Geospatial apps help temper Mother Nature's fury
Agencies find that fusing webs of remote sensors with geospatial management systems is the next best thing to being there
- By John Moore
- Mar 05, 2010
Last year, the southern African nation of Namibia experienced its worst flooding in decades as the Zambesi and other rain-swollen rivers rose more than 25 feet and inundated several regions of the continent. The flooding caused large-scale destruction to homes, schools, health facilities, mahangu and maize fields, and infrastructure.
A result of heavy rains in neighboring Angola and parts of Zambia’s Western Province, the flood also displaced more than 300,000 people and contributed to cholera and other disease outbreaks. The disaster in March 2009 followed a similar season of flooding in 2008.
In this report
Geospatial preparedness checklist
A mappable mashup
This year, Namibian officials hope to get a head start against catastrophic weather situations. Their approach is to create a geospatial application that taps satellite imagery and river-height sensors and get an early read on when and where the flood waters are coming — helping them decide where to deploy the right resources. An international team of experts, including representatives from NASA and the National Oceanic and Atmospheric Administration, are contributing their expertise in satellite mapping and sensor technology.
“Basically, we are trying to build predictive models for floods,” said Dan Mandl, a senior computer engineer at NASA’s Goddard Space Flight Center.
Predicting floods might be a little easier than predicting earthquakes, but Namibia’s project exemplifies an emerging, though largely untested, set of geospatial applications that are still in their technological infancy but promise to have many life-saving uses across the world. Indeed, across all sectors of society, geospatial experts are rapidly establishing a sensory connection between information systems and the real world. And that has big implications for government.
With little fanfare, geospatial information derived from sensors is already helping decision-makers better understand the context of problems that they face on a daily basis. Nearly anything in the physical world that needs to be managed has a location in space and time and can be measured. And it needs to be measured if it is to be managed well.
Sensor technology, a category that includes devices as diverse as satellites and seismic monitors, is leading to sensor integration, which promises to lend geospatial technology more of a real-time edge while providing the benefit of observations made over time — a plus for tracking environmental trends.
DHS considers contracts for collecting disaster-related imagery
Where in the world is the geospatial data?
Web 2.0 increases demand for data visualization tools
NOAA, Google partner on data visualization project
So land-based sensors are proliferating at an exponential rate, and they are connecting not only to orbiting satellites but also to the Web. The resulting ability to fuse location and sensor information to remotely monitor and forecast conditions and operations in the physical world enables managers to do more — and produce results faster and at less cost.
"It really allows the decision-maker to have a real-time, or near-real-time, understanding of what is going on in a location context," said Mark Reichardt, president and chief executive officer of the Open Geospatial Consortium (OGC).
A number of developments have converged to make the latest crop of geospatial applications possible. Key enablers include the growing availability of Web-accessible sensors, proliferation of wireless networks, and development of standards that make sensor integration faster and cheaper to accomplish.
Government agencies are trying to take advantage of the potential of linking sensors and other geospatial data in applications such as emergency management, environmental monitoring and homeland security.
Flood Prediction and Public Health
The trick with sensor integration is finding practical applications, Goddard’s Mandl said. The key question to ask is: Does the system contribute to an organization’s ability to make informed decisions?
“We are doing experiments to find out what is useful and compelling,” he said.
Flood risk management is one promising area. In Namibia’s pilot project, NASA is part of an international team that plans to use satellites to aid in flood prediction and help regional governments deal with flood-related illnesses such as malaria.
The approach relies on observations from the Tropical Rainfall Measuring Mission (TRMM), a program jointly managed by NASA and the Japan Aerospace Exploration Agency. The TRMM satellite uses precipitation radar to measure the intensity and distribution of tropical rainfall. The resulting images are used to create a predictive flood model.
When accumulated precipitation, river flow or both reach a certain threshold in an area — in this case, the Zambesi river basin — other satellites come into play. Images from Earth observation satellites focus on upstream areas, showing the width of the river, and river-height gauges linked to the Internet provide additional confirmation of flood potential. The satellite and ground-based sensors, coupled with historical flooding data, let government officials produce a flood risk map and issue alerts, Mandl said.
Because a flood wave takes several days to move downstream, the prediction system could buy precious time for emergency managers. Additional flood mapping from satellites is used to improve the understanding of the situation on the ground for response and management purposes, said Guido Van Langenhove, head of Hydrological Services Namibia.
Mandl characterized the system as experimental, and he said additional testing will take place this year.
He said he also envisions that flood prediction and mapping could help assess the probability of disease emergencies. Satellite imagery of a flood could be compared with previous flood maps and epidemiological data to issue alerts about the risk of a disease outbreak. For example, government authorities would be able to deploy resources ahead of time in regions vulnerable to a malaria outbreak.
However, fielding such a prediction system faces a number of problems. For example, satellites tend to have specific strengths and weaknesses. Some can produce high-resolution images but can only survey an area once every 16 days. Others can provide daily observations but at a lower resolution. Optical satellites can’t penetrate clouds, so the use of radar satellites might be necessary.
“You need combinations of satellites to get the full picture,” Mandl said.
For the most part, requesting satellite images has been a manual process. But the availability of workflow software and standard interfaces has made it somewhat easier. Campaign Manager, developed at NASA’s Goddard Space Flight Center, lets users send requests to satellites via OGC’s Sensor Planning Service, which is part of the African pilot project. It helps identify the satellites available to observe a given area. Of course, the approach works only for satellites connected to the SPS interface.
In the future, satellite tasking might be more automated. Mandl said the vision is to automatically engage the appropriate satellites, with the TRMM-based flood prediction model serving as the trigger.
However, security remains a problem with automated sensor tasking. For example, access to SPS occurs via the Internet. To counter hacking threats, NASA’s experiment uses multiple-factor authentication. “We can’t expose satellites to the world,” Mandl said.
Closer to earth, the National Weather Service is fusing ground-based radar with geospatial data.
NWS taps about 200 radar stations for sensor data. In 2006, it began offering an enhanced view of its Doppler images. The Radar Integrated Display with Geospatial Elements was NWS’ first integration of data from radar and geographic information systems. RIDGE provides data for the radar displays on NWS Web sites, and customers can also incorporate the data into their own GIS applications.
The latest iteration, RIDGE 2, makes that process easier. Slated to become operational in April, it will provide geospatial data as services. That’s a major change from the previous version, which provided customers with a projection that was difficult to display on their GIS systems.
“We didn’t have a good way for them to interact with the data,” said Paul Kirkwood, chief of the Technology Infusion Branch at the NWS Southern Region Headquarters’ Science and Technology Services Division. “With RIDGE 2...they will be able to better integrate with their own GIS services.”
RIDGE 2 houses radar images, weather warnings — which are displayed as polygons covering the affected area — and precipitation data in a geospatial database. NWS makes the various types of data available as services so customers can choose the data they wish to see on their maps.
Customers take different approaches to accessing the data. Kirkwood said NWS’ core partners, such as the Homeland Security Department and Federal Emergency Management Agency, obtain their data via ESRI’s ArcGIS Server. The public can get a version of the RIDGE 2 data that NWS has created using OpenLayers, which is free, open-source software that displays map data by using a standard Web browser.
Kirkwood cited the vast number of sensors involved as the biggest difficulty in making the geospatial system work. “The largest challenge we have is figuring out how to incorporate 200 radars with an update cycle of up to every 5 minutes,” he said, adding that each radar generates more than 10 services that require updating. “Building an infrastructure to handle this is very challenging.”
NWS is testing ESRI’s ArcGIS Server Image extension to address the updating issue, he said. It enables the processing of large numbers of images, according to ESRI. On an average day, NWS generates and delivers about 24,000 images per hour.
The Pacific Disaster Center in Hawaii harnesses a range of sensor and other data to help government agencies prepare for trouble.
PDC disseminates geospatial data to emergency managers via a controlled-access Web site named Emergency Management Operations (Emops). Emops’ underlying architecture, DisasterAware, pulls data from networks of seismic sensors, water-level gauges, satellite image systems and other authoritative sources, said Chris Chiesa, PDC’s chief information officer.
The data can be viewed on top of a range of base maps, including Google Earth, Bing and Yahoo imagery.
Together, those sensors provide early warning and situational awareness for hazardous events, such as hurricanes, earthquakes, tsunamis, wildfires, floods and even pandemics, Chiesa said.
Emops users include civil defense and humanitarian assistance agencies at the federal, state and county levels. To support their users’ emergency management missions, PDC taps into NOAA’s Pacific Tsunami Warning Center and the sensor network run by the U.S. Geological Survey’s National Earthquake Information Center. PDC also receives data from TRMM and incorporates fire and flood information from NASA’s Moderate Resolution Imaging Spectroradiometer.
PDC makes a public version of the geospatial information available via the Global Atlas on the center’s Web site.
Authorized Emops users will soon be able to contribute to the data collection effort via Apple iPhones. PDC is releasing an iPhone application, iAware, that lets users file reports and upload photos from the field, Chiesa said.
In the meantime, the task of processing data from multiple sensors ranks among the top issues in running Emops.
“A key challenge is to transform the ever-growing data stream into actionable information for effective decision-making, typically in a chaotic and time-constrained environment,” Chiesa said.