Tim Webster

Object-oriented land cover classification of lidar-derived surfaces

Light detection and ranging (lidar) provides high-resolution vertical and horizontal spatial data and has become an important technology for generating digital elevation models (DEMs) and digital surface models (DSMs). The latest terrestrial lidar sensors record intensity and echo information for each pulse in addition to range. In this study, lidar height and intensity data were used to classify land cover using an object-oriented approach. The study area was selected based on the variety of land cover types present and consists of urban, mixed forest, and wetland-estuary coastal environments. Surfaces constructed from the lidar points included DSM, DEM, intensity, multiple echos, and normalized height. These surfaces were segmented and classified using object rule based classification. Ten classes were extracted from the lidar data, including saturated and non-saturated intertidal sediments, saturated or stressed and lush ground cover vegetation, low and tall deciduous and coniferous trees, roads and bare soil, bright-roofed structures, dark-roofed structures, and water. The accuracy of the classification was assessed using independent ground reference polygons interpreted from colour orthophotographs and intensity images. The average accuracy of the 10 classes was 94%, but improved to 98% when the classification results were aggregated into seven classes. The results indicate that accurate land cover maps can be generated from a single lidar survey using the derived surfaces, and that image object segmentation and rule-based classification techniques allow the exploitation of spectral and spatial attributes of the lidar data.

Flood-risk mapping for storm-surge events and sea-level rise using lidar for southeast New Brunswick

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.

LIDAR Validation Using GIS: A Case Study Comparison between Two LIDAR Collection Methods

Abstract In the summer of 2000, the Annapolis Valley of Nova Scotia, Canada was selected for a high‐resolution elevation survey utilizing LIDAR (Light Detection And Ranging). Two different LIDAR systems were used to acquire data for the area. The vertical accuracy specification for the survey called for heights to be within an average of 15 cm of measured GPS heights and 95% of the data to be within 30 cm. Prior to the application of these data to geoscientific problems, extensive validation procedures were employed. High precision GPS and traditional surveys were conducted to collect height validation checkpoints. Two validation methods were developed in a GIS environment that involved comparing the checkpoints to the original LIDAR points and to an interpolated “bald earth” DEM. A systematic height error between flight lines for one of the LIDAR methods was detected that related to the calibration procedures used in the survey. This study highlights the differences between laser systems, calibration and deployment methodologies and emphasizes the necessity for independent validation data.

Airborne Laser-Induced Fluorescence (LIF) Light Detection and Ranging (LiDAR) for the Quantification of Dissolved Organic Matter Concentration in Natural Waters

Analysis of dissolved organic matter (DOM) concentration and composition is essential to quantifying biological and chemical oxygen demand and atmosphere–ocean heat flux exchange in natural waters. However, manual water sampling is costly and time consuming over large areas. The purpose of this research was to analyze the applicability of airborne laser-induced fluorescence light detection and ranging (LiDAR) for the detection of DOM in estuarine ecosystems impacted by agriculture. A fluorescence LiDAR system (Airborne Marine) (FLS-AM) was used to assess the DOM concentration of the Annapolis River and Basin, Nova Scotia, Canada, as well as three rivers and their estuaries in Prince Edward Island, Canada. Two FLS-AM flight missions were conducted in the summers of 2008 and 2009 and the resulting datasets were compared with spectral fluorescence signature (SFS DOM) and dissolved organic carbon (DOC) analysis of in situ water samples. Significant positive correlations were found at five of seven sites between the FLS-AM DOM and SFS DOM relationship which indicates that the FLS-AM sensor is a good surrogate for traditional sample collection of DOM data in estuaries in this region. Positive correlations were also found at all sites between FLS-AM DOM values and DOC. FLS-AM DOM patterns show that DOM values are significantly higher in rivers and estuaries that drain watersheds which are heavily impacted by agricultural practices. The results of this study show that the FLS-AM can be used efficiently as a general indicator for how estuaries are affected by runoff from agricultural watersheds in real time and thus reduce the requirement for traditional water sample collection and laboratory analysis methods.

Flood Risk Mapping Using LiDAR for Annapolis Royal, Nova Scotia, Canada

A significant portion of the Canadian Maritime coastline has been surveyed with airborne Light Detection and Ranging (LiDAR). The purpose of these surveys has been to map the risk of flooding from storm surges and projected long-term sea‑level rise from climate change and to include projects in all three Maritime Provinces: Prince Edward Island, New Brunswick, and Nova Scotia. LiDAR provides the required details in order to map the flood inundation from 1 to 2 m storm surge events, which cause coastal flooding in many locations in this region when they occur at high tide levels. The community of Annapolis Royal, Nova Scotia, adjacent to the Bay of Fundy, has been surveyed with LiDAR and a 1 m DEM (Digital Elevation Model) was constructed for the flood inundation mapping. Validation of the LiDAR using survey grade GPS indicates a vertical accuracy better than 30 cm. A benchmark storm, known as the Groundhog Day storm (February 1–3, 1976), was used to assess the flood maps and to illustrate the effects of different sea-level rise projections based on climate change scenarios if it were to re-occur in 100 years time. Near shore bathymetry has been merged with the LiDAR and local wind observations used to model the impact of significant waves during this benchmark storm. Long-term (ca. greater than 30 years) time series of water level observations from across the Bay of Fundy in Saint John, New Brunswick, have been used to estimate return periods of water levels under present and future sea-level rise conditions. Results indicate that under current sea-level rise conditions this storm has a 66 year return period. With a modest relative sea-level (RSL) rise of 80 cm/century this decreases to 44 years and, with a possible upper limit rise of 220 cm/century, this decreases further to 22 years. Due to the uncertainty of climate change scenarios and sea-level rise, flood inundation maps have been constructed at 10 cm increments up to the 9 m contour which represents an upper flood limit estimate in 100 years, based on the highest predicted tide, plus a 2 m storm surge and a RSL of 220 cm/century.

GIS Modelling of Intertidal Wetland Exposure Characteristics

Abstract Exposure to solar radiation and tidal inundation are important factors for a wide variety of chemical and ecological processes in coastal ecosystems. Accurate quantification of these factors is often difficult on a local scale. To address this research gap, a remote-sensing approach was developed to model inundation and radiation characteristics within an intertidal zone located in the Minas Basin (Bay of Fundy, Nova Scotia, Canada). A light detection and ranging (LIDAR)–derived elevation model was subjected to tidal modelling based on hourly sea level predictions and solar modelling based on sunrise and sunset times for 2009. Model results indicated an intertidal zone of 145.8 km2 with an elevation between −6.9 m and 6.8 m. The intertidal zone was determined to contain three unique wetland classes: (1) 4.4 km2 of high salt marsh, dominated by Spartina patens; (2) 5.0 km2 of low salt marsh, dominated by Spartina alterniflora; and (3) 63.1 km2 of nonvegetated marine flat (73.3 km2 unclassified intertidal). Detailed exposure characteristics were calculated for each of the classes within the intertidal zone at 10-cm vertical intervals. Exposure calculations for 2009 showed that an average of 4.2 km2 of salt marsh were exposed to solar radiation and 8.4 km2 were exposed to the atmosphere each hour. Similarly, 11.7 km2 of marine flat were exposed to solar radiation and 22.9 km2 were exposed to the atmosphere each hour. The developed remote-sensing techniques successfully established intertidal zones, uniquely identified wetland classes, and modelled inundation and solar exposure characteristics within the study area.

Optimization of Data Collection and Refinement of Post-processing Techniques for Maritime Canada's First Shallow Water Topographic-bathymetric Lidar Survey

ABSTRACT Webster, T.; McGuigan, K.; Crowell, N.; Collins, K., and MacDonald, C., 2016. Optimization of data collection and refinement of post-processing techniques for Maritime Canadas first shallow water topographic-bathymetric lidar survey. In: Brock, J.C.; Gesch, D.B.; Parrish, C.E.; Rogers, J.N., and Wright, C.W. (eds.), Advances in Topobathymetric Mapping, Models, and Applications. Journal of Coastal Research, Special Issue, No. 76, pp. 31–43. Coconut Creek (Florida), ISSN 0749-0208. An airborne topographic-bathymetric lidar survey was conducted for five coastal study sites in Maritime Canada in fall 2014 using the shallow water Leica AHAB Chiroptera II sensor. The sensor utilizes near-infrared (NIR) and green lasers to map topography, water surface, and bathymetry, and is equipped with a 60 MPIX camera, which results in 5-cm resolution color and NIR orthophotos. Depth penetration of the lidar sensor is limited by water clarity, and because the coastal zone is vulnerable to reduced water clarity/increased turbidity due to fine-grained sediment suspended by wind-induced waves, several techniques were employed to obtain maximum depth penetration of the sensor. These included monitoring wind speed, direction, and water clarity at study locations, surveying a narrow pass of the study area to assess depth penetration, and quickly adapting to changing weather conditions by altering course to an area where water clarity was less affected by wind-induced turbidity. These techniques enabled 90% depth penetration at all five of the shallow embayments surveyed and up to 6 m depth penetration in the exposed coastal region. Synchronous ground truth surveys were conducted to measure water depth and clarity and seabed cover during the surveys. GPS checkpoints on land indicated that the topographic lidar had an accuracy of better than 10 cm RMSE in the vertical. The amplitude of the green laser bathymetric returns provides information on bottom type and can be useful for generating maps of vegetation distribution. However, these data are not automatically compensated for water depth attenuation and signal loss in post-processing, which results in difficulties in interpreting the amplitude imagery derived from the green laser. An empirical approach to generating a depth-normalized amplitude image which is merged with elevation derivatives to produce a 2-m resolution map product that is easily interpreted by end users is presented. An eelgrass distribution model was derived from the bathymetric elevation parameters with 80% producers accuracy.

RADARSAT‐1 Imagery and GIS Modeling for Mineral Exploration in Nova Scotia, Canada

Abstract Radarsats Application Development and Research Opportunity (ADRO) Project 656, with the Canadian Space Agency, has provided a wealth of Radarsat imagery for the Centre of Geographic Sciences. The focus has been on the validation of Radarsat data pertaining to geological investigations and base metal minerals exploration modeling. Two adjacent, descending Radarsat orbital tracks, comprising four Standard Beam Mode (S2) scenes, were ortho‐rectified, histogram matched, filtered and mosaicked. The imagery encompasses a substantial portion of central Nova Scotia, Canada. The S2 mode Radarsat imagery has been integrated with digital geological, geophysical and stream sediment mineral data layers utilising GIS and image analysis software packages. Within the broader framework of ADRO Project 656, mineral potential models have been constructed for a selected sub‐area in north‐central Nova Scotia. These models, which consist of value‐added Radarsat S2, geological, geophysical and stream sediment mineral concentration data layers, have contributed to a better understanding of the geological nature of central Nova Scotia and facilitated potential identification of mineral emplacement targets.

Remote predictive mapping of aggregate deposits using lidar.

Remote predictive mapping techniques have been used in the Annapolis Valley, Nova Scotia, to map surficial aggregate deposits. High-resolution digital elevation models (DEMs) have been derived from light detection and ranging (lidar) data for two study sites. The lidar surveys were conducted in the spring of 2003 and 2004, ensuring that a significant number of the pulses make it to the ground or near the ground. The glacial landforms were interpreted for each site, and sediment-depth maps were calculated. The sediment depth for the Annapolis Valley site was constrained by drill hole information and pit locations. Several potential target sites have been identified using the DEM data. The North Mountain site represents a network of eskers that rest on the North Mountain Basalt, which dips 6° toward the Bay of Fundy. Sediment depth and volume were calculated for this deposit by constructing a plane representing the bedrock surface and subtracting it from the DEM. Although the landforms on the North Mountain had been previously mapped using traditional methods, details of their morphology and extent were not well resolved. The lidar provided adequate resolution to identify several new potential aggregate deposits. Follow-up fieldwork for the Annapolis Valley site has revealed that some of the targets are comprised of sand, and others do not have enough pure sand to make them economically viable at the present time.

Evaluation of the Traffic Impacts of Mass Evacuation of Halifax: Flood Risk and Dynamic Traffic Microsimulation Modeling

This paper presents a comprehensive evaluation of traffic impacts of a mass evacuation of the Halifax Peninsula under several flooding scenarios. Flood extent and associated damages to the transport network are identified through digital elevation modeling that intersects with the Halifax stream and transport networks. The resulting flood scenarios inform a traffic microsimulation model that uses a dynamic traffic assignment-based microsimulation approach and simulates the evacuation of 34,808 evacuees estimated from the Halifax Regional Transport Network Model. The simulation results suggest that flooding of the links by 7.9 m flood reduces alternative evacuation routes by 31.2%. It takes 15 hours to evacuate 83% of evacuees while the remaining 17% are not accommodated in the network due to reduced network capacity. The number of vehicles in the network has peaked at 13,000 in this flooding scenario. An evaluation of network performance reveals a sustained congestion prevailing from 4th to 7th hour of the evacuation. The novelty of this study is that it develops a comprehensive tool of flood risk and dynamic traffic microsimulation modeling to offer an in-depth evaluation of potential impacts during evacuation. The results will help emergency professionals in evacuation planning and making emergency decisions.