the 
late 
ent 
)act 
fter 
the 
Mea 
een 
eran 
eda GAL 
  
vere 
the 
ures 
'oad 
sure 
sive 
s. 
face 
tion 
ther 
iged 
ing: 
it in 
s on 
: Damage 348968 ment as of 12/01/2010 {based on SsoEys sateiae image) 
  
  
   
  
  
Figure 3. Haiti — Port-au-Prince: damage assessment map as of 
13 Jan 2010 (based on GeoEye satellite image) 
Hal < Port-su-Prince » Host SENSE a55<sament FIP and wx Gang 
7 
       
e bigtenerndoritas dat 
      
  
2d 
  
i: 
  
  
  
Figure 4. Haiti — Port-au-Prince: road damage assessment map 
as of 17 Jan 2010 (based on 0.15 m imagery) 
Haiti - Part-yu-Prirsor - Tomposrary si 
   
  
  
  
  
Figure 5. Haiti — Port-au-Prince: temporary shelter/spontaneous 
camp as of 17 Jan 2010 (based on 0.15 m imagery) 
During the Haiti earthquake response phase, just few products 
were based on SAR data, e.g. a 2D Ground Motion based on 
TerraSAR-X data produced by DLR/ZKI and a multi-temporal 
coherence analysis based on COSMO-SkyMED data produced 
by e-geos in the framework of the G-Mosaic project (most of 
the statistics related to the number and type of map products are 
based on ad-hoc queries in the catalogue service provided by 
ReliefWeb, setting as temporal extent 1 month after the event) 
As far as the approaches based on optical EO data are 
concerned, generally high resolution optical imagery are usedto 
assess the building damage grade, the damages to 
infrastructures or the road network accessibility. Although 
several automatic or semi-automatic techniques exist to identify 
collapsed building or debris after an earthquake, a visual 
interpretation approach was generally adopted in the operational 
context of the Haiti response. 
In both cases (optical and SAR data) the analysis is generally 
based on a multi-temporal analysis between the satellite data 
acquired before and after the event, but in case no suitable pre- 
event data is available the analysis can be carried out on just the 
post event image. 
The selected minimal mapping unit was often the single 
building, especially on the maps based on the high resolution 
aerial datasets, adopting as damage grades the ones proposed in 
the European Macroseismic Scale (EMS-98). However in the 
first hours the analysis were mainly aggregated at grid cell level 
or at building block level. 
When a large amount of data characterized by a high spatial 
resolution is available, and having the emergency response 
phase tight time constraint, a coordinated participatory 
approach, that assigns small grid cells covering the affected 
arcas to several image analysts, is the most effective to assess 
the impact of the event. Effective coordination is especially 
required to allow a joint effort among different entities, like the 
joint remote sensing damage assessment carried out by 
Unitar/Unosat, World Bank and EC Joint Research Centre. 
Earthquake damage assessment obviously needs large or very 
large maps scale datasets (especially concerning built-up areas 
and road network) to act as backdrop of the damage assessment 
maps and/or to perform tailored analyses. The availability of 
this kind of reference dataset is generally critical in developing 
countries such as Haiti, which demonstrates the importance of 
participatory mapping through media as OpenStreetMap or 
Google Map Maker that allowed mapping large parts of Haiti 
within a few days. 
3. SATELLITE/AERIAL BASED DAMAGE 
ASSESSMENT ACCURACY 
As far as satellite based impact assessment is concerned, the 
main issue that requires an in-depth investigation is the 
reliability of the provided information, ic, the expected 
accuracy of the damage assessment according to the technical 
features of the available satellite/aerial imagery (e.g., sensor 
type, ground sample distance, off-nadir angle, and spectral 
resolution). Validation activities related to information derived 
by means of remote sensing are therefore crucial in order to get 
the end users aware of the accuracy they can expect from a 
specific satellite/aerial based analysis, consequently interpreting 
the outputs for a more effective response activity. 
Recent studies (Saito et al. 2010, Booth et al 201 1) highlighted 
that vertical imagery (and in certain conditions also oblique 
ones) may be limiting in discriminating the level of damage of 
some buildings. In the summary of the "2nd International 
Workshop on Validation of geo-information products for crisis 
management" (JRC, Ipsra — Italy, 12-13 October 2010), it is 
explicitly reported that a validation of a joint damage 
assessment (using airborne images) performed with around 
6,000 geo-tagged photos collected in the field gave an overall 
accuracy of only 60%. According to (Corbane et al, 201 1) the 
estimated overall accuracy of the best available spatial 
resolution dataset (0.15 m) is about 65% in comparison to GPS 
geo-tagged pictures acquired during ad-hoc field surveys. A 
497