The NLCD 2001 is created by partitioning the U.S. into mapping zones. A total of 66 mapping zones were delineated within the conterminous U.S. based on ecoregion and geographical characteristics, edge matching features and the size requirement of Landsat mosaics. Mapping zone 56 encompasses whole or portions of several states, including the state of Florida. Questions about the NLCD mapping zone 56 can be directed to the NLCD 2001 land cover mapping team at the USGS/EROS, Sioux Falls, SD (605) 594-6151 or mrlc@usgs.gov.
Each class accuracy is as follows: (Errors of Omission/Commission)
0 Background (N/A)
1 Unclassified (Cloud, Shadow, etc)(N/A)
2 High Intensity Developed (74%/86%)
3 Medium Intensity Developed (50%/88%)
4 Low Intensity Developed (100%/21%)
5 Open Spaces Developed (82%/99%)
6 Cultivated Land (90%/95%)
7 Pasture/Hay (91%/92%)
8 Grassland (79%/61%)
9 Deciduous Forest (N/A)
10 Evergreen Forest (81%/87%)
11 Mixed Forest (N/A)
12 Scrub/Shrub (81%/68%)
13 Palustrine Forested Wetland (88%/92%)
14 Palustrine Scrub/Shrub Wetland (79%/75%)
15 Palustrine Emergent Wetland (88%/99%)
16 Estuarine Forested Wetland (97%/92%)
17 Estuarine Scrub/Shrub Wetland (75%/86%)
18 Estuarine Emergent Wetland (83%/79%)
19 Unconsolidated Shore (100%/100%)
20 Bare Land (83%/91%)
21 Water (99%/96%)
22 Palustrine Aquatic Bed (100%/100%)
23 Estuarine Aquatic Bed (100%/100%)
The validation points were sampled from the field collected dataset of points BEFORE classification. These points were witheld from classification and not viewed or modified by staff involved in the classification process. The sampling procedure involved a spectral/clump-based spatial stratification. The 2001 leaf-off MRLC mosaic was clustered to 100 clusters and clumped. This layer was divided into alternating clumps. The even clumps were sent to potential validation, odd to training. The process was repeated such that 1/4 of the field points were earmarked as validation points. These points were then reviewed for accuracy and placement by independent QC staff. Points that were not correct or too close to feature boundaries were either deleted or moved to more feature-centric locations. No points were added or changed, only moved or deleted. Probably over half of the points were modified. NOAA CSC requires 6 out 9 (2/3) pixels in a 3x3 window of the land cover to correctly match the validation points. The land cover was processed to determine the 2/3 majority call for each pixel and then intersected with the validation points in an error matrix to derive the above numbers.
Post-Processing Steps: None
Known Problems: None
Spatial Filters: None
Conceptually, the descriptive tree is a classification tree generated by using the final minimum-map- unit land cover product (1 acre) as training data, and Landsat and other ancillary data as predictors. The goal of the descriptive tree is to summarize the effects of boosted trees (10 sequential classification trees) into a single condensed decision tree that can be used as a diagnostic tool for the classification process. This descriptive tree can be used to assess the relative importance of each of the input data sets on each land cover class. Such information may also be useful to customize the minimum-mapping-unit classification to meet a user's specific needs through raster modeling. Descriptive trees usually capture 60 to 80% of the information from the original land cover data.
The leaf or terminal nodes of the descriptive tree are assigned to sequential numbers (called node numbers) and mapped across the entire mapping zone on a pixel-by-pixel basis. These node numbers can then be matched with the various conditional statements associated with each respective terminal node. This spatial layer appears similar to a cluster map, but is the result of a supervised classification - not an unsupervised clustering. This node map can potentially be used as input by users to customize NLCD land cover, by linking the spatial extent of an individual node with the rules of the conditional statement.
The Land Cover spatial classification confidence data layer is provided to users to help determine the per-pixel spatial confidence of the NLCD 2001 land cover prediction from the descriptive tree. The C5 algorithm produces an estimate (a value between 0% and 100%) that indicates the confidence of rule predictions at each node based on the training data. This spatial confidence map should be considered as only one indicator of relative reliability of the land cover classification, rather than a precise estimate. Users should be aware that this estimate is made based on only training data, and is derived from a generalized descriptive decision tree that reproduces the final land cover data. However, this layer provides valuable insight for a user to determine the risk or confidence they choose to place in each pixel of land cover.
A logic statement from a descriptive tree classification describes each classification rule for each classified pixel. An example of the logic statement follows:
IF tasseled-cap wetness > 140 and imperviousness = 0 and canopy density < 4, then classify as Water
This logic file can be used in combination with the spatial node map to identify classification logic and allow modifications of the classification based on user's knowledge and/or additional data sets.
Additional information may be found at <http://www.mrlc.gov/mrlc2k_nlcd.asp>.
Pre-processing steps: Each Landsat TM scene was geo-referenced by USGS (United States Geological Survey)/EROS. Then MDA Federal staff verified the scenes for spatial accuracy to within 2 pixels. The data was geo-referenced to Albers Conical Equal Area, with a spheroid of GRS 1980, and Datum of WGS84. The data units is in meters. At-satellite reflectance was performed on each scene and the tasseled cap transformation applied. All of the image data used was Landsat TM 5 or 7. The mosaicked dataset was used for classification.
Classification:
The classification involved automated and manual approaches. Points collected during field work were used as training and also foraccuracy assessment. The training points were used as the dependent variable in the CART (Classification Analysis by Regression Tree) approach. The tasseled cap Landsat TM imagery for three dates were used as the independent variables. Ancillary datasets were also used as independent variables. After many attempts, a rough classification was produced. The urban areas category was derived from the USGS impervious data set (2001-Era data) for zone 56. The result of this produced the provisional classification. Then models were applied to this data to incorporate information from ancillary data. The result of this was the final-no-edits version of the classification. This represented a fully automated product. This product was then altered by hand edits to refine further the classification. This produced the final-with-edits version which is the final version of the classification and is the one described here. Attributes for this product are as follows:
0 Background
1 Unclassified (Cloud, Shadow, etc)
2 High Intensity Developed
3 Medium Intensity Developed
4 Low Intensity Developed
5 Developed Open Space
6 Cultivated Land
7 Pasture/Hay
8 Grassland
9 Deciduous Forest
10 Evergreen Forest
11 Mixed Forest
12 Scrub/Shrub
13 Palustrine Forested Wetland
14 Palustrine Scrub/Shrub Wetland
15 Palustrine Emergent Wetland
16 Estuarine Forested Wetland
17 Estuarine Scrub/Shrub Wetland
18 Estuarine Emergent Wetland
19 Unconsolidated Shore
20 Bare Land
21 Open Water
22 Palustrine Aquatic Bed
Acquisition dates of Landsat ETM+ (TM) scenes used for land cover classification in zone 56 are as follows:
SPRING-
Index 1 for Path 15/Row 41 on 01/09/02 = Scene_ID 7015041100200950
Index 1 for Path 15/Row 42 on 01/09/02 = Scene_ID 7015042000200950
Index 1 for Path 15/Row 43 on 01/09/02 = Scene_ID 7015043000200950
Index 2 for Path 16/Row 40 on 02/17/02 = Scene_ID 7016040000204850
Index 2 for Path 16/Row 41 on 02/17/02 = Scene_ID 7016041000204850
Index 3 for Path 16/Row 42 on 01/13/01 = Scene_ID 7016042100101350
Index 4 for Path 16/Row 43 on 01/05/01 = Scene_ID 5016043000100510
Index 5 for Path 17/Row 40 on 02/24/02 = Scene_ID 7017040000205550
Index 5 for Path 17/Row 41 on 02/24/02 = Scene_ID 7017041000205550
LEAF ON (Summer)-
Index 1 for Path 15/Row 41 on 05/14/01 = Scene_ID 7015041100113450
Index 2 for Path 15/Row 42 on 02/21/00 = Scene_ID 7015042000005250
Index 3 for Path 15/Row 43 on 04/09/00 = Scene_ID 7015043100010050
Index 4 for Path 16/Row 40 on 04/03/01 = Scene_ID 7016040000109350
Index 4 for Path 16/Row 41 on 04/03/01 = Scene_ID 7016041000109350
Index 5 for Path 16/Row 42 on 05/18/00 = Scene_ID 7016042000013950
Index 6 for Path 16/Row 43 on 04/19/01 = Scene_ID 7016043000110950
Index 7 for Path 17/Row 40 on 03/28/02 = Scene_ID 7017040000208750
Index 8 for Path 17/Row 41 on 04/10/01 = Scene_ID 7017041000110050
Index 9 for Path 16/Row 40 on (part) Derived image for cloud replacement
LEAF-OFF (Fall)-
Index 1 for Path 15/Row 41 on 11/20/03 = Scene_ID 5015041000332410
Index 2 for Path 15/Row 41 on 08/26/01 = Scene_ID 5015041000123810
Index 3 for Path 15/Row 42 on 11/06/01 = Scene_ID 7015042000131050
Index 4 for Path 15/Row 42 on 12/19/99 = Scene_ID 7015042009935350
Index 5 for Path 15/Row 43 on 11/06/01 = Scene_ID 7015043000131050
Index 5 for Path 16/Row 40 on 08/25/01 = Scene_ID 7016040000123750
Index 6 for Path 16/Row 41 on 10/17/00 = Scene_ID 5016041000029110
Index 7 for Path 16/Row 42 on 11/13/01 = Scene_ID 7016042100131750
Index 8 for Path 16/Row 43 on 11/13/01 = Scene_ID 7016043000131750
Index 9 for Path 17/Row 40 on 11/07/02 = Scene_ID 7017040000231150
Index 9 for Path 17/Row 41 on 10/06/02 = Scene_ID 7017041000227950
Index10 Various Derived Images for cloud replacement
The completed single pixel product was then generalized to a 1 acre (approximately 5 ETM+ 30 m pixel patch) minimum mapping unit product using a "smart eliminate" algorithm. This aggregation program subsumes pixels from the single pixel level to a 5-pixel patch using a queens algorithm at doubling intervals. The algorithm consults a weighting matrix to guide merging of cover types by similarity, resulting in a product that preserves land cover logic as much as possible.