Tag Archives: LiDAR

Using LiDAR To Detect Changes in Forest Canopy Height Over Time

Background

Coast redwood tree height is of great interest, as these are the tallest trees currently growing on our planet.  There are about 2,000 trees over 100 meters in height, and all of them (except one or two) are coast redwoods.  Then the tallest coast redwood is a little over 116 meters in height.

There are longitudinal studies of redwoods over their range, centering on eleven defined plots, each about 10 m x 1000 m in size, with two each in Jedediah, Prairie Creek, and Redwood National parks and one each in Humboldt Redwoods, Montgomery Woods, Samuel P. Taylor, Big Basin, and Big Creek Parks and Reserves.  All the trees and vegetation in the plots are measured every three to five years and results are tracked over time. 

The overall consensus is the redwoods in the northern plots are holding their own, and in fact growing faster than ever based on dendrochronology studies.  Tan oaks, a companion species to old growth redwoods, are having difficulties due to a spreading root disease (SOD).  Then hemlocks have become susceptible to invasive mistletoe spreading up trunks.  A couple very tall hemlocks in Jedediah Smith upland have fallen or are standing dead.   The largest tree that fell in the plots was a very large Douglas Fir in the Redwood National Park upland plot.

Then there are lists of the tallest redwoods (those over about 105 meters or otherwise locally tall for their area) where these trees are measured every so often, some every year and others perhaps once a decade.

The tree heights for redwoods in the plots and on the tall tree lists are very accurately measured by Laser Range Finders or by crown tape drop (with measuring pole for the tip).  Obviously only so many redwood trees can be assessed in this way.

However, there is another way to track canopy height. Not exact height, but pretty close. It is LiDAR.

LiDAR data is available for many of the areas with tall redwoods.  Availability and download tools are available on the National Map, NOAA Elevation Data, and Open Topography LiDAR Portal.   For   certain areas, LiDAR data is available starting from 2007 (though typically private and not available for download) all the way through 2020 (later years are typically public and available for download).  Be sure to download all points and to work in meters.

I would be remiss not to lobby for public availability of LiDAR data captured for public lands, especially if the LiDAR acquisition was partially or fully publicly funded.

Using LiDAR Data to Measure Canopy Height Over Time

In order to use LiDAR data longitudinally, it is necessary to create LiDAR derivatives for individual surveys, export the height above ground rasters created from these surveys, then compare the height above ground rasters between surveys done at different dates.

To use LiDAR point cloud data to create height above ground, there are several sources with detailed steps.  Let me briefly summarize a method I find works well:

Software – ArcGIS Pro

Filter point cloud to ground points, use Geoprocessing Tool “LAS Dataset to Raster”  to create a dem layer.  Within Tool use Interpolation type “Binning”, Cell Assignment “Average”, Void Fill Method “Natural Neighbor”, Sampling Value “1”, Z Factor “1”.

Next filter point cloud to first return points, use Geoprocessing Tool “LAS Dataset to Raster” to create a dsm layer.  Within Tool use Interpolation type “Binning, Cell Assignment “Maximum”, Void Fill Method “Natural Neighbor”, Sampling Value “1”, Z Factor “1”.

Now that dem and dsm layers have been created, use Geoprocessing Tool “Minus” and subtract the dem layer from the dsm layer.  The resultant layer is the tree height.  A map can be created to show colors by height band.  However, at this point, to access forest height over time, we are interested in exporting this information to a table.

This height layer to table process is a little involved, there is a paper on it, link is here:

https://support.esri.com/en/technical-article/000023109

Since a raster height layer has already been created, we can skip steps 1 and 2 and start with step 3.

Run Geoprocessing Tool “Raster to Point”.  Here the input file is the height above ground layer, the field is “Value”, and an output file will be named and created.  Let’s call this result RasterT_FoundersGrove.

Run Geoprocessing Tool “Add Geometry Attributes”.  Input features are “RasterT_FoundersGrove”, Geometry Properties are “Point x,y,z and m coordinates”, Length Units are “Meters”, Area Units are “Square Meters”, and Coordinate System is “Current Map”.  These attributes are then added to “RasterT_FoundersGrove”. 

Run Geoprocessing Tool “Add Surface Information”.  Input features are again “RasterT_FoundersGrove”.  The input surface is the height above ground layer, for Output Property “Z” is checked (this is height above ground), the Method is forced to “Bilinear” (don’t worry the information behind this Method on point averaging and sampling, it will not be done due to how the height above ground layer was created), and leave Sampling Distance blank.  This adds surface information to RasterT_FoundersGrove.

Run Geoprocessing Conversion Tool “Table to Table”.  Input Rows are again “RasterT_FoundersGrove”.  The output location will default to the project database but can specify any available folder.  The Output Name can be anything but the .csv extension must be explicitly included.  For this example, we can call it FoundersGrove.csv.

The output CSV file will look like this:

OBJECTIDpointidgrid_codePOINT_XPOINT_YZ
656580.0608063421243.6544467320.09380.06080627
666680.1994171421244.65374467320.10480.19941711
676775.6749344421245.65334467320.11675.67493437
68680.0100021421179.68644467318.3550.010002136
69690.0333328421180.6864467318.3670.033332825

Both grid_code and Z are height above ground in meters.  Then POINT_X and POINT_Y are meters east and north in the Universal Transverse Mercator coordinate system.

In a general sense, the 2007 Private LiDAR covers all the northern redwood parks.  Then there is 2017 and 2018 public LiDAR for northern redwood parks south of Eureka.  There is also updated LiDAR, from about 2016, for redwood parks north of Eureka but for now it is private. 

For Founders Grove, there is 2014 and 2018 LiDAR available.  I followed the process noted above and used a max function against the final csv files to find the tallest point within each of 430,000 square meters downloaded for both 2014 and 2018.  I then did these height comparisons:

For points in same square meter where the height was over 70 meters in both 2014 and 2018:

Height ChangeCrown Sq MetersAvg Hgt 2014Avg Hgt 2018ChangeStd Dev of ChgMedian of Chg
Decrease6366285.0482.82-2.222.99-0.83
Increase15362583.6286.312.703.251.48
Total21728784.0385.291.263.880.58
Height Changes in Same Square Meter Canopy Location in Founders Grove, Canopy Over 70 Meters

Then for points in same square meter where the height was over 90 meters in both 2014 and 2018:

Height ChangeCrown Sq MetersAvg Hgt 2014Avg Hgt 2018ChangeStd Dev of ChgMedian of Chg
Decrease1258995.5394.37-1.161.54-0.51
Increase3571894.4795.981.511.610.95
Total4830794.7495.560.811.980.53
Height Changes in Same Square Meter Canopy Location in Founders Grove, Canopy Over 90 Meters

In Founders Grove, the trend is increasing height, of about 0.14 meters per year from 2014 to 2018.  The median height change was an increase of 5.5 inches per year, with 70% of crowns gaining height. 

From this information Founders Grove growth can be assessed as increasing in height, and this is a favorable condition in relation to forest health.

This same comparison can be done over longer periods for many redwood park areas as LiDAR data exists in both the 2007-2010 time frame as well as the 2016-2020 time frame.