Figure 1. DMSP-OLS detection of power outages the night of January 1, 2005 for the northern tip of Sumatra.
Color composite image generated using the percent frequency of cloud-free light detection in 2003 as red, the visible band data from January 1, 2005 as green, and the corresponding thermal band data from January 1 as blue. Areas having heavy cloud cover were masked as black. Areas with power outage on January 1 show up as red. Areas with power present show up as white to yellow.
During the past 100 years electric power has become the principal source of nocturnal lighting on roadways, commercial zones, industrial sites, and residential areas. Typically electric production is centralized and the power is transmitted using a network of power lines and distribution wiring. Because many of the power lines and wires are above ground, they a susceptible to damage, especially during severe weather events, such as hurricanes, typhoons, thunderstorms and in some instances snow or ice storms. Less frequent events, such as earthquakes, can also produce electric power outages, either directly or by earthquake generated tsunamis. Another class of electric power outages, termed brownouts, result from an insufficient capacity to generate electric power. In this case, the electric power supply becomes intermittent.
Reporting on power outages is typically limited to eyewitness reports of individuals within the outage area. Because many types of communication (radio, television, telephone, facsimile) are also impacted by power outages, information on the extent of power outages during a disaster may be very incomplete. Having an observational basis for delineating the extent of power outages could provide valuable information to guide relief, repair and cleanup efforts.
We propose to make use of satellite observed nighttime lights to operationally detect power outages. Initially we will use nighttime satellite data from the Defense Meteorological Satellite Program (DMSP) Operational Linescan System (OLS). Later we anticipate making use of nighttime lights data from the Visible Infrared Imaging Radiometer Suite (VIIRS), scheduled to fly on NASA's NPP satellite in 2007.
The Defense Meteorological Satellite Program maintains a constellation of two satellites in sun-synchronous, near-polar orbit at altitudes of approximately 833 km, an inclination of 98.8 degrees, and an orbital period of 102 minutes. One satellite is in a dawn - dusk orbit, the second in a day - night orbit. The DMSP platforms are three axis stabilized, with roll, pitch, and yaw variations kept to within +/- 0.01 degrees. The NOAA National Geophysical Data Center (NGDC) serves as DoD's archive of data for the DMSP sensors. The U.S. Air Force Global Weather Central sends DMSP data to NGDC using a high speed data line. The DMSP archive was established in March of 1992, and began receiving data on a daily basis in September of 1992 and has operated continuously since that time. At NGDC the DMSP data are decompressed, deinterleaved into separate orbit length files for each sensor.
The DMSP Operational Linescan System (OLS) is an oscillating scan radiometer designed for cloud imaging with two spectral bands (VIS and TIR) at 2.7 km pixel resolution and a swath of 3000 km. The VIS bandpass straddles the visible and near-infrared portion of the spectrum with a full-width-half-maximum (FWHM) of 0.58 - 0.91 um. The TIR band has a FWHM of 10.3 - 12.9 um. The wide swath widths provide for global coverage four times a day: dawn, day, dusk, night. The VIS band signal is intensified at night using a photomultiplier tube (PMT), making it possible to detect faint VNIR emission sources. The PMT system was implemented for the detection of clouds at night. With sunlight eliminated, the light intensification makes it possible to detect city lights, gas flares, and fires.
The potential use of nighttime OLS data for the observation of city lights and other VNIR emission sources was first noted in the 1970's by Croft. NGDC has produced several global inventories of the stable light sources which can be observed with the nighttime OLS data. The procedures involve the detection and geolocation of VNIR emission sources and clouds from a large number of nighttime OLS images and the use of image time series analysis to distinguish stable lights produced by cities, towns, and industrial facilities from ephemeral lights arising from fires and lightning. The time series approach is required in order to ensure that each land area has been covered with sufficient cloud free observations to determine the presence or absence of spatially stable VNIR emission. The stable lights datasets being produced by NGDC are gray scale images depicting the percent frequency of cloud-free light detection and average digital number for lights.
Once the stable lights data set has been derived, it is possible to overlay the lights from an individual orbit to detect new light sources (e.g. fires) or locations where lights are expected, but not observed (power outages). The OLS thermal band can be used to mask out areas with heavy cloud cover, where the detection of lighting is blocked. Typically the results are visually interpreted. Figure 1 shows an example of DMSP-OLS power outage detection from January 1, 2005 on the northern tip of Sumatra - an area devastated on December 26, 2004 tsunami.