1. H07 seminars
Vulnerability -
Data and methods
Climate Risk Management Unit
Institute for Environment and Sustainability
17th May 2016
Vulnerability
to
Drought
Hugo Carrão, Gustavo Naumann
and Paulo Barbosa
2. Context
• Vulnerability to Drought
• DEWFORA (2011-2014, EU FP7 project)
• EUROCLIMA II (started in 2014, Administrative Arrangement DG DEVCO)
• Global Drought Observatory (started in 2015, Administrative Arrangement
DG ECHO)
• A framework for Drought Risk or Likelihood of Drought Impact
EUROCLIMA: Drought risk is not an absolute
measure of actual economic loss or damage to
human health or the environment, but a relative
statistic suitable for ranking regions and prioritize
actions to reinforce mitigation and adaptation plans.
IPCC, 2012: Summary for Policymakers
3. Risk Hazard Exposure Vulnerability
Likelihood of
drought impact.
Probability of
exceeding a
drought event
with a certain
severity.
Propensity of
individuals or
communities to
suffer adverse
effects when
impacted by a
drought event.
Amount of population,
assets or other
valuable elements in
regions where the
probability of drought
occurrence is not null.
Risk or Likelihood
of Drought Impact
• Quantitative statistics varying 0-1;
• The legend breaks are statistical thresholds
and serve only for guidance – percentiles of
respective geographic distribution.
EUROCLIMA II: the case study for Latin America
United Nations International
Strategy for Disaster Reduction
(UNISDR, 2004)
4. Vulnerability to Drought
• Regional unit-free statistic that varies between 0 (min) and 1 (max);
• The values are relative to the most vulnerable geographic region(s)
at a given moment: suitable for ranking and comparison;
• Can be updated in time: appropriate for showing progress in regional
drought mitigation and adaptation plans;
• Composite statistic of three factors, as similar to the Drought
Vulnerability Index (DVI) (Naumann et al., 2014):
• Social;
• Economic;
• Infrastructural.
Naumann, G., Barbosa, P., Garrote, L., Iglesias, A., and
Vogt, J.,2014: Exploring drought vulnerability in Africa:
an indicator based analysis to be used in early warning
systems, Hydrol. Earth Syst. Sci., 18, 1591-1604.
Vulnerability to drought reflects the relative
regional development, namely its economic
capacity, human and civic resources, as well as
physical infrastructures.
5. • World Bank, http://data.worldbank.org/products/wdi
• U.S. Energy Information Administration (EIA), http://www.eia.gov/
• Worldwide Governance Indicators (WGI), http://info.worldbank.org/governance/wgi/index.aspx#home
• Organisation for Economic Co-operation and Development (OECD), http://stats.oecd.org/
• Food and Agriculture Organization (FAO), http://www.fao.org/nr/water/aquastat/main/index.stm
• World Resources Institute, http://www.wri.org/our-work/project/aqueduct
• Global Roads Open Access Data Set (gROADSv1), http://sedac.ciesin.columbia.edu/data/set/groads-
global-roads-open-access-v1
The emphasis in public data ensures that the final result can be tested,
reproduced, and improved with new data by the scientific community and
stakeholders.
Data sources for
computing Factors
6. Proxy Indicators at Country Level
• Agriculture (% of GDP); World Bank
• Poverty headcount ratio at $1.25 a day (PPP) (% of total
population); World Bank
• GDP per capita (current US$); World Bank
• Energy Consumption per Capita (Million Btu per Person); U.S. EIA
• Rural population (% of total population); World Bank
• Improved water source (% of rural population); World Bank
• Refugee population (% of total population); World Bank
• Life expectancy at birth (years); World Bank
• Population ages 15-64 (% of total population); World Bank
• Literacy rate (% of people ages 15 and above); World Bank
• Government Effectiveness; WGI
• Disaster Prevention & Preparedness (US$/Year/capita); OECD
Proxy Indicators at Subnational Level
• Agricultural irrigated land (% of total agricultural land); FAO
• % of retained renewable water; Aqueduct
• Road density (km of road per 100 sq. km of land area); gROADSv1
Social Factor:
Level of well-being of individuals and
communities
Economic Factor:
Economic status of individuals,
communities and nations
Infrastructural Factor:
Infrastructures needed to support the
production of goods and sustainability of
livelihoods
How to convert indicators to a common unit
and aggregate them to derive each factor of
vulnerability?
Proxy Indicators for
computing Factors
7. Factors’ Computation
• Minimum Mapping Unit:
• Sub-national administrative regions;
• Masking sub-national administrative regions:
• >75% hyper-arid conditions;
• Normalization of indicators, as similar as for
the Human Development Index, HDI (UNDP):
• Aggregation of Indicators for each Factor?
• Data Envelopment Analysis (DEA).
Univariate (e.g. Life expectancy)
Multivariate
9. Performance Evaluation (1)
• Outputs are not directly observed but rather inferred – there is no reference for
accuracy assessment;
• Performed a sensitivity analysis with alternative composites based on different
weighting and aggregation schemes of factors, namely:
• (A) arithmetic, (G) geometric, or (P) product of factors
• (W) weighted (i.e. proportional to the # indicators) or (NW) non-weighted
• This sensitivity analysis is like an unsupervised task of inferring the best
clustering technique to describe a hidden structure from raw input data.
• Internal minimum variance criterion of unsupervised cluster stability, i.e. the
minimum distance to the regional median ranks, to evaluate the performance
of alternative composite statistics.
11. Concluding Remarks
• Empirical approach: composite accuracy depends on input data quality;
• Internal performance evaluation: no reference dataset for validation;
• All factors and indicators are equally weighted: expert knowledge can be used to
improve composite statistics;
• Vulnerability focus on agricultural drought: other specific DVIs can be developed
for other sectors of activity (e.g. energy production, river transportation).
12. Thank you for your
attention!
More information:
http://edo.jrc.ec.europa.eu/scado/
hugo.carrao@jrc.ec.europa.eu
Uyuni, Bolivia
Photo credit Carlos Cruz
