Transcript 20041019_csiro_powerpoint_template.pot

Climate change and fire-weather risk in
south-eastern Australia
Kevin Hennessy
www.csiro.au
Insurance Council of Australia (2007)
Bushfire Hobart Feb 1967
Hailstorm Armidale Sep 1996
Hailstorm Sydney Feb 1992
Floods/Hail Melbourne Dec 2003
Hailstorm Sydney Nov 1976
Floods NSW Nov 1984
Cyclone Townsville Dec 1971
Cyclone WA NT QLD March 1973
Hailstorm Sydney Oct 1986
1600
Hail, wind, flood SE Aus Feb 2005
Hailstorm Sydney Jan 1991
Hailstorm Brisbane Jan 1985
Bushfires Vic SA Feb 1983
Cyclone Brisbane Jan 1974
Bushfire Canberra Jan 2003
Hailstorm Sydney March 1990
Cyclone Innisfail March 2006
Cyclone Darwin Dec 1974
Earthquake Newcastle Dec 1989
Hailstorm Sydney Apr 1999
Insured loss in $millions
Observed changes in Australia
Most of the top 20 insured losses have been due to
extreme weather events
1800
Top 20 insured losses 1967-2006
1400
1200
1000
800
600
400
200
0
Black Saturday fires
(7 Feb 2009): 173
dead, 2029
properties destroyed
(30% uninsured),
7000 people
displaced, 61
businesses
destroyed, over 8600
livestock dead (4000
sheep, 4200 beef
cattle, 427 dairy
cows)
Fire risk in Australia
• Fire risk is influenced by a number of factors
including weather, fuels, ignition, terrain, land
management and suppression
• The CSIRO-BoM study assesses potential
changes to one of these factors, fire-weather risk,
associated with climate change
• The most important weather variables are
temperature, humidity, wind-speed and rainfall
Seasonal pattern of fire danger
Fire danger index
McArthur Mark 5 Forest Fire Danger Index (FFDI; Noble et al,,
1980) is defined as:
FFDI = 2exp(0.987logD – 0.45 + 0.0338T + 0.0234V – 0.0345H)
where:
H = relative humidity from 0-100%
T = air temperature oC
V = average wind-speed 10 metres above the ground, in
metres per second
D = drought factor in the range 0-10
FFDI categories
Categories of Fire Danger Rating (FDR). Taken from Vercoe [2003].
Fire Danger
Rating
FFDI range
Difficulty of suppression
Low
0-5
Fires easily suppressed with hand tools.
Moderate
5-12
Fire usually suppressed with hand tools and
easily suppressed with bulldozers. Generally
the upper limit for prescribed burning.
High
12-25
Fire generally controlled with bulldozers
working along the flanks to pinch the head out
under favourable conditions. Back burning may
fail due to spotting.
Very High
25-50
Initial attack generally fails but may succeed
in some circumstances. Back burning will fail
due to spotting. Burning-out should be
avoided.
Extreme
50+
Fire suppression virtually impossible on any
part of the fire line due to the potential for
extreme and sudden changes in fire behaviour.
Any suppression actions such as burning out
will only increase fire behaviour and the area
burnt.
Daily peak FFDIs for large fires
Date
FFDI
13 Jan 1939 (SE Aus)
100
24 Jan 1961 (Dwellingup, WA)
110
7 Feb 1967 (Hobart, Tas)
78
16 Feb 1983 (Deans Marsh, Vic)
100
16 Feb 1983 (Trentham, Vic)
60
16 Feb 1983 (Adelaide to Warrnambool) > 100
8 Jan 1994 (Bankstown, NSW)
88
18 Jan 2003 (Canberra)
115
7 Feb 2009 (Vic)
100-192
Observed changes in FFDI
Large daily, seasonal and annual variability
Fire intensity, fuel load and FFDI
To get a fire of
specified
intensity, you
need more fuel if
the fire-weather
index is lower
Effect of Fuel Load on FFDI value for a fire with an intensity of 3500 kW
m-1, the threshold for ‘uncontrollable’ fires. Adapted from data provided
by Incoll [1994].
Report on climate change and fire-weather
Bushfire Weather in Southeast Australia:
Recent Trends and Projected Climate Change Impacts
C. Lucas, K. Hennessy*, G. Mills and J. Bathols*
Bushfire CRC and Australian Bureau of Meteorology
* CSIRO Marine and Atmospheric Research
September 2007
Consultancy Report prepared for The Climate Institute of Australia
www.bushfirecrc.com/research/downloads/climate-institute-report-september-2007.pdf
Changes in fire weather risk in south-east Australia were
assessed for the period 1973-2007 at 26 sites
Site selection was limited by availability of daily
temperature, rainfall, humidity and wind data
Observed changes in FFDI
• The annual cumulative
FFDI displays a rapid
increase in the late-90s to
early-00s at many
locations.
• Increases of 10-40%
between 1980-2000 and
2001-2007 are evident at
most sites.
• The increases are
associated with a jump in
the number of very high
and extreme fire danger
days.
AP Photo
Observed changes in FFDI
The trend (red number at top left) is 51 FFDI units per year
Observed changes in FFDI
Melbourne airport
The trend (red number at top left) is 23 FFDI units per year
Observed changes in FFDI
The trend (red number at top left) is 51 FFDI units per year
Observed changes in FFDI
The trend (red number at top left) is 24 FFDI units per year
Climate change projections
• Rather than simply extrapolating observed trends, we use
computer models of the climate system driven by scenarios of
greenhouse gas and aerosol emissions, and ozone depletion
• The IPCC emission scenarios have various assumptions
about demographic, economic and technological change
Currently
tracking the
high
scenario
Projected global warming
For the IPCC emission scenarios, 23 climate
models simulate a warming of 0.5 to 1.6°C by
2030, rising to 1.1 to 6.4°C by 2100
A1FI (estimated by CSIRO)
IPCC 2007
Observed trends in temperature
• Global average temperature is tracking the upper end of
the IPCC projections
• The rate of warming since 1975 is almost 0.2oC per
decade
Observed
Rahmstorf et al 2007
Climate change scenarios for 26 sites were
generated using 2 CSIRO climate models
Model selection criteria: good simulation of 1961-1990 mean
temperature, rainfall and MSLP; availability of daily data at “fine”
resolution, e.g. 50 km
Two climate change simulations were suitable: CCAM driven by
the CSIRO Mark2 GCM, and CCAM driven by the CSIRO Mark3
GCM
The mean warming is 0.5-1.5oC by 2020 and 1.5-3.0oC by 2050
CCAM Mark2: rainfall decreases except in autumn in northern
Vic and southern NSW, humidity decreases in spring and
summer and increases in autumn and winter, wind-speed
decreases
CCAM Mark3: rainfall decreases in spring-summer and increases
in autumn-winter, humidity decreases, wind-speed increases
Simulated changes in mean and variability
This study included simulated
changes in the mean and variability
of daily temperature, rainfall,
humidity and wind-speed
Changes in daily decile values for
each calendar month were applied
to observed daily data from 19742007
Observed daily temperature and
rainfall data were considered high
quality
Observed daily humidity data were
acceptable at most sites
Observed wind data were not
homogenised, so there were some
jumps and missing data
% change in average no. of days with very high
Percent changes
the number of
days with
verydanger
high and extreme
fire-weather
–
andinextreme
forest
fire
in 2020
and 2050
2020 and 2050, relative to 1990
2020
2050
Low global
warming
High global
warming
Low global
warming
High global
warming
(0.4oC)
(1oC)
(0.7oC)
(2.9oC)
Very high
+2-13%
+10-30%
+5-23%
+20-100%
Extreme
+5-25%
+15-65%
+10-50%
+100-300%
Extreme fire-weather days (high global warming)
2020
2050
Bendigo:
+53-65%
+135-230%
Melbourne:
+26-38%
+81-136%
Mildura:
+25-38%
+76-120%
Sale:
+15-45%
+80-215%
All changes are
relative to a period
centred on 1990
FFDI > 100
We defined a “catastrophic” fire-weather category for FFDI > 100.
Only 12 of the 26 sites have recorded ‘catastrophic’ fire-weather
days since 1973.
The 2020 low scenarios indicate little or no change, except for a
halving of the return period (doubling frequency) at Bourke.
The 2020 high scenarios show ‘catastrophic’ days occurring at 20
sites, 10 of which have return periods of around 16 years or less.
By 2050, the low scenarios are similar to those for the 2020 high
scenarios.
The 2050 high scenarios show ‘catastrophic’ days occurring at 22
sites, 19 of which have return periods of around 8 years or less,
while 7 sites have return periods of 3 years or less.
Longer fire seasons
Median FFDI for summer, autumn, winter and spring for “now”
(1973-2007), 2020 and 2050 at Melbourne and Canberra
The median FFDI increases in all seasons (mostly spring and
summer), implying longer fire seasons
Adaptation challenges
Longer and more intense fire seasons present significant challenges
for adaptation
•
•
•
•
•
•
•
•
•
•
•
Greater social, environmental and economic costs
More control-burning, smoke pollution and respiratory illness
More volunteer fire-fighters, more pressure on families/employers
Better weather forecasts and early warning systems
New technology for fighting fires
More disaster relief payments & counselling services
Better protection of water catchments & plantation forests
Better emergency plans (stay or go)
Planning guidelines for the urban/rural fringe
Insurance premium incentives for those that reduce their fire risk
Building codes/designs with reduced flammability
Uncertainties
• Quality of observed daily wind and humidity
data at most sites in Australia
• The effect of scenarios based on other
climate models
• Assessment of daily-annual variability in
FFDI, not just annual averages
• Changes in ignition (natural and humaninduced)
• Changes in fuel load, allowing for carbon
dioxide fertilization on vegetation
• Potential impacts on biodiversity, water yield
and quality from fire affected catchments,
forestry, greenhouse gas emissions,
emergency management and insurance
AP Photo & The Age
Research priorities
• Testing and rehabilitation of observed humidity and
wind data (supported by Bushfire CRC and BoM)
• Creation of regional climate change scenarios from
other models (underway)
• Assess fire-weather risk over the whole of Australia
(underway)
• Fine scale fire modelling that captures vegetation and
terrain features and fire management, e.g. using
FIRESCAPE (Sydney basin project)
• Hydrological and ecological modelling to assess
impacts on water and biodiversity