Modeling Remote Sensing Satellite Collection Opportunity Likelihood for Hurricane Disaster Response

Article excerpt


The natural disaster paradigm is often portrayed as a cycle--from the warning stage as the event approaches, the event followed by the response stage, subsequent recovery, and then a relatively more leisurely planning and mitigation stage. Considerable GIScience research has been conducted for the warning, recovery, or mitigation stages. A highly "visible" yet poorly understood or researched problem is the use of remotely sensed imagery during the short response stage. The emergency response stage for most natural disasters is very short. State and local agencies involved in emergency response to natural disasters such as hurricanes have explicitly indicated they need imagery covering the disaster area within three days of the event; and more desirably within 24 hours of the event (Hodgson et al. 2010). During the post-event disaster response remotely sensed derived damage information becomes less important as time passes and in situ data (deemed more accurate) becomes available (Hodgson et al. 2010). Airborne image collections have often been used but suffer from several problems, most noticeably the long collection time required for larger areas. The use of remote sensing satellites carrying high spatial resolution sensors has often been touted (Visser and Dawood 2004; Zhang and Kerle 2008) as the logical response for rapidly collecting post-disaster event imagery for emergency response. While more coarse resolution satellite sensors may provide some information that is useful in emergency response, agencies typically desire high spatial resolution (e.g., 1 m or less) to assess structural and transportation feature damages. Thus, this article focused on the use of satellite imaging opportunities with high spatial resolution sensors on the order of 1.5 x 1.5 m or less.

Satellite image coverage area within a swath (e.g., > 8 km width by an infinite length) is much larger than the footprint of airborne (except the U-2 high altitude aircraft) imaging sensors. Collection times for most remote sensing satellites are -7.5 km/sec, allowing for impact areas to be imaged in a few seconds. The downlink and initial post-processing by satellite image providers can be within a few hours, compared to six to eight hours (or much longer for post-processing and geometric rectification) for airborne imagery collection, processing, and transmittal.

Unfortunately, satellites are maintained on fixed orbits. The repeat interval for remote sensing satellites carrying high spatial resolution sensors, even with pointable sensors, is on the order of several days, depending on the latitude of the area of interest. It may be that the next available overpass is two or more days after the disaster event. Fortunately, more than one operational satellite carries high spatial resolution imagery. The availability of multiple collection platforms may provide the appropriate set of opportunities to collect imagery over disasters at any time. However, these satellites are still on fixed orbits. No known study has examined the collection opportunities provided by high spatial resolution satellites for disaster response. The problem is not simple. Such satellites carrying high spatial resolution satellites are not on systematic repetitive tracks whereby future passes can be made from path-row maps, such as Landsat or SPOT. The pointable (off-nadir in particular) sensor systems onboard many earth resources satellites today offer somewhat flexible collection opportunities. Prediction of future collection opportunities would require modeling satellite orbital trajectories and somehow incorporating unsystematic maneuvers where the satellites are repositioned. Furthermore, the observational record of many high-impact but low-frequency events, such as hurricanes and earthquakes, is weakly suited to developing well fitted spatial/temporal probability functions for predicting future occurrences.

The goal of this research was to develop a generic approach for empirically assessing the likelihood of collecting satellite imagery given a spatial/temporal distribution of disaster locations. …