Forecasting numbers of people affected annually by natural disasters up to 2015 | Forecasting | Regression Analysis

Please download to get full document.

View again

of 10
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information Report
Category:

Documents

Published:

Views: 15 | Pages: 10

Extension: PDF | Download: 0

Share
Related documents
Description
Using data from the Centre for Research on the Epidemiology of Disasters (CRED) in the University of Louvain, the analysis in this paper identifies an underlying upward trend in the numbers of people affected by climate-related disasters since 1980, and makes a projection of the numbers that are likely to be affected by 2015. For the purpose of this analysis, six types of natural disaster types, as defined by CRED, are categorised as climate-related: droughts, extreme temperatures, wildfires, storms, floods, mass movements (wet). This paper projects that, by 2015, on average over 375 million people per year are likely to be affected by climate-related disasters. This is over 50 per cent more than were affected in an average year during the decade 1998-2007. Limitations in the quality and coverage of data currently available, coupled with the natural volatility in numbers affected in a given time frame, will limit the robustness of statistical forecasting models. Nonetheless, this analysis provides a broad-brush indication of the rising scale of humanitarian need due to climate-related disasters in the relatively near future.
Transcript
  Forecasting the numbers of people affected annually by natural disasters up to 2015 Shamanthy Ganeshan, Oxfam GB, Research Intern (Economist) & Wayne Diamond, Oxfam GB, Global Research Methods Adviser Introduction The vast majority of people affected by natural disasters suffer from climate-related disasters. To make this clear, between 1980 and 2007, 98% of all those affected by natural disasters suffered from climate-related ones (as opposed to, for example, earthquakes). This analysis therefore considers only trends in climate-related disasters, but the results for all natural disasters would be virtually identical (see Technical annex, Table 2 ). This analysis does not include the numbers of people affected by conflicts. This might be an area for future research, but would face significantly greater challenges in identifying an underlying trend and projecting it into the future. Headline result By 2015, on average over 375 million people per year are likely to be affected by climate-related disasters. This is over 50% more than have been affected in an average year during the last decade. In figures: Numbers of people affected by climate-related natural disasters  Annual average, 1998–2007 243 million Forecast, 2015 375 million % Change Increase by 54% -252575125175225275325375 Current Average & Forecast for 2015 average    P  e  o  p   l  e  a   f   f  e  c   t  e   d   b  y   C   l   i  m  a   t  e  -  r  e   l  a   t  e   d   d   i  s  a  s   t  e  r  s   (   M   i   l   l   i  o  n  s   )  Annual  Average1998-2007 Annual  Average 2015   Source:  CRED EM-DAT Global natural disaster occurrence and impact: 1980–2007.  See also Chart 1  & Chart 2   Oxfam GB  April 2009 Page 1 of 10  The CRED EM-DAT Database The Centre for Research on the Epidemiology of Disasters (CRED) maintains a publicly accessible database on emergency events. The EM-DAT (Emergency Events Database) is a comprehensive database carrying data for various types of natural disasters by both country and date (year & month) going back to 1900. It is generally accepted that EM-DAT data are more reliable from the beginnings of the 1980s onwards. For the forecast shown here we have only used quarterly data from after 1980. Three groups of disasters are distinguished in EM-DAT: natural disasters, technological disasters and complex emergencies. Natural disasters are in turn categorised into five main groups (biological, climatological, geophysical, hydrological and meteorological) and then into 11 main types (see Technical annex, Table 1). For the purposes of this paper Oxfam has categorised 6 of these 11 natural hazard types as being climate-related. Predicted changes in the global climate could be expected to increase the frequency and severity of these natural hazards. That is not to say, however, that the projected increase in the numbers of people affected by these climate-related disasters should be solely attributed to climate change. There are a number of other factors involved, including the greater number of people who are likely to be vulnerable to those disasters because of, for example, their location or their poverty. For a disaster to be entered into the database at least one of the following criteria must be fulfilled: ten or more people reported killed; 100 people reported affected; declaration of a state of emergency; or a call for international assistance. The main sources for events listed are UN agencies, but information also comes from national governments, insurance organisations, and the media. Between 1980 and 2007 EM-DAT contains records of over 6,500 climate-related natural hazard events, with an average of 343 events per year over the past decade (77% of all natural hazard events, see Technical annex, Table 2a ). The EM-DAT database not only gives counts of the numbers of disaster events over time, but also the total number of those affected by the event ( Technical annex, Table 2b ). The total number of people affected by an event includes those who suffered physical injuries or illness, as well as those made homeless or who otherwise required immediate assistance during a period of emergency. Between 1998 and 2007 in an average year, some 243 million people were affected by climate-related disasters (98% of all the people affected by natural disasters in this period). The forecasting model There is a large number of factors that can affect vulnerability of different populations to natural disasters, including the rising number of disasters themselves, population growth, including in vulnerable areas, and the range of factors that can make people more vulnerable to the disasters that occur. For that reason forecasting the number of people that are affected by natural disasters is an imprecise science, and the figures presented here should be treated accordingly. The EM-DAT data shows a significant variation in the number of people affected from one year (or one quarter) to another. This reflects a number of extremely large natural disaster events that have a significant impact on the quarterly and annual totals. These extreme events seem to happen at regular intervals (see Charts 1 & 2), so removing them from the forecast model altogether would not give an accurate impression of the numbers of people affected by climate-related natural hazards. One approach to forecasting with ‘noisy’ data is to use a statistical ‘smoothing’ technique to even out the extremes of highs and lows in the data so as to give a Oxfam GB  April 2009 Page 2 of 10  clearer picture of any underlying trend. Smoothing tends to reduce the impact of extreme events on the overall forecast. The 2015 forecast predictions shown here were made using a simple two-step process: ã ‘smoothing’  the very volatile historical quarterly data on numbers of people affected in climate-related disasters; ã using a technique called Linear Regression  to fit a straight trend-line through the smoothed data so as to get a reasonable forecast up to 2015. The type of smoothing used here is called ‘Double Exponential Smoothing’, which takes into account historic data from across the time series but gives more importance (a higher weighting) to more recent disaster events than to those longer ago (see Technical Annex for more information). When using Linear Regression to estimate a time-series trend, it is possible to calculate the upper and lower probable ‘range’ of a future forecast. These are known as upper and lower ‘confidence intervals’. The convention is to estimate a 95% confidence interval. (See Technical annex, Table 3 ). In other words, we can be 95% confident that the number of people affected by climate-related natural disaster in 2015 will be between 336 million and 413 million in an average year.  As noted above, the data shows a significant variation in the number of people affected from one year to another. Such ‘volatility’ in the data means that different forecasting models could lead to different results. It also means, however, that a highly sophisticated model is unlikely to be more precise than the relatively simple model used here. The headline result above, therefore, is a reasonable forecast; it is not the only possible forecast that could be made from the data available. Part of this volatility may be due to the way data on disasters is collected and recorded. It is likely that the reporting of natural hazards (and especially estimates of the numbers of people affected), as well as data collection standards, definitions and sources, have changed considerably both over time and in different locations. Different types of disasters may also be reported more or less consistently over time. In addition estimates of people ‘affected’ by an event are likely to be less consistent and more volatile than counts of people actually killed. It is reasonable to assume that more recent data is likely to be more accurate than older data (hence the use of Exponential Smoothing before forecasting, and also restricting this analysis to data from post-1980). © Oxfam GB April 2009 This paper was written by Wayne Diamond and Shamanthy Ganeshan. It is published as background to the Oxfam International report The Right to Survive , see www.oxfam.org . uk/right-to-survive  The information in this publication is correct at the time of going to press. The text may be used free of charge for the purposes of advocacy, campaigning, education, and research, provided that the source is acknowledged in full. The copyright holder requests that all such use be registered with them for impact assessment purposes. For copying in any other circumstances, or for re-use in other publications, or for translation or adaptation, permission must be secured and a fee may be charged. Email publish@oxfam.org.uk. For further information on the issues raised in this paper please email: enquiries@oxfam.org.uk Oxfam is a registered charity in England and Wales (no. 202918) and Scotland (SCO 039042). Oxfam GB is a member of Oxfam International. www.oxfam.org.uk Oxfam GB  April 2009 Page 3 of 10  Chart 1:  People affected by climate-related disasters 1980 to 2007 (millions, Quarterly & Year-to-Date Actual & Smoothed Trend) 50100150200250300350400450500550600650700    1   9   8   0   1   9   8   1   1   9   8   2   1   9   8   3   1   9   8   4   1   9   8   5   1   9   8   6   1   9   8   7   1   9   8   8   1   9   8   9   1   9   9   0   1   9   9   1   1   9   9   2   1   9   9   3   1   9   9   4   1   9   9   5   1   9   9   6   1   9   9   7   1   9   9   8   1   9   9   9   2   0   0   0   2   0   0   1   2   0   0   2   2   0   0   3   2   0   0   4   2   0   0   5   2   0   0   6   2   0   0   7     P    e    o    p     l    e    a     f     f    e    c    t    e     d     b   y     C     l     i    m    a    t    e       ‐     r    e     l    a    t    e     d     d     i    s    a    s    t    e    r    s     (    M     i     l     l     i    o    n    s     ) Quarterly   Data Year ‐ to ‐ Date   (4   Qtrs) Smoothed   Data   (4   Qtrs)   Source:  CRED EM-DAT Global natural disaster occurrence and impact: 1980–2007. Notes: Year-to-date (4 Qtrs) shows the sum of four consecutive quarters. Events with unknown starting month are averaged over four quarters. Smoothed trend series estimated using Double Exponential Smoothing (with a smoothing weight equivalent to using 112 quarters). Oxfam GB  April 2009 Page 4 of 10
Recommended
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks