Post-Extreme Event Damage Assessment and Response for Highway Bridges (2016)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

13 could have been avoided if more objective assessment tech- niques had been used (Phares et al. 2004). This becomes more important specifically where regionally distributed extreme events require the minimums of connectivity and accessibility in the transportation network. Hand-Held Nondestructive Testing Nondestructive methods using handheld devices were ranked second by the state bridge engineers for damage assessment. Forty-six percent of responding states mentioned that they use one or more of NDT techniques for damage assessment. However, it can be surmised that some of the respondents did not differentiate between the application of the techniques after extreme events for rapid detection and operational condi- tions. The NDT methods mentioned by state bridge engineers included magnetic particle testing (seven states); dye pen- etrant testing, ultra-sonic testing, and sounding with a hammer (four states each); chain drag (two states), and rebar scanner (one state). Sonar Surveys Sonar surveys were ranked as the third most used technique by state bridge engineers (26%). The survey for hydraulic engineers separated the sonar surveys into the following categories: portable (35%), manned and unmanned (21%), and fixed (14%). As such, the responses from hydraulic engineers revealed that among these three categories, sonar surveys are used by 70% of the respondents, making it the second most used technique by hydraulic engineers. Sonar systems or echo sounders can be portable, fixed, or manned or unmanned systems. For fixed sonar scour moni- toring, the device is mounted on the pier or abutment for con- tinuous streambed elevation data collection. The mechanism includes a transducer that sends signals that are reflected back, from which the elevation of the streambed can be esti- mated. The advantage of fixed sonar is that it provides continu- ous data during and immediately after high floods, assisting with estimating the extent of scour depth. The application of fixed sonar surveys may be limited, considering its low depth tolerance and that installation depth and the resolu- tion distance between the head and the interface affects the reading. Therefore, if the instrument is not mounted at a loca- tion directly above the point of interest it will provide a false This chapter provides a brief overview of the damage-detection technologies identified by the state bridge and hydraulic engi- neers. A short synopsis of the applications of these techniques, their capabilities, and possible shortcomings in rapid damage detection is provided. This is followed by an overview of more advanced technologies that have been developed or are under development and that have the potential to contribute to the emergency damage assessment techniques. The goal of this section is to initiate discussion on the new opportunities, their applications, and the possible ways they can be inte- grated into the current practice for bridge damage assessment in emergency conditions. CURRENT PRACTICE Visual Inspection A review of the responses provided by the state bridge and hydraulic engineers shows that visual inspection is the most commonly used method of damage detection among DOTs (100% of respondents from both surveys). All of the respon- dents reported that they follow some form of cursory or at- arm-reach visual inspection that helps them assess the extent of bridge damage. Most of DOTs identified visual inspection as a satisfactory tool for making immediate decisions on the structural integrity of a bridge. Bridge inspection personnel must be properly trained to identify any possible damages to a bridge quickly and accurately. Based on what can be learned from the visual inspection, the bridge inspection personnel can then determine whether a more in-depth inspection is required. The reasons that make visual inspection a desired approach are the immediate results, lower costs, and limited amount of preparation or special skills required. However, there have been concerns regarding the reliability of the visual inspections resulting from different factors such as variations in state prac- tices and inspector’s use of inspection methods (Moore et al. 2001). Furthermore, such an inspection is only able to detect visible larger defects, which opens up the possibility of serious invisible defects going undetected. This is in addi- tion to the shortcomings of the method in hard to reach sites where access routes have been demolished in the aftermath of an extreme event. Furthermore, the results are highly dependent on the experience and training of the inspector, which increases the possibility of judgmental errors, underlining the importance of the training of the inspection crew. In addition, the subjec- tivity of the method may result in unnecessary closures that chapter three BRIDGE DAMAGE DETECTION TECHNIQUES: CURRENT PRACTICE AND FUTURE DIRECTIONS

14 reading (Briaud et al. 2011). An approach to overcome this limitation would be to use side scan; however, high costs limits its use. Portable sonar survey systems can be divided into sector side scan, scanning, and multi-beam sonars. The primary difference between scanning and multi-beam sonar is the elimination of the time lag required to rotate the transducer from one side of the vessel to other. Because of its position (behind and below the vessel), side scan is one of the most accurate systems for mapping a large area of the riverbed. It can quickly and efficiently generate detailed images of the channel bottom regardless of the water clarity. Because of their portability, they have the advantage of being deployed at any location of interest during floods and provide accu- rate streambed elevations, in addition to collection of large amounts data rapidly and over a large area. The accuracy of sector and multi-beam sonar depends on the position or array of the transducer. Error in the transducer angle is reflected in a substantial error in the displacement measurement. Water temperature, acoustic frequency, high turbidity, air entrainment, and heavy suspended-sediment loads are other factors that could affect the accuracy of mea- surements. Considering the large amounts of data that could be collected during each survey, fast computational processing of data and high capacity storage is required. Measurements made at acute angles to the streambed can be used to measure the scour under ice and debris jams. A disadvantage of the side scan is that it is labor intensive and most do not provide depth information. Also, it cannot provide detailed visual images of the vertical components of submerged structures. Because of safety issues, unmanned systems are recommended during floods. Figure 5 is an example of an image created using sonar surveys in addition to the schematics of the technique in fixed sonar surveys. Photogrammetry Photogrammetry was identified by 17% of responding bridge engineers as a solution for measuring bridge deformations. In this technique three-dimensional geometry of physical objects is obtained by measuring and analyzing two-dimensional photographs taken from different perspectives (Jiang et al. 2008). It can be used for monitoring structures under load- ing. Traditionally, linear variable differential transformers (LVDTs) have been used for deflection measurement of bridge elements. However, their application is limited due to the restriction of the accuracy of the measurement to the linear range of the sensor and the capability of only provid- ing one-dimensional information (Valença et al. 2010). With the advancement of digital cameras, photogrammetry offers a suitable alternative for bridge deformation monitoring by providing visual recording during testing (video camera). Furthermore, the technology is low-cost compared with LVDT. When compared with LVDT test results, photogram- metry has yielded high accuracy in a laboratory setting (Woodhouse et al. 1999; Mills et al. 2001; Whiteman et al. 2002). In most cases, the field applications for bridge defor- mation measurements have been successful (Jáuregui et al. 2006; Detchev et al. 2011). The accuracy of photogrammetry in bridge deformation monitoring depends on camera positioning and the placement of target points. The camera can be stationary or movable. Two target types are necessary: static targets and movable targets. Before making the measurement, the camera is calibrated for radial lens distortion coefficients, principal distance, principal point offset, and sensor size determina- tions. Figure 6 is the schematic for test setting in the field applications. Inclinometer and Tiltmeter The use of an inclinometer or a tiltmeter was reported by 17% of the responding bridge engineers. These devices are used to measure the inclination of the entire bridge or the element it is attached to resulting from loading. Single axis and dual axis tiltmeters have been used in the field (Figure 7). An accu- racy of 0.005° was reported by Kulchin et al. (2004). Kosnik and Ng (2010) reported the use of a tiltmeter to monitor scour at Stony Creek Bridge, California. The advantages of these instruments compared with other scour monitoring devices are invulnerability to damage caused by floating debris, the FIGURE 5 Sonar surveys: (left) an example of river bed surveyed using the technology; (right) basics of the technology (Browne 2010).

15 robustness of the measurements even in high turbulence conditions or with accumulation of debris, capability of tak- ing measurements from both longitudinal and transverse inclination, and ease of installation. However, establishing a critical tilt angle at which the alarm would go off is hard as the bridge is always moving due to different type, of environ- mental and loading conditions (temperature variation, wind, and hydraulic forces). Ground Penetrating Radar Fifteen percent of the responding bridge engineers reported using ground penetrating radar (GPR) for damage detection after extreme events (Figure 8). This technique operates by using electromagnetic signals that are sent by a source antenna through the solid structure before the reflected pulses are picked up again by a receiving antenna for later processing. Photogrammetry target FIGURE 6 (Left) Application of photogrammetry on bridges (Albert et al. 2002); (right ) example of the photogrammetry targets. FIGURE 7 (Left) Tiltmeter in the field; (right) close up of the tiltmeter box (Hunt 2009). FIGURE 8 (Left) GPR device; (right) application in the field (Alani et al. 2013).

16 The data obtained from GPR systems are used to generate the tomography of the bridge deck. The quality of the tomog- raphy depends on the data sampling, the longitudinal res- olution, and the distance between successive scan lanes, and in case of multiple antenna arrays, the lateral resolu- tion will depend on the distance between two consecutive antennas (Benedetto 2013). The application of technique in extreme conditions appears to be limited to instances where damage to the bridge decks is expected during such events as a fire. Accelerometers Accelerometers are being used by 13% of the responding bridge engineers. Such devices have been used to monitor the vibration of a structure. Data acquisition can be done wirelessly or through a hardwire connection. A network of accelerometers is used to collect information on bridge behavior under loading conditions (Figure 9). California DOT (Caltrans) has a network of instrumented bridges to monitor their performance during earthquakes. A series of different analyses such as modal analysis, Fast Fourier Transforms, and nonlinear time history analysis could be used to mea- sure the natural frequency of a bridge, detect and locate damage, and even predict the remaining life capacity. In addition, the displacement (deflection) and velocity can be extracted by integrating the acceleration, assuming a lin- ear response. With the development of more robust and less expensive packages, the technology is low-cost and installa- tions relatively simple. Despite these advantages, issues such as maintenance, numerous false errors, and the availability of a continuous power supply were reported by many of the state engineers as disadvantages that have hindered a wider application of this technique. Other limitations include the requirement for placement of the sensor away from cracks and the treatment of ambient noises for the low to medium frequencies. Float-Out and Tethered-Buried Switches The use of float-out and buried switches was reported by 7% of responding hydraulic engineers. These devices are transmitters buried at predefined depths. Once the scour reaches its buried depth, the devices float-out to surface and their information is transmitted to a receiver that registers the location of the scour information. Because these devices are buried at specific loca- tions they can only provide local scour depth information. In addition, the internal radio transmitter is only triggered when the instrument is in the horizontal position. As such, the local scour depth before the release of the device is not known until the scour reaches a depth greater than the datum (Figure 10). During installation more than one float-out device can be inserted in the same hole at various depths. A float-out device is easy to install and inexpensive. Its use is limited because of maintenance issues as well as the reporting of numerous false alarms. The device cannot be checked for operational capabil- ity and the on-board power must be reliable for long periods of time without use. A tethered-buried switch is a type of float-out device that gives an indication of the scour development around a bridge foundation. It is hardwired to the data acquisition system and the electrical switch is triggered when the rod is turned into a horizontal position as the scour depth reaches the buried depth of the device. The electrical switch can also be triggered when the wire is broken as a result of an act of vandalism or damage from debris. Tethered-buried switches can only provide maxi- mum local scour depth. The advantages of this instrument include ease of installation, low costs, and an open-ended life span. Sliding Magnetic Collars Although this technique was not mentioned by any of the hydraulic engineers, it was confirmed that it is being used by FIGURE 9 (Left) A wired accelerometer installed on the bridge; (right ) an accelerometer kit in the field.

17 many state DOTs. A buried or driven sliding magnetic col- lar is a hollow pipe that is driven or drilled into the stream- bed at scour-critical locations. A sliding magnetic collar is placed on the pipe and rests on the streambed. The elevation of the collar is detected by a probe outfitted with a magnetic switch (Figure 11). The downward movement of the collar as a result of erosion triggers the magnetic switches located at constant intervals along the pipe. The collar depth reading can be manual or automated (Lu et al. 2008). Sliding collars are considered a fixed monitoring scour device and can only provide maximum scour depth. If the scour hole refills at flood recession, these instruments become buried. A continuous monitoring of the streambed can be obtained dur- ing and after a major flood event. Also, the scour depth reading is only provided at the location of the device. As such, many devices must be placed around the pier in order to measure the global scour effect. However, the device is not capable of registering the refilling of the hole that occurs during flood FIGURE 10 (Left) Schematics of the float-out device; (right) close up of the float-out devices (Hunt 2009). FIGURE 11 (Left) Schematics of the mechanism (Haas et al. 1999); (right) a sliding magnetic collar at the bridge site (Avila 2006).

18 recession. Compared to other forms of scour monitoring, the magnetic collars are more effective for floods with high sedi- ment concentration. However, the conduit required for manual reading in the sliding magnetic collar is susceptible to damage from debris and ice accumulation (Lagasse et al. 1999). FUTURE DIRECTIONS A review of survey responses and the follow-up interviews revealed that most states, with a few exceptions, do not take advantage of emerging technologies for post-extreme event damage assessment. Disaster response requires an assessment of damaged infrastructure as quickly as possible; however, data collection can be dangerous in an area after a natural disaster. In addition, it may not be possible to rally enough survey teams to cover a large disaster area. Traditional disaster assessment practices involve both detailed and rapid ground surveys, but these practices can be limited by timeliness. Infor- mation provided by remote sensing technologies has proven to be beneficial to detecting and locating damage. Remote sens- ing is used to acquire data on the structures without making physical contact and it is a commonly used method in disaster assessment and recovery, especially for buildings and geotech- nical features, as a result of its adeptness over traditional site observation. Technologies associated with remote sensing are able to obtain post-disaster data over a large area more rap- idly compared with in-field surveying. The speed of remote sensing allows for quicker decision making and quicker action on necessary repairs on infrastructure. Spatial data, including maps, aerial photography, satellite imagery, global positioning system (GPS) data, and rainfall data is an important compo- nent of disaster management. The different data types require a co-registering process that will bring them to a common map-basis. Image registration or co-registration is a process of converting different sets of data into one coordinate system, which is necessary to integrate and compare the data. The vol- ume of data can be overwhelming if handled through manual methods; however, gathering and organizing technologies such as remote sensing and geographic information systems (GIS) have improved in efficiency. In this section three major techniques of remote sensing will be discussed: satellite imag- ery, Light Detection and Ranging (LiDAR), and unmanned aerial vehicles (UAVs). Satellite Imagery Modern high-resolution satellites have progressed greatly in imagery over the past two decades. Optical satellites and syn- thetic aperture radar (SAR) satellites offer cutting edge data. The satellites are able to provide quick damage detection and change in elevation of an area’s surface. Both optical and SAR satellites have their strengths and weaknesses when utilized in damage assessment; however, when combined the pair can remedy each other’s shortcomings. Optical remote sensors measure radiation caused by reflected sunlight. Optical sensors utilize visible, near-infrared, and shortwave-infrared sensors to detect solar radiation. The sen- sors use the data collected from solar radiation reflected from targets on the ground to create images. Different materials absorb and reflect solar radiation at different wavelengths so targets can be determined based on spectral reflectance signa- tures. Although it is a common method in disaster assessment and recovery, the efficiency of the optical imagery acquisition process is often diminished during the early stages of a large disaster, such as an earthquake, because of communication interruption. SAR is a sensor mounted on a moving platform used to gen- erate two-dimensional or three-dimensional (3D) representa- tions of objects. It is an advanced form of an oblique viewing (side-viewing) radar. Consecutive pulses of radio waves are transmitted from the radio to detect a target, and then the echo, or bounce, is recorded. As SAR moves, the antennae location corresponding to the target changes accordingly. SAR employs the flight path of a satellite to create a long synthetic antenna to generate a high-resolution image of the earth’s surface derived from the overlaying of several images. SAR data have lower spatial resolution compared with optical remote sensing data; however, in recent years the spa- tial resolution has been greatly enhanced. The improvement in SAR image resolution has resulted in an increase in its use of post-disaster damage assessment. Also, high-resolution SAR data have become more easily accessible to the public through the Internet. Satellites, such as the Cosmo-SkyMed and TerraSAR-X, are able to obtain impressive images at the decimeter scale. SAR satellite images can be used to compare pre- and post-disaster images to map out destruction areas and recovery efforts. SAR technology has been extensively used to estimate the regional damage to buildings after different natural disas- ters (e.g., Matsuoka and Yamazaki 2004, 2010). Except for a few studies, the technology has not been widely used for bridge damage detection after extreme events. One of the instances of using SAR for bridge damage detection was a research project funded by Caltrans that studied the appli- cation of remote sensing technologies in post-disaster damage assessment of highway systems. The proposed approach uses a two-phase damage detection algorithm devel- oped for highway bridges: Bridge Hunter and Bridge Doctor. Bridge Hunter locates the bridges of interest in remote sens- ing coverage. For this purpose the state bridge management systems available at the time such as PONTIS were used to enhance the accuracy of the collected bridge data from the imagery technique. Considering the large inventory of bridges, the process was automized through locating bridges from fed- erally managed databases, such as National Bridge Inventory, using a dynamic segmentation process. For this purpose, the procedure starts with registration (through template matching), subsetting, clipping, and then locating bridges on airborne and satellite imagery. The algorithm generates a suite of clipped images accompanied by the bridge attribute information. The

19 second phase of Bridge Doctor diagnoses the bridge damage state of “health” in terms of the spectral change within a tem- poral sequence of imagery acquired before and after the event. In this approach, changes between before and after scenes were identified in terms of difference and correlation values, mea- sured along a central transect running along the highway and across the designated bridge. Figure 12 shows the application of this process for a sample bridge. Optical Sensors versus SAR Optical sensors employ wavelengths near that of visible light or 1 micron, whereas SAR uses a wavelength of 1 cm to 1 m. The difference in wave lengths between technologies yields the results of the data. SAR is able to collect data despite cloud cover, but optical sensors cannot make observations above cloud cover. The wavelengths also affect how objects on the earth’s surface are portrayed because the scattering and reflection mechanisms are dependent on frequency and wavelength. Optical sensors rely on the sun’s illumination to make observations; conversely, SAR has its own illumina- tion source through the radio waves transmitted by the anten- nae. The difference in illumination reliance affects the image. Optical sensors are only able to capture images in the day- light, whereas SAR can produce images at any time. Optical sensors provide data looking straight down, but SAR is side- looking. SAR depicts angles to an object differently than opti- cal sensors because it assesses range rather than angle in the direction perpendicular to the line of flight. The side-looking capability of SAR is able to provide object heights. In comparison, SAR exceeds optical remote sensors in valuable data acquisition. SAR is an all-day, all-weather method that has the power to offer timely remote sensing data in the early response stages following a natural disaster. It is important to note that severe weather conditions, such as heavy rainfall, can affect the quality of the SAR data. SAR sensors are particularly sensitive to surface variations. SAR is one of the only remote sensing tools that is independent of weather conditions and sun illumination. Light Detection and Ranging (LiDAR) This method was identified by three of the state bridge engi- neers surveyed in this study. LiDAR is a remote sensing technology that calculates ranges (distance) by illuminating FIGURE 12 Analytical procedures employed by the bridge damage detection methodology, illustrated using east- and westbound lanes of Los Angeles bridge 53-1797, on the I-5 at Gavin Canyon. (a) Delineation of a transect that runs along the highway and across the bridge, overlaid on a USGS 1-m aerial photograph. Blue squares demarcate the SPOT pixels used in analysis. (b) Difference between before and after SPOT images. (c) Correlation between before and after SPOT images.

20 the target with a pulsed laser and then assessing the reflected light. The light pulses create precise, 3D data of the target. Airplanes and helicopters are the most common platforms for LiDAR; however, it can be placed on a ground vehicle for data collection. LiDAR provides fast data acquisition and process- ing. Similar to SAR technology, LiDAR acts independently of weather and sunlight. Aerial LiDAR technology can also pen- etrate tree canopies. On the other hand, LiDAR is dependent on a GPS signal and productivity can be adversely affected by a low GPS signal. In comparison with other remote sensing technologies, LiDAR is an economical option as it relates to speed, accuracy, and density. LiDAR has the capability to cover a large ground area while maintaining high resolution and accuracy in its images. Mobile LiDAR is a game-changer in ground-based, post-disaster data collection. Gong and Maher (2014) collected data from the aftermath of Hurricane Sandy. The results of the data were uti- lized to chart new approaches for hurricane damage assessment of properties and infrastructures. Although road conditions and GPS outages affected the overall efficiency of the mobile LiDAR system, the preliminary information offered an innova- tive way in damage assessment. LiDAR scanning is a key remote sensing tool utilized to produce reliable data regarding bridge damage. It creates point cloud data from a bridge surface. The data are used to derive 3D geometric information to determine surface irregularities. These data can reveal the concrete deterioration and clearance height changes. LiDAR scanning originally used the spatial data to determine bridge defects, which is only beneficial to flat planes or intersections of multiple planes. By using LiDAR reflectivity, damage detection automation may be improved. The reflectivity value in the laser beam of LiDAR scanning is the light intensity recording of the reverting laser beam (Bian et al. 2012). In a single scan point, the row and column, polar coordinates, Cartesian coordinates, and reflectivity value are obtained. For a scan, dark scan points indicate small reflectiv- ity values and brighter scan points are associated with higher values. LiDAR reflectivity data have been used in previous stud- ies to separate building outlines from surrounding trees, but has yet to be used for bridge damage detection. Automated damage detection on bridges was previously based on a pho- tography imaging technique. The LiDAR reflectivity value is processed in a way similar to that of the photography imaging technique. By using histograms and statistic characteristics of the reflectivity data from a bridge surface, defects can be identified on the structure. The distribution in histograms of the reflectance value can depict the differences in surfaces. Bian et al. (2012) conducted an experiment on collected LiDAR images from a bridge deck. The five-step process was applied to two damaged areas on the bridge, a deck joint area and a reinforced concrete beam. Intensity histograms and standard variances clearly illustrated the damaged areas. The result proved that reflectivity-based methods are an effective process in determining bridge damage. LiDAR data defective detection methods on bridges previously used strictly spa- tial information that limited the process to flat surfaces only. Bian et al. (2012) applied reflectivity values to the existing processes of LiDAR data to study a possible improvement in the damage recognition procedure in bridge inspection. The experiment determined that the reflectivity-based method can be used on curved surfaces. This study showed that reflectivity values in LiDAR data can support automated defect detection in bridges by merging it with image processing algorithms. Unmanned Aerial Vehicle (UAV) UAV is commonly known as a drone. It is often flown remotely or by a pre-programmed flight path because there is no human onboard to pilot the aircraft. There are a variety of design plat- forms for UAVs including fixed-wing and quadcopter. For remote sensing purposes, digital sensors are attached to the underbelly of the UAV for image and video acquisition. Satel- lite images are often utilized in post-disaster management to analyze the damage to buildings and infrastructure. Unfortu- nately, such images are not always available for early impact analysis. In this case, UAVs are an efficient tool for delivering high-resolution images after natural disasters. The use of UAV in disaster management has only recently been implemented and is still an active area of research. The low-cost, flexibility, and ease of operation make for an efficient tool in post-disaster assessment. UAVs have many advantages over traditional aircraft and have become especially valuable, not only in damage assess- ment applications, but to military and civil applications as well. UAVs can perform at a wide range of altitudes. They are also able to adapt to areas depending on mission type. Lastly, UAVs are inexpensive compared with their traditional aircraft counterparts. Regardless of the benefits, the technology has not made much of an impact on the damage assessment of infrastructure systems so far. This is primarily the result of factors such as insurance issues, regulatory issues, and lack of safe communication frequencies. Although UAVs are efficient in data acquisition for post- extreme event management, they have also been shown to be useful in rapid response data collection. Advancements in UAV features have allowed for semi-automated and fully automated map creations. The rapid creation of UAV imagery- derived maps is beneficial to disaster response because they hold up-to-date spatial information. Suzuki and Miyoshi (2008) studied two methods in the production of real-time hazard maps utilizing UAV data. The first method creates a mosaic image of optical imagery collected in a video mode; the second method orthorectifies the imagery, projects them onto a map, and enables images to become integrated in a GIS database (Adams and Friedland 2011). Through an experi- ment conducted in Kanagawa, Japan, the results concluded that the second method offered a better user interface for sharing information and developing disaster recovery plans.

21 Marenchino (2009) also conducted an experiment that demon- strated the rapid mapping potential of employing imagery col- lected by UAVs. Through several studies, UAVs have shown an opportunity for use in rapid response in disaster recovery. In recent years, UAV imagery has been used more frequently for post-extreme event damage detection. Both automatic and manual UAV imagery assessments have been analyzed. Hurricane Katrina devastated the Mississippi Gulf Coast in 2005 and UAVs were used to inspect structural damages within a short time of the event (Pratt et al. 2006). This experiment included a helicopter UAV equipped with a digital camera with photo and video capability. The UAV was also equipped with a GPS-based flight hold feature that allowed the UAV to hover in place over the operator’s desired location. This helped the operator to set the UAV into a semi-autonomous flight setting so that the camera could be focused to capture the data. The technology offered promising results in post-disaster data col- lection except for a few downfalls such as obstacle avoidance, site access, sensor coverage, and weather conditions. Weather conditions especially affect UAV performance. Downdrafts, wind shear, and turbulence reduced UAV stabilization and therefore reduced imagery quality. Nevertheless, this was an early experiment with UAV technology and new systems have been improved to address these weaknesses. Following Hurricanes Wilma and Ike, both UAVs and USVs (unmanned sea-surface vehicles) were used to conduct post-disaster assessments of bridges, seawalls, and piers. According to Murphy et al. (2008), the Hurricane Wilma assessment is considered to be the “first known demonstra- tion of a UAV and USV cooperation.” The UAV and USV were designed to be launched simultaneously while indepen- dently surveying damage; however, this proved to be diffi- cult for the USV operator to control. The problem was fixed by using the UAV to relay real-time video of the USV to the USV operator. This experiment led the way to more coop- erative UAV and USV experiments. It also initiated a study in USV to UAV to operator communications. Following the Hurricane Wilma investigation, a study also using the UAV and USV was conducted. This study showed that the UAV could be employed to create maps for the USV navigation before it is launched. As the technology supporting UAVs grow and develop, their use in post-disaster management will become more common. UAVs offer satisfactory imagery that is useful for rapid damage assessment after extreme events at a low cost. As studies on UAVs continue, lessons and opportunities for application of UAVs in the post-disaster field will become more common.