This week, lab focused on working with LiDAR data to understand forest structure in the Shenandoah region of Virginia. After decompressing the .laz file, I loaded the LAS dataset into a Local Scene in ArcGIS Pro. The result was a dense floating point cloud showing elevation returns, styled in red-yellow-blue based on height in feet (Figure 1, top). While visually striking, the terrain wasn’t fully clear until I generated a DEM to create a ground surface from the filtered LiDAR ground points (Figure 1, bottom).
The canopy density calculation seemed to have been affected by processing limitations within the virtual workspace, possibly due to the size and complexity of the LiDAR data. This likely caused the layer to render mostly low-density values, with the high-density canopy areas not fully represented. Despite this, the resulting map still highlights key features like roads and clearings where vegetation is sparse (Figure 2). This experience underscores the importance of system resources and careful setup when working with large 3D datasets in ArcGIS.
I also calculated canopy height by subtracting the DEM from the DSM, producing a height raster with a maximum tree height of 163 ft and some negative values down to -5 ft. The cells with values under 0 feet are primarily located along roads and slopes, with fewer than expected near streams, likely due to LiDAR return errors or edge effects. The histogram of vegetation heights shows a roughly normal distribution, peaking around 50–60 feet with a mean of 54.4 feet and a standard deviation of 20.4 feet. The data generally stops around 110–120 feet, making the 163-foot tree an outlier. This distribution illustrates the forest’s vertical structure and variation in canopy height (Figure 3).
No comments:
Post a Comment