First, each Landsat image was delineated into regions in an unsupervised classification. Second, each region was assigned a land cover type label in a supervised classification. Regions classified as conifer cover types were also assigned labels for tree size class. Regions classified as shrub or tree land cover types also had canopy cover assigned based on MNDVI values (derived variable from TM bands 3, 4, & 5). A total of 41 land cover type classes, 4 tree size classes, and 3 canopy cover classes were labeled across the six scene area. Accuracies were estimated for the land cover type and tree size classifications of each Landsat image. Third, the regions from all six Landsat images were edgematched together into the CICPSILC2Z database. Below is a description of how the original Landsat images were delineated into ground feature regions, how the regions were labeled to land cover type, tree size, and tree canopy, and how the six Landsat images were edgematched into one database.
Image Landsat Image-Date Covertype-Acc Treesize-Acc SCENEPOLY-ID ------------------------------------------------------------------------------- P40R29 5 7/31/92 57.1% 66.3% 48000001 .. P40R30 5 8/03/93 74.2% 49.1% 24000001 .. P41R28 5 7/20/91 57.6% 42.9% 47000001 .. P41R29 5 7/31/95 51.0% 44.1% 58000001 .. P41R30 5 7/31/95 57.5% 65.2% 23000001 .. P42R29 5 7/27/91 56.3% 65.8% 19000001 ..
Each image was run through an unsupervised classification to classify the approximately 56 million pixels into 120-130 spectral classes. Using an iterative classification process with spectral signature files, each pixel was assigned to a spectral class. The pixels were then merged into regions (raster polygons) through a multi-stage merge algorithm using a minimum map unit of 2 hectares. The final unsupervised classification resulted in most regions ranging in size from 0.1 hectare to 200 hectares. A small number of regions (about 1% of all regions in each scene), ranged from 200 hectares to 10,0000 hectares. Each region was assigned a set of attributes for TM channels 1-7 from the imagery, along with elevation, aspect, and slope from the 30 meter digital elevation model (DEM); these attribute values were calculated based on the mean values of the 30 meter pixels within each region.
Vegetation overstory (land cover type) was labeled using a supervised classification. The land cover type supervised classification was performed using a combination of manual and training data classifications. The following classes were manually labeled: urban, irrigated and non-irrigated agriculture, recently burned areas, water, mines, snow, and clouds. The other land cover types were labeled using training data and spatial classifiers. The training data were screened for data attribute accuracy, positional accuracy, and agreement with the imagery. The land cover type labels were filled into the COVERTYPE field of each scenes region grid.
COVERTYPE DESCRIPTION ------------------------------------------------ 1100 Urban or Developed Lands 2010 Agricultural - Dry 2020 Agricultural - Irrigated 3100 Upland Grassland 3110 Altered Grassland 3180 Mesic Montane / Subalpine Grassland 3210 Mesic Shrub 3301 Mountain Mahogany 3304 Bitterbrush 3318 Shadscale 3350 Big Sagebrush Steppe 4101 Aspen 4200 Conifer Forest (undetermined type) 4203 Lodgepole Pine 4204 Whitebark Pine 4205 Limber Pine 4206 Ponderosa Pine 4207 Grand Fir 4208 Subalpine Fir 4212 Douglas-fir 4223 Douglas-fir/Lodgepole Pine 4225 Douglas-fir/Grand Fir 4230 Douglas-fir/Ponderosa Pine 4260 Mixed Whitebark Pine Forest 4270 Mixed Subalpine Forest 4280 Mixed Mesic Forest 4290 Mixed Xeric Forest 4300 Mixed Broadleaf / Conifer Forest 4402 Burns - Moderate 4403 Burns - High 5000 Water 6110 Conifer Riparian 6120 Broadleaf Riparian 6130 Broadleaf / Conifer Riparian 6210 Grass Riparian 6310 Shrub Riparian 7300 Rock 7301 Basalt 7800 Barren 9100 Snow 9800 Cloud 9900 Cloud Shadow
Conifer tree size labels were assigned using a training data classification. The training data were screened for data attribute accuracy, positional accuracy, and agreement with the imagery. Due to a high level of disagreement between the training data tree canopy call and the imagery, the tree canopy labels were based on MNDVI splits within the regions classified as conifer cover types. Shrublands also had canopy cover assigned based on MNDVI splits within the regions classified as shrub cover types. The tree size labels were filled into the TREESIZE field, the tree canopy cover and shrub canopy cover labels were filled into the CANOPYCOVER field of each scenes region grid.
TREESIZE DESCRIPTION ------------------------------------- 1 Sapling 1.0-4.9" DBH 2 Pole 5.0-12.0" DBH 3 Medium 12.1-20.0" DBH 4 Large 20.1-99.9" DBH
TREECANOPY & SHRUB CANOPY DESCRIPTION ------------------------------------- 2 10-40% Cover 3 41-70% Cover 4 71-100% Cover
An accuracy assessment for land cover types classified with training data was performed by randomly selecting and witholding 20% of the training data set. A similar accuracy assessment was done for the conifer tree size classification. No estimate of accuracy was performed on the manually classified land cover types.
Edgematching was done by selecting regions from only one of the scenes in overlap areas. First, the imagery was visually analyzed to select cloud free or snow free areas in overlap zones. Next, the images were ranked based on overall image quality. Finally, after the land cover type classification, regions were selected that created a consistant covertype pattern across the transition zone from one scene to another. The end result was a region grid with 2,458,634 regions.
Image Landsat Date Priority Number of Regions Used ------------------------------------------------------------- P41R29 5 07/31/95 1 443,207 P41R30 5 07/31/95 2 420,982 P41R28 5 07/20/91 3 545,126 P40R29 5 07/31/92 4 386,486 P42R29 5 07/27/91 5 380,607 P40R30 5 08/03/93 6 282,226 ------------------------------------------------------------- Total Regions in Database 2,458,634
Please see the report "Landsat Vegetation Mapping of the Southwest and Central Idaho Ecogroups" submitted to the Boise National Forest 14 November 1997 for further details.
The manually classified land cover types did not have an accuracy assessment performed.
Users accuracies of land cover types were:COVERTYPE P40R29 P40R30 P41R28 P41R29 P41R30 P42R29 -------------------------------------------------------------- 3100 34.5 11.1 56.2 59.2 14.3 78.6 3110 91.7 0.0 75.0 NA 100.0 50.0 3180 71.4 83.3 14.3 54.3 72.7 100.0 3210 NA 0.0 80.0 91.3 33.3 55.6 3301 41.4 NA 60.0 16.7 NA NA 3304 NA NA NA NA 0.0 NA 3318 66.7 NA NA NA NA 3350 77.8 94.1 80.0 67.2 90.8 66.7 4101 50.0 37.5 NA 20.0 0.0 33.3 4203 43.9 16.7 51.7 59.8 0.0 31.2 4204 0.0 NA 0.0 52.9 NA NA 4205 100.0 NA NA NA NA NA 4206 90.9 NA 53.8 46.2 72.1 51.7 4207 NA NA 35.3 0.0 NA 0.0 4208 NA 16.7 NA NA 33.3 NA 4212 70.5 57.1 68.5 35.3 52.2 33.3 4223 24.1 28.6 50.0 48.6 15.4 44.4 4225 NA NA NA 16.7 NA 42.9 4230 20.0 NA 14.3 40.2 32.1 44.1 4260 37.5 0.0 40.0 64.1 50.0 50.0 4270 45.5 29.2 64.7 27.6 19.3 69.2 4280 NA NA 50.0 21.9 NA 40.0 4290 NA NA NA 64.0 NA NA 4300 20.0 50.0 NA 23.5 11.1 50.0 6110 66.7 100.0 20.0 66.7 33.3 NA 6120 66.7 40.0 0.0 21.4 0.0 66.7 6210 37.5 50.0 66.7 0.0 0.0 0.0 6310 40.9 53.3 50.0 85.2 33.3 66.7 7300 63.2 70.0 92.5 69.3 70.6 92.3 7800 22.2 100.0 0.0 45.2 33.3 50.0 -------------------------------------------------------------- Overall 57.1 74.2 57.6 51.0 57.5 56.3
(NA - covertype not mapped in scene)
Users accuracies of conifer tree size classes were:
TREESIZE P40R29 P40R30 P41R28 P41R29 P41R30 P42R29 ------------------------------------------------------------- 1 Sapling 16.7 40.0 14.3 31.2 55.6 85.7 2 Pole 73.6 53.6 56.5 51.2 40.0 59.1 3 Medium 62.8 65.4 50.0 43.0 75.0 63.3 4 Large 65.5 41.9 33.3 37.8 100.0 70.6 ------------------------------------------------------------- Overall 66.3 49.1 42.9 44.1 65.2 65.8
BUILD_VAT: built an ArcInfo value attribute table (VAT) to hold all items (fields) for which descriptive data could be developed for each region. Filled in VALUE, COUNT, LINK, HECTARES, PERIMETER, TM1, TM2, TM3, TM4, TM5, TM6, TM7, MNDVI, ELE, SLP, and ASP fields.
Scene 2 - Low (10-40%) 3 - Mod (41-70%) 4 - High (71-100%) --------------------------------------------------------------- P40R29 < 47 47-85 > 85 P40R30 < 82 82-105 > 105 P41R28 < 110 110-245 > 245 P41R29 < 65 65-185 > 185 P41R30 < 95 95-160 > 160 P42R29 < 73 73-200 > 200The CANOPYCOVER field was filled in based on the MNDVI
Scene 2 - Low (10-40%) 3 - Mod (41-70%) 4 - High (71-100%) --------------------------------------------------------------- P40R29 < 150 150-235 > 235 P40R30 < 157 157-228 > 228 P41R28 < 230 230-390 > 390 P41R29 < 170 170-390 > 390 P41R30 < 220 220-325 > 325 P42R29 < 280 280-440 > 440
LIFEFORM DESCRIPTION --------------------- A Agriculture B Barren G Grass / Forb N Not applicable R Rock S Shrubland T Forest U Urban W Water
TREESIZE DESCRIPTION ------------------------------------- 0 NOT CONIFER 1 Sapling 1.0-4.9" DBH 2 Pole 5.0-12.0" DBH 3 Medium 12.0-20.0" DBH 4 Large 20.1-99.9" DBH
Shrub: 2 - Low (10-40%) 3 - Mod (41-70%) 4 - High (71-100%)
---------------------------------------------------------------
P40R29 < 47 47-85 > 85
P40R30 < 82 82-105 > 105
P41R28 < 110 110-245 > 245
P41R29 < 65 65-185 > 185
P41R30 < 95 95-160 > 160
P42R29 < 73 73-200 > 200
Tree: 2 - Low (10-40%) 3 - Mod (41-70%) 4 - High (71-100%)
---------------------------------------------------------------
P40R29 < 150 150-235 > 235
P40R30 < 157 157-228 > 228
P41R28 < 230 230-390 > 390
P41R29 < 170 170-390 > 390
P41R30 < 220 220-325 > 325
P42R29 < 280 280-440 > 440
Code Definition ----------------- 0 Not Shrub or Conifer 2 Low 10-40% 3 Mod 41-70% 4 High 71-100%
ASP Description -------------------- 0 Flat 1 North 2 Northeast 3 East 4 Southeast 5 South 6 Southwest 7 West 8 Northwest
TREESIZE DESCRIPTION ------------------------------------- 1 Sapling 1.0-4.9" DBH 2 Pole 5.0-12.0" DBH 3 Medium 12.0-20.0" DBH 4 Large 20.1-99.9" DBH
Image Landsat Date Priority Number of Regions Used ------------------------------------------------------------- P41R29 5 07/31/95 1 443,207 P41R30 5 07/31/95 2 420,982 P41R28 5 07/20/91 3 545,126 P40R29 5 07/31/92 4 386,486 P42R29 5 07/27/91 5 380,607 P40R30 5 08/03/93 6 282,226 ------------------------------------------------------------- Total Regions in Database 2,458,634Attribute_Definition_Source none Attribute_Domain_Values Unrepresentable_Domain 0 - Not Used 1 - Use in overlap area based on dominance order