Text Size:   A+ A- A   •   Text Only
Find     
Site Image
Indicator E.a. - Data Availability and Analysis
Data Availability
One inventory of historic data for tree species is available from the Current Vegetation Survey of the USDA Forest Service - Region 6 and the BLM in western Oregon and data from the Forest Inventory and Analysis (FIA) Program (Table 1).  Another database containing species occurrence data and associated environmental variables for understory plant species is available for analysis. 
 
Primary sources of plot data: 
  • Periodic FIA inventories on nonfederal lands in Oregon and Washington, using data available in the Integrated Database (IDB) compiled by the PNW’s FIA Program.
  • Current Vegetation Survey (CVS) for National Forest lands in Region 6 and BLM lands in western Oregon, which we obtained from Region 6/BLM. These data include the full grid intensification on BLM lands and two measurement occasions for most CVS plots on National Forests.
  • FIA’s Annual Inventory data. Ecology plot data.  Ecology plot sources were primarily Region 6, USDA Forest service, but with several additional sources including BLM.
 

Data Analysis
 
An analysis project was undertaken by the Landscape Ecology, Modeling, Mapping and Analysis (LEMMA) Program at Oregon State University to integrate these existing inventory data to generate detailed maps of existing forest vegetation and land cover across all ownerships in the Pacific Coast States.
 
The mapping analysis was separated into eight ecoregions across Oregon and the imputation of forested attributes from sampled locations to the greater forested ecosystems consisted of a mathematical technique known as (Gradient Nearest Neighbor, or GNN; Ohmann and Gregory 2002).  The GNN technique was used to map detailed vegetation composition and structure for areas of forest and woodland. GNN uses multivariate gradient modeling (Canonical Correspondence Analysis (CCA)) to integrate data from field plots (Table 1) with satellite imagery and mapped environmental data. A suite of fine-scale plot variables is imputed to each pixel in a digital map, and regional maps can be constructed for many of the same vegetation attributes available for FIA plots.\
 
The creation of maps based on georeferenced field samples, remote sensing, and multivariate statistics provides the information basis by which changes in the amount and distribution of ecological features of forested ecosystems that are likely to come about by disturbance or changes in climate. 
 
Table 1.  Data sources and dates of collection that were integrated into the GNN Analysis 


Data Source
Assessment Date
BLM Western Oregon3/7/1997 - 10/27/2001
FIA Eastern Oregon7/24/1998 - 12/10/1999
FIA Eastern Oregon Juniper Inventory5/1/1999 - 11/1/1999
FIA Western Oregon5/22/1995 - 12/29/1997
Jackson/Josephine Counties Inventory9/29/2003 - 2/25/2007
USFS R6 Continuous Vegetation Survey -
Occasions 1 and 2
4/10/1993 - 11/28/2004
 
 
Two GNN models and associated map products were produced for each modeling region, a GNN “species model” and a GNN “species-size” model. 
 
Species model: Response variables used in model development were basal area by tree species. Landsat, disturbance, and ownership variables were not included as explanatory variables. This model provides the most accurate spatial predictions of distributions of individual species and of community types that are defined based on species composition. Stand structure variables are not attached to this grid.

Species-size model: Response variables used in model development are basal area by species and size-class. This model weighs information on both species composition and stand structure, and resulting maps are useful for applications where elements of both are important – especially where it is important to maintain the covariance among these vegetation elements (e.g., if tree lists are to be input into simulation models such as the Forest Vegetation Simulator).
Digital GNN imputation maps are provided as 30-m-resolution ArcGIS grids, where the grid value is a unique plot number that links to the plot database. Selected vegetation variables from the plot database are joined as items in the grid to facilitate viewing and exploratory spatial analysis. Metadata for the vegetation variables are included with the grids and in the plot database. Dates for maps developed from GNN species-size models are determined by the vintage of the satellite imagery used in their development. For Oregon, all GNN species-size maps are based on 2000-2001 imagery.

The distributed products contain ‘masked’ versions of the GNN maps, where areas of nonforest land cover developed from ancillary data sources have been embedded in the GNN grids. The GNN models apply only to forest land (areas currently or with the potential to support at least 10% tree cover).  The masks in Oregon and Washington use the Ecological Systems grids.
The GNN maps are evaluated in several ways, as described in Ohmann and Gregory (2002).
 
Accuracy assessment products for each modeling region include the following:
  • For GNN species model:
    • kappa statistics for individual tree species (from cross-validation).
  • For GNN species-size model:
    • correlation coefficients for core vegetation variables (from cross-validation)
    • confusion matrix for vegetation class (from cross-validation)
    • kappa statistics for vegetation class (from cross-validation)
    • area distribution for vegetation class (compares the distribution of forest area among vegetation classes, as predicted from GNN and as estimated from the systematic sample of grid plots, to provide a regional view of overall accuracy).
 
Results of the Accuracy assessment can be accessed with the following link: http://www.fsl.orst.edu/lemma/main.php?project=imap&id=home