I'm pleased to have the opportunity to come talk to you today about thinking
about who's at risk and how one can use new tools to do rapid mapping of
population exposure or potential population exposure. I don't have to
remind you that last year was a pretty intense year on a number of fronts. we
had this unusual hurricane season, but also fires and earthquakes and all sorts
of other disasters. Of course an interesting thing about hurricanes that
I don't have to tell you is it there's a extended period of time when you're sort
of wondering which way is the hurricane going to go, and at various points in
times the entire Atlantic Coast population was potentially at risk of
being hit, but of course in actuality only a small proportion of those people
actually experienced the landfall or associated wind and flooding damage. So
as most of you I'm sure working the field have a inherent understanding of
who and where the population centers are, but of course it's always good to see it
on the map and to be able to explain where population centers are and who's
vulnerable to external audience or to local and regional decision makers. So
having data first off on just population distribution is useful from that
perspective, and this kind of data can be used of course both by managers
and by the press and members of the public but one also might be interested
in more details. Not just where population lives, but you know where are
the elderly, what's the housing stock that they live
in that might be more vulnerable to a storm or certain kinds of damage; where
are the urban areas in the coastal zone that's low in elevation; a lot of
different data may be relevant to identifying specific populations, and of
course infrastructure is really critical. Transportation infrastructure things
that certainly have secondary impacts if they are affected by a severe storm, dams
power plant,s even in the case of Puerto Rico there was a decommissioned nuclear
power plant there, cultural sites, of course hospitals, and other emergency
facilities. These are important infrastructure and of course within the
U.S. government there are groups, the Highfield Group and others, that collect
data on infrastructure, geospatial data, and make it available. Of course, I bet
many of you, if you were involved, probably also use Google Maps or other
tools site view to get a sense of what's actually there. But so putting, getting
access to this data and putting it together is an important way of
understanding the context of any disaster that's imminent or after its
occurred. But more than looking at pretty maps, you'd also like to be able to give
get a quick estimate of, first off, how many people is that, so if you have a
National Weather Service warning area, this is about a week ago in Michigan,
it's a little polygon that appears from the Weather Service when there's a River
outlook flooding. You know wouldn't you just like to be able to go to the map and
see well how many people actually live in that and watch it as it moves across
the landscape and get that total. Or if you see an event like this, fires in
Southern California, wouldn't you like to be able to kind of get a sense, "Well is
that a really heavily populated area where those fires are occurring or not.
You know where is the plume that's aerosol optical depth from a NASA MODIS
satellite showing overlaid with the extent of fires detected by satellite,
and then the population. So being able to put these different kinds of data
together to get a sense of what the risk is where is the potential associated
effects of disaster is helpful. There was just a earthquake off the coast of
Alaska, and of course there was concern about a tsunami. One question you know
how many people are within a certain distance of the epicenter, but in the
case of a tsunami you might want to refine that query to kind of understand
how many people live along the coast that could conceivably be hit by a wave.
So you know the ability to go from a simple query like a circle to a more
complex spatial query like a polygon that you just draw in a client allows
you to make a somewhat more focused or get a somewhat more focused answer. Now
these are not as accurate as if you had to as if you used a geospatial
information system went and downloaded it and made the query in the
tool but remember only GIS experts can do that. They can you know it takes some
time; they can't do it fairly instantaneously through a simple
interface. This tool that I'm talking about, which is developed by our center,
is an interactive map. It also accesses data from a range of sources. One of the
ones here is called EOnet; it comes from the NASA earth data website and allows
you to look at historical hazard events and for example you can see tracks of
prior severe storms and hurricanes or fires and allow you to do a little bit
of retroactive analysis. And the way in which you can think about these tools,
you know, would be we were actually interested in feedback. Here's a case
where there was a Weather Service warning in an area along the i-10
corridor between two major cities, so you're not just worried about the direct
impact of that hazard on a point but what how important is that point for example and
disrupting a major transportation quarter. We're sitting here
in the DC area, how many people actually live inside the beltway and what would
happen if an event occurred here. As we as you think about these data, what one
wants to start looking at is the way in which hazard occurrence may interact or
or have the way in which they may not be totally independent of the where people
live, and so one obvious question after many of the recent hurricanes is,
for example, are the elderly particularly concentrated in areas of flooding. Some
of my colleagues took a look at this using the type of data I've been showing
and and didn't find a correlation in the Houston area between where the elderly
live and the flooding. But it's a important question to ask and certainly
one that's relevant in disaster response. And of course as we move not only from
looking at potential exposure to actual impacts there are opportunities to bring
in new types of data. This I'm sure you've seen after Puerto Rico was data
from a relatively new satellite and an
instrument called VIIRS which looks at nighttime lights and can detect the
changes essentially in the electrical grid before and after Hurricane Maria.
And we've been working with Miguel Ramon at NASA Goddard to try to improve this
as an indicator and then incorporate this into kind of a near real-time
service to make it available as a as a useful tool for measuring impact. You can
actually already see this data in NASA's worldview client in which you see our
original population data fading into the night time lights data so you can get a
sense of where there are lights, where there are people or not. You know it's
nice for a center like ours to build a client not everybody will come to your
site and try to use a client what's more important is how do you get the data out
and into tools that the community already uses. So for example we've been
working with a tool called GeoCollaborate operated by the All Hazards
consortium which is an industry consortium that helps in utilities
coordinate fleet movements after a disaster. So after all the hurricanes
actually I guess there were two phases of intervention in Puerto Rico where
utility trucks were shipped to Puerto Rico in January, and those trucks have to
go between states, they have to deal with issues related to weigh stations, and
finding housing for the crews, and dealing with blocked roads, so the
all-hazards consortium has developed this tool called GeoCollaborate which
helps pull that information together in a collaborative way. What we did this
past fall is add this little button here on the bottom left which is
an icon with people, and the point is it accesses the data service that I showed
you previously in order to do something like draw polygon around part of Puerto
Rico and return the estimated population. So it's just a simple way of making the
same type of query through a tool that industry and others use and get access
to you know just a very basic piece of information about population exposure. Of
course not everybody using this is sitting at a desk with a web browser. How
do you give this kind of capability to people out in the field who might be
using mobile apps and also think about how to take advantage of the location
services that you get when you're in a mobile platform? So you might be at a
point where you want to know the distance away from a facility or from an
event, and of course if you go to Google Maps, you'll just get the road distance
not the line of sight distance. So this is a very simple query that just allows
you to say, "well how far am i away from say a nuclear power plant." Or as with the
case of the earthquake, you know, what are the how many people are living within a
particular distance of the epicenter of an earthquake or some other event. And
you may also want to quickly pull up information about those facilities, so it
happens we have access to a nuclear power plant database that pulls
information in from the International Atomic Energy Agency and that gives you
sort of detail, the imagery of the particular site. There's actually
operating information about it and then you can do that little query about how
many people live within a particular radius. That's where I live, where you
know it matters whether my kids' school was within 10 kilometers or
10 miles are out and I have one kid in, and one kid out which meant they got
evacuated to totally different places, but that's important information to know
f you're in the field. And just to start wrapping up, there are a bunch of other
kinds of data that may be relevant, which some of the scientific community is
producing and part of our interest is trying to figure out what information is
really relevant to the user to; the Emergency Management community, and what
could be folded in. So for example, we recently released a data set also
developed at NASA Goddard which is about the percent impervious surface in any
particular area based on 30 meter resolution Landsat data, and yeah one of
the questions is, "What is that, how can that be useful for people who are
thinking about flood response?" and you know how much asphalt in an area might
affect flood levels for example, or does it have to be plugged into hydrologic
model that quickly uses that information to estimate what the the flood levels
might be. And so on that vein we we're recently also developed a global data
set which has estimates of not just the total population but specific age groups
and the sexes of area males versus females because those are important in
thinking, of course, about vulnerability. If they're excessive number of elderly
or children or women of childbearing age those can be important in thinking about
vulnerability to disaster. In the U.S. census, of course, there are many useful
variables thinking both about individuals and households and other
kinds of differences that may be important in certainly in planning. One
of the things we're doing with the New York State is mapping
literally every building in the state trying to detect using imagery
whether the roof is peaked or flat but useful in integrating into a decision
support tool that may be interested in potential flooding impacts or wind or
other kinds of damage. And of course we have a Android version of the mobile app
that I mentioned under development. So just to summarize you can find these
mapping tools on our NASA SEDAC website. They are free, open to the public. You
can also go directly to them or use that QR code to go directly to the iTunes
Store and download it. And you can also see the fleet response working groups
geo collaborate tool they keep a record of many of the recent events and some of
the data that were delivered and how these were used by industry. and of
course, you can go to the NASA site to find a whole range of data from the data
systems that take data and deliver it to users from satellites. We're certainly
interested in feedback on how to make these tools and services more relevant
to emergency managers and hope to be able to improve them and work with the
community to make them more effective. Thank you
Không có nhận xét nào:
Đăng nhận xét