Logo GRF IDRC 2012

Conference Agenda

Overview and details of the sessions of this conference. Please select a date or room to show only sessions at that day or location. Please select a single session for detailed view (with abstracts and downloads if available).

 
Session Overview
Session
THU4.3: Monitoring and modelling for risk management
Time: Thursday, 30/Aug/2012: 1:00pm - 2:30pm
Session Chair: Ali ASGARY, York University
Session Chair: Artur PINTO, European Commission - DG JRC
Location: Sertig

Session


Presentations

Informed response via satellite based technologies

Iain Hay MACINNES

DigitalGlobe, United Kingdom

DigitalGlobe owns and operates the most agile and sophisticated constellation of high-resolution commercial earth imaging satellites. QuickBird, WorldView-1 and WorldView-2 together are capable of collecting over 500 million km2 of quality imagery per year with intraday revisit around the globe. Attending this presentation, you will learn how DigitalGlobe’s satellite constellation is used to monitor natural and manmade major disasters, including earthquakes, tsunamis, floods, tropical cyclones and fires. Digital Globe also monitors civil unrest, refugee displacement and military operation on a global scale.

DigitalGlobe’s analysis team uses a number of international sources to quickly identify crisis events around the globe. Once identified, imagery of the affected areas is collected. Maintaining the largest commercial image library in the world enables us to compare pre and post event images to effectively determine the scale and impact of a disaster. DigitalGlobe’s FirstLook Service supports the rapid acquisition, processing and disseminating of imagery via our web based technologies. This presentation provides a brief overview of DigitalGlobe’s use of open source methodologies to improve response times for acquisition and delivery of very high resolution satellite imagery in the immediate aftermath of major disasters. It will highlight the need for alert based messaging formats to improve responsiveness.


Identifying landslides using binary logistic regression and landslide detection index techniques

Wentao YANG, Ming WANG, Peijun SHI

Beijing Normal University, China, People's Republic of

Rapid urbanization and limited suitable land around the world has been pushing more population under the risk of various geo-hazards. Among those geo-hazards in mountainous regions, landslide is one of the most important hazards that can be triggered by earthquake, heavy rainfall and human activities. Landslide mapping plays an essential role in landslide susceptibility, hazard and risk studies by providing landslide inventory. Although it is very time- and labor-consuming, traditional manual interpretation of high resolution aerial images is a commonly used method. To promote the efficiency of landslide mapping, in this paper a landslide index and regression method are used to automatically detect landslides occurred during the 2008 Wenchuan Earthquake in Pingwu, Sichuan Province, by using two SPOT5 images pre and post the event. These images are first pan sharpened, ortho-rectified, co-registered and relatively corrected for atmospheric influence. Then, layers of NDVI difference, image spectral angle and principal components are obtained as candidate useful landslide identification layers. To pick out useful layers for landslide identification and establish the binary logistic regression model, landslides inventory in a training area is used to calculate the correlations between landslide and the above candidate layers. Highly correlated ones are selected and used to establish logistic model to detect landslides. Then, landslide detection index is developed in the same region to map landslides. To compare the efficiency of these two techniques, results from both methods are verified using existing landslide inventory. Final results indicate that regression models and landslide detection index both can be dependently used to map landslides using images pre and post earthquake.


Satellite application for non-structural flood risk management in Pakistan

Lubna RAFIQ1, Thomas BLASCHKE2

1SUPARCO-Pakistan Space Agency, Pakistan, Islamic Republic of; 2Z_Gis, Centre for Geoinformatics- University Salzburg Austria

Flood management procedures rely on the recording of the hydrological parameters of a flood event, its modelling and short-term and long-term forecasts. Methods and models are well established yet it is still difficult to predict the risk levels of such extreme events. For the non-structural management of flood related risks, precise flood prone areas need to be known and a detailed analysis of both societal and environmental aspects of such flood events needs to be conducted. Vulnerability was assessed using the census data in order to conduct statistical and GIS analysis, along with using the Delphi method to assign weights to the respective vulnerability indicators. Multi temporal Landsat MSS, TM and ETM images were utilized in order to calculate the flood inundation extent for four selected flood events. SPOT- 4 multi spectral high resolution imagery was found to be very useful in identifying population location (in terms of built up area) as well as agricultural units within the flood zone. Non-structural flood risk methodology is illustrated by using the Chenab and Jehlum River floods of 1976, 1988, 1992 and 2010 (Jhang and Sargodha districts of province Punjab, Pakistan) as a case study. Additionally the analysis of the vulnerability of critical infrastructures (schools & hospitals) within flood hazard zones provides indicators for the degree of spatial exposure to disaster. Findings of the study may help in the planning and management of the flood plain area of Jhang Tehsil in order to mitigate future flood events accordingly.


Estimating casualties in future earthquakes for preparedness: probabilistically or deterministically?

Max WYSS

WAPMERR, Switzerland

There are many large cities located in earthquake prone areas for which an estimate of the extent of future disasters is necessary for mitigation and preparedness. The probabilistic method works as follows. One estimates the seismic hazard by Peak Ground Acceleration (PGA, given in meters/seconds2) expected to be an upper bound not to be exceeded by a certain percentage in a period of specified length at the location in question. Next, one calculates the probability of casualties, if the expected upper bound PGA should occur. In the deterministic approach, one assumes a worst case scenario by estimating the maximum credible PGA resulting in the city at risk from the maximum credible earthquake on a nearby fault. Comparing the results of the probabilistic approach with the observed casualties, we have found that it underestimated the fatalities in the most deadly recent earthquakes by two to three orders of magnitude, rendering the probabilistic method useless. Using the deterministic approach to calculate casualties in real time for earthquakes worldwide and in scenarios for future disastrous earthquakes, we have found that our estimates are usually correct within factors of two to ten. We advocate that cost effective deterministic estimates of future human losses in large cities should be carried out as follows. Construct a city model in which population and building stock characteristics are estimated separately for each administrative district (alternatively neighborhoods with uniform building stock). Estimate the position and magnitude of the maximum credible earthquake on faults within approximately 50 km. The calculated losses in each district measure the catastrophic event civil protection has to prepare for. The probability that this worst case will happen is not calculated rigorously. For cities for which most experts agree that future large earthquakes are likely, this advocated worst case approach should be standard.


Using dasymetrics to address the aggregation error in spatial data: a multi-criteria approach for flood vulnerability assessment using spatial data

Paul KAILIPONI1, Duncan SHAW2

1University of Manchester, United Kingdom; 2Warwick University, United Kingdom

Multi-criteria decision analysis (MCDA) methods that use spatial data can assess risk/vulnerability for emergency management. These risk/vulnerability assessments are used by emergency managers to support resource allocation and disaster mitigation projects. Improvements can be made to risk/vulnerability assessments by utilizing a spatial disaggregation/aggregation technique known as dasymetrics. The dasymetric process allows for improved comparison of spatial data that is aggregated to dissimilar (non-commensurate) areas. Two error terms will be discussed that can quantitatively assess data improvement using dasymetrics against assumptions of homogenous distribution. This combination of methods will be illustrated through a flood vulnerability analysis according to Environment Agency (EA) regulations in the United Kingdom (UK). The case study shows how the use of dasymetrics can change the results of the flood vulnerability assessment. Improved spatial techniques can substantively improve the identification of vulnerable areas to flood hazard and support precautionary actions. This advancement in the combination of multi-criteria risk assessments and spatial data can be generalized to any hazard that can be spatially represented. The inclusion of the dasymetric process to MCDA is especially suited to emergency management due to its reliance on data aggregated to spatial areas (polygons).


Early detection, surveillance of wildfires and the integration into fire management systems

Joachim Franz DREIBACH

Fire Watch international AG, Switzerland

During the last years, worldwide more and more optical systems (video based) are installed for the detection and surveillance of forest fires. The precocious detection of wildfires, determining their precise location, and the rapid alert of intervention teams all play major roles in reducing the size and extent of fires and therefore limiting their effect and damage. The evaluation of the yearly fire reports provides figures, which do not show much positive results and improvements to areas operating such technologies. No reduction of the average annual burned areas monitored by video systems. On the opposite, the French Center of Research and Experience has been observing over the past 10 years a reduction of the average annual burned area by 50% while at the same time the average temperature has been rising slightly and forest are increasingly dry with rain become more and more scarce. In this means, there seems to be a difference between the capability of early detection and simply a way for visual monitoring of a starting fire. There are actually no common agreed specifications or parameters defined to qualify a optical detection system for a qualified detection system or to classify a technical solution to a monitoring system. The presentation shows capabilities of different, common used technologies for wildfire surveillance and successfully integrated solutions for early detection and integration into fire management procedures. And, after 10 years of evaluation, such long term results may be used to establish first parameters and procedures for a international standardisation of such technology.


Natural disaster mitigation and earth observations: a Group on Earth Observations perspective.

Francesco GAETANI, Douglas CRIPE

GEO Group on Earth Observations, Switzerland



The presentation will provide an overview of key drivers as well as technical and scientific trends in the Societal Benefit Areas of Disasters and Water (including Flood and Droughts) of the Global Earth Observations System of Systems (GEOSS), being developed by the Group on Earth Observations (GEO). Specific emphasis will be given to the role of Earth observations (EO) in achieving the related GEOSS Strategic targets, through activities of the GEO 2012-2015 Work Plan.

Disaster Risk Reduction (DRR) can be achieved if science is successful in providing society with clear and detailed information on the potential risk it is facing. In fact, objective and reliable information on hazards, vulnerability and exposure, presented through an analysis of expected impacts for given Risk Scenarios, is instrumental for triggering and, more importantly, sustaining the political will and economic strength needed to achieve adaptation and mitigation. In this framework, EO have the powerful capacity to represent and describe complex dynamics and processes by means of detailed, objective and up to date risk assessment maps. Additionally, EO have an important role to play in supporting the scientific community through the development of large-area (seismic, landslides, flooding, and wildfire) vulnerability modeling and mapping.

A further key role of EO is in dynamic risk assessment, especially when properly assimilated in mathematical models or systems, which can in turn be used to feed the Early Warning operational chain. Finally, in real-time emergency and response phases EO from geostationary and low earth orbit satellites can be coupled with meteorological forecasts and observations to track/monitor events, measure or evaluate their magnitude and expected impacts and, most importantly, define meaningful and near real time event Scenarios, which can support decision makers in managing resources and organizing emergency plans.



 
Contact and Legal Notice · Contact Address:
Conference: GRF IDRC 2012
Conference Software - ConfTool Pro 2.6.49+TC
© 2001 - 2012 by H. Weinreich, Hamburg, Germany