Use and utility of rasterized heatmaps - Part 1

High-resolution heatmaps are a useful tool for field foresters, inventory managers, and landowners. They can be used for a wide variety of planning and management activities, including stand delineation, identification of failed plantations, timber trespass, and harvest planning.

In this two-part blog series, we’ll walk through some of the useful ways you can leverage heatmap information to do more than you can with a traditional cruise workup.

SilviaTerra’s CruiseBoost and Basemap remote-sensing assisted inventory systems produces pixel-level treelist estimates of trees by species and diameter across a full area of interest. In both cases, these estimates are used to generate rasterized heatmaps in addition to other summaries.
What are heatmaps?

SilviaTerra’s inventory estimates are rendered into rasterized ‘heatmaps’ at a 15 x 15 m resolution. These .tif files provide a smooth raster surface of quantitative estimates. They may represent forest inventory information, such as total basal area per acre or quadratic mean diameter, or the basal area or number of trees per acre for an individual species. Heatmaps are also generated to summarize other variables of interest, such as volume, value, or estimates of standing carbon.

SilviaTerra heatmaps are delivered in EPSG 2163 – US National Atlas Equal Area – projection, unless otherwise noted.

How do you work with heatmaps?

Heatmap .tif files can be read into any GIS system, such as ESRI software or QGIS, and manipulated and used with the same set of tools as any other quantitative raster file.

What can heatmaps tell you that stand summaries can't?

Rasterized heatmaps provide sub-stand information at a very fine scale. These data layers can be used to identify areas of distinctly high or low stocking within a stand. They provide a great deal more detail than a stand-level summary alone. Stand summaries are by nature averages of all the values within a stand; a heatmap provides a surface of information that gives a more complete picture of the stand.

image2

In this image, the southernmost stand has some very patchy areas to the east of the road - important for a manager to know about, but information that would be lost in a stand-level summary.

Visualization of structure and heterogeneity

In a traditional cruise, plots are installed in a systematic (grid) or random, more scattered sample across a stand. This type of sample gives a forester some information about the distribution of stocking within a stand. However, it’s also common that while cruising, you might take note of distinct patches, or holes, that the cruised plots simply miss. This is one of the greatest benefits of using remote sensing imagery to generate wall-to-wall estimates of stocking.

Each pixel of information is represented and included in the final inventory – so the final heatmaps “fill in the gaps” between plots, accounting for distinct areas not captured by the traditional measurements. Because of this, heatmaps can be particularly useful in visualizing the spatial distribution of stocking within a stand. Heatmaps enable you to identify particularly homogenous or heterogenous patches within stands, providing useful information that simply isn’t available from a stand-level summary of plot data.

Stand (re) delineation

Rasterized heatmaps are readily used for preliminary stand delineation or re-delineation, based on the pixel-level inventory estimates. Manual delineation can easily be performed by adjusting existing polygon boundaries or creating new polygons following demarcations visible in the raster files. There are also a range of more automated GIS tools that can be used for this process, such as clumping approaches and watershed identification toolchains.

Using multiple raster layers as inputs to this type of delineation process is recommended for a more robust and data-driven delineation process. Ways to do this may include performing a principal components analysis on a set of rasters (such as total basal area as well as basal area or quadratic mean diameter of a few selected species of interest). Another way to leverage the data of multiple layers is to create a red-green-blue color composite using 3 different raster layers, which creates a more colorful and interpretable visualization of three different components. This composite approach creates an excellent starting point for manual delineation.

image4