Mark Stafford Smith, Science Director

1
Anatomy of an integrated analysis involving
adaptive capacity
Climate Adaptation National Research Flagship

• Mark Stafford Smith, Science Director
• Climate Adaptation Flagship
• GEOSS/IPCC Workshop, Geneva, 1 Feb 2011

2
Topics

• Relating an experience A multi-level analysis of
  drivers of migration from drylands globally
• Project, not yet public, for UK Foresight process
• Focus here on process and experience not results
• (Really only proof of concept)
• Characteristics
• Linked some environmental and social drivers
• Considered adaptive capacity and multiple levels
  explicitly
• Needed to focus on consistent biome
  within-country
• Trying to detect and forecast trends over time
• Project to 20302060
• Implications for data??

3
Drylands x countries
From http//geodata.grid.unep.ch/
4
Land cover x aridity zone x country polygons
Global Land Cover Facility, U Maryland
5
Conceptualisation - 1

• More movement is likely where there is
• more long term trend to less (environmental)
  resources per head
• less national capacity and interest to invest in
  dryland regions
• poorer investment outcomes in dryland regions
• poorer recent or current environmental
  conditions,
• all exacerbated by greater inequality

6
Conceptualisation - 2
7
Conceptualisation - 3
Pressure to migrate

• urban
• GDP/capita
• Corruption idx

• GINI index

• Drought idx 1y
• Drought idx 10y

• Child mortality
• Road density (this polygon)

Trend in environmental services per head

• Popn increase
• Trend in NPP/capita

8
Slow variable trends in environmental services

• Trend in NPP
• 1980-2000 AVHRR NDVI-derived NPP (Prince and
  Goward 1995 GloPEM)
• Recognising MODIS would be better in the long
  run
• Averaged across each polygon
• Future NPP explored 5 DGVMs (Sitch et al. 2008)
• V. variable performance in drylands much
  coarser resolution, so some polygons had to be
  dropped
• Population GPW from CIESIN
• Allocation gridding algorithm to assign
  population values to grid cells may be least
  accurate in drylands
• NPP/population decadal trend
• Created ratios within a polygon over time
• Nb avoided comparing across space

9
Trend in NPP/hd in drylands 85-90 to 95-00
10
Environmental impacts

• Drought index (Sheffield Wood)
• Indicator of acute drought and short-term
  changes in production capital
• Independent of NPP dataset
• Looked within polygon at periods gt12 months in
  its own lowest decile
• Assumes local society in balance with the
  polygons long-term median index
• Generally seems good but poor in hyperarid
• Long-term! 1948-2000 at 1 resolution
• But not yet available for future runs at higher
  resolution

11
Social projections

• Country GDP (SRES) and population urbanisation
  UN projections
• Actually usually false resolution in databases
  since projected regionally
• Ie. not even at country level, let alone drylands
  within country
• Used as indicators of proportional change, not
  absolute
• No future projections of other indicators
• Sensitivity analysis instead
• (still useful for decision-making)

12
Case studies

• Easy 50 more if we could have gone back another
  decade in NPP

13
Projected migration intensity A1, 2030
14
Issues

• Need long time series, unavoidably
• Case studies over decades, detection of change
  in variable environments
• Historical and projections data need to be
  compatible
• Problems with definition typologies (cf.
  forest)
• Partially avoided by only looking at changes over
  time within one pixel (what is one pixel?.)
• Data sets tuned for a particular purpose
• e.g. tuned for C mitigation dont do drylands
  well
• Adaptive capacity invariably multi-scaled!
• Sub-national social data hard to come by
• Not commensurate with environmental data
• In space, in time, in collection units

15
Some implications for adaptation research

• Matched nested data sets
• In space and time
• Multiple levels, multiple scales
• Accessibility
• Documentation of data-set prejudices
• What purpose in mind when it was cleaned up, etc?
• Commensurate sampling
• Especially social lt-gt environmental datasets
• Need to make mature learn to walk before we run
• Work through in systematically chosen set of case
  studies

16
Resolving antagonistic paradigms

• Adaptation bottom-up local/regional/sectoral
  responses
• Participatory ownership vital
• Need a structured approach to extrapolation/scalin
  g up

Generalisations global statements
Typology of diverse systems x
Categories of regional GEC impacts
Complex sets of case studies without
generalisability
?Broadly predictable sets of responses
17
Directions for the workshop?

• Long-term architecture and indicators needs
• To deliver data for adaptation investment
  evaluation for decision-makers (e.g. Adaptation
  Fund, nations)
• What key decisions?
• What key information for these decisions at what
  scales?
• What architecture to aim towards? (i.e. Tues
  talk!)
• Short term delivery to IPCC AR5
• Published description of needs promote to
  parties
• 2-3 proof-of-concept case studies, ??written up
  in time
• Adaptations to changing water availability by
  basin?
• Adaptation within mitigation actions of REDD??
• Adaptive DRR preparations for one class of
  disasters?
• Monitoring to a purpose!! but not just
  mitigation
• Biophysical and social data at multiple scales,
  including in situ, developed through
  demonstrators?

18
Climate Adaptation Flagship

Climate Adaptation Flagship
Director Andrew Ash 61 07 3214 2234 /
andrew.ash_at_csiro.au Science Director Mark
Stafford Smith 61 0408 852 082 /
mark.staffordsmith_at_csiro.au