GIS applications using LiDAR

Module 2 for GIS Applications returns us to take a more in depth look at the use of LiDAR, a topic briefly covered in previous courses. LiDAR (LIght Detection And Ranging) uses lasers, usually in the visible or near-infrared portion of the spectrum, calculates heights by measuring distances between a scanner and target area. The high energy pulses record the reflected response of different objects in a very narrow wave length.  This produces a point cloud, where the masspoints associated with each return are distributed throughout the target area at various densities. The densities vary depending upon the materials encountered by the laser pulses.

Interpolation processes of individual masspoints creates a digital surface model (DSM), which shows the elevation characteristics of natural phenomenon and man-made structures. Another procedure with the removal of LiDAR masspoints in the first, intermediate and/or the last returns produces a bare-Earth digital terrain model (DTM). LiDAR is also effective at measuring water depth relative to the water surface.

Digital Elevation Models (DEM) provide elevation data in a raster format. DEMs with 3, 10 and 30 meter resolution are available for most of the United States. The USGS catalogs DEMs and makes this data available through the Earth Explorer and National Map portals. Furthermore, states also provide DEM datasets through Clearinghouse Websites. 

LiDAR data is often delivered in a standard LAS format, which defines the file structure, content, storage order, naming, codes, etc. The standard LiDAR exchange file uses the .las file extension. The Lab for this week's module utilizes a LiDAR dataset provided by the Virginia Geographic Information Network (VGIN) covering a section of the Shenandoah Mountains at Big Meadows.

North view toward along the Shenandoah Mountains from Skyline Drive. Photo by Andy Field.

Our textbook GIS Fundamentals references discrete-return LiDAR as the most common collection system. This system records records specific values for each laser pulse downward. It produces a point cloud consisting of X, Y, and Z coordinates along with the intensity, scan angle, return order and other information.

Points in the LiDAR point cloud are assigned to feature types such as structures or vegetation. Standard codes identify ground, buildings and water. These are derived by return strength, point order, local slope, etc.

LiDAR data is widely used to estimate vegetation characteristics such as tree height, forest density, growth rates and forest type. The initial part of our lab focuses on calculating forest height starting with the conversion of the LAS file into both a DEM and a Digital Surface Model (DSM).

We first use the Point File Information geoprocessing tool in ArcGIS Pro on the LAS file to summarize the file with values for the minimum bounding rectangle, number of points, the average point spacing, and the min/max z-values. DEM and DSMs are created next using the LAS Dataset to Raster geoprocessing tools. With these two rasters, a calculation using the Minus tool outputs a raster populated with estimated heights.

The LAS file for Big Meadows at Shenandoah National Park and the compiled DEM

The subsequent step in Lab calculates the biomass density of the forested area in question. MultiPoint files for the ground and vegetation are created using the LAS to Multipoint geoprocessing tool on the point file previously created. These are then processed using the Point to Raster tool to output respective rasters.

Continuing with geoprocessing, both ground and vegetation multipoint rasters are further processed using the IS NULL tool to produce a boundary file where similar to Boolean, 1 is assigned to all values that are not null . The Con tool then juxtaposes the IsNull rasters with the Multipoint rasters for both sets so that if a value of zero is encountered, it is accepted as a true value and values of 1 are in turn pulled from the original multipoint rasters. This produces rasters of the cell counts.

The output tree height estimation and the statistics of raster cells (points) versus height

Working forward to calculate the density of the returns, the Plus tool combines the counts for both the vegetation and ground. This results in a raster where all cell counts are assigned an integer value from zero to 23. After converting the raster values from integer to float (decimals), the tree canopy density can finally be calculated. This is completed by using the Divide geoprocessing tool on the raster of the cell counts for vegetation and the combined vegetation/ground counts raster with float values.

Forest Canopy Density of the Big Meadows area of Shenandoah National Park derived from LiDAR