Impact of Climate Change on Road Infrastructure

1
Impact of Climate Change on Road Infrastructure

• Mark Harvey
• mark.harvey_at_infrastructure.gov.au
• www.bitre.gov.au

2
Austroads project published in 2004

• Austroads AP-R243/04  Impact of Climate change
  on road infrastucture
• Downloadable for free from Austroads website,
  with volume of appendices AP-R243/04A
• http//www.austroads.com.au
• Also downloadable for free from BITRE website
  (main volume only) http/www.bitre.gov.au
• http//www.bitre.gov.au/publications/92/Files/clim
  ate_change.pdf

3
Project aims

• assess likely local effects of climate change for
  Australia for the next 100 years, based on the
  best scientific assessment currently available
• assess the likely impacts on patterns of
  demography and industry, and hence on the demand
  for road infrastructure
• identify the likely effects on existing road
  infrastructure and potential adaptation measures
  in road construction and maintenance, and
• report on policy implications arising from the
  findings.
• Note The project was not concerned with impacts
  of transport emissions on climate change.

4
Project structure

• BITRE coordinated the project and prepared the
  Executive Summary, Introduction and the Policy
  Implications chapters.
• The CSIRO Division of Atmospheric Research ran
  its global climate change models to produce
  forecasts of climate on a grid of about 50
  kilometres up to 2100.
• The resultant data was passed on to three
  consultants to assess its implications.

5
Project structure
6
Project structure continued

• The Monash University Centre for Population and
  Urban Research investigated the likely effects on
  population settlement patterns and demographics.
• ARRB Group used these population projections to
  forecast changes in road transport demand.
• calculated changes to an index of climate from
  the CSIRO data
• road demand and climatic indexes were together
  used in pavement deterioration models to predict
  the implications for pavement deterioration and
  maintenance expenditure needs.

7
Project structure continued

• Australian Bureau of Agricultural and Resource
  Economics (ABARE) employed its hydrologicaleconom
  ic model of the Murray-Darling basin to forecast
  implications of climate change for salinity and
  agricultural production in the region, and
  related this to road infrastructure.
• Multi-disciplinary project involving experts from
  a range of fields.

8
Project structure continued
CSIRO Regional climate change forecasts to 2100
Monash University Centre for Population and Urban
Research Impacts on population projections
ABARE Impacts on salinity and agriculture in
Murray-Darling basin
ARRB Group Impacts on demand for roads
ARRB Group Impacts on road pavements
BITRE report and summary
9
Not covered in the study

• Local flooding implications
• requires a catchment hydrological model to
  predict flooding heights, durations and water
  velocities, and
• an area topology model to relate flood heights to
  local road infrastructure.
• Salinity and impacts of agricultural industries
  outside the Murray-Darling Basin
• The CSIRO models do not forecast sea level rises
  or the likelihood of changes in storm activity.

10
Emissions forecasts

• International Panel on Climate Change (IPCC)
  emissions scenario selected
• A2 scenario
• high scenario chosen to provide strong contrast
  with current conditions
• predicated on global population of 15 billion in
  2100
• rate of CO2 release grows, increasing to nearly
  fourfold by 2100.

11
IPCC emission scenarios
A2 scenario (red line) used in this study.
12
CSIRO Atmospheric-Ocean Global Climate Change
Model

• global circulation model with atmospheric,
  oceanic, sea-ice and biospheric submodels
• globe divided up into a grid comprised of 300 km
  squares
• 9 layers of atmosphere, each block having
  parameters such as temperature, air pressure,
  wind velocity, water vapour content
• 12 layers of ocean
• time step of 30 minutes
• run from 1870 to 2100
• suite of properties (temperature, moisture) saved
  for6-hourly intervals for the 230 years
• took three months on a supercomputer

13
CSIRO Conformal-Cubic General Circulation Model

• Results from global model used to nudge more
  detailed model for Australia
• wind speeds outside Australia adjusted to make
  consistent between models
• grid of about 50 km squares for Australia and
  lower resolution for rest of the world (up to
  about 800 km for the other side of the globe)

14
Method of deriving detailed forecasts

• outputs monthly means of average, maximum and
  minimum temperatures, precipitation, solar
  radiation, potential and actual evaporation for
  each grid point
• converted to
• local temperature change per degree of global
  warming for temperature
• percent for rainfall, radiation evaporation
  change per degree of global warming
• used to derive forecast for any IPCC scenario for
  any grid point over the next 100 years.

15
Key findings temperatures

• average annual temperatures increase by 2 to 6C
  by 2100
• Tasmania coastal zones least affected, inland
  areas most affected
• more extremely hot days and fewer cold days, for
  example
• average number of summer days over 35C in
  Melbourne to increase from 8 at present to 10-20
  by 2070
• average number of winter days below 0C in
  Canberra to drop from 44 at present to 6-38 by
  2070

16
Average annual temperature base (2000) and 2100
climate
5 15 17 19 21 23 24 25 26 27 28 29 30 32 34 40
Base (2000) climate
2100 climate
17
Temperature changes year 2100 relative to base
climate
18
Key findings rainfall and evaporation

• general reduction in rainfall except for the far
  north where there will be significant increases
• where average rainfall decreases, more droughts
• where average rainfall increases, more extremely
  wet years
• in the north, more intense tropical cyclones,
  more severe oceanic storm surges, more frequent
  and heavier downpours
• evaporation to increase over most of the country
  adding to moisture stress on plants and drought

19
Average annual rainfall base (2000) and 2100
climate
10 15 20 25 30 35 40 50 60 70 100 125 200 gt500
Base (2000) climate
2100 climate
20
change in average annual rainfall 2000-2010
134 -9 -12 -14 -25
21
Sea level rise

• Not predicted in CSIRO modelling.
• IPCC projects rise of 9 to 88 cm by 2100
• 0.8 to 8.0 cm per decade

22
Impact on population and settlement patterns
methodology

• undertaken by Monash University Centre for
  Population and Urban Research (Dr Bob Birrell)
• population projections developed for Australia as
  a whole, States and major metropolises (based on
  ABS mid-range projections supplemented by ANU
  demographic projection software).
• adjustments made to the projections for the eight
  major metropolitan regions for climate change
• using expert judgement supported by a comfort
  index (function of temperature and humidity). A
  comfortable climate is a major driver of internal
  migration.

23
Population base case without climate change

• total fertility rate will fall to 1.6 and net
  overseas migration 90,000 per year over the 21st
  century
• total population 19.1m in 2000 to 27.3m in 2100
• greater concentration in four major metropolises
  Sydney, Melbourne, Brisbane and Perth
• other growth outside the four cities is in
  non-metropolitan Queensland and WA.
• significant increase in share of population in
  Queensland
• for planning purposes, need to take base case,
  then adjust for climate change

24
Population climate change impacts

• of the eight metropolitan regions assessed, only
  Darwin and Melbourne gain population from climate
  change
• even though hotter, wetter Darwin less
  attractive, higher rainfall should promote
  agricultural production
• but note the contrary view from the recent
  Northern Australia Land and Water Taskforce
  report lack of suitable soils high evaporation
  and lack of dam sites limits water storage
• Losers Adelaide (water supply), Cairns (less
  attractive climate) and Perth (water and climate)
• coastal areas of NSW and Victoria more attractive
  climate
• hotter, drier climate in inland areas will may
  have adverse impact on agriculture

25
2100 population without and with climate effects
Selected Statistical Division of 2000 population without climate change Adjustment factor with climate change Climate change factors driving population change
Sydney 159 1.00 temps higher but not expected to affect population growth
Melbourne 125 1.15 temperatures higher resulting in more attractive climate
Brisbane 211 0.96 temperatures higher resulting in less attractive climate
Moreton 305 0.98 temperatures higher resulting in less attractive climate
Adelaide 63 0.79 restricted water supply, especially in spring
Perth 195 0.88 less attractive climate restricted water supply
Darwin 275 1.34 temps high but heavy rainfall drives increased agriculture
ACT 93 1.00 temps higher but not expected to affect population growth
Cairns 279 0.83 temperatures higher resulting in less attractive climate
26
Note Climate change is not the most important
influence on population patterns.

• range of projections for 2100 compared with 2000
• without climate change -63 Adelaide to 305
  Darwin
• with climate change -50 Adelaide to 369 Darwin

27
Impact on road demand Methodology

• passenger and freight tasks considered separately
• base-case forecasts developed
• cars a function of population, per capita car
  ownership
• freight a function of population, per capita
  freight, average payload (trend to larger
  vehicles)
• converted to equivalent standard axel loads for
  pavement impacts
• ARRB used a gravity model to estimate impacts of
  climate change on traffic.
• If population at A increases by 100a and
  population at B by 100b due to climate change,
  then traffic between them increases by
  100(1a)(1b)-1.

28
Impact on road demand 2100 conclusions

• 60 additional traffic (total vehicles passengers
  and freight)
• dramatic increase in Queensland, moderate in
  Syd-Mel corridor, decline around Adelaide,
  increase in Perth urban only slight rise in Perth
  intercapital traffic
• proportion heavy freight vehicles will rise from
  12.1 to 13.9
• total road freight to rise by 112 from 2000 to
  2100
• average payload to increase by 25, most in next
  decade
• equivalent standard axles per articulated truck
  to double due to higher mass limits
• total ESA-kms on National Highway to rise by 230
• due to freight growth, higher mass limits and
  payloads

29
Impact on pavement performance methodology

• climate represented by Thornthwaite moisture
  index
• a function of precipitation, temperature and
  potential evapo-transpiration. Index varies from
  100 on Cape Yorke Peninsula to -50 in central
  Australia.
• used a National Highway System road database
• pavement models estimate present value of
  life-cycle road agency costs (maintenance and
  rehabilitation) and road-user costs (travel time
  and vehicle operation)
• select treatment options and timings to minimise
  present value of costs subject to specified
  constraints on maximum roughness and annual
  agency budgets.

30
Pavement Life Cycle Costing (PLCC) model

• 60 year analysis period, 7 real discount rate
• National Highway System split into 60 sections
  with similar climate characteristics, traffic
  levels, vehicle mixes and pavement
  characteristics
• pavement deterioration a function of pavement
  age, cumulative equivalent standard axle loads,
  Thornthwaite index, and average annual
  maintenance expenditure.

31
HDM4 model

• much more detailed pavement deterioration
  algorithm covering roughness, rutting, cracking,
  potholing, ravelling, strength etc and
  consequently much more detailed data requirements
• case studies of 8 road segments analysed in
  detail
• one segment from each state and territory located
  in or near a metropolitan area
• data inputs that vary with climate are
  site-specific changes in Thornthwaite index,
  traffic levels and per cent heavy vehicles

32
Other points to note

• Note only trucks cause pavement wear, not cars
• but cars impact on the models because increased
  roughness adds to road user costs for cars
• Limitations
• effects of floods, severe storms and sea-level
  rise not taken into account
• no allowance for expansion of lane-kilometres
• design pavement strengths assumed to remain
  unchanged
• road agencies may not minimise present value of
  costs due to budget constraints causing
  maintenance to be deferred and higher than
  economically warranted maintenance standards in
  some areas for social and equity reasons

33
Thornthwaite moisture index base (2000) and 2100
climates
-45 -30 -15 0 20 40 60 80 gt100
Base (2000) climate
2100 climate
34
Changes in Thornthwaite moisture index 2000 to
2100
35
Changes to Thornthwaite index

• tendency to a drier climate overall (negative
  change in Thornthwaite Index)
• central area of Australia relatively unchanged
• localised areas where the changes are greatest
  include
• south-west of Western Australia
• north-east Victoria and southern NSW
• south-west Tasmania
• top-end Queensland.

36
Optimal road agency costs (PLCC model)
State Optimal Agency Cost (million) Optimal Agency Cost (million) Change
State Base Climate 2100 Climate Change
NSW 72.3 90.1 25
VIC 32 37.6 18
QLD 82 124.2 51
WA 48.3 56.1 16
SA 27.6 23.4 -15
TAS 6.5 6.8 5
NT 17.9 37.3 108
ACT 0.6 0.7 17
Total 287.3 376.1 31
37
Comparing optimal road agency costs

• comparison is not with and without climate change
• but 2000 traffic volumes and climate
• with 2100 climate-adjusted traffic volumes and
  climate
• Northern Territory and Queensland experience
  large increases
• primarily due to population growth but wetter
  climate contributes.
• South Australia declines due to smaller
  population and drier climate.

38
Maintenance rehabilitation funding split

• maintenance routine and periodic maintenance
  (pothole patching, kerb and channel cleaning,
  surface correction, resealing)
• rehabilitation chipseal resheeting, asphalt
  overlays, stabilisation, pavement reconstruction
• for Australia as a whole, no predicted change in
  3565 split
• rehabilitation proportion to rise (maintenance
  proportion to fall) significantly for Tasmania
• converse for WA
• reflects differences in pavement age
  distributions and life times

39
HDM4 results road agency costs
Base climate 2100 traffic changes only 2100 traffic changes only 2100 climate traffic changes 2100 climate traffic changes
Cost ('000) Cost ('000) Change (3/1) Cost ('000) Change (5/1)
Col number 1 2 3 4 5
ACT 97.8 97.8 0 97.8 0
NSW 46.2 72.7 57 72.8 58
NT 176.6 176.9 0 177.1 0
QLD 83.6 103.1 23 106.1 27
SA 99 96.8 -2 96.8 -2
TAS 140.2 159.5 14 159.4 14
VIC 128.7 175.5 36 177.3 38
WA 205.9 244.5 19 244.1 19

• undiscounted total costs per kilometre over
  20-year period
• Virtually all the changes are from population
  growth leading to traffic increases, not climate
  change.

40
Impact on salinity in the Murray-Darling Basin
methodology

• ABARE Salinity and Land-use Simulation Analysis
  (SALSA) model
• network of land management units linked through
  overland and ground water flows
• hydrological rainfall, evapo-transpiration,
  surface water runoff, irrigation, ground water
  recharge/discharge rates, salt accumulation in
  streams and soil
• climate projections incorporated by changing
  rainfall and evapo-transpiration
• rate of flow of groundwater depends on hydrolic
  gradients
• very flat in lower parts of the catchment

41
Impact on salinity in the Murray-Darling Basin
methodology continued

• land-use allocated to maximise economic return
  from use of agricultural land and irrigation
  water
• relationship between yield loss and salinity for
  each agricultural activity
• land-use can shift with changes in salinity and
  water availability
• costs of salinity measured as reduction in
  economic returns

42
Catchments in the Murray Darling Basin covered by
the SALSA model
43
Impact on salinity in the Murray-Darling Basin
Key findings
Units Base scenario Without climate change With climate change
Year 2000 2100 2100
Net production revenue m, npv 3827 3718 3400
Area of high water tables 000 ha 1137 5341 4404
SALT CONCENTRATION
Darling below the Macquarie mg/L 152 277 483
Murray below the Murrumbidgee mg/L 141 181 198
Murray below the Darling mg/L 226 301 343
Murray at Morgan mg/L 313 445 548
SURFACE WATER FLOWS
Darling below the Macquarie confluence GL 7345 7784 6060
Murray below the Murrumbidgee confluence GL 8128 9040 5259
Murray below the Darling GL 6789 7720 4435
Murray at Morgan GL 3827 3718 3400
44
Impact on salinity in the Murray-Darling Basin
comparisons

• area affected by high water tables
• base case rise from 1.1m hectares in 2000 to
  5.3m hectares in 2100
• climate change rise to 4.4m hectares in 2100
• climate change mitigates salinity problems but
  nowhere near sufficient to reverse the rising
  trend
• net production revenue
• base case falls by 3 due to high water tables,
  shift from pasture to cropping
• with climate change falls by 11 due to reduced
  surface water flows, switching from irrigated to
  dryland activities
• less demand for road transport

45
Impact on salinity in the Murray-Darling Basin
comparisons

• higher water tables are bad for road pavements,
  but this is a problem in both the base and
  climate change scenarios
• slightly less with climate change
• reduced surface water flows make salt
  concentrations higher in rivers which reduces
  yields from irrigated production
• and rusts steel reinforcing in concrete
  structures in riverine environments such as
  bridges and culverts.

46
Summing up uncertainty

• high level of uncertainty about
• IPCC emissions forecasts
• CSIRO estimates of climate impacts
• consultants forecasts
• uncertainties built upon uncertainties
• numerical results are broad indicators that tell
  a story

47
Summing up demand for roads

• Higher car and truck traffic from population
  growth is the main driver of investment and
  maintenance needs for roads.
• Large changes are forecast without climate
  change.
• strong growth for SE Queensland, Cairns, Darwin,
  Brisbane, Sydney, Melbourne, Perth
• decline for Adelaide and inland areas
• Climate change adds to forecasts for Darwin and
  Melbourne and reduces forecasts for Adelaide,
  Perth and Cairns.

48
Summing up road design and maintenance

• less rainfall should slow pavement deterioration
• but effects so small as to have negligible impact
  on costs
• exception for far northern parts of Australia,
  which are forecast to become wetter. Capacities
  of culverts and waterways may prove inadequate.
• sea-level rise a concern for low-lying roads in
  coastal areas
• changed and frequencies of floods in some areas
• requires modelling of individual catchments to
  forecast impacts

49
Overall conclusion

• Changes affecting road infrastructure will occur
  regardless of climate change.
• Climate change is just another factor in the mix,
  and usually not the most important.
• The main impacts on road infrastructure may come
  from changes in flood heights and frequencies,
  and sea-level rise with storm surges, which were
  not addressed in detail in the project.
• impacts vary greatly between locations

50
Subsequent research ARRB Climate change
framework for Queensland Department of Main Roads
in 2008

• report not published, but a summary is available
  in a conference paper by Evans, Tsolakis and
  Naude http//www.patrec.org/web_docs/atrf/papers/2
  009/1737_paper66-Evans.pdf (ATRF Conference 2009)
• comprehensive list of potential impacts on road
  infrastructure and operations
• detailed review of (short- and long-term) climate
  change forecasts for Queensland
• framework to assess risks, and to assist in the
  planning of climate change mitigation and
  adaptation responses.

51
ARRB Framework

• Four impacts relevant to Queensland
• temperature changes (increases in very hot days)
• rainfall changes (reductions and increases) and
  flooding
• rising sea levels with storm surges
• increase in cyclone frequency and intensity.
• Phases of framework to identify investment
  priorities
• identify climate change effects
• geographic scale, certainty, timeframe
• determine impacts on transport
• adaptation strategies
• planning and project evaluation

52
Other research underway

• Austroads project Impact of climate change on
  road performance, undertaken by ARRB
• software to provide climate information from 1960
  to 2099 by GPS coordinates based on CSIRO
  modelling
• minimum and maximum daily temperatures, rainfall,
  Thornthwaite moisture index
• pre-2007 based on historical meteorological data
• Climate Futures Tasmania Infrastructure project
• World Road Association (PIARC) Technical
  Committees C.3 (natural disasters), D.2 (road
  pavements), D.3 (bridges), D.4 (geotechnics and
  unpaved roads)
• all have working groups on adaptation to climate
  change