Skip Maine state header navigation

Agencies | Online Services | Help

Skip First Level Navigation | Skip All Navigation

Home > Explore! > Coastal Marine Geology > LIDAR - Casco > Methods


Assessment of LIDAR for Simulating Existing and Potential Future Marsh Conditions in Casco Bay

Methodology

Real Time Kinematic Global Positioning System Surveys

location map
Figure 2		 With guidance from staff at Casco Bay Estuary Project (CBEP), MGS selected two separate marsh areas for groundtruthing studies within Casco Bay, and received data from the CBEP for a third location as part of its efforts. The study areas were selected to coincide with available FEMA LIDAR data, and were meant to cover different geographic regions of Casco Bay (Figure 2). They included:
  1. finger marsh system at Cousins River, Yarmouth,
  2. fringing marsh system at Back Cove, Portland; and
  3. finger marsh system at Thomas Bay, Brunswick

At all field sites, GPS data was collected using Ashtech Z-Xtreme RTKGPS rover and base units with Ashtech Geodetic IV antennae and Tripod Data Systems (TDS) SurveyPro software and stored on a TDS Ranger data collector. Horizontal Dilution of Precision (HDOP) and Vertical Dilution of Precision (VDOP) values were monitored in the field to ensure that values, to the maximum extent practicable, remained between 1 and 2 (ideal to excellent), and not record values above 3 (good), if at all possible. These grades of precision are required for highest or most demanding data needs. Data was transferred to a personal computer (PC) via TDS SurveyLink software into formats compatible with GIS analysis. Raw GPS data was stored in SurveyLink proprietary (.RAW) formats.

GPS points were converted to UTM NAD83 Zone 19 (GEOID99 orthometric elevations were converted to GEOID03 orthometric elevations using the geoidal separation between the 2 geoids for the 2 study locations provided from the NOAA National Geodetic Survey (2009)). Thomas Bay, Brunswick GPS data was provided in GEOID03.

Cousins River, Yarmouth.

There were several goals in undertaking a field study at the Cousins River site in Yarmouth. This study was considered a "pilot" in determining whether or not the LIDAR datasets could be used for successfully mapping existing marsh features and subsequently simulating the potential impacts of sea level rise. In that regard, our goals for the study were to:

  • identify and map the extent of the Highest Annual Tide (HAT, 6.4 ft NAVD), which occurred on January 12, 2009 and reached a verified elevation of 6.56 ft NAVD;
  • capture additional elevation data of the existing marsh surface while in the field;
  • compare the field identified HAT with that derived from 2006 FEMA LIDAR data; and
  • compare field measured marsh elevations with those extracted from LIDAR topographic data;
  • compare the horizontal spatial difference between the HAT derived from LIDAR and the field mapped HAT.

In order to meet the first goal, field studies were conducted during early January, 2009, under difficult conditions, including deep snow and extensive ice cover on both upland and marsh surfaces. Horizontal and vertical survey control, along with base point establishment, was completed using numerous survey benchmarks within the area. Control work and base point setting was completed from January 9 through January 12, 2009. Due to snow, ice and stormy weather conditions, field GPS data was collected on January 13, 2009 - a day after the higher tide was measured. However, field efforts needed to be undertaken as close to the time of highest tide regardless of field conditions in order to capture the higher tidal event. Data was collected on a cold, windy day with considerable cloudiness.

tidal data Jan 2009
Figure 3		 Verified tidal data available from NOAA CO-OPS showed that the high tide reached an elevation of 6.56 ft NAVD at 11:42 am on January 12, 2009 (Figure 3). This was slightly higher than the predicted HAT, which was 6.4 feet (based on Prince's Point, Yarmouth; Dickson, 2009).

Field data collected included the demarcation of the "wet-dry" line, which marked the extent of the HAT from January 12, 2009, on both the north and south sides of Interstate 295, in addition to elevation data along the high marsh surface on the northern side of I-295.

Although difficult to continuously delineate due to snow and ice cover and ice rafting, there was a notable wet-dry line observed in the field that could be followed somewhat consistently during field efforts, especially on the south side of I-295. On the north side, field definition of the boundary was more difficult. Examples of the visible wet-dry line representing the limits of the HAT are shown in Figure 4 and Figure 5.

wet-dry line south of I-295
Figure 4		 wet-dry line north of I-295
Figure 5

Marsh elevations were also recorded, to the maximum extent practicable. The ice surface was too thick to break through on a consistent basis, so random points were recorded at locations where there were apparent holes in the ice, or where the survey pole could be punched through to the frozen marsh. It was noted that GPS signal quality varied during the survey, especially adjacent to tree-lined uplands.

Cousins River data points
Figure 6		 A total of 118 GPS points were collected to define the field extent of the HAT, while an additional 182 points defined the elevation of the high marsh, and 43 points were collected to inspect elevations of the "low marsh". GPS signal quality was relatively good, but not ideal. The resulting data points for all surveyed elevations are shown overlain on 2001 aerial orthophotographs (MEGIS, 2001) in Figure 6.

Back Cove, Portland

The field study at Back Cove was undertaken to inspect the elevations of an "urban" marsh. Back Cove has a relatively large area of fringing marsh along Baxter Boulevard, and low marsh comprises a much higher percentage of the marsh than at the Cousins River site. The purposes of the study here included:

  • identify and map in the field the high marsh-low marsh boundary;
  • capture additional elevation data of the existing marsh surface and some surrounding uplands;
  • compare the field identified boundary between the high and low marsh vegetation with LIDAR derived mean high water (4.21 ft NAVD);
  • compare field measured marsh elevations with those extracted from LIDAR topographic data.

Horizontal and vertical survey control, along with base point establishment, was completed using numerous survey benchmarks within the area. Control work and base point setting was completed in late August 2009, and field GPS data was collected on September 2, 2009.

Field data collected included the demarcation of the boundary between the high marsh (dominated by Spartina patens and Juncus gerardi) and low marsh (dominated by Spartina alterniflora). In some areas, the boundary was difficult to delineate where pockets of low marsh incurred into dominant areas of high marsh; in these areas, best field judgment was used. The boundary between these two marsh types is generally approximated by the "mean high water" line (MHW) (calculated to be 4.21 ft NAVD in 2009 using data from NOAA CO-OPS and the Portland tide gauge; Dickson, 2007; 2009).

We also collected elevation data along shore perpendicular transects, which started in the uplands (usually across Baxter Boulevard), and continued into the low marsh. Due to softness of the marsh surface, it was impossible to continue surveying into lower portions of the low marsh and onto the adjacent mudflats.

Back Cove data points
Figure 7		 GPS quality during the survey was very good. A total of 212 GPS points were collected to define the field extent of the boundary between the low marsh and high marsh, and an additional 102 points were collected as part of 8 roughly equally spaced shore-perpendicular transects (40 points were in the "upland" and 62 points extended onto the high and low marsh surfaces). Resulting survey data points are shown overlain on 2001 aerial orthophotographs in Figure 7.

Thomas Bay, Brunswick

Thomas Bay data points
Figure 8		 Collected GPS data was provided to MGS by the Estuary Partnership from survey work completed on May 6, 2009. Data was collected using a Trimble 5700 GPS receiver with Zephyr antenna. Base station data was provided by the Maine Technical Source Continuous Operating Reference Station (CORS, point YMTS), which was held as a fixed control point. A total of 100 points were collected by the University of Southern Maine GIS laboratory, mostly within the marsh area adjacent to the north and south side of Adams Road in Brunswick (Figure 8).

FEMA LIDAR Data Processing

For each of the study site locations, available LIDAR data from the FEMA LIDAR dataset were processed so that later comparisons between field collected GPS elevations and associated LIDAR elevations could be made. Initial analysis of the data showed that the bare-earth .xyz files were actually at approximate 2 foot (not 2 meter) spacing and represented ground elevation data with vegetation removed. Data was provided in NAD1983 Maine State Plane Coordinate System (feet, Maine West, 1802), and referenced to NAVD88 (feet) orthometric elevations. ESRI ArcToolbox was used to transform data to UTM NAD83 Zone 19.

Bare Earth XYZ files

Applicable bare-earth .xyz files were converted to ASCII raster format using an executable file. An input format (including header file, no data), and desired output format (including cell size, and floating point or integer output) were specified; two meter (6 foot) output point spacing and floating point output were used. Once this was completed ArcToolbox>Spatial Analyst Tools>Extract>Extract Values to Points command was used to find the nearest LIDAR data points to the collected GPS data. No interpolation was used; the nearest central grid value was taken.

Triangulated Irregular Network (TIN) files

It was determined in discussions with Dr. Matthew Bampton and Dr. Vinton Valentine of the GIS Department at the University of Southern Maine, who had preliminarily inspected the LIDAR datasets, that extracted nodes from the TIN files may better represent bare earth topographic values for comparing LIDAR with GPS elevation values. Using available TIN data, elevation data nodes were extracted using ArcToolbox>3D Analyst Tools>Conversion>from TIN>TIN node command.

This creates a 2D feature class of the extracted TIN nodes with an optional spot elevation field associated with each node. After nodes were extracted, the nearest node to each GPS point was joined using ArcToolbox>Analysis Tools>Overlay> Spatial Join.

Tidal Elevation Processing

Additional LIDAR processing was completed to derive the Highest Annual Tide (HAT) and Mean High Water (MHW) boundaries for comparison with field collected GPS data for the Cousins River and Back Cove study sites, respectively. Applicable bare earth raster datasets were queried to find all cells with values below the HAT (6.4 ft for the Cousins River) and MHW (4.21 ft for Back Cove) elevations. The same routines were carried out for creating the existing and future marsh surfaces, albeit at specified elevation ranges, discussed later in the report.

LIDAR and GPS Data Comparison

Vertical Comparison

Collected GPS elevation data was subtracted from the nearest extracted LIDAR data points (both TIN nodes and raster values) to determine the difference in elevations.

Horizontal Comparison

We also compared the horizontal spatial variation of a field delineated boundary, such as the HAT, and the same boundary defined using extracted LIDAR data. Extracted LIDAR data was used to create a polyline representing the landward limits of the subject boundary and was compared with the GPS delineated boundary. To do this comparison, the ArcGIS Digital Shoreline Analysis System (DSAS) extension, developed by the USGS (Thieler et al., 2009) was used. A baseline was set on the marsh side of the delineated boundaries, and transects were cast at 5 m intervals in a landward direction along the length of the boundaries. The net difference between the two "shorelines" indicates the horizontal offset between the two boundaries. In areas where the LIDAR delineated boundary was marshward of the GPS delineated boundary, net difference values were negative; where the GPS delineated boundary was located marshward of the LIDAR boundary, net difference values were positive.

Slope Analysis

For analysis at the pilot site (Cousins River), ESRI Spatial Analyst was used to calculate the slopes, in degrees, of the LIDAR data (using the gridded raster as an input). An output raster of slope data was created.

Hawth's Tools (Beyer, 2009) ArcMap extension was then used to analyze adjacent upland slopes where DSAS transects were cast. Transects were cast at lengths of 20 m, at a 5 m interval, in order to match the DSAS transect locations and capture the slopes of the adjacent uplands. The tool calculates a Length Weighted Mean (LWM, in degrees), Maximum, Minimum, and Standard Deviation slope values.

Last updated on March 4, 2010