent of drought as it does not factor in the effects of radiation, vapour-pressure deficit and wind speed. By using a physically-based estimate of PE based on the Penman-Monteith equation instead, Sheffield et al. (2012) showed that there has actually been little change in drought over the past 60 years (Fig.4). As such, due to the lack of direct observations of drought, it has been difficult to determine the global trends of droughts. The IPCC SREX report (2012) could at best conclude that there is only a medium confidence that some regions of the world have experienced more intense and longer droughts since 1950s.

Figure 4 A) PSDI calculated using the Thornthwaite PE method is shown in blue and using the Penman-Monteith method is shown in red. B) Area in drought (PSDI < -3). The shading shows the range derived from uncertainties in precipitation (both) and radiation (Penman-Montieth only). Source: Sheffield et al. (2012)

Secondly, there are also large uncertainties involved with our understanding of how climate change will affect droughts in the future. There have been preliminary studies based on model simulations that suggest that aridity increases and because severe by the 2060s over most of Africa, southern Europe, Middle East, most of Americas, Australia and Southeast Asia, while central and northern Eurasia, Alaska and northern Canada and India become increasingly wetter (Fig.5) (Dai 2011). Such results broadly correspond to the modelling results by Hirabayashi (2008) that suggest that drought frequency from 2070 to 2100 increases over North and South America, central and southern Africa, the Middle East, southern Asia and central and western Australia. Meanwhile, the eastern part of Russia shows significant decreases in drought days due to increases in precipitation and flood flows. 

Figure 5 Mean annual self-calibrated PDSI (sc-PDSI) using the Penman-Monteith PE method for years 2060-2069 and 2090-2099. These values were calculated using the 22-model ensemble-mean surface air temperature, precipitation, humidity, net radiation and wind speed from the IPCC AR4 using the SRES A1B 21st century simulations. Red to pink areas are extremely dry conditions while blue colors indicate wet areas relative to the 1950-1979 mean. Source: Dai 2011. 

However, these projections do not consider the impacts of changes in large-scale atmospheric and ocean circulations under climate change as there is insufficient knowledge of how these have an impact on droughts. Moreover, it is also not known how the behaviour of plant transpiration, growth and water use efficiencies will change under increasing CO₂ emissions, which would consequently affect soil moisture storages (IPCC 2012). Therefore, unfortunately, it is still hard to determine how trends in droughts will evolve in the future under climate change.

Documentary: Earth Under Water by National Geographic

As a follow up to the previous post on coastal flooding, here's a documentary produced by National Geographic which looks at the impacts of predicted future sea level rise on different cities and regions around the world. They also explore the ways that societies have been trying to adapt to such changes by beefing up on their sea defenses. 

Coastal Floods and Climate Change

In the previous post, we looked at the impacts of climate change on river floods. So in this post, we will turn our attention towards another category of floods – coastal floods. Across the globe, numerous coastal cities and towns are susceptible to such floods that create massive destruction and havoc each time they hit. For example, when Hurrican Sandy made landfall last year, large portions of Lower Manhattan was flooded with seawater as a record storm surge of 4.15m swept into New York City. A year after the floods, New York City is still trying to rebuild homes, buildings and transport systems that have been damaged, while many business and building owners are still battling to make insurance claims or gain compensation from the government. Meanwhile, other low-lying regions like the Maldives and Ganges Delta are under the threat of being inundated as sea levels continue to rise in the 21^st century.

Seaside Heights, New Jersey flooded after Hurricane Sandy made landfall. Source: AP

Although these coastal cities and towns are vulnerable to these devastating floods, they are ironically home to large concentrations of population. 13 out of 20 of the most populated cities in the world are coastal cities, including Shanghai, Guangzhou and Jakarta (Hanson et al. 2011). Moreover, many of these cities are also major national and regional economic hubs as they house key ports responsible for handling a large proportion of global seaborne trade. Therefore, the impacts of climate change on the risk coastal flooding have major social and economic consequences on these societies.

Sea level rise

One of the main causes of coastal inundation is the rise in sea levels. Sea level rise is caused by the combined effects of (1) the thermal expansion of seawater due to ocean warming; and (2) input of water from land ice melt and land water reservoirs.

From 1880 to 2009, it is estimated that sea level rise is about 210mm (Church and White 2011). Tide gauge measurements since the 19^th century has shown that sea level rose by an average of 1.7 ± 0.3mm/year since the 1950s. As seen from Fig.1, however, from the early 1990s onwards, the mean rate of sea level rise increased to 3.3 ± 0.4mm/year (Nicholls and Cazenave 2010). Evidently, sea level rise has accelerated in recent years. Moreover, it is critical to note that since the early 1990s, global average sea level rose at a rate near the upper end of the sea level projections for both the Intergovernmental Panel on Climate Change’s 3^rd and 4^th Assessment Reports (Church and White 2011).

Fig.1 Global mean sea level trend from the late 18th to early 21st century. The red curve is based on the tide gauge measurements while the black curve is based on the high-precision altimetry record from 1993 to 2009. As can be seen, there is a close match between tide gauge and altimetry records during that period suggesting that tide gauge records are reliable. Source: Nicholls and Cazenave (2010). 

However, it is important to note that sea level is not rising uniformly across the world as seen in Fig.2. Local factors can either amplify or increase the effects of eustatic sea level rise. These factors include the isostatic rebound of the Earth’s crustal due to the unloading of ice (as evident over Scotland), the subsidence of land due to the withdrawal of groundwater as well as the subsidence of deltaic regions due to the shortage of sediment supply caused by upstream damming. Nonetheless, based on Fig.1, it still seems that there is still a larger proportion of regions that have experienced sea level rise between 1992 and 2009, compared to regions that have experienced a decline in sea level for that period.

Fig.2 Regional sea level trends based on satellite altimetry records from 1992 to 2009. Source: Nicholls and Cazenave (2010).

Modelling studies carried out by Jevrejeva et al. (2012) have shown that sea level rise ranges from 0.57 to 1.10m by 2100, with the maximum rate of sea level rise reaching 17mm/year by 2100 under the RCP8.5 pathway (Fig.3). Moreover, rising sea levels will also elevate storm surges brought about by cyclonic activity, which is another cause of coastal floods (Brecht et al. 2012). Therefore, the impacts of climate change on the risk of coastal floods cannot be underestimated.

Fig.3 Sea level projections till 2100 with RCP scenarios: red - RCP3PD, blue - RCP4.5, green - RCP6.0 and black - RCP8.5. The shaded regions of similar color around the projections represent the upper (95%) and lower (5%) confidence levels. Source: Jevrejeva et al. 2012.  

Coastal flood risks

Two recent reports have attempted to quantify the losses associated with the exposure of people and assets to coastal flood risks under future sea level rise, socio-economic changes and subsidence of land. 

-----

Population exposure

Currently, the total number of people currently exposed to coastal flood risks is about 38.5 million, which is about 0.6% of the global population. Hanson et al. (2011) estimates that this figure will increase by more than threefold to around 150 million people for the period 2070-2080 due to population growth, sea level rise and subsidence of the coastal areas. Asia will be most vulnerable to coastal floods due to the high population density along the coast. The Asian cities that are within the top 20 cities with the largest population exposure to coastal flood risks in the period 2070-2080 include Mumbia, Guangzhou, Shanghai, Ho Ching Minh City, Bangkok and Tokyo (Fig.4). 

Fig.4 Map showing the top 20 cities for exposed populations under the FAC scenario for the period 2070-2080. The FAC scenario takes into account population and economic growth, natural subsidence/uplift, global sea level rise and potential human-induced subsidence. As can be seen, there is a disproportional number of Asian cities within the top 20 cities. Source: Hanson et al. 2011.  

Assets exposure

A recent report by Hallegatte et al. (2013) assessed the average annual flood losses in 136 coastal port cities and estimated that at present, the aggregated average annual flood losses (AAL) in these cities is about US$6 billion per year. Guangzhou tops the list with an AAL of US$687million, which is about 1.32% of the city’s GDP (Fig.5). In fact, most of the losses are concentrated within a small number of cities, with Guangzhou, Miami, New York and New Orleans accounting for 43% of the global losses.

Fig.5 The top 20 cities where the economic average annual losses with respect to local GDP are the largest in 2005. Source: Hallegatte et al. (2013).

The report suggested that future losses associated with coastal floods will increase dramatically under sea level rise, socio-economic changes and subsidence of land. They projected that if no adaptations were made to maintain or reduce the flood risks, the projected increase in average losses will rise to more than US$1 trillion per year. On the other hand, if adaptations were carried out, the losses will be lower, reaching between US$60 and 63 billion per year.

-----

Given that coastal flood risks will inevitably increase in future due to the combined effects of sea level rise, population growth in coastal areas as well other isostatic effects such as subsidence of land, coastal cities must attempt to adapt to the increase in coastal flood risks. Failure to adapt will result in massive losses both in terms of deaths and economic losses. Moreover these studies have not factored in other causes of coastal flooding including storm surges and tsunamis. Hence, future coastal risks might be even larger than what has been projected. However, Hallegatte et al. (2013) highlighted a grim reality that even though improvements in flood defence infrastructure can help to reduce the risk levels and reduce the number of floods, but when an extremely large flood overwhelms these infrastructure, the resultant losses will still increase. Therefore, there is a limit to these flood defence infrastructure and effort must be channelled to other forms of adaptations as well. This includes effective land use planning, early warning systems, flood evacuation plans and post-disaster relief response, which would help communities be more resilient towards coastal floods. 

Floods and Climate Change

Credits: Getty images

Now that we have spent quite a fair bit of time discussing the links between climate change and hurricanes, let’s focus the spotlight on floods! A point to note is that there are many types of floods, including riverine floods (overtopping of river banks), coastal floods (usually caused by rising sea levels and storm surges), urban floods (could be caused by overflowing of storm drainage systems), etc. In order for the discussions to be more focused, we shall just look at rain-fed river floods in this post and coastal floods in the following.

First, let’s start off with the Disaster Bites for the Colorado Floods 2013: 

Disaster Bites: Colorado Floods 2013

Video credits: ABC News

• Began September 9 2013
• Boulder county, Colorado was the worst hit out of the 14 counties affected by the floods
• At least 8 lives were lost and thousands of homes destroyed
• More than half a year’s worth of rain fell within three days
• 1 in 1000 year storm event, 1 in 100 year flood event
• High rainfall due to an active southwest monsoon and a presistent broad area of low pressure at upper levels of the atmosphere. This low-pressure area helped pull the moisture out of the tropics and into Colorado.
• Moisture was forced up the Rocky Mountains by the southwesterly winds to form orographic rain.
• Floods were exacerbated by the long-term drought in the Colorado River basin, which hardened the soil and reduced the infiltration capacity of the ground. 

More information:

http://www.huffingtonpost.com/2013/09/14/colorado-flooding-climate-change_n_3926284.html

http://news.nationalgeographic.co.uk/news/2013/09/130913-colorado-flood-boulder-climate-change-drought-fires/

http://www.cnbc.com/id/101039317

As seen above, one of the primary causes of the Colorado floods was the massive amount of rainfall that hit the region within a short period of time. Boulder County’s total 3-day (10^th to 12^th September) rainfall was 12.30 inches (312.42mm). This has far exceeded the highest recorded rainfall in the city for any month since records started in 1897. The previous record rainfall was 9.59 inches (243.59mm) back in May 1995. Could this be part of a trend of more frequent intense precipitation events associated with climate change, which are causing more frequent extreme floods?

Theory

In theory, climate warming will result in the intensification of the hydrological cycle according to the Clausius-Clapeyron relation that suggests that the atmosphere’s water holding capacity increases with temperature (Fig.1). Hence, as the air gets warmer, the increased moisture in the atmosphere will favour heavier precipitation events. Modelling studies such as those done by Stephen and Ellis (2008) suggest that precipitation would increase by 1-3%/K. Given that precipitation is one of the key drivers of river floods, an increase in precipitation, ceteris paribus, would lead to increased likelihood of such floods (Kundzewicz et al. 2010).

Fig.1 As temperature increases, the vapor pressure increases exponentially. Source: Ohlone College

Observed Trends

Temperature and Precipitation

Hansen et al. (2012) has shown that the distribution of seasonal mean temperature anomalies has shifted towards higher temperatures, especially in summer, likening it to the ‘loading of the climate dice’. They suggest that the chances of unusually warm seasons have greatly increased in the past 30 years. According to the Clausius-Clapeyron relation, the warmer temperatures ought to result in heavier precipitation. Indeed, the IPCC AR4 Report (Trenberth et al. 2007) highlighted that there had been increases in the frequency of heavy precipitation events over the second half of the 20^th century over many land areas, particularly in many regions of North America.

Runoff and floods

There have been a number of flood events in recent years where the river flow records have been unprecedented. Kundzewicz et al. (2010) highlighted several examples including the 2002 flood in Central and Eastern Europe where the Vltava River exceeded a flow rate of 5000m³/s for the first time in the last 175 years; in fact the flow rate has never reached 2500m³/s in the 60 years. However, global analyses of runoff trends for the 20^th century (Bates et al. 2008) have concluded that there is great variation in annual runoff over different regions, with the high latitudes and large parts of the USA experiencing an increase in runoff and southern Europe, West Africa and southernmost South America experiencing a decline in runoff. The same report pointed out that observed changes in runoff might not be consistent with changes in precipitation due to the competing effects of evaporation, effect of human interventions such as dam construction as well as poor data quality for some rivers.

Interestingly though, two recent reports have attempted to draw the links between anthropogenic greenhouse gas emissions and the hydrological cycle. First, modelling results by Min et al. (2011) suggest that anthropogenic greenhouse gas emissions have contributed to the observed intensification of heavy precipitation events over approximately two-thirds of the Northern Hemisphere land area. Second, Pall et al. (2011) concluded that their modelling studies show that anthropogenic greenhouse gas emissions had increased the risk of occurrence of floods in England and Wales in autumn 2000. These studies are part of a growing number of studies, known as attribution science, that are attempting to attribute specific climate and weather phenomenon to anthropogenic climate change.

Projections

With the projected changes in temperature and precipitation under climate change, it is expected that river discharge and flood risk would likely change as well. Modelling done by Hibarayashi et al. (2013) using 11 AOGCMs participating in the CMIP5 suggest that for the projected period of 2071-2100, flood frequency increases across large areas of South and Southeast Asia, Northeast Eurasia, eastern and low-latitude Africa and South America. Meanwhile, areas like northern and eastern Europe, Central Asia, central North America and southern South America see a decrease in flood frequency (Fig.2).

Fig.2 Projected return period of the 1971-2000 100-year flood projected onto the 2071-2100 period for 29 selected river basins under RCP 8.5. A) Basin map of 29 selected rivers. The color of each basin represents the multimodel median return period at basin outlets. B) The height of the grey box indicates the interquartile range (75th-25th percentile) and the solid line represents the median value. The dashed line represents the maximum and minimum return periods. Directions of change in return period and model consistency are indicated. Source: Hibarayashi et al. 2013

Nonetheless, floods are complex phenomena and we need to consider other factors that may amplify or diminish flood risks including changes in land cover such as deforestation as well as the alteration of flow regimes by activities including reservoir impoundment. Moreover, although not covered in this discussion, changes in atmospheric circulation associated with climate change such as ENSO may also have implications on flood frequency and extent. Hence, a flat-rate statement on the change in flood risk in future is hard to be made. However, this does not mean that land use planners can put off thinking of ways to minimise the exposure of people and assets to flood impacts until studies produce more concrete results, for we never know when another 1 in 1000 storm event or 1 in 100 flood event might hit us again.