File Geodatabase Feature Class
Tags
hierarchical
classification, tree dominance type, tree canopy cover, R1-VMap, lifeform,
Biology, Ecology, and Biophysical, tree size, Northern Rockies
This dataset was produced for use at project levels of analysis and planning (in some cases additional work would be needed for site specific or project level work).
One of the most fundamental information needs to support ecosystem assessment and land management planning is consistent, continuous, and up to date vegetation data of sufficient accuracy and precision. The Northern Region Existing Vegetation Mapping Program (VMap) database and map products help meet this information need and provides the Northern Region with a geospatial database of existing vegetation produced using consistent analytical methodology according to the Existing Vegetation Classification and Mapping Technical Guide (Brohman and Bryant, 2005) to support the Region 1 Multi-level Classification, Mapping, Inventory, and Analysis System, R1-CMIA (Berglund et. al., 2009).
The Custer Gallatin National Forest (CusGal) VMap database provides four primary map products; lifeform, tree canopy cover class, tree size class, and tree dominance type to support mid and base-level analysis and planning. VMap uses the Region 1 Existing Vegetation Classification System (R1-ExVeg) (Barber, et.al. 2009) in its map unit design. The R1-ExVeg system describes the logic for grouping entities by similarities in their floristic characteristics. This has been an iterative process in Region 1 as different classification schemes have been tested and evaluated for utmost utility by end users. The system was designed to allow consistent applications between Regional inventory and map products within the R1-CMIA framework.
VMap is a remote sensing derived product. As such, it uses a combination of high resolution airborne imagery and a nationally available digital elevation model (DEM). The image data (i.e., pixels) are put through a process of aggregation to derive spatially cohesive units (i.e., polygons). A small portion of these polygons are then sampled in the field to determine their composition and through spatial statistics, un-sampled polygons are given labels based on an analysis of the sampled polygons. Draft map products are then reviewed and appropriate changes are made in the labeling algorithms. Final results are then used to populate the VMap database.
An accuracy assessment was conducted to provide a validation of the data, giving an indication of reliability of the map products, so that managers are fully informed throughout the decision making process. Estimates of overall map accuracy and confidence of individual map classes can be inferred from the accuracy assessment error matrix derived from the comparison of known reference sites to mapped data. These accuracy assessment results are relevant to the entire CusGal as a whole and the resulting map products show exceptional accuracies, ranging from 75-90%, depending on the attribute in question. These accuracies exceed National Map Accuracy standards.
There are no credits for this item.
The USDA Forest Service manages resource information and derived data as a service to USDA Forest Service users of digital geographic data. The USDA Forest Service is in no way condoning or endorsing the application of these data for any given purpose. It is the sole responsibility of the user to determine whether or not the data are suitable for the intended purpose. It is also the obligation of the user to apply those data in an appropriate and conscientious manner. The USDA Forest Service provides no warranty, nor accepts any liability occurring from any incorrect, incomplete, or misleading data, or from any incorrect, incomplete, or misleading use of these data.
Extent
| West | -111.797320 | East | -102.758805 |
| North | 46.360103 | South | 44.290318 |
| Maximum (zoomed in) | 1:5,000 |
| Minimum (zoomed out) | 1:150,000,000 |
publication date
Monday-Friday, 8am-4:30 pm (MST)
email preferred
The USDA Forest Service manages resource information and derived data as a service to USDA Forest Service users of digital geographic data. The USDA Forest Service is in no way condoning or endorsing the application of these data for any given purpose. It is the sole responsibility of the user to determine whether or not the data are suitable for the intended purpose. It is also the obligation of the user to apply those data in an appropriate and conscientious manner. The USDA Forest Service provides no warranty, nor accepts any liability occurring from any incorrect, incomplete, or misleading data, or from any incorrect, incomplete, or misleading use of these data.
This dataset is in the public domain, and the recipient may not assert any proprietary rights thereto nor represent it to anyone as other than a dataset produced by the USDA Forest Service, Northern Region.
The USDA Forest Service manages resource information and derived data as a service to USDA Forest Service users of digital geographic data. The USDA Forest Service is in no way condoning or endorsing the application of these data for any given purpose. It is the sole responsibility of the user to determine whether or not the data are suitable for the intended purpose. It is also the obligation of the user to apply those data in an appropriate and conscientious manner. The USDA Forest Service provides no warranty, nor accepts any liability occurring from any incorrect, incomplete, or misleading data, or from any incorrect, incomplete, or misleading use of these data.
<15 meters
Create derivatives of NAIP imagery. The derivatives are used to provide spectral and texture-based information that is useful for landcover mapping. All of these derivatives used in the mapping process are degraded to 10 meters in order to allow for faster processing and less variability in the dataset. Image Derivatives MEANIHSC1 - NAIP CIR intensity MEANIHSC2 - NAIP CIR hue MEANIHSC3 - NAIP CIR saturation MEANIHSR1 - NAIP RGB intensity MEANIHSR2 - NAIP RGB hue MEANIHSR3 - NIAP RGB saturation MEANCNDVI - NAIP CIR normalized difference vegetation index MEANCPCA1 - NAIP CIR 1st principal component MEANCPCA2 - NAIP CIR 2nd principal component MEANCPCA3 - NAIP CIR 3rd principal component MEANNAIP1 - NAIP band 1: red MEANNAIP2 - NAIP band 2: green MEANNAIP3 - NAIP band 3: blue MEANNAIP4 - NAIP band 4: near infrared
Create image segmentation: Image segmentation is the process of combining pixels within digital images into spatially cohesive regions. These individual regions are called image objects and represent distinct areas within the image. The resulting image objects are inherently more data rich than individual pixels, and form the building blocks upon which image classifications are built (Haralick and Shapiro 1985, Ryerd and Woodcock 1996). Ultimately, the raster-based image objects are converted to vector-based polygons with associated image statistics as attributes. The segmentation process is performed using a proprietary software package, Definiens’ eCognition, and is based on the local variance structure within imagery and User defined shape indices. These image objects effectively depict elements of vegetation and other patterns on the landscape (McDonald et al. 2002).
Field Data Collection: Ground or other reference data is used to build the relationships between the observed phenomena and the spectral and biophysical information derived from remotely sensed and ancillary data. Collectively, ground and other reference data are known as training data because they are used to construct algorithms that relate observations to quantified variables and are used to interpret and label unsampled areas within a study area. Thus, they “train” the algorithm to distinguish between and label the unknown areas within a modeling area. Specifically, training data was collected to distinguish Lifeform, Treecanopy, Treesize, and Dominance.
Lifeform Classification Mapped lifeform is derived from photo/image-interpretation and abundance is determined using species canopy cover, with a minimum of 10% canopy cover needed to assign dominance. Mapped lifeforms include Tree, Shrub, Herbaceous, Sparsely Vegetated, and Water with precedence order being tree, shrub, herbaceous in the lifeform key.
Canopy Cover Tree canopy cover is modeled as a binary response function based on NAIP reflectance, primarily in the green, red, and near-infrared bands of the image. A multistep process model is utilized that first masks out all shadow areas within the image, then calculates a quotient index of green band reflectance compared with the global mean of all image bands to identify “all” canopy (both live and dead portions), and then utilizes two band ratios (the Normalized Difference Vegetation Index and the Simple Ratio) to identify and extract the “live” portion of that spectrum. The difference between the “live” and the “all” allows for an estimation of stand disturbance when summarized to the VMap polygon. The “live” portion by itself is summarized to the VMap polygon to provide the percent tree canopy cover estimate that is the basis for the Tree Canopy Cover Class value in the database. After the canopy value is summarized to the VMap polygon, it is reviewed and edited to fix heavily shadowed areas.
Tree Size Tree Size Class is modeled from field data collected to quantify the basal area weighted average tree diameter at breast height (BAWDBH) per the Region 1 Exiting Vegetation Classification System. BAWDBH is estimated for each variable radius plot to the nearest 1”. This data is then brought in and sampled against a NAIP image derivative stack which is used to calibrate a logistic regression model, with the resulting surface being a prediction of BAWDBH for each pixel in the modeling area. This surface is then summarized to the VMap polygons to obtain a BAWDBH estimate for each polygon that becomes the basis for the Tree Size Class assignment. It was found that the best results were obtained through a majority assignment for the Seedling/Sapling and Large tree size classes, while a mean calculation was employed to identify the Small and Medium size classes.
Tree Dominance Similar to tree size, Tree Dominance type was modeled using a logistic regression function in the RandomForests software based on abundance information collected at the field plot level. A separate surface was built for each species modeled abundance and then all of the layers were aggregated. The species with the highest predicted abundance for any given point was given the label for that point and the aggregate surface was then summarized to the VMap polygons for a dominance type label based on R1 Existing Vegetation Classification System tree dominance type rules.
Accuracy Assessment An independent accuracy assessment of the VMap products was conducted across the entire CusGal. This accuracy assessment provides a validation of the data, giving an indication of reliability of the map products, so that managers are fully informed throughout the decision making process. Too often vegetation and other maps are used without a clear understanding of their reliability. A false sense of security about the accuracy of the map may result in an inappropriate use of the map and important decisions may be made based on data with unknown and/or unreliable accuracy. Estimates of overall map accuracy and confidence of individual map classes can be inferred from an error matrix derived from the comparison of known reference sites to mapped data. Overall the resulting CusGal map products show exceptional accuracies, ranging from 75-90% depending on attribute. These accuracies exceed National Map Accuracy standards which are typically 65% minimum for most attributes. Please refer to Custer Gallatin National Forest VMap Accuracy Assessment; Version 15 (Brown, 2015) for complete details.
This layer contains the elevation information for each sub-path model. Illumination, slope and aspect were derived from the DEM.
In machine learning, a random forest is a classifier that consists of many decision trees and outputs the class that is the mode of the classes output by individual trees. The algorithm for inducing a random forest was developed by Leo Breiman and Adele Cutler, and "Random Forests" is their trademark. The term came from random decision forests that was first proposed by Tin Kam Ho of Bell Labs in 1995. The method combines Breiman's "bagging" idea and Ho's "random subspace method" to construct a collection of decision trees with controlled variations. The Random Forests classifier is part of the open source statistical package R. The USDA Remote Sensing Applications Center created a CartTools Pythogn program script (contact Bonnie Ruefenacht for the latest version of this script 801-975-3828) which accesses Rs Random Forest to create prediction surfaces.
VMap models for processing are based on watershed boundaries. They are restricted in size for better mapping precision and also to keep within the size limit.
These are the four channel NAIP image data.
ground condition
M-F, 8am-4pm (MST)
Internal ESRI number
ESRI
Sequential unique whole numbers that are automatically generated.
Internal ESRI number
ESRI
Coordinates defining the features.
Polygon unique identifier
Area of the polygon in acres
Enumerated_Domain 3100 HERB - Herbaceous Enumerated_Domain 3300 SHRUB - Shrubland Enumerated_Domain 4000 TREE - Tree Enumerated_Domain 5000 WATER - Water Enumerated_Domain 7000 SPVEG - Sparsely Vegetated Enumerated_Domain 7100 URBAN - Urban areas Enumerated_Domain 8600 DECIDTREE - Decidous Tree
Enumerated_Domain 3160 GRASS-DRY - Dry grass or herbaceous types Enumerated_Domain 3190 GRASS-WET - Wet grass or herbaceous types Enumerated_Domain 3320 SHRUB-XERIC - Xeric shrub type Enumerated_Domain 3330 SHRUB-MESIC - Mesic shrub type Enumerated_Domain 5000 WATER - Water Enumerated_Domain 7000 SPVEG - Sparsely vegetated Enumerated_Domain 7100 URBAN - Urban areas Enumerated_Domain 8015 MX-PIPO - Ponderosa pine dominated (>40% relative cover) Enumerated_Domain 8025 MX-PSME - Douglas fir dominated (>40% relative cover) Enumerated_Domain 8045 MX-LAOC - Western Larch dominated (>40% relative cover) Enumerated_Domain 8055 MX-PICO - Lodgepole pine dominated (>40% relative cover) Enumerated_Domain 8065 MX-ABLA - Subalpine fir dominated (>40% relative cover) Enumerated_Domain 8075 MX-PIEN - Englemann spruce dominated (>40% relative cover) Enumerated_Domain 8125 MX-PIAL - Whitebark pine dominated (>40% relative cover) Enumerated_Domain 8155 MX-PIFL2 - Limber pine dominated (>40% relative cover) Enumerated_Domain 8165 MX-POPUL - Cottonwood dominated (>40% relative cover) Enumerated_Domain 8175 MX-POTR5 - Aspen dominated (>40% relative cover) Enumerated_Domain 8185 MX-JUNIP - Juniper dominated (>40% relative cover) Enumerated_Domain 8400 IMIX - Shade-intolerant conifer mix (no single species >40% relative cover) Enumerated_Domain 8500 TMIX - Shade-tolerant conifer mix (no single species >40% relative cover) Enumerated_Domain 8600 HMIX - Hardwood mix (no single species >40% relative cover) Enumerated_Domain 8900 TRANSITIONAL FOREST - Forested areas in GRASS or SHRUB due to distrubance
Enumerated_Domain 3160 GRASS-DRY - Dry grass or herbaceous types Enumerated_Domain 3190 GRASS-WET - Wet grass or herbaceous types Enumerated_Domain 3320 SHRUB-XERIC - Xeric shrub type Enumerated_Domain 3330 SHRUB-MESIC - Mesic shrub type Enumerated_Domain 5000 WATER - Water Enumerated_Domain 7000 SPVEG - Sparsely vegetated Enumerated_Domain 7100 URBAN - Urban areas Enumerated_Domain 8010 PIPO - Ponderosa pine dominated (>60% relative cover) Enumerated_Domain 8020 PSME - Douglas fir dominated (>60% relative cover) Enumerated_Domain 8040 LAOC - Western Larch dominated (>60% relative cover) Enumerated_Domain 8050 PICO - Lodgepole pine dominated (>60% relative cover) Enumerated_Domain 8060 ABLA - Subalpine fir dominated (>60% relative cover) Enumerated_Domain 8070 PIEN - Englemann spruce dominated (>60% relative cover) Enumerated_Domain 8120 PIAL - Whitebark pine dominated (>60% relative cover) Enumerated_Domain 8150 PIFL2 - Limber pine dominated (>60% relative cover) Enumerated_Domain 8160 POPUL - Cottonwood dominated (>60% relative cover) Enumerated_Domain 8170 POTR5 - Aspen dominated (>60% relative cover) Enumerated_Domain 8180 JUNIP - Juniper dominated (>60% relative cover) Enumerated_Domain 8400 IMIX - Shade-intolerant conifer mix (no single species >60% relative cover) Enumerated_Domain 8500 TMIX - Shade-tolerant conifer mix (no single species >60% relative cover) Enumerated_Domain 8600 HMIX - Hardwood mix (no single species >60% relative cover) Enumerated_Domain 8900 TRANSITIONAL FOREST - Forested areas in GRASS or SHRUB due to distrubance
Enumerated_Domain 3160 GRASS-DRY - Dry grass or herbaceous types Enumerated_Domain 3190 GRASS-WET - Wet grass or herbaceous types Enumerated_Domain 3320 SHRUB-XERIC - Xeric shrub type Enumerated_Domain 3330 SHRUB-MESIC - Mesic shrub type Enumerated_Domain 5000 WATER - Water Enumerated_Domain 7000 SPVEG - Sparsely vegetated Enumerated_Domain 7100 URBAN - Urban areas Enumerated_Domain 8010 PIPO - Ponderosa pine dominated (>60% relative cover) Enumerated_Domain 8013 PIPO-IMIX - Ponderosa pine intolerant conifer mix (>40% relative cover) Enumerated_Domain 8020 PSME - Douglas fir dominated (>60% relative cover) Enumerated_Domain 8023 PSME-IMIX - Douglas fir intolerant conifer mix (>40% relative cover) Enumerated_Domain 8040 LAOC - Western Larch dominated (>60% relative cover) Enumerated_Domain 8043 LAOC-IMIX - Western Larch intolerant conifer mix (>40% relative cover) Enumerated_Domain 8050 PICO - Lodgepole pine dominated (>60% relative cover) Enumerated_Domain 8053 PICO-IMIX - Lodgepole pine intolerant conifer mix (>40% relative cover) Enumerated_Domain 8054 PICO-TMIX - Lodgepole pine tolerant conifer mix (>40% relative cover) Enumerated_Domain 8060 ABLA - Subalpine fir dominated (>60% relative cover) Enumerated_Domain 8064 ABLA-TMIX - Subalpine fir tolerant conifer mix (>40% relative cover) Enumerated_Domain 8070 PIEN - Englemann spruce dominated (>60% relative cover) Enumerated_Domain 8074 PIEN-TMIX - Englemann spruce tolerant conifer mix (>40% relative cover) Enumerated_Domain 8120 PIAL - Whitebark pine dominated (>60% relative cover) Enumerated_Domain 8123 PIAL-IMIX - Whitebark pine intolerant conifer mix (>40% relative cover) Enumerated_Domain 8150 PIFL2 - Limber pine dominated (>60% relative cover) Enumerated_Domain 8153 PIFL2-IMIX - Limber pine intolerant conifer mix (>40% relative cover) Enumerated_Domain 8160 POPUL - Cottonwood dominated (>60% relative cover) Enumerated_Domain 8170 POTR5 - Aspen dominated (>60% relative cover) Enumerated_Domain 8180 JUNIP - Juniper dominated (>60% relative cover) Enumerated_Domain 8183 JUNIP-IMIX - Juniper intolerant conifer mix (>40% relative cover) Enumerated_Domain 8400 IMIX - Shade-intolerant conifer mix (no single species >60% relative cover) Enumerated_Domain 8500 TMIX - Shade-tolerant conifer mix (no single species >60% relative cover) Enumerated_Domain 8600 HMIX - Hardwood mix (no single species >60% relative cover) Enumerated_Domain 8900 TRANSITIONAL FOREST - Forested areas in GRASS or SHRUB due to distrubance
Enumerated_Domain 4001 CTR 10-24.9% - CTR 10-24.9% Enumerated_Domain 4002 CTR 25-39.9% - CTR 25-39.9% Enumerated_Domain 4003 CTR 40-59.9% - CTR 40-59.9% Enumerated_Domain 4004 CTR >= 60% - CTR > 60% Enumerated_Domain 3100 HERB - Herbaceous Enumerated_Domain 3300 SHRUB - Shrub Enumerated_Domain 5000 WATER - Water Enumerated_Domain 7000 SPVEG - Sparsely vegetated Enumerated_Domain 7100 URBAN - Urban areas Enumerated_Domain 8600 TREE-DECID - Deciduous Tree Enumerated_Domain 8900 TRANSITIONAL FOREST - Forested areas in GRASS or SHRUB due to distrubance
Enumerated_Domain 4100 DBH 0-4.9" - Basal area weighted average diameter 0-4.9" Enumerated_Domain 4200 DBH 5-9.9" - Basal area weighted average diameter 5-9.9" Enumerated_Domain 4300 DBH 10-14.9" - Basal area weighted average diameter 10-14.9" Enumerated_Domain 4400 DBH 15-19.9" - Basal area weighted average diameter 15-19.9" Enumerated_Domain 4500 DBH >= 20" - Basal area weighted average diameter >= 20" Enumerated_Domain 3100 HERB - Herbaceous Enumerated_Domain 3300 SHRUB - Shrub Enumerated_Domain 5000 WATER - Water Enumerated_Domain 7000 SPVEG - Sparsely vegetated Enumerated_Domain 7100 URBAN - Urban areas Enumerated_Domain 8600 TREE-DECID - Deciduous Tree Enumerated_Domain 8900 TRANSITIONAL FOREST - Forested areas in GRASS or SHRUB due to distrubance
Average elevation of the polygon in meters
Enumerated_Domain 0 - Flat (slope < 10%) Enumerated_Domain 1 - North (338-360 & 0-22 degrees) Enumerated_Domain 2 - Northeast (23-68 degrees) Enumerated_Domain 3 - East (68-112 degrees) Enumerated_Domain 4 - Southeast (113-157 degrees) Enumerated_Domain 5 - South (158-202 degrees) Enumerated_Domain 6 - Southwest (203-247 degrees) Enumerated_Domain 7 - West (248-292 degrees) Enumerated_Domain 8 - Northwest (293-337 degrees)
Average percent slope of the polygon
Length of feature in internal units.
Esri
Positive real numbers that are automatically generated.
Area of feature in internal units squared.
Esri
Positive real numbers that are automatically generated.
Percent Tree Canopy for polygons where Lifeform = 4000
Average Diameter Breast Height for polygons where Lifeform = 4000
M_F, 8am-4pm (MST)
email preferred