|Description and Attributes of the Digital Land Cover
The cover was created in ARC/INFO 7.0 format.
For Version 5 of the land cover:
Number of Arcs: 71,395
Number of Polygons: 14,740
Number of Nodes: 57,588
Number of Tics: 25,872
Number of Arc Segments: 696,938
Number of Polygon Labels: 14,739
Fuzzy tolerance: 150 ft
Dangle length: 150 ft
Projection: State plane
Spheroid: Clarke 1866
Table 1: Land cover attributes (.pat file)
Item Name Width Type Redefined as
AREA 4 F
PERIMETER 4 F
COVER# 4 B
COVER-ID 4 B
REGIONP 2 I PRIMARY
ZONEP 2 I PRIMARY
PRIM 3 I PRIMARY
PRIMOCC 1 I
REGIONS 2 I SECONDARY
ZONES 2 I SECONDARY
SECO 3 I SECONDARY
SECOOCC 1 I
REGIONT 2 I TERTIARY
ZONET 2 I TERTIARY
TERT 3 I TERTIARY
TERTOCC 1 I
SOURCEA 2 I
SOURCEB 2 I
CONFIDENCE 1 I
ANALYST 1 I
DATE 8 D
REMARKS 200 C
A standard Arc cover attribute for each polygon
Units: square feet, since projection units are in feet.
1 ha = 107,639.1 sq. ft
1 ac = 43,560 sq. ft
1 sq. ft = 2.295684*10-5 ac
1 sq. ft = 0.09290304 sq. m
A standard Arc cover attribute for each polygon.
A standard Arc cover attribute; the internal number for each polygon. This attribute changes whenever topology changes.
A standard Arc cover attribute; the external identification number for each polygon. This is a unique number for each polygon which does not change when topology changes.
REGIONP, REGIONS, REGIONT
The ecoregion to which the polygon belongs, identified by a 2-digit code. REGIONP, REGIONS, and REGIONT are the ecoregions to which the primary, secondary, and tertiary covers, respectively, belong, but since each polygon can only belong to one ecoregion, REGIONP = REGIONS = REGIONT. The ecoregion and vegetation zones attributes were repeated three times for each polygon so that primary, secondary, and tertiary land cover could be redefined to include ecoregion and vegetation zone.
6 Outer Olympic Peninsula
7 Puget Trough
9 Northeast Corner
10 Okanogan Highlands
11 Columbia Basin
12 East Central Cascades
13 Southeast Cascades
14 Blue Mountains
15 Northeast Cascades
16 Northwest Cascades
17 Southwest Cascades
18 Willamette Valley
19 Inner Olympic Peninsula
98 Beyond the borders of Washington
ZONEP, ZONES, ZONET
The vegetation zone to which the polygon belongs, identified by a 2-digit code. ZONEP, ZONES, and ZONET are the zones to which the primary, secondary, and tertiary covers, respectively, belong, but since each polygon can only belong to one zone, then ZONEP = ZONES = ZONET. The ecoregion and vegetation zone attributes were repeated for each cover in the polygon so that primary, secondary, and tertiary land cover could be redefined to include ecoregion and vegetation zone.
Steppe zone codes:
11 Festuca idahoensis/Rosa nutkana (Blue Mountains Steppe)
12 Festuca idahoensis/Symphoricarpos albus (Palouse)
13 Artemisia tripartita/Festuca idahoensis (Three-tip Sage)
14 Festuca idahoensis/Hieracium cynoglossoides (Klickitat Meadow Steppe)
15 Purshia tridentata/Festuca idahoensis (Bitterbrush)
16 Artemisia tridentata/Agropyron spicatum (Central Arid Steppe)
17 Agropyron spicatum/Festuca idahoensis (Wheatgrass/Fescue)
18 Agropyron spicatum/Poa sandbergii (Canyon Grasslands)
19 Artemisia tridentata/Festuca idahoensis (Big Sage/Fescue)
Forest and high-elevation zone codes:
21 Ponderosa Pine
23 Interior Douglas-fir
24 Grand Fir
25 Cowlitz River
26 Willamette Valley
27 Low-elevation Lava Flows
28 High-elevation Lava Flows
29 Woodland/Prairie Mosaic
30 Sitka Spruce
31 Puget Sound Douglas-fir
32 Olympic Douglas-fir
35 Western Hemlock
36 Interior Western Redcedar/Western Hemlock (east Cascades)
37 Interior Western Redcedar (Northeast)
38 Interior Western Hemlock (Northeast)
40 Silver Fir
71 Blue Mountains, High Open Forests
72 Blue Mountains, High Basalt Ridges
81 Subalpine Fir
82 Mountain Hemlock
83 Whitebark Pine (incompletely delineated, and treated with Alpine/Parkland)
90 Permanent Ice/Snow
PRIM, SECO, TERT
PRIM is the land cover that occupies the greatest surface area of the polygon. SECO, if necessary, is the land cover that occupies the second greatest surface area of the polygon. TERT, if necessary, is the land cover that occupies the third greatest surface area of the polygon. PRIM, SECO, and TERT are all identified by a 3-digit code.
The first digit of the cover code indicates one of 9 Level 1 categories. A Level 1 category was always assigned:
1-- Bare ground
4-- Open Water
5-- Wetlands, riparian
7-- Hardwood forest
8-- Mixed hardwood/conifer forest
9-- Conifer forest
The second and third digits each indicate Level 2 categories within the Level 1 categories. Digit 2 is Attribute A and digit 3 is Attribute B. Attributes A and B are independent of one another and, if the attribute could not be confidently assigned or the cover had multiple (mixed) attributes, the attribute was assigned a 0.
1-- Bare ground (<10% vegetation cover; excludes water, fallow fields, pavement, etc.)
General type (Attribute A)
1- Rocky ground
Form (Attribute B)
Sample combinations used:
110 Bare rocky/stony ground
111 Bare cliffs
130 Sand (to date, always beach sand)
Level of development (Attribute A)
0- Unknown/mixed levels
1- High density (>60% of surface covered by pavement, structures, etc.)
2- Medium density (30-60% surface development)
3- Low density (10-30% surface development)
Type of development (Attribute B)
-0 All types
-1 Primarily residential
-2 Primarily business/industrial
-3 Mining operations
-5 Predominant cover is mowed grass (golf courses, sports fields, airports)
-6 Predominant cover is planted trees among mowed understory (wooded city parks)
231 Low density residential (e.g., suburban housing w/large lots)
212 High density business/industrial (e.g., core Seattle downtown area)
Irrigation status (Attribute A)
0- Unknown or a combination of irrigation types
Crop type (Attribute B)
-0 Unknown or a mix of crop types
-1 Maintained pasture (i.e., seeded, mowed, hayed, etc.)
-2 Herbaceous row crops
-4 Soil Conservation Reserve fields (usually Crested Wheatgrass)
312 Irrigated row crops
320 Non-irrigated crops of unknown type or a mix of types
4-- Open water
Fresh or salt water (Attribute A)
0- Variety of salinity levels
1- Fresh water
2- Salt water
3- Brackish water
Water body type (Attribute B)
-0 Mix of types (usually combination of rivers and lakes)
-1 Lakes (including shoreline)
-2 Municipal ponds
-3 Channeled scabland and steppe kettle ponds
-4 Fish hatcheries
-5 Large rivers
-7 Irrigation canals
-8 Irrigation ponds
411 Freshwater lakes, including shorelines
415 Rivers large enough to have visible open water at our resolution
430 Brackish water (always the mouths of large rivers)
Water flow type (Attribute A)
0- Mixed or unknown type
1- Estuarine marsh
2- Freshwater marsh
4- Vernal pools or small ponds (interchangeable with 413)
5- Beach/ocean shoreline
-0 Mixed or unknown dominant vegetation form
-1 Sparse emergent, submerged, or floating
-2 Herb, shrub dominated
-3 Hardwood dominated
-4 Conifer dominated
Sample code combinations:
512 Estuarine marshes, usually grassy
523 Freshwater marsh, dominated by hardwood trees
530 Riparian vegetation, variety of communities
Water/wetland notes: The distinction between open water and wetlands is not sharp. The wetland category "riparian," for example, refers to the vegetation along a river and the open water category "river" refers to the open surface water of a river. In practice, a narrow river would be labeled "riparian" with the polygon including the shoreline vegetation and the open river, while a wide (> 50 m across) river in a polygon would be labeled with the open water category "river" but would include at least part of riparian vegetation on the shore. Thus the distinction between open water and wetland is often a quantitative difference in the proportion of open water to emergent vegetation, rather than a qualitative categorical difference.
Seral stage or disturbance status (Attribute A)
0- Unknown or a mix of seral types
1- Disturbed by logging, burning, or heavy grazing
2- Undisturbed or climax type
Vegetation physiognomy (Attribute B)
-0 Combination of types or type unknown
-1 Sparse vegetation (10-30% cover)
-2 Grass or forb dominated (<10% shrub or tree cover)
-3 Shrub savanna or mix herbs and shrub patches (shrub cover 10-25%; tree cover <10%)
-4 Shrub dominated (shrub cover >25%; tree cover <10%)
-5 Tree savanna (tree cover 10-25%)
613 In a forest zone, usually a logged or burned area regrown to a mix of herbs and shrubs; in a steppe zone, usually indicates a shrub savanna dominated by exotics due to heavy grazing
605 Tree savanna, disturbance status unknown
620 A variety of apparently natural communities (i.e., no sign of recent logging, burns, heavy grazing); often used to designate a mix of alpine/subalpine meadows, parklands, and shrub fields.
7-- Hardwood forest (overstory generally > 75% hardwood)
8-- Mixed hardwood/conifer forest
9-- Conifer forest (overstory generally >75% conifers)
Seral stage (Attribute A; same for all forest classes)
0- Mix of seral (i.e., successional) stages or seral stage unknown
6- Early seral
7- Mid seral
8- Late seral
Canopy closure (Attribute B; same for all forest classes)
-0 Mixed closure, patchy closure, or closure unknown
-1 Open canopy (25-60% overstory cover)
-2 Closed canopy (generally >60% overstory cover)
Sample code combinations
962 Early seral, closed canopy, predominantly conifer forest
860 Early seral, mixed forest, with patchy closure (e.g., tree stands mixed with shrubby regrowth) or a mix of closed and open stands.
901 Conifer forest, open canopy, unknown seral stage.
Forest note 1: Conifer forest is the most common forest type in Washington State. Hardwood and mixed forests are often (but not always) an early seral stage in a succession that eventually leads to conifer domination. Thus, a mid- or late-seral hardwood forest is, in many cases, something of a contradiction. If these terms are applied to hardwood forest, they are generally meant to indicate that the hardwoods are probably large and mature.
Forest note 2: Seral stage is most easily determined when canopies close early and evenly and when the terrain is not too rugged. Under those conditions, the evenness of the top of the tree canopy is roughly correlated to seral stage. In dry forests, forests on uneven terrain, or forests on deeply shadowed slopes, seral stage was either not labeled (e.g., 901, 902 classes) or is unreliable unless the satellite data were supplemented by other sources of information. Most east-side forest classes were not labeled with seral stage because they are relatively dry with uneven canopy closure. High elevation forest types, which are often on rugged terrain, tended to be labeled as late seral unless there was some obvious sign of recent logging or fire.
Forest note 3: Seral stage and tree size. We initially used a classification that substituted tree size for seral stage, but that type of system is difficult to apply across large elevation gradients, and forest size classes are exceedingly difficult to determine with satellite data. In forests at low elevations on moderate terrain, "early seral" corresponds roughly to sapling and pole-sized trees; "mid seral" to large pole and small saw timber; and "late seral" to large saw timber and "old growth." However, at high elevations, late-seral forest may consist of short, stunted, pole-sized subalpine trees. Additionally, "tree" generally applies to plants taller than 5 meters. Thus a Douglas-fir plantation of 2-meter seedlings, if classified correctly, would be labeled 614 (a non-forested, disturbed shrub land) rather than a forest. However, some latitude was applied to subalpine forests, where trees hundreds of years old are often less than 5 meters; if classified correctly, these stands would be labeled forests.
PRIMARY, SECONDARY, TERTIARY
These are redefined items.
PRIMARY = REGIONP + ZONEP + PRIM
SECONDARY = REGIONS + ZONES + SECO
TERTIARY = REGIONT + ZONET + TERT
Each item is a 7-digit integer.
SECONDARY = 1221901:
ecoregion (REGIONS) = 12,
zone (ZONES) = 21,
secondary cover (SECO) = 901.
In words: A polygon in the East-central Cascade region, in the Ponderosa Pine zone, with a secondary cover of open-canopy conifer forest.
PRIMOCC, SECOOCC, TERTOCC
The occupation class of the primary, secondary, and tertiary covers, respectively; a 1-digit integer which indicates a range of proportional cover:
1 1%-5% (Midpoint - 2.5%)
2 5%-25% (Midpoint - 15%)
3 25%-50% (Midpoint - 37.5%)
4 50%-75% (Midpoint - 62.5%)
5 75%-95% (Midpoint - 85%)
6 95%-100% (Midpoint - 97.5%)
The primary and, if applicable, the secondary source of cover interpretation; a 2-digit integer.
0 Based primarily on the appearance of the satellite image.
1 DeLorme (1995) Atlas information.
2 USGS 1:100,000 scale map.
3 USGS 1:12,000 scale maps.
4 Air photo interpretation.
5 Ground visit.
6 Daubenmire, 1970.
7 Daubenmire and Daubenmire 1968.
8 Hanford Site map, provided by Hanford researchers.
9 Interpolated through clouds, based on neighboring vegetation and topography.
10 Nisqually Delta vegetation map, provided by the Nisqually National Wildlife Refuge.
11 Washington Department of Natural Resources maps of vegetation on State Parks, determined by air photos and ground surveys.
13 James Powell's ground data from southwestern Washington.
14 Makah Indian Reservation supplied map of vegetation on the tip of Cape Flattery.
15 Klickitat County Oak map, Audubon Society
21 Okanogan National Forest notes
This attribute was added late in the project and, consequently, is usually blank (i.e., 0).
0 No confidence level applied
1 Very confident (site visit or highly reliable source).
2 Moderately confident.
3 Not too confident, but not totally uncertain.
4 Low confidence.
5 Best guess.
A code for the person assigning the label; a 1-digit integer.
0 Unidentified, but usually Kelly Cassidy
1 Kelly Cassidy
2 Mike Smith
3 Jon Titus
4 Jane Cassady
The date the polygon was labeled (not the date of the information on the label). This attribute was added mid-way through the project, so polygons without a date were labeled prior to mid-October, 1994.
A written description of the polygon and comments. This attribute was added late in the project and is often blank.
Comments and Evaluation of Land Cover Mapping by Region
Region 6 (Outer Olympic Peninsula)
Outside of the Olympic Park, spatial resolution of polygons is moderately good. Most cuts, even those below the minimum mapping unit, were delineated. Olympic Park polygons are split primarily along elevation contours to allow zone labeling. Shadowing in steep terrain caused some problems in identifying seral stage. Puffy clouds over Cape Flattery interfered with labeling in that area, and we relied on maps supplied by the Makah Indian Reservation. All offshore islands large enough to see with satellite imagery were delineated.
Region 7 (Puget Trough)
In general, there are few unusually large polygons in this region, but the extreme fragmentation of this entire region necessitates a smaller polygon size. Ideally, this region should be delineated so that polygons are as close to the minimum mapping unit as possible. Because of the extreme fragmentation, a large proportion of polygons have a tertiary cover; in many, the tertiary cover is greater than 5%. Polygon borders are more subjective in this region than in any other; roads were often arbitrarily used for lack of any better criteria. The heterogeneity of each polygon means that it is very difficult to adequately represent polygon composition with three class codes. Because the extent of any single cover type within a polygon is usually small, most cover types could be expected to be strongly influenced both by other cover types within the polygon and by neighboring polygons. These edge effects are especially important for vertebrate modeling and care should be taken when associating vertebrates with habitats in this region. A useful distinction in this area would be better differentiation of mixed and hardwood forests into riparian, oak/conifer, and successional alder. Less critical hardwood identification would include birch, madrone, and a few other types that cover small areas. Also desirable would be better identification of prairie/conifer mosaics (which would probably require much ground work) and better separation of young conifer forest from shrubs and broadleaf trees.
Region 9 (Northeast Corner)
Few of the polygons in this region need further spatial delineation, but labeling with respect to forest successional status and agricultural subclass could be improved. Generally, seral stage of east side forests is difficult or impossible to determine because of the more open canopies at all seral stages compared to west-side forests. Some coarse seral stage differentiation might be possible in the more mesic zones because canopy closure is more complete. The west side of this region was covered by a scene from late May, too early for the forest to be at its peak. Difficulties caused by the early scene are separation of low-elevation, open-canopy forests from cuts and interference of snow with upper-elevation classes.
Region 10 (Okanogan Highlands)
Spatial delineation of polygons is good, but the scenes for this region were both from May. It is particularly difficult here to distinguish open forest from cuts and steppe. We purchased a scene taken in October of 1992 but that scene was not used in classification. Even with a good scene, differentiation of forests by seral stage is probably impossible because of the variability of canopy closure.
Region 11 (Columbia Basin)
The central part of the Basin is covered by a scene from May when vegetation was at its peak. Differentiation of agricultural classes by irrigation status and crop type is good, but some of the agricultural polygons are very large. Polygon size presents a problem in vertebrate modeling because, all other things being equal, larger polygons would be expected to include more species. Thus, even though the large agricultural polygons are usually very uniform, more spatial refinement is desirable. Straight lines between cover types are often encountered in the Basin; these lines are usually the result of land management by section block and are only rarely an artifact of the classification techniques.
Classification of steppe vegetation by degree of disturbance, i.e., whether vegetation is mostly native grasses and shrubs is desirable, as is classification by structure (grass, shrub savanna, shrub, etc). In most cases, classification to these levels is dubious, so users of the land cover map should not rely too heavily on the difference between, say 603 (a shrub savanna of unknown disturbance status) and 612 (a disturbed grassland). Soil signatures dominate in the Basin, especially on the west side where the available scenes were from August. A user should also keep in mind that grasslands are defined as having <10% shrub cover. The majority of Basin vegetation has between 5% and 15% shrub cover and so falls along the line between grasslands and shrub savanna as defined by the UNESCO classification. Naturally occurring, "true" shrubland (>25% shrub cover) is rare in the Basin except in drainages and in the lowest, hottest part of the Basin. Overgrazing and fire suppression have also created more shrubland. Shrubland in drainages may have been labeled as riparian.
Region 12 (East Central Cascades)
This complex region could use a thorough review by persons intimately familiar with its forests. Moisture flow through Snoqualmie Pass makes some of these forests more mesic than those to the north and south. The scene that covers this region is from August and is nearly ideal for mid- and high-elevation forests, though shading is a problem. The complex topography and rainfall patterns make zonal labeling a challenge. In some areas, seral stage labeling is probably fairly accurate; in others, the complex topography and shadowing turned it into guesswork; seral stage of most forests is unlabeled. The late date allowed alpine/subalpine areas to be differentiated with more confidence than in other areas.
Region 13 (Southeast Cascades)
Ground data for oak and conifer forests in Klickitat County were very good, so the oak, oak/pine, and savanna cover probably has a high accuracy. Ground data did not supply adequate information about conifer species or agricultural classes. Vegetation zone labels for conifer forests outside of the Ponderosa Pine zone need more investigation, especially near the Cascade crest and for the locations of the Grand Fir and Western Hemlock zones.
Region 14 (Blue Mountains)
The most common delineation of a polygon is the outline of a slope. I have a high confidence that the polygons labels are correct, but the heterogeneity between draws and humps on slopes needs to be kept in mind when interpreting classes. The primary cover of a slope will typically be the vegetation between draws and the secondary cover will typically be the vegetation in the draws. (The order can also be reversed.) Both primary and secondary cover need to be known to form a good mental picture of a particular slope. A user should also bear in mind that zone boundaries in the Blues are blurred at best. The patches of Subalpine Fir, Engelmann Spruce, and Lodgepole Pine are distinct on satellite imagery, but they may be included in polygons with Ponderosa Pine and grasslands that occur within a few feet on a southern slope with shallow soil.
Region 15 (Northeast Cascades)
Many of the polygons in this region are unacceptably large and need more spatial refinement. The west part of this region was covered by a July scene; the date of the scene was ideal for all but the highest elevation forest. The east part of the region was covered only by a scene from May. The early date caused many problems distinguishing open forest from cuts, the canopy closure of high forest, and the extent of subalpine forest and alpine zones. In addition, a small part of the region was under cloud cover, so zonation and cover were deduced from location and elevation. The majority of these high, cold, dry mountains are in the Subalpine Fir zone or the Interior Douglas-fir zone. Most of the Subalpine Fir zone is included in the Pasayten Wilderness. Age cannot be reliably determined from signature alone in these forests, but the high forests, if closed, were usually assumed to be late seral because of their location. This assumption may not be valid, especially on south slopes where fire frequency is high.
Region 16 (Northwest Cascades)
This region contains a number of large polygons that, in reality, cover several zones (although each polygon is labeled with only one zone) and include many structural and seral classes, particularly in Whatcom and Skagit Counties. Lack of time was particularly a factor in this region; the terrain is so rugged, more ground data is needed to separate the influence of shadow, seral stage, and steep slopes. We were fortunate in having a relatively recent classification of the North Cascades National Park and vicinity as a guide (Jim Agee, personal communication). That classification effort had used a large number of ground data points.
Region 17 (Southwest Cascades)
Spatial division of polygons is fairly good, with a moderate to high degree of confidence in labels. Many of the cuts in the Gifford-Pinchot National Forest are smaller than our minimum mapping unit, so the polygons give a misleading impression of cover uniformity in places. Zonations need more work in the transition area between east-side and west-side forest types.
Region 18 (Willamette Valley)
Polygon delineation is about as fine spatially as is practical, but, like all developed areas, fragmentation is extremely high, which makes boundaries difficult to discern. We needed more information about crop type to improve agricultural labeling. Also, the high fragmentation and small patch size of remnant native prairies (if there are any) would require ground surveys.
Region 19 (Inner Olympic Peninsula)
Spatial breakdown of polygons is good, with moderate to high confidence in labels. Part of this region is highly fragmented and the same comments apply as for the fragmentation in the Trough, but the problem is minor compared to the fragmentation of the Trough. This region includes the dry northeast Olympic Peninsula. The dry montane forest is given its own zone (32), which is currently dominated by Douglas-fir. The driest high forests are dominated by Subalpine Fir rather than the Mountain Hemlock that dominates most other west-side subalpine forests. The Elwha Valley is transitional between the dry northeast and the wet side of the Olympics. I excluded the Silver Fir zone from the Elwha Valley, but it may actually be a small component.
References for Land Cover Mapping Sources
Daubenmire, R. 1970. Steppe vegetation of Washington. Originally Agriculture Experiment Station Publication XT0062. Reprinted in 1988 as EB1446, United States Department of Agriculture, Washington State University Cooperative Extension, College of Agriculture and Home Economics, Washington State University, Pullman, 132 pp.
Daubenmire, R., and J. B. Daubenmire. 1968. Forest vegetation of eastern Washington and northern Idaho. Technical Bulletin 60. Washington Agricultural Experiment Station, College of Agriculture, Washington State University, Pullman, 104 pp.
DeLorme Mapping Company. 1995. Washington Atlas and Gazetteer. DeLorme Mapping, Freeport, Maine. 120 pp.