Pacific Southwest Region

Compiling Vegetation Inventories

Vegetation inventories are compiled from two main sources of information, an existing vegetation map covering the entire extent of a National Forest, and a set of field sample plots. Following is a brief description of the inventory compilation process for linking sample plots to a vegetation map, using the vegetation map to post stratify the sample plots, and developing a strata layer to serve as the area expansion of the compiled inventory.

Table of Contents

• Existing Vegetation Maps
• Sample Area of Existing Vegetation Map on the Lassen National Forest
• Sample Plots
• Lassen National Forest FIA Plot Layout
• Linking Sample Plots to Vegetation Maps
• Developing a Population Strata Map and Inventory
• Sample Area of Strata on the Lassen National Forest

Existing Vegetation Maps

Existing vegetation maps form the basis for acreage estimates for land and water, forest and non-forest land, and defining sample populations for the stratification of ground plots into areas of similar vegetation types, tree sizes and canopy closure classes. The vegetation maps are also used to target additional sample plots for rare forest types and plantations.

Vegetation maps are developed from multi-source imagery (30-meter TM imagery and 5-meter Spot imagery), field reconnaissance, map notes, photo editing, image classification for life form and forest size, forest canopy modeling for forest cover and GIS modeling for vegetation typing attributes. The Calveg classification system is used to describe the floristic type of vegetation. Image segmentation is used to delineate vegetation stands or regions with a minimum mapping unit of 2.5 acres. Water bodies are incorporated from 1:24,000 scale USGS 7.5 minute quadrangle map base series cartographic feature files through GIS integration. Reforestation areas are incorporated from a 1:24,000 scale map of forest plantations based on agency records. Updates for changes are guided by change detection projects using TM imagery from two dates, as well as agency records of forest fires, fuel treatments, harvest, and reforestation activities. An example of the existing vegetation map developed for an area on the Lassen National Forest from the process described is shown below.

• Downloadable Existing Calveg Mapping Tiles

top

Sample Area of Existing Vegetation Map on the Lassen National Forest

More Info on the CALVEG classification and mapping program, and Existing Calveg Mapping Tiles for National Forests in the Pacific Southwest Region.

top

Sample Plots

FIA (Forest Inventory and Analysis) sample plot locations are based on a systematic sample design consisting of cluster plots placed roughly 3.4 miles apart, one plot every 7,400 acres. The FIA grid design was established by the Pacific Northwest Research Station as the basis for sampling all forestlands within the State of California. On National Forest lands in the Pacific Southwest Region, all 3.4-mile sites were installed, on both forest and non-forest lands. Additional plots were installed to sample forest types that had too few samples from the 3.4-mile grid design. This occurs when a particular forest type on a National Forest has fewer than 40,000 acres. The existing vegetation map is used to identify areas of rare forest types, and is the basis for optimally allocating additional intensification ground samples on a 2.4-mile, 1.7-mile, or .85-mile grid within these mapped areas. A sample of 3.4-mile plus intensification plots for the Lassen National Forests is shown in the following graphic.

top

Lassen National Forest FIA Plot Layout

Each cluster plot consists of 5 independent subplots, sampling 2.5 acres and is installed in the field without rotation to avoid different vegetation conditions. If part of a plot crosses a vegetation type, forest size and or forest canopy cover class indicating a change in condition, portions of the cluster plot are assigned to different conditions by mapping, referred to as the mapped plot design. If a change occurred on a cluster plot, mapping was only done between subplots, and not within a subplot. At each subplot measurements of tree frequency, live or dead status, species, diameter, crown ratio, and damages were recorded. A sub-sample of trees was measured for age and height. Understory vegetation life form, species, cover and height as well as ground cover were estimated. Additional samples of down log frequency, species, large end size and length was measured along with estimates of fine forest fuels and fuel bed depth. More Info on the FIA plot design and measurement attributes used in these inventories.

top

Linking Sample Plots to Vegetation Maps

Sample plots were linked to existing vegetation maps by assigning the vegetation map label from the map, where the plot is located. This is done through a GIS overlay of plot locations mapped from GPS and map vegetation polygons. Plots that fall near stand boundaries are reviewed for possible split conditions using aerial resource photos and heads up GIS views of vegetation maps and imagery. If the field installation did not recognize a split condition, then the plot is assigned to only one stand, and one map label. If the field installation mapped two or more conditions, then the best map labels are assigned to each split condition as indicated by the maps, photos and imagery. Approximately 10 percent of the cluster plots sample split conditions, where subplots are assigned to different forest populations in the vegetation map.

top

Developing a Population Strata Map and Inventory

Once map labels are assigned to each cluster plot/split condition, post stratification logic is assigned to the existing vegetation map. The goal is to obtain 6 or more cluster plots or 30 or more subplots per vegetation map stratum population to form a statistical sample for each vegetation condition. This in turn allows for estimating the area weight on the map that each sample cluster plot represents in the sample area expansion.

Using the post stratification process of assigning plots to map strata populations results in detailed population estimates of vegetation types, and reduces variance of overall estimates of basal area, number of trees, volume, growth and mortality. This is also useful in habitat modeling and estimating old growth forests since like vegetation types, forest size and condition are key attributes in estimating serial stages of vegetation conditions. The grid systematic sample, while a good design for tracking trends in changes over time, and for sampling common vegetation types and conditions adequately, can completely miss, and often inadequately samples rare forest types and structural conditions. Using vegetation maps to target additional plots compensates for these rare conditions, similar to what can be obtained with a stratified random sample design. Even with intensification, some rare forest types, size and canopy closure combinations have so few acres, they are still missed, and end up lumped into strata of the most similar type or structural condition.

The post stratification process starts with assigning an initial map stratum label to vegetation polygons that have the same life form, most similar floristic type and forest productivity. Some rare forest conditions, even with intensification plots, may have no further stratification for forest size and canopy cover characteristics. If more than 6 cluster plots occur within a general forest type, then forest stand size classes are used to further break out seedling, sapling, and pole sizes from small, medium and large sizes. More detailed breakdown of size classes is done if enough plots area available to do so. Within size classes, further breakdowns based on canopy closure classes of 20 to 50 percent are done, but only if enough plots still exist to create a meaningful statistical estimate. To assure consistent links of plots to a final strata map, all map labels are assigned to strata in a relational data base, and then the same logic is used assign strata labels to the plots, using the linked vegetation map label assigned to each cluster plot/split condition. A vegetation map label to strata crosswalk report is available for each National Forest. Example report for the Lassen National Forest.

After assigning all map labels to strata labels, a GIS coverage is created by clipping the existing vegetation coverage with an administrative National Forest lands ownership coverage and then dissolving on like strata labels. The result is a simplified strata map where all polygons of the same stratum label represent the area of a vegetation population that is estimated by a set of sample plots with the same stratum label. An example map is shown below.

top

Sample Area of Strata on the Lassen National Forest

• More Info on inventory strata maps
• Download Vegetation Inventory Data

top

USDA Forest Service - Pacific Southwest Region
Last Modified: Friday, 13 April 2012 at 13:20:48 EDT