Dear James,

First of all thanks alot for your comments, and also thanks are due to
Iian
Kelman for the insightful review of my proposal, for the providing some
good
reference and for getting my attention to some important points I should
further consider, and finally for Anshu Sharma for commenting on my
research
questions and for pointing me to the UNESCO conference.

So, back  to your comments, I think that remote sensing can be useful in
more other things in the urban environment than what you have limited it
to
for the detection of the roofing materials of the buildings. Besides that,
the remote imagery can be very powerful for the detection of urban and
neighborhood quality indicators which correlate in away or another with
vulnerability levels (either positive or negative). Take for example the
percentage of trees and green parks in the local neighborhood. In most
cases, and this is more acute of course in developing world... this can
give
us a clue of the social level in this area Sam is for the quality of
street
pavements. Another are is the accessibility and "walkability" of the urban
areas during emergencies which can also be detected by remote imagery.
Building density can of course be detected using texture analysis
techniques. Further, some research have shown a great potential for
applying
remote imagery to provide good estimates for population that are close so
far to the real census (See for example Lo 1995 in Int. J. or Remote
Sensing), others have used the remote sensing to detect the urban heat
island phenomenon and which itself tells us about how the urban areas are
organized. And even more, some authors have gone to further extent and
tried
to use remote sensing to uncover the spatial extent of fertility
transition
(See Weeks et al 1999), and of course you know the importance of the shade
component which can tell us more about building heights. I can give you
more
and more examples, but you can see (People and Pixels 1998) if you have
more
intrest in such a topic.

However, I totally agree that remote sensing cannot tell us how individual
behave and it is limited to some degree in the detail that it can provide
us
with, but if we can follow the scene model and build our own "spectral"
library from various features in the urban area under consideration, with
the new innovative techniques such as spectral mixture modeling and with
the
new high resolution of multispectral imagery, it can still provide us with
a
rich set of  information that are useful for understanding variability of
vulnerability over space, especially in such places as developing
countries
where it is not easy to have such information over large scales, or even
found, it will probably being up-to-date.  As for the GIS, of course it
has
the powerful to combine all variables together, and with what you called
"invisible" features that can not easily detected by passive remote
sensing
(I use passive, because radar remote sensing comes with its own set of
measures). Combining them together, and based on our criteria, (which is
usually region, or area specific) we can assign vulnerability levels, and
then using  the power of local spatial statistics, we can  examine if such
levels are clustered or not, and how this is reflected with our knowledge
about social structures and people in such areas.

Again, I would like to thank you for being helpful and I hope I have
clarified some of the concerns you raised.


Tarek



*****************************************************
Tarek Rashed
PhD Graduate Student & Teaching Assistant
Department of Geography
San Diego State University
San Diego, CA 92182-4493  USA
Phone: 619-265-8541; FAX: 619-594-4938;
e-mail: [log in to unmask]
*****************************************************

----- Original Message -----
From: James Lewis <[log in to unmask]>
To: <[log in to unmask]>
Sent: Tuesday, November 09, 1999 10:11 PM
Subject: Re: Requesting feedback


>
> I am wondering how far remote sensing of built environments would
> accurately reflect social vulnerability.
>
> Buildings are not always used for their intended purpose; neither
building
> use nor density of occupation would be apparent. Building use may not be
> uniform; there may be mixed use with only partial occupation.
>
> Building age, construction quality and material, and existing condition
> would all impact on physical and social vulnerability - but would not
> appear in remote sensing ?
>
> How would vulnerability to various hazards be identified; eg:
earthqaukes,
> floods, cyclones ?
>
> There may be habitation other than in buildings; eg: railway wagons;
sewer
> pipes (above ground); "invisible" (to GIS) shanty towns/biddonvilles -
or
> people without buildings at all ?
>
> Conversely, in the 1980s, Julio Kuroiwa surveyed parts of Lima using
> buildings as indicators of vulnerability.
>
> Also in the 1980s, fieldwork in Sri Lanka noted that buildings had the
> capacity to protect each other (eg: from cyclone damage), especially in
> urban areas. Also, that there were more people in some rural areas than
> there were/are in obvioulsy urban areas.
>
> With Best Wishes for a challenging project:
>
>
>
>




%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%