Transcript A Comparative Analysis of Global and National Datasets

Improving the Quality,
Coverage and Accuracy of
Disaster Data:
A Comparative Analysis of
Global and National Datasets
Presentation by Working Group 3
to the Sixth Meeting of the
Inter-Agency Task Force
Geneva 24 – 25 October, 2002
Working Group 3
Plan 2002 - 2003
Sub-Working Groups:
1.
2.
3.
Improving Disaster Impact Data and
Analysis
Tools and Best Practices for Risk and
Vulnerability Analysis at the Local and
Urban Levels
Indicators, Models and Data-sets for Risk
and Vulnerability Indexing
Improving Disaster Impact Data
and Analysis
Chair: Maxx Dilley, IRI Columbia University
 Activity 1: Comparative analysis of Global
and National Datasets (IRI Columbia
University, CRED, LA RED, UNDP)
 Activity 2: Linking climate and disaster
databases – jointly with WG1(IRI Columbia
University, CRED, Munich Reinsurance, LA
RED, ADPC, European Commission DG Joint
Research Centre, UNDP)
Disaster Data
 Accurate and reliable disaster data is
essential for achieving all the goals and
objectives of the ISDR:
- Risk and vulnerability analysis
- Early warning systems
- Response preparedness
- Adaptation to climate change
Key Challenges
 Global datasets are missing substantial
numbers of disasters at the national level due
to deficiencies in international reporting
 National datasets capture a greater
proportion of the total losses but most
countries do not maintain consistent and
comparable records
 Variations in methods and standards make
comparison difficult
 Economic losses are inadequately captured
and recorded
Complementary Initiatives
 ProVention consortium has compared 3
global data sets (NatCat, SIGMA, EM
DAT) across four countries
 ISDR IATF Working group 3 has compared
one global data set (EM DAT) and national
data sets (DesInventar) across 4 further
countries
Working Group 3 Study
 Compared 149 records in the CRED EM DAT
dataset with 19,004 records in the DesInventar
database for the period 1970 – 2000 (*)
 Covered Chile, Colombia, Jamaica and Panama.
Very different countries in a single region
 Used No. of deaths and No. of affected people
as surrogate loss indicators
 Study commissioned to OSSO, Universidad de
Valle, Colombia – winner of 1996 Sasakawa
Prize
(Panama for the period 1996 – 2001)
Methodology
 National disaster records classified into 3
categories:
- Those that correlated with international
reports in EM DAT
- Those that fulfilled EM DAT criteria (more
than 10 deaths or 100 affected people) but
were not captured by international
reporting
- Small scale events with less than 10 deaths
or 100 affected people
The Comparison
Number of deaths:
Events
with
Events
with
more Events with
Deaths
as
% in Country
Database
less than
10
more
than
10
more
than
10
than 10
deaths
Events
with
with and/or
Events with
deaths
and/or
deaths
and/or
100Events
more
than
more 100
thanaffected
less than 10
100affected
affected
10 deaths
and/or 100
affected
10 deaths
and/or 100
affected
deaths
and/or 100
affected
reporting
international
reporting
international
reporting
Not
captured
Captured
by
Not captured
Captured
by
Not captured
Not captured
by
international
international
by international
international
Captured
by
by reporting by
reporting
reporting
international
Chile
10%
7%
7%
7%
83% 83%
10% 10%
Jamaica
11%
22%
22%
22%
67% 67%
11% 11%
Panama
1%
15%
1%
1%
15% 15%
84% 84%
Colombia
85%
6%
85%
85%
6% 9%
6%
9%
The Comparison
Number of affected:
Events
Events
with
with
more Events with
People
in Country
Database
more
than Affected
10
than
deaths
10as %more
than10
10
less
than
Events
with
with and/or
Events with
deaths
and/or
and/or
100Events
deaths
more
than
more 100
thanaffected
less than 10
100affected
affected
10 deaths
and/or 100
affected
10 deaths
and/or 100
affected
deaths
and/or 100
affected
reporting
international
reporting
international
reporting
Captured
by
Not captured
Captured
by
Not
captured
Not captured
Not captured
international
by international
international
by
international
Captured
by
by reporting by
reporting
reporting
international
Chile
13%
13%
71%
71%
16% 16%
13%
71%
Jamaica
87%
87%
11%
87%
2% 2%
11% 11%
Panama
3%
3%
3%
81%
81% 81%
16% 16%
Colombia
8%
8%
8%
85%
85% 85%
7% 7%
Conclusions
 In the countries studied international reporting is not
capturing a significant proportion of either deaths or
affected people (*)
 In 3 of the 4 countries small scale events accounted
for significant proportions of disaster death but not
people affected (*)
 For every disaster captured by international reporting
there are approximately 16 medium scale events that
fulfill EM DAT criteria that are not being captured. In
the 4 countries a total of 87 EM DAT events could be
correlated but a total of 2,467 other events could have
been captured.
(* except when a unique catastrophic event has occurred in the reporting period)
Conclusions
The results cannot be extrapolated globally
but indicate that there is a serious problem of
reporting disaster occurrence and loss that:
 Would seem to be underestimating real
losses in many countries
 Could lead to skewed and incorrect
conclusions and projections in disaster
reduction and adaptation to climate change
applications.
Recommendations
 In order to underpin the achievement of
the goals and the objectives of the ISDR,
Working Group 3 proposes a range of
inter-institutional activities to promote and
facilitate the building of a multi-tiered
global system of disaster reporting and
data sets.
Recommendations
 The consolidation of a system for creating
a unique global disaster identifier – GLIDE, which can link national and global
datasets, and particularly its incorporation
in national datasets.
 Development of common reporting
standards and protocols for both national
and international datasets
Recommendations
 Development of national datasets in areas
where these do not currently exist.
 Development and promotion of methods
and standards for capturing economic loss
 Capacity building and training in all the
above areas
Next Steps
 Working Group 3 will host the next
meeting of the CRED Technical Advisory
Group in early 2003 and will agree on an
action plan to put into practice the
activities mentioned above.