home > data and maps > foundational data layers

Foundational Data Layers

These data provided here are for the western Mogollon Plateau area only. Please contact us to receive a copy of the available data products on CD or DVD for the White Mountains region in eastern Arizona (GIS Data List) or the North-central New Mexico area (GIS Data List). Additional descriptions and maps of the available GIS data for these study areas can be found in their respective Data Atlas. These are downloadable from our documents web page under the "Major Reports" section.

These data layers are the property of the Forest Ecosystem Restoration Analysis (ForestERA) project, the Ecological Restoration Institute (ERI), and Northern Arizona University (NAU). They are provided free of charge to the public. However, we request that users carefully review the metadata provided with the layers and consult with the ForestERA project in advance of using these data in publications and/or presentations to ensure that the strengths and limitations of the data are considered.

Publications and presentations that include these data should acknowledge the ForestERA project. Citations for published manuscript that describe the creation and assessment of these layers are listed with their appropriate spatial layers below and on the documents page. Remaining layers should be cited using the following references:

Hampton, H. M., Y. Xu, J. W. Prather, E. N. Aumack, B. G. Dickson, M. M. Howe, and T. D. Sisk. 2003. Spatial tools for guiding forest restoration and fuel reduction efforts. Proceedings of the 2003 ESRI Users Conference, San Diego, CA. Retrieved [month day, year], from http://gis.esri.com/library/userconf/proc03/p0679.pdf.

Sisk, T.D., H.M. Hampton, B.G. Dickson, Y. Xu, M. M. Howe, and J. Palumbo (2004). Forest Ecological Restoration Analysis (ForestERA) Project: Data derived from foundational data layers. Retrieved [month day, year], from http://forestera.nau.edu/data_downloads.htm

Dominant Overstory Vegetation
Canopy Cover
Basal Area and Tree Density
Digital Elevation Model
National Hydrography Database
Terrestrial Ecosystem Survey Data

Dominant Overstory Vegetation

General description

This is a layer representing the dominant overstory vegetation types across the region. Our layer contains a total of nine vegetation classes. Areas without trees (shrubland, grassland, and bare soil) are lumped into a single “open” category. Other vegetation classes are based on the dominant tree species in the area. Mixed classes are identified when there are co-dominant tree species.

Layer creation

This layer was developed from Landsat 7 Enhanced Thematic Mapper (ETM) imagery using a classification tree methodology and training data from over 1100 ground locations. We used See-5 software (Rulequest Research) along with 27 predictor variables derived from the ETM imagery and/or a Digital Elevation Model to develop the model. The layer has a resolution (pixel size) of 90m (0.8 ha or 2 acres).

Layer accuracy

We found a misclassification error rate of 5% and a Kappa value of 0.9 in an internal accuracy assessment. Cross-validation, and external accuracy checks, which are considered better indicators of accuracy, gave misclassification error rates of 24% and 26% respectively, and Kappa values of 0.57 and 0.54 respectively. These results indicate good correspondence between actual vegetation types on the ground and predicted vegetation types in our layer.

Canopy Cover

General description

Percent canopy cover is a measure of the total amount of the landscape covered by canopy foliage (trees). It does not include small gaps within tree canopies (this is defined as canopy closure) as these cannot be detected in satellite photography. Values range from zero to nearly 100% within the assessment area.

Layer creation

This layer was developed from a mosaic of Digital Orthophoto Quads (DOQs) taken primarily in 1997. The layer was developed using a type of advanced exploratory data analysis, in which each 1m pixel in the image is classified as canopy foliage, shadow, or ground vegetation. The pixels of canopy foliage are then aggregated across larger areas to determine percent cover. As provided, this layer has a resolution (pixel size) of 90m (0.8 ha or 2 acres) to match the resolution of other data layers.

Layer accuracy

For accuracy assessment we obtained canopy cover measurements taken on 200 ground plots from 18 locations spread across the assessment area. We used linear regression to determine the relationships between these ground measurements and the values from our predictive layer at the same locations. This analysis indicated that a highly significant relationship exists between the two measures (r² = 0.545, P < 0.001), and that the canopy cover estimates in the predictive layer are nearly unbiased in relation to the ground estimates (slope = 1.02). Based on the actual differences in the data between the two estimates we have determined that values for 50% of the pixels in the DOQ derived canopy cover layer will be within 9% of the actual values for canopy cover in those areas, and values for over 80% of the pixels will be within 16% of the actual values.

Citation: Xu, Y., J. W. Prather, H. M. Hampton, E. N. Aumack, B. G. Dickson, and T. D. Sisk. 2006. Advanced exploratory data analysis for mapping regional canopy cover. Photogrammetric Engineering & Remote Sensing 72: 31-38.

Basal Area and Tree Density

General description

Basal area is a commonly used measure developed by foresters as an index of the amount of woody material present. It is the total cross-sectional area of trees in a stand. Tree density is a count of the number of tree stems (> 1” diameter) in a given area. Our basal area and tree density layers were developed from satellite imagery taken in year 2000. The predicted values for basal area range from 0 to over 300 ft2/acre (0 to over 70 m2/ha) and the predicted values for tree density range from 0 to nearly 700 trees/acre (0 to nearly 1700 trees/ha) within the assessment area.

Layer creation

These layers were developed from Landsat 7 Enhanced Thematic Mapper (ETM) imagery using a regression tree methodology and training data from over 560 ground locations. We used Cubitst software (Rulequest Research) along with 27 predictor variables derived from the ETM imagery and/or a Digital Elevation Model to develop the layer. These layers have a resolution (pixel size) of 90m (0.8 ha or 2 acres).

Layer accuracy

For accuracy assessment we obtained canopy cover measurements taken on 567 ground plots from 63 locations spread across the assessment area. We used linear regression to determine the relationships between these ground measurements and the values from our predictive layer at the same locations. These analyses indicate that highly significant relationships exist between actual basal area on the ground and basal area values in the predictive layer (r² = 0.508, P < 0.0001), and between actual tree density on the ground and tree density values in the predictive layer (r² = 0.584, P < 0.0001). The slope of the basal area regression line is 1.09 and the slope of the tree density regression line is 0.99 indicating that the predicted values for these attributes in our layer are nearly unbiased with relation to the actual values on the ground.

We estimated uncertainty in the predictive layers by assessing the differences between ground measurements for each attribute and the values from each predictive layer. For basal area, this analysis indicated that over 50% of the predicted values lie within 17.5 ft2/acre (4 m2/ha) of the actual value and over 80% of the predicted values lie within 30.5 ft2/acre (7m2/ha) of the actual value. For tree density, this analysis indicated that over 50% of the predicted values lie within 35 trees/acre (80 trees/ha) of the actual value and over 80% of the predicted values lie within 60 trees/acre (150 trees/ha) of the actual value.

Basal Area
Tree Density

Digital Elevation Model

General description

A Digital Elevation Model is a computer map of elevation across a land surface. In addition to elevation, a number of other terrain attributes can be derived using the DEM. These include slope, aspect, and various measures of roughness. Elevation ranges from approximately 4300 ft (1300 m) to over 12,000 ft (3650 m) across the assessment area. Slope ranges from zero to over 70 degrees in the assessment area.

Layer creation

Digital Elevation Models are created by the United States Geological Survey using interpolation techniques and one of two elevation data sources; either digital line graph (DLG) hypsographic and hydrographic data, or various types of photogrammatic imagery and specific ground locations where elevation has been precisely measured. We obtained a 30m DEM for this region from the USGS. As provided, this layer has a resolution (pixel size) of 90m (0.8 ha or 2 acres) to match the resolution of other data layers.

Layer accuracy

The Digital Elevation Model and all of its derivatives meet USGS standards for accuracy. DEM data accuracy is derived by comparing linear interpolation elevations in the DEM with corresponding map location elevations and computing the statistical standard deviation or root-mean-square error (RMSE). For the DEMs used in this project, 90 percent of the predicted elevation values are expected to lie within 23 feet (7m) of the actual value and the remaining 10 percent are expected to lie within 50 feet (15m) of the actual value.

National Hydrography Database

General description

The National Hydrography Database (NHD) is a newly combined dataset that provides comprehensive coverage of hydrographic data for the United States. Although based on a relatively low resolution (1:100,000-scale) data, the NHD is designed to incorporate and encourage the development of higher resolution data.

Layer creation

The NHD was put together by the U.S. Environmental Protection Agency (USEPA) and the U.S. Geological Survey (USGS). It combines elements of USGS digital line graph (DLG) hydrography files and the USEPA Reach File (RF3). The DLG files contribute a national coverage of millions of features, including water bodies such as lakes and ponds, linear water features such as streams and rivers, and also point features such as springs and wells. From RF3, the NHD acquires hydrographic sequencing, upstream and downstream navigation for modeling applications, and reach codes. The reach codes provide a way to integrate data from organizations at all levels by linking the data to this nationally consistent hydrographic network. The feature names are from the Geographic Names Information System (GNIS, see Appendix). The map shows perennial streams listed in the NHD and springs from the GNIS database.

Layer accuracy

We have not undertaken an accuracy assessment on this layer. It is assumed to meet accuracy standards of the agencies that provided it.

Terrestrial Ecosystem Survey Data

General description

Terrestrial Ecosystem Survey (TES) data describes predicted and surveyed soils, potential climax vegetation, and predicted limitations of soil and vegetation characteristics for selected land uses. We developed this layer from data obtained from the individual National Forests within the assessment area.

Layer creation

Terrestrial ecosystems are defined by the interaction of soil, climate and vegetation. In its original form the TES data is in a polygon (vector) format. The USDA Forest Service delineated mapping unit boundaries by stereoscopic examination of aerial photographs on the basis of differences in topography, geology and vegetation. From these polygon data, and in consultation with FS soil scientists, the ForestERA team has generated a preliminary grid (raster) map of mollisol soils in the assessment area. Mollisols are soils that are normally generated in grassland and savannah areas. Thus, they may be used to identify where those vegetation types have occurred in the past.

Layer accuracy

We have not undertaken an accuracy assessment on this layer. It is assumed to meet accuracy standards of the agency that provided it. Each National Forest has its own survey methods and types of data it associates with each terrestrial ecosystem, so building a consistent data layer across the study site is not always possible.

Last updated April 10, 2007