Flood-Risk Mapping for Storm-Surge Events and Sea-Level Rise Using Lidar for Southeast New Brunswick

Home , Bouctouche, Le Pays de la Sagouine, Richibucto, Shediac

Can. J. Remote Sensing, Vol. 32, No. 2, pp. 194–211, 2006 Flood-risk mapping for storm-surge events and sea-level rise using lidar for southeast New Brunswick

Tim L. Webster, Donald L. Forbes, Edward MacKinnon, and Daniel Roberts

Abstract. Coastal flooding from storm-surge events and sea-level rise is a major issue in Atlantic Canada. Airborne light detection and ranging (lidar) has the spatial density and vertical precision required to map coastal areas at risk of flooding from water levels typically 1–2 m higher than predicted tides during storm surges. In this study, a large section of the New Brunswick coast along Northumberland Strait was surveyed in 2003 and 2004 using two lidar systems. Water levels from a major storm-surge event in January 2000 were surveyed using a global positioning system (GPS) and used as a benchmark for flood-risk maps. Maps of flood depth were also generated for all water levels and used for socioeconomic and ecosystem impact assessment. Flood-risk maps were constructed using standard geographical information system (GIS) processing routines to determine the spatial extent of inundation for a given water level. The high resolution of the lidar digital elevation model (DEM) captured embankments such as raised roadbeds that could prevent flooding inland. Where connectivity was present due to culverts or bridges, the DEM was notched across the roadbed to simulate the connection between the ocean and upstream low-lying areas in the GIS. An automated routine was then used to generate maps of flood extent for water levels at 10 cm increments from 0 to 4 m above mean sea level. Validation of the flood-risk and flood-depth maps for the January 2000 storm-surge water level by field visits indicates that the simulations are generally accurate to within 10–20 cm. The lidar data were also used to evaluate the potential for overtopping and dune erosion on a large coastal spit, La Dune de Bouctouche. This showed a high vulnerability to storm damage for critical habitats on the spit. The lidar- derived maps produced in this study are now available to coastal communities and regional planners for use in the planning process and to assist in development of long-term adaptation strategies. 211 Résumé. Les inondations côtières dues aux ondes de tempête et aux crues des eaux de la mer constituent une menace significative au Canada Atlantique. Les données levées par lidar (« light detection and ranging ») aéroporté ont la densité spatiale et la précision verticale nécessaires pour la cartographie de zones côtières exposées au risque d’inondation en raison de crues causées par des vagues de tempête qui, d’une façon générale, produisent des niveaux d’eau dépassant d’un ou deux mètres le niveau prévu de la marée. Au cours de cette étude en 2003 et 2004 et à l’ aide de deux systèmes lidar on a effectué l’arpentage d’un assez grand tronçon de la côte du Nouveau Brunswick s’étalant le long du détroit de Northumberland. Les niveaux d’eau résultant d’une onde de tempête qui s’est produite en janvier 2000 ont été relevés à l’aide du GPS (« global positioning system ») et ont servi de repères pour la cartographie des zones exposées au risque d’inondation. Des cartes de la profondeur des crues ont été également dressées pour tous les niveaux d’eau afin de déterminer l’impact socio- économique et les répercussions environnementales des inondations. Les cartes des zones exposées aux inondations ont été dressées à travers un travail de préparation d’un système d’information géographyique (SIG) ordinaire pour déterminer l’étendue spatiale de la zone inondée pour un niveau donné d’eau. La haute définition du modèle altimétrique numérique (MAN) lidar a permis de capter des levées de terrain, telles que des assiettes de voirie remontées, susceptibles d’empêcher l’inondation de l’arrière-pays. Là où il existait une connectivité due à la présence de buses ou de ponts, le MAN a été entaillé à travers l’assiette de voirie pour simuler le raccord entre l’océan et les zones de terrain bas se trouvant en amont dans le SIG. Nous nous sommes ensuite servis d’une routine automatisée pour dresser des cartes de l’étendue de l’inondation pour des différents niveaux d’eau augmentés de 10 cm à chaque incrément, partant de zéro jusqu’à quatre mètres au-dessus du niveau moyen de la mer. La validation de la cartographie des zones exposées aux inondations et de la profondeur des eaux d’inondations effectuée au courant de visites sur le terrain donne à entendre que les simulations sont en général exactes dans les limites de 10 à 20 cm. Des donées lidar ont été utilisées pour évaluer la probabilité que les vagues de tempête puissent atteindre un niveau plus haut que celui de la crête de la dune ou qu’il y ait érosion du relief de sable à la Dune de Bouctouche, grande flèche littorale du Nouveau Brunswick. Il est apparu que des habitats essentiels de la flèche

Received 30 September 2005. Accepted 21 February 2006. T.L. Webster,1 E. MacKinnon, and D. Roberts. Applied Geomatics Research Group, Centre of Geographic Sciences (COGS), Nova Scotia Community College, Annapolis Valley Campus, 295 Main St., Middleton, NS B0S 1M0, Canada. D.L. Forbes. Geological Survey of Canada (Atlantic), Natural Resources Canada, Bedford Institute of Oceanography, Dartmouth, NS B2Y 4A2, Canada. 1Corresponding author (e-mail: [email protected]).

littorale étaient très vulnerables aux dégâts causés par les intempéries. Des cartes issues de la technologie lidar sont maintenent prêtes à être mises à la disposition des collectivités côtières et des planificateurs régionaux pour faciliter le processus de planification et pour appuyer le développement de stratégies d’adaptation à long terme.

Introduction Webster et al. States have been reported by Sallenger et al. (1999) and Krabill et al. (1999), among others. Preliminary trials in Atlantic For coastal communities, the risk of storm-surge flooding Canada were reported by O’Reilly (2000), and subsequent with long-term sea-level rise is a particular concern. The experience was described by Webster et al. (2002; 2003; Intergovernmental Panel on Climate Change (IPCC) has 2004a; 2004b), O’Reilly et al. (2005), and Webster and Forbes predicted a rise in global mean sea-level from 1990 to 2100 of (2006). Most of the coast of the conterminous United States has between 0.09 m and 0.88 m, with a central value of 0.48 m now been mapped using this technology (Brock et al., 2002). (Church et al., 2001). Relative sea-level rise at any one place is Stockdon et al. (2002) used lidar to map the shoreline and a combination of changes in regional sea-level and vertical detect changes. ground motion. Subsidence in coastal areas will result in more In this paper we describe a recent study using two lidar rapid relative sea-level rise. This is of particular importance in systems to generate high-resolution digital elevation models the Canadian Maritimes, where postglacial isostatic (DEMs) and associated flood-risk maps for the coast of New adjustments are ongoing and much of the region is subsiding Brunswick along Northumberland Strait (Figure 1) (Grant, 1980; Shaw et al., 1998). Superimposed on sea-level (MacKinnon, 2004). This region was highlighted in a national rise, large storm waves and runup riding on high storm-surge study by Shaw et al. (1998) as being vulnerable to sea-level water levels pose additional coastal erosion and flooding rise. This paper represents one part of a multidisciplinary hazards. Other factors being equal, coastal erosion will increase project led by Environment Canada to derive information to aid with a rise in mean sea-level (Cowell and Thom, 1994; local resource managers in identifying areas prone to coastal Leatherman et al., 2003). On dune-backed coastal beaches and flooding and erosion and the possible socioeconomic and barriers, erosion often takes the form of dune-front scarping, ecosystem impacts of sea-level rise and storm-surge events. and backshore flooding may occur if runup and overwash The maps depicting areas vulnerable to sea-level rise have been exploit low points in the dune crest to create breaches and used to assess potential impact to ecosystems such as salt washover channels. Sea-level rise can also threaten ecosystems, marshes and coastal dunes and cultural features such as homes, especially areas such as coastal wetlands and marshes that are businesses, municipal infrastructure, and historic sites in low- dependent on specific tidal range conditions (Nicholls et al., lying areas. This project builds on the experience gained from 1999). the successful application of lidar technology to storm-surge A storm surge is defined as the difference between observed flood-risk mapping on Prince Edward Island (McCulloch et al., water level and the level predicted for the astronomical tide. 2002; Webster et al., 2001; 2003; 2004a; 2004b; Webster and Surges are caused by atmospheric low-pressure systems and Forbes, 2006). high winds associated with storms. Storm surges along the east coast of Canada can add2mormore of water to the predicted water level along the coast (Parkes et al., 1997; Parkes and Methods Ketch, 2002). Coastal communities are most vulnerable if a Lidar systems storm surge peaks during a high-tide event, especially during large spring high tides. With accelerated sea-level rise, the Lidar mapping systems consist of three components, the extent and frequency of flooding will increase in the future, as laser ranging system, an inertial measurement unit (IMU), and will the impact of related factors such as shoreline erosion. the global positioning system (GPS), integrated to provide the Thus, there is a demand for information that predicts areas at horizontal coordinates and elevation of each laser return. risk from such events, at present and into the future, as a basis Earlier lidar systems were capable of recording a single return, for sustainable coastal zone management, effective planning, either the first or last pulse reflected from the target. The latest and the development of appropriate adaptation strategies lidar sensors are capable of recording multiple returns, (McLean et al., 2001). typically at least the first and last pulse, with some sensors Because storm surges in eastern Canada are typically less recording up to four intermediate returns, and in most cases the than2minheight, technologies with vertical precision intensity of the reflected pulses. significantly finer than this must be employed to generate Various studies have been reported on the calibration and flood-risk maps of sufficient resolution. Airborne light systematic errors of lidar systems (Kilian et al., 1996; Burman, detection and ranging (lidar) is a technology that offers the 2000; Filin, 2001; 2003a; 2003b; Katzenbeisser, 2003) and the vertical precision and high spatial detail required for this accuracy of laser altimetry data (Huising and Gomes Pereira, purpose. A general overview of airborne laser scanning 1998; Kraus and Pfeifer, 1998; Crombaghs et al., 2000; Schenk technology and principles is provided by Wehr and Lohr et al., 2001; Maas, 2000; 2002; Ahokas et al., 2003; Artuso et (1999). Applications to coastal process studies in the United al., 2003; Bretar et al., 2003; Elberink et al., 2003; Kornus and

Figure 1. Study area location of the 2003 and 2004 lidar surveys for flood-risk mapping of the southeast New Brunswick coast. Background image is a Landsat true colour composite of maritime Canada.

Ruiz, 2003; Hodgson et al., 2003; 2005; Hodgson and sophisticated Mark II system was used to acquire data over the Bresnahan, 2004; Hopkinson et al., 2005; Webster, 2005). remaining three polygons (Figure 1). Surveys were acquired Although lidar technology has improved over the years and near low-tide conditions where possible, and GPS baselines systems are capable of providing elevations of the earth’s were kept below 50 km to minimize aircraft trajectory errors. A surface to decimetre-level precision, systematic and random large coastal spit, La Dune de Bouctouche, was surveyed in errors can still occur and independent validation is both 2003 and 2004 to assess changes in the morphology of the recommended. Classification of the lidar returns into “ground” dune following a storm in February 2004 (Roberts et al., 2005). and “non-ground” (i.e., vegetation or buildings) points is an The Mark I is a first-return lidar system that was originally important aspect of the lidar processing stream and can affect designed for corridor mapping and was mounted on a pod fixed the accuracy of derivative products such as a DEM. The correct to the underside of a Bell 206L helicopter (Figure 2A). The definition of coastal features such as wharves and steep rock lidar operated at a 10 kHz laser repetition rate along with the embankments in the lidar point data is critical to the derivation scanning mirror operating at 15 Hz to direct the laser pulses of an accurate DEM for flood-risk mapping. across the swath. At a flying altitude of 300 m, the laser beam had a ground footprint diameter of 9 cm and an average post Lidar surveys spacing of 60 cm (Figure 2A). The narrow laser beam In this study, Terra Remote Sensing Inc. (Sidney, B.C.) was facilitated penetration to the ground in vegetated terrain. The contracted to acquire lidar data for coastal areas of southeast survey was conducted from 23 May to 3 June 2003. New Brunswick. The total study area consisted of 278 km2. The The Mark II system is capable of recording the first and last requested lidar point density on open ground was a posting returns and the intensity of one of the returns and was mounted every 60 cm. The absolute accuracy of the lidar data for this on a pod that was fixed to the underside of a Bell 206L study was specified to be ±30 cm in the vertical. Two lidar helicopter (Figure 2B). It was decided to acquire the intensity systems were used to acquire data during leaf-off conditions on alternating returns, thus every second first and last return following snowmelt in the spring of 2003 and 2004. In 2003, would record the intensity. The lidar operated at a 40 kHz laser the Mark I system was used to acquire data over four of the repetition rate along with the scanning mirror operating at seven survey polygons (Figure 1). In 2004, the more 17 Hz to direct the laser pulses across the swath. At a flying

Figure 2. Lidar configurations. (A) Mark I system used in 2003. (B) Mark II system used in 2004. AGL, above ground level. altitude of 600 m, the laser beam had a ground footprint Division, Natural Resources Canada, www.geod.nrcan.ca/ diameter of 27 cm and an average post spacing of 45 cm index_e/products_e/software_e/gpsht_e.html). (Figure 2B). Because the Mark II sensor is capable of Lidar data are typically delivered in ASCII files consisting of recording the first and last returns, the wider laser beam could x, y, z information. There is no standard format for lidar data, still partially penetrate the vegetation canopy and reflect off the although a new proposed binary format known as LAS has ground or near-ground surface and be measured as the last recently been published (Schuckman, 2003). In this study the returning pulse. The survey was conducted from 26 April to data were delivered as ASCII files separated into 4 km × 4 km 1 May 2004. The Mark II has improved precision to better than tiles based on the Universal Transverse Mercator (UTM) grid. ±15 cm in the vertical as a result of the higher precision laser The 2003 data files have data fields for each lidar point, ranging system and IMU. including easting and northing positions based on UTM zone 20 (WGS84), ellipsoidal and orthometric heights, the GPS time Lidar preprocessing for every laser shot, and the flight line number. The last two Hodgson et al. (2005) provide an overview of the general fields facilitated testing for systematic errors between adjacent procedure used by many of the automated routines employed to flight lines or strips (e.g., Maas 2000; 2002). The 2004 data classify the lidar point cloud data. They point out that most have the same fields as the 2003 data, with additional fields lidar data providers consider the details of this process including echo number and the intensity of alternating returns. proprietary and do not report the specifics of the parameters The echo number is an integer code that defines whether the used for the classification. In this study, the lidar point cloud return was the first or last of multiple returns or a single return. was classified into ground and non-ground categories using Terrascan™ software by the data provider, who did not provide Lidar GIS processing the parameters used in the classification. The ASCII lidar data were imported into the ESRI ArcGIS™ For many coastal applications, including flood-risk mapping, system (ESRI, Redlands, Calif.) utilizing Arc Macro Language the lidar elevations are transformed from ellipsoid heights to (AML) automated routines. Ground and non-ground point orthometric heights. Orthometric heights are based on the geoid, layers were built in the GIS system. Various surfaces were an equipotential surface defined by the earth’s gravity field, constructed from the lidar points. Digital surface models approximately equal to mean sea level. To obtain orthometric (DSMs) were constructed from the lidar data by utilizing the heights, an adjustment must be made for the local vertical ground and non-ground points. The surface was linearly separation between the ellipsoid and the geoid. The difference interpolated toa1mgrid from a triangular irregular network between the WGS84 ellipsoid and the Canadian Geodetic Vertical (TIN). The ground points were used to construct a DEM using Datum of 1928 (CGVD28) is obtained using the HT1_01E model, the same method. The DSMs and DEMs were examined in which has since been replaced by HTv2.0, with an accuracy of combination with the classified lidar points, and classification ±5 cm with 95% confidence in southern Canada (Geodetic Survey errors were identified. For example, the edges of wharves were

© 2006 CASI 197 Vol. 32, No. 2, April/avril 2006 classified as non-ground, therefore the initial DEM did not data with the lidar ground points and derived DEMs. Coastal accurately represent these features. The non-ground points that features (e.g., wharves) in the DEM were compared to digital were incorrectly classified for these areas were used to orthophotographs to ensure an accurate representation. construct a new DEM that more accurately represented the The absolute accuracy of the lidar data was validated using coastal features following a method described in Webster et al. two techniques: (i) comparing proximal lidar points to (2004a). It was determined that the most effective method to checkpoints, and (ii) comparing the lidar DEM to checkpoints represent the intensity information was derived by interpolating (Webster, 2005). The validation checkpoints consisted of real- the intensity values of all the lidar returns (ground and non- time kinematic (RTK) GPS surveys of roads and wharves and ground). The exploitation of the intensity data is beyond the traditional surveys utilizing a total station for dune transects scope of this paper; however, using these data in combination and under the forest canopy. with the DSM, the DEM, the normalized height grid (DSM – DEM), and echo information, Brennan and Webster (2005) Water levels for flood modelling demonstrated that accurate land cover classification based The study area in southeast New Brunswick has sustained a entirely on lidar is possible. number of damaging storms over the past few years (Forbes et One issue with lidar data collected in the coastal zone is the al., 2004). The 21–22 January 2000 storm produced one of the specular reflection over flat water and intertidal areas. Only highest storm-surge water levels on record for the returns within a narrow angle off nadir were recorded over Northumberland Strait region (Parkes and Ketch, 2002; Forbes these areas, often having anomalously high intensity values. et al., 2004). Unfortunately the tide gauge at Escuminac, the The variable density of returns over water causes the lidar only permanent gauge remaining in the study area, did not surfaces to have large triangles that do not accurately represent capture this event, having frozen prior to the highest water level the surface. We have used a combination of lidar point density during the storm. This surge produced a record water level at maps to constrain the surface over water and assign a constant Charlottetown, Prince Edward Island, where the tide gauge has elevation for the areas of no data between flight lines based on been operating since 1911 and records exist back to 1904 the water level observed from the nadir data. (Parkes et al., 2002). Staff from Environment Canada visited DEM and DSM grids were imported into PCI Geomatica™ several coastal communities in the aftermath of the storm and software (PCI Geomatics, Richmond Hill, Ont.) to construct marked the high-water position using nails in power poles and colour shaded-relief models. Surfaces were “illuminated” from other wooden structures. These high-water marks and others the northwest using an zenith angle of 45° with a five times identified on buildings were surveyed with RTK GPS in 2002 vertical exaggeration applied. To ensure that adjacent lidar and used to estimate the water level associated with the January study area polygons (Figure 1) have a consistent colour range, 2000 storm-surge event. It is estimated that the water level the maximum elevation range of the surveyed region was associated with the January 2000 storm reached 2.5 m above determined and used to scale the data. The maximum elevation CGVD28 (Figure 3). Three water levels were selected for used was 38 m for the DEM and 58 m for the DSM. initial flood modelling. These were the January 2000 storm level and 50 and 70 cm above that level. Tide-gauge records Lidar validation from Charlottetown indicate that mean relative sea-level is The accuracy of lidar data depends on the removal of the rising at a rate of 32 cm per century (Parkes et al., 2002; Forbes systematic errors associated with the system (Filin, 2001; et al., 2004). This value is a combination of regional sea-level 2003a; 2003b). Several researchers have examined the issues of rise from global warning and crustal subsidence, estimated at lidar validation and have highlighted the potential for errors 15 cm per century for this area. The rate of relative sea-level between flight lines or strips (Kilian et al., 1996; Huising and rise diminishes to the west and north along the New Brunswick Gomes Pereira, 1998; Crombaghs et al., 2000; Maas, 2000; coast as a function of lower rates of subsidence (Koohzare et 2002; Schenk et al., 2001; Latypov and Zosse, 2002; Ahokas et al., 2005). Due to the uncertainties in the estimates of future al., 2003; Bretar et al., 2003; Elberink et al., 2003; Kornus and sea-level rise, it was decided to construct additional flood-risk Ruiz, 2003; Webster, 2005). Many of the studies have dealt maps using water levels at 10 cm increments up to a maximum with individual flight strips, where the overlapping areas are of 4 m above CGVD28 in addition to the three benchmark compared either as points or as interpolated surfaces. To levels described previously. evaluate the possible error sources between strips, the GPS time tag or flight line number for each lidar point is used in the Flood-risk modelling validation procedure. In this study the absolute accuracy was In the present study of coastal flooding, it was decided to use required for accurate flood-risk mapping, therefore extensive existing GIS and image-processing capabilities combined with ground control using GPS and traditional survey methods was the lidar DEM to model the potential areas of flooding. Galy used in the analysis. In all cases the HT1_01 model was used to and Sanders (2002) presented a similar approach where they transform the GPS ellipsoidal heights into orthometric heights used a DEM derived from interferometric synthetic aperture for comparison with the lidar data. A methodology described in radar (InSAR) data to map flood risk along the River Thames in Webster (2005) was used for comparing the GPS validation the United Kingdom. It is assumed that a given water level from

198 © 2006 CASI Canadian Journal of Remote Sensing / Journal canadien de télédétection the storm-surge event would form a horizontal flood surface the coast where stream channels or valleys exist but appear in extending landward from the open ocean. Thus hydraulic the DEM to be blocked from the ocean by the elevated roadbed, effects, associated time lags, or flood expansion or dampening to determine if a bridge or culvert may be present. Where it was were not considered in the modelling effort. This assumption is thought that a culvert or bridge exists, then the area upstream of justified with the results of water levels measured by RTK GPS the road was included in the flood extent (see MacKinnon, for the January 2000 storm which show a nearly consistent 2004). This proved to be a very time consuming task, especially height of near 2.5 m above CGVD28 throughout most of the when constructing flood extents for every 10 cm increment surveyed area (Figure 3). Mean sea-level is approximately from 0 to 4 m above CGVD28. 19 cm above CGVD28. In the present study, we modified this methodology to Webster et al. (2004a) discussed the requirement to ensure automate the computation of the 10 cm flood limits. The that areas in the DEM below a given flood level are included in method involved visually evaluating the DEM using a criterion the flooded polygon only if they have free connection to the similar to that described by Webster et al. (2004a). The DEM ocean (i.e., low-lying areas behind an embankment would not was notched across the roadbed where bridges or culverts were be flooded unless there was a channel allowing water to enter). thought to be present. The base of the notch was assigned an In the earlier Charlottetown study, engineers were consulted on elevation equal to that of the nearest upstream part of the the types of culverts used at different locations when deciding channel or basin landward of the road. The automated routine whether to include upstream low-lying areas in the flood extent made use of this modified DEM with “hydraulic pathways” to (Webster et al., 2004a). In the present study, the area of interest ensure areas with free connection to the ocean were included. was significantly larger, and specific information on culvert To determine the flood extent, a script was written in Python™ locations was not readily available during this phase of the and executed in the ESRI ArcGIS™ system. The script project. As a result, we did a visual interpretation of areas along involved applying a threshold to the modified DEM at a given

Figure 3. Reference marks for water levels associated with the January 2000 storm-surge event surveyed using RTK GPS. A water level approximately 2.5 m above CGVD28 was determined from the survey data.

© 2006 CASI 199 Vol. 32, No. 2, April/avril 2006 flood level. The resultant binary raster was then vectorized and foredune crest elevation (or the beach crest where dunes are perimeter and area attributes were computed. Since the absent) was obtained from shore-normal transects modified DEM always contained a significant portion of the superimposed on the DEM at 50 m intervals along the beach. area that represented the ocean, the largest polygon (area) The upper limit of the wave-formed beach was also determined generated for a given flood level would represent the by visual inspection of each transect profile. The resulting continuously connected coastal flooded area. Smaller low- alongshore elevation profiles for barrier and dune crest and top lying inland polygons would not have free connection to the of beach provide a basis for estimating the extent and location ocean and thus would not be selected during this process. The of flooding and overtopping under storm-surge conditions. The original DEM, without the notches, was used again to evaluate spit segments subject to flooding in the January 2000 storm, each flood level to determine where the roadbed would be with sea ice present and no waves, can be determined by flooded at the various water levels. This was important because superimposing the surge water level on the barrier profile. For the water level at which a road will become impassable due to other storms, such as the storm on 29 October 2000, when large flooding is critical information for emergency-measures waves were present (Forbes et al., 2004), the water level organizations and in the design of adaptation strategies. The provides a highly conservative estimate of overtopping results of this analysis are merged with the flood extents potential. Although deepwater wave data are now available for derived from the notched DEM to produce the final flood- some events from 2003–2005, we have no data on setup and extent maps. breaking wave conditions from which to compute runup. For In addition to the flood-extent maps, Webster and Forbes purposes of discussion, we add 1.5 m to the peak water level for (2006) highlighted the importance of flood depth for the large storm-surge events to obtain a conservative estimate of assessment of potential damage associated with storm-surge runup and the extent of overtopping. It should also be noted that flooding. Flood-depth maps were generated for the various where the resulting runup level lies above the top-of-beach flood levels by subtracting the DEM from the flooding level profile, there is a high potential for dune-front erosion. over the area flooded.

Flood visualization Results The flood extents at 10 cm increments are visualized as vector The lidar surfaces, DSM, DEM, and intensity maps and the overlays in the GIS system for ease and convenience of use by flood-risk extent and flood-depth maps have been passed to the the non-GIS specialist. In addition, because the surge events are other members of the project team who are currently using dynamic, the 10 cm flood extents are used to construct animation these data to evaluate potential socioeconomic and ecosystem sequences. The animations are created using both planimetric impacts (e.g., Chouinard et al., 2004; Daigle, 2004; DeBaie, and perspective views of the terrain. The animation sequences 2004; Hanson, 2004; Vasseur and Délusca., 2004) and to advise involve overlaying the flood extents on colour shaded-relief on possible adaptation strategies (Nichols and Sutherland, models of the lidar DSM, orthophotography, or high-resolution 2004). In the following section, we present examples of the satellite imagery such as QuickBird™. PCI Geomatica™ flood-risk mapping from the lidar surveys with the validation software is used for the visualization and to generate the results. individual frames for the animations. In previous studies (Dickie, 2001; Webster et al., 2004a; 2004b; MacKinnon, 2004), a solid Surface models and validation: 2003 lidar data colour was used to represent the water level. In this study we MacKinnon (2004) inspected all of the lidar data acquired in modified this method to allow the water to be transparent so that 2003 and highlighted two main types of problems: (i)afew the underlying features affected by the floodwater are still small missing areas within the survey polygons, and (ii) lidar visible. An EASI script is used to automate the process once a point classification errors. The areas that were missed in 2003 desirable perspective view is defined. The individual frames are were acquired during the 2004 mission. In addition to the annotated to describe critical water-level events such as the absolute accuracy of the lidar information, classification errors ® January 2000 storm surge and combined to build a Windows between ground and non-ground targets along the coast are AVI file. critical for this type of flood-risk mapping application. The classification errors are associated with coastal features that Evaluating the potential for coastal barrier overwash have abrupt vertical relief relative to that of the surrounding La Dune de Bouctouche is one of the largest natural barrier terrain. The erroneously classified points were integrated into beach and dune systems along the coast of New Brunswick and the final DEMs used for flood-risk mapping. The data provider hosts several species at risk, including Gulf of St. Lawrence was informed about this error and instructed to carefully aster (Symphyotrichum laurentianum) and piping plover inspect the classification results along the coast to ensure such (Charadrius melodus). Lidar data were acquired over this area features are accurately classified in subsequent datasets. In in both 2003 and 2004 and used to generate DEMs gridded at subsequent data delivered, this problem was not observed and 1 m horizontal resolution. To evaluate the potential for dune- coastal features were accurately classified. face erosion or overtopping under storm conditions, the

Table 1. Summary statistics of 2003 lidar DEMs and GPS checkpoints. Mean difference, Standard deviation, RMS error, No. of Lidar polygon GPS – DEM (m) GPS – DEM (m) GPS – DEM (m) GPS points Cap Pelé 0.04 0.11 0.11 20 748 Cormierville 0.21 0.12 0.24 2 426 Bouctouche 0.12 0.10 0.16 3 125 Cap Lumière 0.07 0.12 0.14 810

Table 2. Summary statistics of 2004 lidar DEMs and GPS checkpoints. Mean difference, Standard deviation, RMS error, No. of Lidar polygon GPS – DEM (m) GPS – DEM (m) GPS – DEM (m) GPS points Cap Jourimain 0.10 0.09 0.14 1645 Shemogue 0.11 0.07 0.10 3333

The 2003 lidar data met the specifications based on the study, field validation data concentrated on assessment of the comparison with the GPS and total-station survey data with the height, and no data were collected specifically to validate the derived DEMs. A systematic pattern is observed in the data horizontal position of the data. The intensity surfaces were over flat areas such as water, however. This pattern, also found assessed for relative accuracy, however, and compared with in lidar data from other systems and data providers (D. Whalen, 1 : 10 000 scale base maps. The relative accuracy was assessed Geological Survey of Canada, personal communication, 2005), qualitatively by examining linear features such as parking lots. has been termed the “wood-grain effect.” It is caused by a In a few cases it appeared the motion of the aircraft was not combination of the limited precision of the laser ranging system completely resolved because the intensity images showed wavy and the motion of the aircraft (Figure 4). The variance of the features that should have been straight. pattern is within the specifications of the survey, but the pattern The intensity data were used in combination with the DEM detracts from the fidelity of the DSM. There were no systematic and DSM to assess the moisture content and vegetation cover errors detected between strips or flight lines in these data. on the dune system in Kouchibouguac National Park. The lidar The summary statistics for the lidar DEMs compared with data were capable of highlighting spruce trees that occur on the GPS surveys are presented in Table 1. The mean difference in dune system and different moisture levels in the sediments height between the GPS and lidar DEMs for the 2003 survey (Figure 7). Although the intensity of the reflected laser pulses polygons is 0.11 m, with a mean standard deviation of 0.11 m near the edge of the scan tends to be lower than nadir pulse for and a mean root mean square (RMS) error of 0.16 m. some lidar systems, the intensity values for these data are very The results of the total-station transects across dunes within consistent across the swath, with the exception of nadir scans the Cap-Pele lidar polygon show general good agreement over flat water, thus allowing intensity mosaics to be (Figure 5). Where the profiles differ, the lidar DEM is constructed without significant artefacts between flight lines. consistently too high by a few decimetres. We attribute this bias to the dense growth of dune grass (Ammophila breviligulata) Flood-risk maps growing on the backslope of the dune and to some The notched lidar DEMs are used initially to construct flood- accumulation of sand on the upper dune face between the dates risk maps for three water levels. The levels are based on the of the survey and the lidar acquisition. The dune crest location January 2000 storm, determined to be 2.5 m above CGVD28, and height are accurately captured by the lidar survey and the same storm superimposed on increased relative sea- (Figure 5). level 50 and 70 cm above the mean water level in 2000. Initial flood modelling was carried out for the Pointe-du-Chêne area, Surface models and validation: 2004 lidar data near Shediac, because this built-up area was inundated during The 2004 lidar data met the specifications based on the the January 2000 storm-surge event (O’Reilly et al., 2005) and comparison with the GPS and total-station survey data with the people had to be evacuated in snowstorm conditions using derived DEMs (Table 2). The mean difference in height front-end loaders. In the summer of 2004 this site was visited between the GPS and lidar DEMs for the 2004 survey polygons with the flood-risk maps to verify the extent of the flooding is 0.11 m, with a mean standard deviation of 0.08 m and a mean with local residents. The high resolution of the lidar DSM RMS of 0.12 m. No systematic patterns were detected in the allows one to easily navigate to specific road intersections 2004 lidar surfaces (Figure 6). associated with the flood limit. The qualitative assessment The horizontal accuracy of lidar data is challenging to indicated that the flood extent derived for this event matched evaluate if only height information is supplied. Data from the flood limit observed by the local residents (Figure 8A). sensors capable of recording the intensity of the return pulse Flood-depth maps are constructed to better estimate the can be used to assess the positional accuracy, however. In this potential economic impact of storm-surge events (Figure 8B).

Figure 4. Example of 2003 lidar data with “wood grain” pattern and location of the GPS survey check data. (A) Lidar DEM of Bouctouche Bay with GPS checkpoints (magenta triangles). (B) DSM of Bouctouche Bay with the direction (arrow) of the perspective view of the sea-level rise simulation depicted in Figure 9.

Figure 5. Total-station dune survey transects at l’Aboiteau in the Cap Pelé lidar polygon. (A) Lidar DEM with transect locations (solid circles). (B) Comparison of survey and lidar DEM elevations for transect 1 at the east end of the dune. Vertical bars denote standard deviation (SD). (C) Comparison of survey and lidar DEM values at transect 3 on the central part of the dune.

Figure 6. Lidar DEM (2004 data) and GPS validation points for the large embayments with salt marshes in the Shemogue area.

Visualization of flooding The flood-depth maps also proved useful for validating the results of the flood modelling. Staff from Environment Canada The 10 cm increment flood extents are available as vector visited several communities throughout the larger study area to polygons. We have created animations to visualize flooding as evaluate the flood-extent and flood-depth maps. They reported a result of a storm-surge event superimposed on sea-level rise. a general agreement of within 10 cm of the flood depth These have a background image consisting of a lidar colour simulated for the January 2000 storm compared with those shaded relief (CSR) or satellite image draped over the DSM to measured in the field. construct a perspective view. This is a very effective way to Additional flood-risk maps were generated for 10 cm present the flood-risk modelling results in a way that is readily increments in water level to allow for a more precise estimate of understood by the general public. Figure 9 presents an example the areas potentially affected by future sea-level rise and storm- from the community of Bouctouche in the New Brunswick surge events. Other members of the project team are calculating study area. return probabilities for the various water levels. Histograms of the area of flood extent are assessed to determine if there are Potential for overtopping and dune erosion specific water-level thresholds for specific regions. In addition, The alongshore profile of beach and foredune crest other ecosystem impacts can be assessed with these flood-risk elevations was developed as described earlier using an map sequences. approach similar to that presented by Elko et al. (2002a; The flood-risk map sequences at 10 cm elevation increments 2002b). The dataset for La Dune de Bouctouche is almost can be used for assessing ecosystem impacts. Salt marshes are 12 km in length, with a gap of about 1 km where the 2004 lidar sensitive to the tidal range and frequency and duration of coverage missed the outer beach (Figure 10). The x axis inundation. They will migrate upslope as relative sea-level rises represents distance alongshore (downdrift) from the proximal (Shaw and Ceman, 1999; Donnelly et al., 2004). The extent of point of detachment (at the Irving Eco-Centre) to the outer end landward migration is dependent on local slope and land-cover of the spit (Figure 10A). Blue squares in Figure 10B represent conditions as well as the species composition and habitat transects without dunes, where the crest elevation is determined conditions (Donnelly and Bertnes, 2001; Bertnes et al., 2002). by wave runup dynamics. The orange squares and line denote The flood-risk maps are used to determine the area available for the morphological break in slope at the base of the foredune, salt-marsh migration as a function of the rising sea-level and representing the top of the beach at the end of April 2004, at the associated tidal range. time of the lidar data acquisition (Figure 10B). The blue-green triangles and line represent the foredune crest elevation, almost invariably the highest elevation on the transect (Figure 10B).

Figure 7. Example of the 2004 lidar DEM and intensity data for Kouchibouguac National Park. The imagery represents the southern tip of the North Richibucto Dune. (A) DEM of the dune. (B) Lidar intensity image of the dune with various features annotated.

The highest dunes at La Dune de Bouctouche exceed 6 m 1.60 km of the spit, range from 2.0 to 3.3 m, with most lying above CGVD28, and the modal dune crest elevation is 4.0 ≤ between 2.2 and 2.8 m above mean sea level. The top-of-beach hDC < 4.5 m (mean = 3.94 m). Dune crest elevations below 3 m (base-of-dune) elevations range from about 1.3 to >3.0 m and occur in some areas, notably between 7 and 8 km downdrift, show distinctive patterns of alongshore variability on various representing a total distance of 1.45 km. The dune crest falls to segments of the spit. The most common runup (base-of-dune) ≤ ≤ below2matthedistal tip, where the minimum crest elevation elevations are 1.5 hR < 2.0 m (3.35 km) and 2.0 hR < 2.5 m is 1.8 m. Barrier crest elevations without dunes, representing (2.75 km). Overall, the barrier crest elevations are bimodal,

≤ ≤ with maxima in the classes 2.0 hC < 2.5 m and 4.0 hC < (Figure 10B), shows that some sections were overtopped by 4.5 m. storm-surge flooding to a level of 2.5 m in the storm of 21– Preliminary analysis of overtopping potential for La Dune de 22 January 2000 (when sea ice prevented wave development). Bouctouche, based on the lidar DEM crest elevations Some of the lower parts of the spit near the proximal end had

Figure 8. (A) Flood-risk map of the three water levels for Pointe-du-Chêne area. Lidar DSM with water levels overlain. The water levels correspond to the January 2000 storm (red line); the January 2000 storm plus 100 years of sea-level rise at a rate of 50 cm per century (yellow line); and the January 2000 storm plus 100 years of sea-level rise at a rate of 70 cm per century (white line) (orthophoto provided by Service New Brunswick). (B) Flood-depth map of the Pointe-du-Chêne area for the January 2000 storm water level.

foredunes at least 3 m high in January 2000, prior to the damaging storm of 29 October 2000 (Forbes et al., 2004). In the latter storm, the surge peaked at 2.42 m chart datum at Escuminac (about 1.7 m above mean sea level) and was accompanied by very high waves that caused significant damage along the New Brunswick coast. Adding a conservative 1.5 m to the maximum storm-surge elevation to account for wave setup and runup, all parts of the spit below 3.2 m would have been subject to overtopping, and severe dune cut would have occurred along almost the entire length of the spit (Figure 10B). This is indeed what happened, as demonstrated by the complete removal of 3 m high dunes on the proximal spit (Forbes et al., 2004). This dataset highlights the increased vulnerability of barriers along the southeast New Brunswick coast following strong storm events in October 2000, November 2001, and December 2004, among others. The crest on parts of La Dune de Bouctouche is now considerably lower, leaving it exposed to potential overtopping in less severe storms.

Summary and discussion The lidar surveys of 2003 and 2004 in southeast New Brunswick provided the high-resolution point sampling along the coast required to construct DEMs that enabled flood-risk and flood-depth maps to be built to model the impact of water levels up to 4 m above CGVD28. The data met the vertical specifications for the study. The flood extents and flood depths associated with the January 2000 storm surge have been qualitatively validated by field visits to sites where the water levels are known. The lidar DSM maps, in combination with the flood-extent and flood-depth maps, facilitated this work. The flood extents generally agree, and the flood depths are typically within 10 cm of observed water levels where observed in January 2000. The modification of the lidar DEM by notching the roadbed to allow a hydraulic pathway to low-lying upstream areas has facilitated the use of automated scripts in the GIS to generate flood extents at water levels every 10 cm up to 4 m. This allowed near-continuous flooding simulations and animations. The ability to generate a full sequence of water levels allows the emergency-measures and planning officials access to more information for use in designing adaptation plans. The lidar DEMs proved valuable for analysis of flooding potential in natural estuarine settings as well, where the data are being used to evaluate the potential for salt-marsh migration or coastal Figure 9. Perspective view of lidar DSM for Bouctouche Bay. squeeze. (A) Normal tidal water level during the lidar survey. The historic The repeat lidar surveys for La Dune de Bouctouche Acadian village of le Pays de la Sagouine is located on the island highlighted the dynamic morphology of this system and centred in the bay south of the bridge. (B) Simulated water level changes that occurred over 1 year (Roberts et al., 2005). for the January 2000 storm-surge event. The arrow indicates the nail that represents the surveyed water level of the January 2000 Extraction of crest elevations from the lidar DEM enabled storm-surge event (Figure 3). (C) Simulated water level of 4 m an assessment of overtopping potential and risk of dune erosion above CGVD28. This represents an approximate upper limit for under a variety of storm conditions. This analysis demonstrated the risk of flooding over the next 100 years. that the sequence of storms since 2000 has partly cut down the crest elevation, resulting in higher vulnerability for the spit, including threats to low-lying areas occupied by the

Figure 10. (A) QuickBird™ near-infrared false colour composite image of La Dune de Bouctouche acquired in December 2004 (includes material copyright 2004 DigitalGlobe Inc., all rights reserved). (B) Barrier crest profile for La Dune de Bouctouche, showing top of beach (base of dune) in orange, barrier crest without dunes in blue, and foredune crest in blue-green. Broken lines in grey show peak storm-surge water levels for January 2000 (2.5 m above mean sea level) and October 2000 and December 2004 (1.7 m above mean sea level). Higher broken line in black represents a conservative estimate of wave runup during the latter two storms. threatened Gulf of St. Lawrence aster. The impact on habitat DEMs are valuable tools for assessing the potential impacts of availability for piping plover is less clear. Overall, at least accelerated sea-level rise under future climate change. 1.75 km of the ~12 km long spit has crest elevations below 2.5 m CGVD28, the peak surge level during the storm of 21– 22 January 2000. The damaging storm of 29 October 2000, Acknowledgements with runup estimated up to 3.2 m or higher, overtopped about This project was funded under the Climate Change Action 3 km of the spit and reached the base of dunes almost Fund (CCAF), managed by the Climate Change Impacts and everywhere else. Very similar conditions pertained during the Adaptation Program of Natural Resources Canada. The lidar storm of 27 December 2004, but a concentration of slush in the data were acquired with funding from Environment Canada and nearshore may have reduced the runup height. Nevertheless, the ACOA Atlantic Innovation Fund (AIF) through the Nova slush was deposited to a depth of at least 0.5 m over the barrier Scotia Community College (NSCC). Validation data were crest in the low area extending about 1 km from the start of the acquired by the Applied Geomatics Research Group (AGRG, spit (Figure 10B). NSCC) and the Geological Survey of Canada (GSC). In summary, this study demonstrates multiple applications of QuickBird™ imagery for the New Brunswick coast was airborne lidar topographic mapping to assess flooding, acquired in collaboration with Gavin Manson (GSC) through overtopping, and erosion hazards. Furthermore, the resulting GRIP funding from the Canadian Space Agency. Partial support was also provided through the Interdepartmental

Recovery Fund managed by Environment Canada in Church, J.A., Gregory, J.M., Huybrechts, P., Kuhn, M., Lambeck, K., Nhuan, collaboration with Al Hanson (Canadian Wildlife Service). We M.T., Qin, D., and Woodworth, P.L. 2001. Changes in sea level. In Climate change 2001: the scientific basis. Contribution of Working Group I to the thank David Finley and Bernie Conners of Services New Third Assessment Report of the Intergovernmental Panel on Climate Brunswick for supplying survey control information and the Change. Cambridge University Press, Cambridge and New York. digital orthophotography. We would like to thank our partners Chapt. 11, pp. 639–693. on the CCAF team, especially Réal Daigle, Project Manager, Cowell, P.J., and Thom, B.G. 1994. Morphodynamics of coastal evolution. In for his coordination efforts and overall support. AGRG Coastal evolution: Late Quaternary shoreline morphodynamics. Edited by students from 2002–2005 assisted with fieldwork and data R.W.G. Carter and C.D. Woodroffe. Cambridge University Press, reduction. David Frobel, Gavin Manson, and Paul Fraser Cambridge and New York. pp. 33–86. (GSC) assisted with the field surveys. Jennifer Strang (formerly Crombaghs, M., Brügelmann, R., and de Min, E.J. 2000. On the adjustment of GSC) generated the barrier beach and dune crest profile data. overlapping strips of laser altimeter height data. International Archives of We thank Bob Maher and Roger Mosher (AGRG) and Gavin Photogrammetric Engineering and Remote Sensing, No. 33, No. B3/1, Manson and Dustin Whalen (GSC) for discussions and support pp. 230–237. of the project. We would like to thank an anonymous reviewer Daigle, R. 2004. LIDAR: digital elevation modeling. In Climate change and for comments and the guest editor of the special issue, Chris coastal communities: concerns and challenges for today and beyond. Hopkinson, for comments that have improved the manuscript. Southern Gulf of St. Lawrence Coalition on Sustainability, Final Report of This is contribution number ESS 2005738 of the Earth Sciences Workshop, Bouctouche, N.B. Abstract, p. 17. Sector of Natural Resources Canada. DeBaie, L. 2004. New Brunswick sea level rise project: socio-economic impacts. In Climate change and coastal communities: concerns and challenges for today and beyond. Southern Gulf of St. Lawrence Coalition References on Sustainability, Final Report of Workshop, Bouctouche, N.B. Abstract, Ahokas, E., Kaartinen, H., and Hyyppa, J. 2003. A quality assessment of p. 19. airborne laser scanner data. In 3-D reconstruction from airborne Dickie, S. 2001. Storm-surge inundation mapping for Charlottetown, Prince laserscanner and InSAR data. Edited by H.-G. Maas, G. Vosselman, and A. Edward Island: analysing flood risk using high-resolution imagery. Streilein. Institute of Photogrammetry and Remote Sensing (IPRS), GITC, Applied Geomatics Research Group (AGRG), Nova Scotia Community Wageningen, The Netherlands. pp. 1–7. College (NSCC), Middleton, N.S. Artuso, R., Bovet, S., and Streilen, A. 2003. Practical methods for the Donnelly, J.P., and Bertnes, M.D. 2001. Rapid shoreward encroachment of salt verification of countrywide terrain and surface models. In 3-D marsh cordgrass in response to accelerated sea-level rise. Proceedings of reconstruction from airborne laserscanner and InSAR data. Edited by H.- the National Academy of Sciences, Vol. 98, pp. 14 218 – 14 233. G. Maas, G. Vosselman, and A. Streilein. Institute of Photogrammetry and Remote Sensing (IPRS), GITC, Wageningen, The Netherlands. pp. 14–19. Donnelly, J.P., Cleary, P., Newby, P., and Ettinger, R. 2004. Coupling instrumental and geological records of sea-level change: evidence from Bertnes, M.D., Ewanchuk, P.J., and Silliman, B.R. 2002. Anthropogenic southern New England of an increase in the rate of sea-level rise in the late modification of New England salt marsh landscapes. Proceedings of the 19th century. Geophysical Research Letters, Vol. 31, doi: National Academy of Sciences, Vol. 99, pp. 1395–1398. 10.1029/2003GL018933. Brennan, R., and Webster, T.L. 2005. An evaluation of LiDAR intensity for Elberink, S.O., Brand, G., and Brügelmann, R. 2003. Quality improvement of landcover classification. In Proceedings of the 26th Canadian Remote laser altimetry DEMs. In 3-D reconstruction from airborne laserscanner Sensing Symposium, 14–16 June 2005, Wolfville, NS. Canadian and InSAR data. Edited by H.-G. Maas, G. Vosselman, and A. Streilein. Aeronautics and Space Institute (CASI), Ottawa, Ont. Abstract: Institute of Photogrammetry and Remote Sensing (IPRS), GITC, Bretar, F., Pierrot-Deseilligny, M., and Roux, M. 2003. Estimating intrinsic Wageningen, The Netherlands. pp. 51–58. accuracy of airborne laser data with local 3-D-Offsets. In 3-D Elko, N., Sallenger, A.H., Jr., Guy, K., Stockdon, H.F., and Morgan, K.L.M. reconstruction from airborne laserscanner and InSAR data. Edited by H.- 2002a. Barrier island elevations relevant to potential storm impacts: 1. G. Maas, G. Vosselman, and A. Streilein. Institute of Photogrammetry and Techniques. US Geological Survey, Open-file Report 02-287. 6 pp. Remote Sensing (IPRS), GITC, Wageningen, The Netherlands. pp. 20–26. Elko, N., Sallenger, A.H., Jr., Guy, K., and Morgan, K.L.M. 2002b. Barrier Brock, J.C., Wright, C.W., Sallenger, A.H., Krabill, W.B., and Swift, R.N. island elevations relevant to potential storm impacts: 2. South Atlantic.US 2002. Basis and methods of NASA airborne topographic mapper LIDAR Geological Survey, Open-file Report 02-288. 4 pp. surveys for coastal studies. Journal of Coastal Research, Vol. 18, pp. 1–13. Filin, S. 2001. Recovery of systematic biases in laser altimeters using natural Burman, H. 2000. Adjustment of laser scanner data for correction of surfaces. International Archives of Photogrammetry, Remote Sensing and orientation errors. International Archives of Photogrammetry and Remote Spatial Information Sciences, Vol. 34, No. 3/W4, pp. 85–91. Sensing, Vol. 33, No. B3/1, pp. 125–132. Filin, S. 2003a. Analysis and implementation of a laser strip adjustment Chouinard, O., Délusca, K., and Tramblay, M. 2004. Erosion and climate model. In 3-D reconstruction from airborne laserscanner and InSAR data. change: perceptions of group representatives in southeastern New Edited by H.-G. Maas, G. Vosselman, and A. Streilein. Institute of Brunswick. In Climate change and coastal communities: concerns and Photogrammetry and Remote Sensing (IPRS), GITC, Wageningen, The challenges for today and beyond. Southern Gulf of St. Lawrence Coalition Netherlands. pp. 65–70. on Sustainability, Final Report of Workshop, Bouctouche, N.B. Extended Abstract, pp. 11–12. Filin, S. 2003b. Recovery of systematic biases in laser altimetry data using natural surfaces. Photogrammetric Engineering & Remote Sensing, Vol. 69, No. 11, pp. 1235–1242.

Forbes, D.L., Parkes, G.S., Manson, G.K., and Ketch, L.A. 2004. Storms and Sensing Annual Conference, Washington, D.C. American Society of shoreline retreat in the southern Gulf of St. Lawrence. Marine Geology, Photogrammetry and Remote Sensing (ASPRS), Bethesda, Md. p. 13. Vol. 210, pp. 169–204. Leatherman, S.P., Douglas, B.C., and LaBrecque, J.L. 2003. Sea level and Galy, H.M., and Sanders, R.A. 2002. Using synthetic aperture radar imagery coastal erosion require large-scale monitoring. Eos, Transactions of the for flood modelling. Transactions in GIS, Vol. 6, No. 1, pp. 31–42. American Geophysical Union, Vol. 83, No. 2, pp. 13–16.

Grant, D.R. 1980. Quaternary sea-level change in Atlantic Canada as an Maas, H.-G. 2000. Least-squares matching with airborne laserscanning data in indication of crustal delevelling. In Earth rheology, isostasy and eustasy. a TIN structure. International Archives of Photogrammetric Engineering Edited by N.A. Morner. John Wiley and Sons, London, UK. and Remote Sensing, Vol. 33, No. B3/1, pp. 548–555.

Hanson, A. 2004. Ecosystems. In Climate change and coastal communities: Maas, H.-G. 2002. Methods for measuring height and planimetry Concerns and challenges for today and beyond. Southern Gulf of St. discrepancies in airborne laserscanner data. Photogrammetric Engineering Lawrence Coalition on Sustainability, Final Report of Workshop, & Remote Sensing, Vol. 68, No. 9, pp. 933–940. Bouctouche, N.B. Abstract, p. 26. MacKinnon, E. 2004. Three-dimensional flood modelling with LIDAR. Hodgson, M.E., and Bresnahan, P. 2004. Accuracy of airborne lidar-derived Applied Geomatics Research Report, Applied Geomatics Research Group elevation: empirical assessment and error budget. Photogrammetric (AGRG), Nova Scotia Community College (NSCC), Middleton, N.S. Engineering & Remote Sensing, Vol. 70, No. 3. pp. 331–339. 200 pp.

Hodgson, M.E., Jensen, J.R., Schmidt, L., Shill, S., and Davis, B. 2003. An McCulloch, M.M., Forbes, D.L., Shaw, R.W., and the CCAF A041 Scientific evaluation of LIDAR and IFSAR derived digital elevation models in leaf- Team. 2002. Coastal impacts of climate change and sea-level rise on on conditions with USGS Level 1 and Level 2 DEMs. Remote Sensing of Prince Edward Island. Edited by D.L. Forbes and R.W. Shaw. CD-ROM. Environment, Vol. 84, No. 2, pp. 295–308. Geological Survey of Canada, Open File 4261, xxxiv + 62 pp. and 11 supporting documents. Hodgson, M.E., Jensen, J., Raber, G., Tullis, J., Davis, B.A., Thompson, G., and Schuckman, K. 2005. An evaluation of lidar-derived elevation and McLean, R.F., Tsyban, A., Burkett, V., Codignotto, J.O., Forbes, D.L., terrain slope in leaf-off conditions. Photogrammetric Engineering & Mimura, N., Beamish, R.J., and Ittekkot, V. 2001. Coastal zones and marine Remote Sensing, Vol. 71, No. 7, pp. 817–823. ecosystems. In Climate change 2001: impacts, adaptations and vulnerability. Contribution of Working Group II to the Third Assessment Hopkinson, C., Chasmer, L.E., Sass, G., Creed, I.F., Sitar, M., Kalbfleisch, W., Report of the Intergovernmental Panel on Climate Change. Cambridge and Treitz, P. 2005. Vegetation class dependent errors in lidar ground University Press, Cambridge and New York. pp. 343–379. elevation and canopy height estimates in a boreal wetland environment. Canadian Journal of Remote Sensing, Vol. 31, No. 2, pp. 191–206. Nicholls, R.J., Hoozemans, F.M.J., and Marchand, M. 1999. Increasing flood risk and wetland losses due to global sea-level rise: regional and global Huising, E.J., and Gomes Pereira, L.M. 1998. Errors and accuracy estimates of analyses. Global Environmental Change, Vol. 9, pp. S69–S87. laser data acquired by various laser scanning systems for topographic applications. Journal of Photogrammetry and Remote Sensing, Vol. 53, Nichols, S., and Sutherland, M. 2004. Building on local knowledge: No. 5, pp. 245–261. community-based adaptation strategies for sea level rise. In Climate change and coastal communities: Concerns and challenges for today and beyond. Katzenbeisser, R. 2003. On the calibration of LIDAR sensors. In 3-D Southern Gulf of St. Lawrence Coalition on Sustainability, Final Report of reconstruction from airborne laserscanner and InSAR data. Edited by H.- Workshop, Bouctouche, N.B. Abstract, p. 29. G. Maas, G. Vosselman, and A. Streilein. Institute of Photogrammetry and Remote Sensing (IPRS), GITC, Wageningen, The Netherlands. pp. 59–64. O’Reilly, C. 2000. Defining the coastal zone from a hydrographic perspective. Backscatter, Vol. 11, No. 2, pp. 20–24. Kilian, J., Haala, N., and Englich, M. 1996. Capture and evaluation of airborne laser scanner data. International Archives of Photogrammetric Engineering O’Reilly, C.T., Forbes, D.L., and Parkes, G.S. 2005. Defining and adapting to and Remote Sensing, Vol. 31, No. B3, pp. 383–388. coastal hazards in Atlantic Canada: facing the challenge of rising sea levels, storm surges and shoreline erosion in a changing climate. In Ocean Koohzare, A., Vaní ek, P., and Santos, M. 2005. Compilation of a map of Yearbook Volume 19. Edited by A. Chircop and M.L. McConnell. vertical crustal movements in Eastern Canada using spline polynomials. University of Chicago Press, Chicago, Ill. pp. 189–207. Elements, Vol. 23, pp. 30–33. Parkes, G.S., and Ketch, L.A. 2002. Storm-surge climatology. In Coastal Kornus, W., and Ruiz, A. 2003. Strip adjustment of LIDAR data. In 3-D impacts of climate change and sea-level rise on Prince Edward Island. reconstruction from airborne laserscanner and InSAR data. Edited by H.- Edited by D.L. Forbes and R.W. Shaw. CD-ROM. Geological Survey of G. Maas, G. Vosselman, and A. Streilein. Institute of Photogrammetry and Canada, Open File 4261, Supporting Document 2. 87 pp. Remote Sensing (IPRS), GITC, Wageningen, The Netherlands. pp. 47–50. Parkes, G.S., Ketch, L.A., and O’Reilly, C.T. 1997. Storm-surge events in the Krabill, W.B., Wright, C., Swift, R., Frederick, E., Manizade, S., Yungel, J., Maritimes. In Proceedings of the Canadian Coastal Conference 1997. Martin, C., Sonntag, J., Duffy, M., and Brock, J. 1999. Airborne laser Canadian Coastal Science & Engineering Association, Guelph, Ont. mapping of Assateague National Seashore Beach. Photogrammetric pp. 115–129. Engineering & Remote Sensing, Vol. 66, pp. 65–71. Parkes, G.S., Forbes, D.L., and Ketch, L.A. 2002. Sea-level rise. In Coastal Kraus, K., and Pfeifer, N. 1998. Determination of terrain models in wooded impacts of climate change and sea-level rise on Prince Edward Island. areas with airborne laser scanner data. Journal of Photogrammetry and Edited by D.L. Forbes and R.W. Shaw. CD-ROM. Geological Survey of Remote Sensing, Vol. 53, No. 4, pp. 193–203. Canada, Open File 4261, Supporting Document 1. 33 pp. and 5 attachments. Latypov, D., and Zosse, E. 2002. LIDAR data quality control and system calibration using overlapping flight lines in commercial environment. In Roberts, D., Webster, T.L., and Forbes, D.L. 2005. Multi-temporal analysis of Proceedings of the American Society of Photogrammetry and Remote La Dune de Bouctouche, NB, using high-resolution LiDAR. In

Proceedings of the 26th Canadian Remote Sensing Symposium, 14–16 June Nova Scotia, Canada. In GIS for coastal zone management. Edited by D. 2005, Wolfville, N.S. Canadian Aeronautics and Space Institute (CASI), Bartlett and J. Smith. CRC Press, Boca Raton, Fla. Chapt. 15, pp. 195–218. Ottawa, Ont. Abstract: Wehr, A., and Lohr, U. 1999. Airborne laser scanning — an introduction and Sallenger, A.B., Jr., Krabill, W., Brock, J., Swift, R., Jansen, M., Manizade, S., overview. Journal of Photogrammetry and Remote Sensing, Vol. 54, Richmond, B., Hampton, M., and Eslinger, D. 1999. Airborne laser study pp. 68–82. quantifies El Niño-induced coastal change. Eos, Transactions of the American Geophysical Union, Vol. 80, pp. 89, 92.

Schenk, T., Seo, S., and Csatho, B. 2001. Accuracy study of airborne laser altimetry data. International Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol. 34, No. 3/W4, pp. 113–118.

Schuckman, K. 2003. Announcement of the proposed ASPRS binary lidar data file format standard. Photogrammetric Engineering & Remote Sensing, Vol. 69, No. 1, pp. 13–19.

Shaw, J., and Ceman, J. 1999. Salt-marsh aggradation in response to late- Holocene sea-level rise at Amherst Point, Nova Scotia, Canada. The Holocene, Vol. 9, pp. 439–451.

Shaw, J., Taylor, R.B., Forbes, D.L., Ruz, M.H., and Solomon, S. 1998. Sensitivity of the coasts of Canada to sea-level rise. Geological Survey of Canada, Bulletin 505.

Stockdon, H.F., Sallenger, A.H., List, J.H., and Holman, R.A. 2002. Estimation of shoreline position and change using airborne topographic LIDAR data. Journal of Coastal Research, Vol. 18, No. 3, pp. 502–513.

Vasseur, L., and Délusca, K. 2004. Potential impacts of climate change on La Dune de Bouctouche and Shediac. In Climate Change and Coastal Communities: Concerns and Challenges for Today and Beyond. Southern Gulf of St. Lawrence Coalition on Sustainability, Final Report of Workshop, Bouctouche, N.B. Extended Abstract, pp. 44–49.

Webster, T.L. 2005. LIDAR validation using GIS: a case study comparison between two LIDAR collection methods. Geocarto International, Vol. 20, No. 4, pp. 11–19.

Webster, T.L., and Forbes, D.L. 2006. Airborne laser altimetry for predictive modeling of coastal storm-surge flooding. In Remote sensing of aquatic coastal ecosystem processes: science and management applications. Edited by L.L. Richardson and E.F. LeDrew. Springer, New York. Chapt. 7., pp. 157–182.

Webster, T.L., Dickie, S., O’Reilly, C., Forbes, D.L., Thompson, K., and Parkes, G. 2001. Integration of diverse datasets and knowledge to produce high resolution elevation flood risk maps for Charlottetown, Prince Edward Island, Canada. In CoastGIS2001, 4th International Symposium on Computer Mapping and GIS for Coastal Zone Management, Halifax, N.S. Vol. 2, 27 pp.

Webster, T.L., Forbes, D.L., Dickie, S., Covill, R.A., and Parkes, G.S. 2002. Airborne imaging, digital elevation models and flood maps. In Coastal impacts of climate change and sea-level rise on Prince Edward Island. Edited by D.L. Forbes and R.W. Shaw. CD-ROM. Geological Survey of Canada, Open File 4261. Supporting Document 8. 36 pp.

Webster, T.L., Dickie, S., O’Reilly, C., Forbes, D.L., Parkes, G., Poole, D., and Quinn, R. 2003. Mapping storm-surge flood risk using a LIDAR-derived DEM. Elevation (Supplement to Geospatial Solutions and GPS World), May 2003, pp. 4–9.

Webster, T.L., Forbes, D.L., Dickie, S., and Shreenan, R. 2004a. Using topographic lidar to map flood risk from storm-surge events for Charlottetown, Prince Edward Island, Canada. Canadian Journal of Remote Sensing, Vol. 30, No. 1, pp. 1–13.

Webster, T.L., Christian, M., Sangster, C., and Kingston, D. 2004b. High- resolution elevation and image data within the Bay of Fundy coastal zone,

View publication stats