EH Home About Us Communities Radiation Medicine Water Sewage Programs Newsroom Data

EH PrepCast : GIS and Response

EH PrepCast #3

This is a complete transcript of EH PrepCast #3 : GIS and Response

The participants were Richard Butgereit, GIS Administrator for the Florida Division of Emergency Management; Calvin Desouza, GIS Manager, Florida Division of Environmental Health, inside the Department of Health; and Jennifer Williams, Hazard and Vulnerability Analyst, Florida Division of Environmental Health, inside the Department of Health

Mitch Stripling, Florida Division of Environmental Health, moderates.

CALVIN: Thank you.

MITCH: So the first thing that I want to ask is for somebody to discuss a little bit of the history of GIS for a lay person. How does it work, how has it been used in the modern world. Calvin.

CALVIN: Well, you briefly mentioned earlier the first history of GIS was of Dr. John Snow in London when he mapped out a cholera epidemic. In London people were dying from cholera. So what he did he took the location of all the deaths, mapped, overlay with a layer of the streets and saw that there was a correlation between where people were getting their water from and where the deaths were occurring. So what he then did, he had the county officials remove all the well pump handles and that actually made the deaths subside, because people were going different locations to get their water from. That is the earliest known occurrence of GIS. Now GIS is a huge computer based system where you have computers and servers, it is on the Internet, it is on your phone. It is not just layers on a map. When people think of GIS they just think of a sheet of paper, but now the outcome of GIS isn't normally a map. You can use it to do analysis. You can do it to find populations and you are not really trying to make a paper.

MITCH: So what about response? How is it being used in the response world?

CALVIN: In the response world I feel like it is used for more planning than in actual response, and I think the word GIS gets most of its venom from. If you, for an event that you know is going to happen ahead of time, like a hurricane, you can plan out and see where your affected populations are going to happen. So ahead of time you can do pre-planning where you can send down things like, send water trucks, send down sanitary packs which are portable toilets, hand washing stations and you can already have the trucks prepped and ready to go. When the actual event occurs we can overlay the path of if say a storm over on top of facilities, our environmental facilities of our affected populations to see how many people could you affected by the event.

MITCH: Richard, what is the first time that GIS was used in a hurricane response?

RICHARD: FEMA rolled out for their response for Hurricane Andrew using GIS in the field in 1992. Since then, you know, the technologies have grown and the efforts have been supported through the four phases of emergency management; preparedness, response, mitigation and recovery.

MITCH: Is there anything, you said the technology has improved. Since the 2004 season, what do you think the major sort of improvements in GIS for responses have been?

RICHARD: Better GPS units, better ability to have data running from mobile applications, just an explosion of data. Much more aerial photography, much more up-to-date data, roads, networks, all these things.

MITCH: And what is GPS just for those who might not know?

RICHARD: Global positioning systems.

MITCH: Okay. And that connects you up to the satellite systems?

RICHARD: Using satellites, it reports your, triangulates your position on the face of the earth and plugs that into live mapping systems. My wife has that in her car now.

MITCH: Well, what do you think, tell me some way that GIS has been used in a response that you guys have known personally that you have done. I remember a lot of maps in those years, 2004, 2005.

CALVIN: Well, in 2004 and 2005, one of the most common requests we were getting was show us the location of migrant labor camps. Show us the location of RV parks and trailer parks, because those are the people whose homes most likely in the storms were no longer there. So we could see where destruction was taking place. So when we were getting those requests a lot of times people were saying, put all the dots on the map for us. There comes an issue of scale when you are making a statewide map of putting all the dots on the map. That is not the best way to show the location. So what we were doing, we were aggregating the number of the facilities up to a geographic area. Maybe ZIP codes, maybe census tracts so we could see there were this number of affected trailer parks or migrant labor camps in this census tract or this geographic boundary.

MITCH: He mentioned the issue of data. Richard, what have you seen about the data? I mean, it seems like GIS is only as good as the underlying data. So you said data has gotten better. How good do you think that data is?

RICHARD: Well, better, but we need a lot more of it and the problem is in some cases there is almost too much data. We have so many different competing sources that are creating data and you can go look at this explosion of online web mapping applications through Google and Yahoo and Microsoft, these major competing companies and it is hard to drill down and get to the authoritative data sources through these things. You know, the data is available through GPS units in your car like you mentioned or on your mobile phone and, you know, who is really responsible for creating the location of that fire station or this shelter and, you know, is there consistency across the applications. So there is a challenge with, you know, kind of too much data at some point and really authoritative data sources.

MITCH: So the trick is to find the good data. I know, well, you were going to talk, Calvin.

CALVIN: Well, Mitch, what Richard was saying, there is no such thing as good data. A lot of times all the data starts from Tiger data which is from the U.S. Census Bureau and then it is cleaned up and repackaged and is sold under a different name. So when we get our data like for our roads data it might be a year old by the time we get it which isn't that good for planning emergency situations. So if we had a better relationship with our DOT or our local transportation, maybe we can get, maybe a weekly refresh of the data, because when you want a geocode to put points on the maps, like for our RV parks and our trailer parks we do a processing known as geo-coding where we take the address, when we get the address it generates a latitude and longitude value. If we don't have the most up-to-date roads layer we can't geocode.

RICHARD: There are also complications or challenges to statewide datasets like roads, in that jurisdictional and regulatory agencies have jurisdictional boundaries, just different regulatory agencies. Even the DOT here in Florida is not responsible from all of the roads. They are responsible for the State roads and then you have the Federal highway systems and the Federal DOT and what they're doing and, of course, the local governments and the road building there. So, you know, how do you get all of this different data into one place and make it up-to-date and readily to be used.

MITCH: You know, that question is one that I think Jennifer Williams has been working on a lot. I want to bring her into the conversation. Hazard and vulnerability assessment is one way that GIS is being used now that it was not even a few years ago. So can you tell us a little bit about what that is and how it uses GIS

JENNIFER: Hazard and vulnerability assessments overlay threats and hazards to a community and the vulnerable populations. Public health hazard and vulnerability assessments are unique because they look at, they look at populations that will be affected that have health outcomes, negative health outcomes. People like who are medically dependent on dialysis, oxygen dependent, socioeconomic status indicators, things like economics or poverty, their lack of transportation, their age and the elderly and the very young are most vulnerable.

MITCH: So you are getting all this data, you are trying to figure out what exactly vulnerability means. How do you map it? I mean, how do you put that on a map in a way that tells you how vulnerable a specific area might be?

JENNIFER: Well, CDC has helped us out with one tool and it is a human vulnerability assessment index and it takes a lot of the census data and creates community-wide indexes that show us where most vulnerable populations are. And then you can take those indexes and overlay them, say with flood zones or the path of a hurricane or a tornado, a chemical spill or a radiological event, and through some modeling, generate maps and lists of populations who may need the most help.

RICHARD: Some of the ways how you can model hazards and vulnerability to certain locations is like the tornadoes. You can take the point locations where the tornadoes actually down and if there is data available on the tracts, how long they were on the ground, but you can't, how do you work with that at a community level is you might use a grid or a raster-based analysis that would, you know, take the points and turn them into a cell value. So that if you entered a latitude and longitude or a ZIP code you could pull up a map that would show you based upon the current history of tornadoes where the most intense ones that touched down and where they touched down the most and then have kind of a grided cell values to determine, you know, where you may -- areas in Florida may be more susceptible to damaging tornadoes.

MITCH: So NOAH has one of those for hurricanes and tropical storms, right?

RICHARD: Right.

MITCH: But there is not one for tornadoes?

RICHARD: There is. We have several different projects in the Division of Emergency Management. The local mitigation strategies project is one where we are trying to create authoritative, definitive data sources for hazard analysis, like tornadoes, earthquakes, storm surge, wildfires, and, of course, we work with many different partners like with wildfires, with the Division of Forestry, storm surge with the National Hurricane Center and their SLOSH model and with Tsunamis and earthquakes we also work with NOAH and the National.

MITCH: National Oceanic and Atmospheric Administration.

RICHARD: There we go. And the USGS is another partner who helps us, you know, doing these natural analyses that we can then turn into local results. In fact, here in Florida right now, NOAH has just updated some digital elevation models and bathometers and terrestrial areas to plug into Tsunami updates.

MITCH: I don't even know what that means, but that sounds really, really impressive.

RICHARD: Tsunamis, waves generated by earthquakes.

MITCH: Not the Tsunami part, I understand all too well about the Tsunami and earthquakes. I mean, that point is just something -- what do you think is the biggest problem for GIS? I mean, all these advances seem to be happening really rapidly. All of you guys are working in geographic information. What kind of obstacles are you hitting?

CALVIN: I think one of the biggest things is getting the information to the proper people at the proper time and another thing is having them ask for the proper thing. When they say, show me a list of this. If I give it to them it is really not the information that they need to see. So it is on providing what they ask you for and then give them actually the information that they want or they should have asked you for. Another big issue in GIS is data connectivity. If there is a response somewhere, how do we get this information back to the people in the field who need it the most. And in the case with the hurricanes, the Internet was down and so is electricity. So the people can't access our network. So later on in the 2005 season we were doing things where we were giving them (INAUDIBLE) greater projects where we can load all of our facilities and all of our data information in a stand alone application that did not require a network connectivity. That way it could be used out in the field. Another issue is when we go up to the EOC and do mapping we have a firewall that is between the two of us. So the Department of Health can't share their data with the EOC, which is a pretty big thing. There is no essential data location for all of the State departments to post their data. We get data from AHCA but that is because I made a friend with the lady of AHCA and that is the reason why we get our data from.

RICHARD: There is a lot of ad hoc or personal, you know, one-on-one communication between State agency and other governmental units that would be responding in the EOC and, you know, that works to a degree, but as Calvin has mentioned, you know, we don't have sustainable structures for communication and breaking down some of the barriers to communicating with data. Another project that we are working on here in Florida right now is a strategic plan for statewide GIS coordination. Florida had some coordinated bodies and councils back in 1999 and 2000, but since then we have just been working, you know, informally amongst the agencies and accomplishing, meeting these challenges that keep rising to us.

MITCH: What do you think about the other side? So we have one problem which is sharing data in between State agencies. You talked a little bit about authority problems sharing good data with the public.

RICHARD: Yes.

MITCH: In terms of sharing data with the public, do you think there is a role to do more of that?

RICHARD: Absolutely, and I think this explosion of technology that the public is interfacing with, the Google Earth and Virtual Earth applications, you know, these are really good at helping you find hotels, get driving directions, but they're not going to tell you when you should evacuate or if you should evacuate. There is data that we can build and analyses and models and we can make them available through those easy to use, on easily understood applications, but there is still a challenge to be met there of going and turning data from, you know, an elevation upgrades or light our data collection, like ranging and detection mapping efforts to how do you get from high resolution elevation data into a SLOSH model and how is that actually output and turned into an evacuation area on the ground.

MITCH: Tell me what a SLOSH model is.

RICHARD: SLOSH stands for sea, land, something or another. I really don't know.

MITCH: We are drowning in acronyms. Give me a general sense of what it is.

RICHARD: The SLOSH model is used by the National Hurricane Center to model storm surge and inundation. So it is the computer system and data that they use to come out and figure out when and if a hurricane approaches what areas are likely to be flooded by that event. It gets used in the common term of storm surge zones. So that a category one storm surge zone is delineated on a map and you would expect the areas to the landward side of that to be flooded in a hurricane in the areas, you know, if it was -- and then in category, two, three, four and five all the way up.

MITCH: What about in hazard and vulnerability assessment? What do you think the value is in getting that information out to the public, or is it something that agencies should sort of use to make their own judgment about this stuff?

JENNIFER: Well, without, without scaring the public, I think you can share that information. There are security issues as well as, you know, we are very cautious about what types of information we want to make available to the public as far as chemical facilities or sites where radiologicals are stored, things like that. But using that caution, I think that the public could make use of HVAs and GIS knowing maybe where, say during a (INAUDIBLE), knowing where there are volunteer neighbor quarantines or where they can get, you know, medicines or pharmaceuticals. Where there are shelters in the closest proximity to their home or things like that.

RICHARD: In the security issues, I am trying to here in Florida kind of get some new vocabulary and get people to use the same vocabulary for critical infrastructure being these kind of facilities and data that are, you know, could be used by terrorists s or perhaps a terrorist wish list and those are the things that we want to keep secure, and obviously in the hands of the right people and out of the hands of the wrong people. But then we have other datasets that are called critical facilities and those are really the ones that are more important to emergency management. Shelters and where is food and where is water and where are communication towers and things that we definitely need to have available to help us do our job of responding to disasters, but we have got to be careful because you have a little bit of over reaction to the domestic security side, and if you don't start putting out where shelters are, well, that won't help the public who needs to find the shelters. You know, so you can't just lock up all the data. You need to walk a fine line between the domestic security concerns, and then obviously the public, the information that the public needs.

MITCH: And it seems to me that in my history looking at responses, if it doesn't get send out efficiently it is going to go out inefficiently through the press and that information might be incorrect.

RICHARD: Absolutely. And like I mentioned earlier, you know, there is layers on maps available through the Internet right now of inventories of schools and churches and other places where shelters may be located, but that doesn't mean the shelter is open. That doesn't mean that that particular shelter has been opened by an emergency manager in that county in response to that event.

MITCH: We have talked about some of the problems and issues that you guys are wrestling with and some solutions that you are working on. You know, let's look 10 years out or towards some better future of disaster response. What would be your kind of wish list for the way that you could have GIS working in a disaster response and recovery mode?

RICHARD: First of all, I think it needs to be more integrated. GIS is sometimes it is on beast and sometimes it is just information management, you know, or just information. We need to move from having GIS as it is own separate discipline to really just be information. To get there though is, means sometimes we like GIS is often thrown into information technology groups or divisions or, you know, where that is located in an agency. Sometimes that is convenient for people to do good GIS work, other times to do good GIS work you need to claim that you are separate from IT. So it is just whatever is advantageous at the moment is how you can claim it, but I think that eventually, you know, we have some sayings around the EOC, everything happens somewhere and we are there to help you figure out and get information about where that happened. Integration of GIS into our whole business planning, our local mitigation strategies, our State mitigation planning, our regional evacuation studies and plans.

CALVIN: Okay. I am working with Jennifer. We are trying to get our data more out to the public. We are currently working on a hazard and vulnerability website. That way all the information we have we can distribute out to the public. What we are going to be using is this product called ArcGIS Server. Currently there is no state data clearing house for our environmental health data, but if we can put this on an ArcGIS server and then create a web-based mapping application, that web-based mapping application can then be used by people in the EOC. So we are just trying to bridge the gap between all our different agencies. Other things we are working on is trying to get mobile applications out into the field. If there's an event, we're going to send disaster response teams into the field to see how people are doing and what kind of things do they need. Currently, we're working on Tablet PCs where they can go out and fill out checkboxes. Do these people need medicine? Do they need bleach to clean out their wells? We're trying to get that information up to Tallahassee so we can help them put in a proper request with the EOC. That way we can get the people the help they need. I haven't mentioned it but that's at a cost, and that's what we're working on right now. We're working on it now, so probably by next summer it will be up and running.

JENNIFER: Also, I'd like to see -  not being a GIS technician myself - tools for the counties that are just very user friendly. Without having a lot of training or any training at all.

RICHARD: Some tools we'd like to see rolled out in the future are the implementation of the US National Grid. This is particularly useful for search and rescue, and it was events like Katrina have definitely shown how important it is to be able to pinpoint your location, when you have a rescue swimmer hanging from the bottom of a helicopter with an axe so he can touch down on a roof and let people who are trapped in the attic out  - it's very important that you are at the right house. And those differences of five and ten meters can become very important. Latitude and Longitude is the system most people are used to for measuring coordinates. The US National Grid offers some advantages over that system. Including some scalability and the ability to finer pinpoint locations. This is also a system a lot of our response guys are already familiar with. It's based on the military system, so when your National Guard guys are rolling up and doing their jobs, this is how they're already reporting locations - or fairly similar to it. So, further implementation of the National Grid, getting more tools and the ability to communicate with that number system. And quit passing around addresses that may or may not be found by the particular geocoding system you're using. And even if they do - every sungle system out there finds my house on the wrong side of the street and up about three houses.

MITCH: I think we have some of our facilities that show up off the coast of Africa - am I right about that?

CALVIN: Yeah, if it can't find it, it's going to assign a Latitude/Longitude of 0, 0 - and some people don't know that that's a valid location. It's a coordinate off the coast of Africa. It's a common problem we've had.  In previous instances, we would geocode facilities, give them to our people as a list and tell them to go out. Well, if there are no streets signs or anything, it made it harder, even if they had latitude and longitude coordinates, if they weren't trained to use the GPS, that wasn't any benefit to them. What Richard was saying about the National Grid, it's a grid based system, so you can just look at it and figure out where it's going.

MITCH: You know what's interesting, I think if you asked a layperson how you make GIS better, they would talk about toys. Better GIS units for a car or something. But you guys have all talked about integration, a better backend and authority of the data.

RICHARD: Oh, we like gadgets, too.

MITCH: Okay. I don't want to get in the way of gadgets.

CALVIN: A lot of what we're doing in GIS is moving away from desktop applications where we have to spend hours and hours training people how to use them, to server based resources. If we get a hundred requests to make a certain map, we shouldn't have to make that map a hundred time. What we can do is make a model, and we put that on the server, and then they just have to press a button. Say click this button and give me the list of all the migrant labor camps in my county. By putting it out there, we make a product that we and other agencies can consume...

JENNIFER: Over and over

CALVIN: Over and over again...

MITCH: You know, one particular model that I wanted to brag on is the wild bird disposal model. Could you talk a minute about how that walks?

CALVIN: Yeah, me and my colleague Judith Ippolito, we got a request from a county health department that they were trying to find the location to bury birds in case of an outbreak of Avian Flu. We were given the criteria that the bird had to be buried at least two feet above the seasonal high water table, not within the hundred year flood plain, and that it had to be covered with two feet of soil after it was buried. So, we used GIS - we took data from the USGS, all this data - and we made locational areas all over the state of where birds can be buried in case of an outbreak of pan flu. The issue was that people in South Florida were saying that they don't have areas where they can bury birds because the water table is so high. Because the water table might be two or three feet underground. So, we have to find other ways - do we incinerate the birds? do we ship the birds to another county?  And that's one of the models that we have available.

MITCH: So, to me, that's something that is really practical and that there's no way that you could do without a Geographic Information System - blending several different kinds of data in a way that allows you to have a real response impact.

RICHARD: Also with models...models can be very complex. I think people need to look more at the circle of models to data or maybe data into models, but then made readily available through applications. We sometimes build great applications or tools, but if they're not using good data, then it doesn't do you much good. So there's definitely this vicious circle there, and I think we move back and forth, sometimes investing in the data, but not creating a tool to make that readily available or usable. Or we build these tools and then you don't have good data to distribute with it. An example where we're getting this more right is the Florida HAZUS user's group. And HAZUS is an application, kind of an extension for ArcGIS that FEMA distributes for emergency managers and others for working with disasters and doing planning and also helping make plans for response with these tools. So with the HAZUS User's Group we're working together to get good data to put into this tool that works well. We're also holding training classes and communicating better to support each other through the use of that application . So it's just a real good example of coordination here in Florida.

MITCH: That's great. The eternal pendulum swinging between data and tools. It sounds like we're now at a place to really bring this data together to feed these great tools that you're building. Thank you, Calvin. Thank you, Jennifer. Thank you, Richard. Thank you very much for joining us for the EH Prepcast.