Page 1
Land Cover Mapping for the Southwest Regional GAP Analysis Project
Tenth Biennial Forest Service Remote Sensing Applications Conference, RS-2004, Salt Lake City, Utah
John Lowry and R. Douglas Ramsey
Remote Sensing/GIS Laboratory
Utah State University
Logan, Utah
Page 2
Presentation Overview
• Project Background & Objectives
• Mapping Methodology
• Training Data Collection Approach
• Current Status & Preliminary Results
Page 3
• State-based vegetation classification systems (cover type legends)
• State-based mapping methods
• State-based mapping area
A R I Z O N A
1999
52 Classes
N E W M E X I C O
1996
42 Classes
U T A H
1995
36 Classes
C O L O R A D O
2000
52 Classes
N E V A D A
1997
65 Classes
I. Project Background & Objectives
Page 4
• 40 Mapping zones
• Spectrally consistent
• Eco-regionally distinct
• Labor divided among 5 state teams
UTNV
CO
AZ NM
Page 5
NVC Formation
NVC Alliance
NVC Association
Gap Analysis ProgramMRLC 2000
Proposal
~1,800 units
National Park Mapping
~ NVC Class/Subclass
~10units
NatureServe Ecological Systems
~5,000 units
~700 units
(Natural/Semi-natural types)
~300 units
(Slide Courtesy Pat Comer, Nature Serve)
Thematic Target LegendDeveloped with NatureServe
Page 6
Groups of plant communities and sparsely vegetated habitats unified by similar ecological processes, substrates, and/or environmental gradients...and spectral characteristics.
Ecological Systems
Page 7
Elevation Landform
Predictor Datasets: DEM derived
Page 8
July-Aug Sept-Oct
ETM Bands 5, 4, 3 ETM Bands 5, 4, 3
Predictor Datasets: Imagery Derived
Page 9
• Data-mining software for decision-making and exploratory data analysis
• Identifies complex relationships between multiple independent variables to predict a single categorical class
• Predictor variables may be categorical or continuous
• Recursively “splits” the predictor data to create prediction rules or a decision tree.
• Software packages available: See5, SPLUS, CART
II. Mapping Methods: Classification Trees
Page 10
Mining the Predictor Layers
Fall Brightness
Summer NDVI
Elevation
Landform
Etc….
Output table
SAMPLE SITESImagery: Landsat 7 ETM (1999-2002) for spring, summer & fall
NDVI, SAVI, Brightness,Greeness, Wetness, Landsat 7 Bands
DEM: Elevation, Aspect, Slope, Landform
Vector: Geology, Soils
Meteorological : DAYMET
Page 12
0.2 0.3 0.4 0.5
FALL 1999 NDVI
1500
2000
2500
3000
ELE
V
grass
wyoming
mountain
juniper
mountain
g
g
g g
gg
g
g
gg
g
g
gg
g ggg
ggg
gg
g
g
g
g
g g
gggggg
g
j
jj
j
j
jj
j jjjj
jj j
j
j
j
j
j
j
jj
jj
j
j
j
jj
j
jj
j
j
j
j
j
j
j
m
m
m
m
m
m
mm
mm
mm
m
m
mm
mm
m
m
m
m
m
m
m
m
m
mm
m
mm
mmm
m m
m
m
mm
mm
mm
m
m
m
m
m
ww
ww
www
w
w ww
wwww
w ww
ww
w
w
ww w
www
www
w
w
ww
ww ww
ww
www
ww
ww
w ww
ww
ww w
w w
w
w
w
ww
w
www www
w
ww
ww
ww w
www
www
w
ww w
w
w
w www
w ww
w w
ww ww
www
ww
w
w
ww w w
ww
www
w wwwww
Simplified Example: Splits on 2 variables
Page 13
|FA99ND<0.24685
ELEV<1515.5
ELEV<1931.38
ELEV<1935.83
grass
wyoming mountain
juniper mountain
Simplified Example: Tree output for 2 variables
Page 14
Example: Rules Output
See5 [Release 1.17] Wed Apr 23 13:42:02 2003 Options: Rule-based classifiers Class specified by attribute `dep' Read 7097 cases (10 attributes) from t3.data Rules: Rule 1: (17, lift 45.4) band01 = 1 band03 > 115 band03 <= 122 band05 <= 81 band06 <= 1419 -> class 1 [0.947] Rule 2: (9, lift 43.6) band01 = 1 band02 <= 102 band03 > 115 band03 <= 118 band04 <= 117 band06 <= 1419 -> class 1 [0.909] Rule 3: (6, lift 42.0) band01 = 13 band03 <= 110 band05 <= 73 band07 = 4
| Generated with cubistinput by EarthSat| Training samples : 10260| Validation samples: 2565| Minimum samples : 0| Sample method : Random| Output format : See5 dep. |h:/mgzn_5/trainingdata/mrgpts1.img(:Layer_1) Xcoord: ignore.Ycoord: ignore.band01: 1,2,-30 |h:/mgzn_5/img_files/sum30cl.img(:Layer_1)band02: continuous. |h:/mgzn_5/img_files/subrt.img(:Layer_1)band03: continuous. |h:/mgzn_5/img_files/sundvi.img(:Layer_1)band04: continuous. |h:/mgzn_5/img_files/fandvi.img(:Layer_1)band05: continuous. |h:/mgzn_5/img_files/fabrt.img(:Layer_1)band06: continuous. |h:/mgzn_5/img_files/elev.img(:Layer_1)band07: 0,1,2,3,4,5,6,7,8,9,10. |h:/mgzn_5/img_files/landf.img(:Layer_1) dep: 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20.
|h:/mgzn_5/trainingdata/mrgpts1
Page 15
Boosting (iterative tree’s try to account for previous tree’s errors)—C5
Different over-fitting issues associated with each tree tend to be averaged out.
Multiple Tree Approaches MNF1<=2
8
MNF3<=19 MNF13>5
6
MNF1>19
MNF16<=54
decid.
shrub
MNF3<=38
MNF1<=28
MNF1<=25
MNF3<=24
MNF8<=28
decid. shrub
MNF11<51
MNF17>56
shrub cedar
decid.
P. pine
cedar
MNF2<=43
cedar
cedar P. pine
MNF1<=28
MNF3<=19 MNF13>5
6
MNF1>19
MNF16<=54
decid.
shrub
MNF3<=38
MNF1<=28
MNF1<=25
MNF3<=24
MNF8<=28
decid. shrub
MNF11<51
MNF17>56
shrub cedar
decid.
P. pine
cedar
MNF2<=43
cedar
cedar P. pine
MNF1<=28
MNF3<=19 MNF13>5
6
MNF1>19
MNF16<=54
decid.
shrub
MNF3<=38
MNF1<=28
MNF1<=25
MNF3<=24
MNF8<=28
decid. shrub
MNF11<51
MNF17>56
shrub cedar
decid.
P. pine
cedar
MNF2<=43
cedar
cedar P. pine
MNF1<=28
MNF3<=19 MNF13>5
6
MNF1>19
MNF16<=54
decid.
shrub
MNF3<=38
MNF1<=28
MNF1<=25
MNF3<=24
MNF8<=28
decid. shrub
MNF11<51
MNF17>56
shrub cedar
decid.
P. pine
cedar
MNF2<=43
cedar
cedar P. pine
MNF1<=28
MNF3<=19 MNF13>5
6
MNF1>19
MNF16<=54
decid.
shrub
MNF3<=38
MNF1<=28
MNF1<=25
MNF3<=24
MNF8<=28
decid. shrub
MNF11<51
MNF17>56
shrub cedar
decid.
P. pine
cedar
MNF2<=43
cedar
cedar P. pine
V
O
T
E
(Slide Courtesy Bruce Wylie, USGS EDC)
Page 16
Imagine CART Module (USGS Eros Data Center)—See5-Imagine Integration
Page 17
Legend
Desc
cool aspect cliffs, scarps, cirques, canyons
gently sloping ridges and hills
hot aspect cliffs, scarps, cirques, canyons
moderately dry slopes
moderately moist steep slopes
nearly level plateaus or terraces
toe slopes, bottoms, and swales
valley flats
very dry steep slopes
very moist steep slopes
III. Training Data Collection
Opportunistic, ground-based sampling, stratified by digital landform model
Page 18
Percent ground cover by dominant species is recorded through ocular estimation. Only the top 4 species of each of 4 life forms are recorded
Page 20
~3000 Air Photo Interpretation Sites from USFS Photos
! !!! !! !!!! !!!! ! !! !! !!! !! !!! !!! !!! ! !! ! !! ! !! !!! !! ! ! !!! !!! !! !! !!! ! !! !! !!! !!! ! !!! ! !! ! ! !!! !!!! !!! !!!! !! !!!!! !!! ! !! !! ! !!!! !!!! !! !! !!!! !!! ! !! !! ! !!! !!! !! !! !!!! !!! !! !!! !!!! !!!! !!! !!!!! !! !! !!!!! !!!! ! !! !! !! ! ! !!! !!! !! !!!!! !! !!! !!!!!!! ! ! !! !! !!! !! !!!! ! ! !! !! !!!! ! !! !!! !! !! !! !! !! !! !! !!! ! !! !! !! ! !! !!!! ! !! ! !! ! !!! ! ! !! ! ! !!!!! !!! !!!! !!!! !!! ! !!!! !! ! !!!!! !!! ! !!! !!! !!! !! ! !! !! !! ! !! ! !!! ! !! ! ! !! ! !!! !! !!! !! !!! ! !! ! ! !! !! !!! !!!! ! !!! !! !! ! !!! ! ! ! !! ! !! !!!! !!! !!! !!!! !! !! !! !!!!! ! ! !!! !! ! !! !!!! ! !! !!! !!! ! ! !! !!! !!! !!!! !!!! !! !! !! ! !!! ! !! ! !! ! !! !!! !! !! !! !! ! !!! !!! !!!! !!! !!!! !! ! !!! !! !!!!!! !! ! !!! ! ! !! ! !! !! !! !! ! ! !!! !! ! !! !! !! ! ! !!! !!! !! ! !!! !! ! !! !! ! !! !!!! ! !! !!! !! !! !! !! !! !! !! !! !!! ! ! !!! !! !! !! !!! !!!!! !!!!!!!
!!!!!!!!! !!!!! !!! !!! !!! !!!!!!! !!! !!!! !!!! !!!! !!! !!!!!!!!!!!!! !!!!!! !! !!!! ! !! !! !!!! !!!! ! !!! !!! ! !!!!! !! !!! !!!! !! !!!!!! !!! !! !! !! !! !!! !!!! !!! !!! !! !! !! !! !!! ! !!!!!! !! !! ! !! !! !! ! ! !!! ! !!! !!! ! !!!! ! !!!! ! !! !! !! !! !! !! !!! !! !!! ! !!!! !!! !!! !!! !!!! !!!! !! !!! !!! !!! ! !! !!! ! !!! !! ! !! !!!! ! !! !! !!! !! !!! !! ! !!! !!!! !! ! !!! !! ! !! ! !!! !!! !!! !! !!!! ! !!! !!! !! !!! !!! ! !!!! ! !! ! ! ! !!!!! ! !! !! ! !!! !!!! !! !! !!! !!! ! !!!! !! ! !! !!! !!! !!! !!! !! ! !!! !!! !!! ! !!! !! ! !!! ! !! !! !!! !! ! !!! !! !! !! !! ! !!! !!!!! !!! !!!!!!! !!! !! !!! !! ! !!! !! ! !! !! !! !! !!! !! ! !! !! !!! !!! !! !!!!! !!! !!!!! ! !! ! !!!! !!!! ! !!! !!!! !! !!! !! !!!! !! !!!! !! !!! !!!! !!!!!! ! !! ! !!!! !!! !!!!! !!! !! !!!! !! ! !! !!!! ! !! ! !!!! !!! !! !!! ! !! !! !! !!! !!! !!!!!! !!!!!! !!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!! !!!!! !!!!! !! !!!!! !!! !!!!! !! ! !!! !! !! !! !!! !!!!! !! !!! !!!! !!!!!! !!! ! !! !!! !!!! !!!!!!!!!!!!! !! !!! !! !!!!!!!!!!! !!!!!!!! !! !!!!!!!!!!!!!!!!!!
!! !! !!!!!! !! ! !!!!!! !! ! !!! !! !! !!! !! !! !! !! !!!! !!!!! !! !!! !!!!!! !!!! !!! !!!! !! !! ! !!!!! !!!!! !!! ! !! ! !! !! !!!!!! !!! ! ! !! !! !! !! !! !! !!! !!!! !!!! !!!!!!! !!! !! !!! !!! !! ! !! ! !! ! !!!!! !! ! !! !!! !!! !! !! ! !! ! !! !!!!!!! ! !! ! !!! !!! !! !!!!!!! !! !!!! !! !!! !!!! !!!! !!! !!! !!!! !! !! ! !!! !! ! !! !!!!! !!! !!!!!!! !!!! !! !!! !!! !!!! !!!! !!! ! !!! !! !!! !!! !!!!! !!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!! !!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!! !!!! !! !! !! !! !!!! ! !! ! ! !! !!!! !!!!!! ! !!!! !!! ! !! !!!! !!!!! !!! !!!!!! !! !!! !! ! !! ! !! !!!! ! ! !!!!! ! !! !! !!! ! ! !! !! ! !! !!!! ! !!! !!! ! !!! !!!! !!! !! !! ! !!! ! !! !!!!!! !! ! !!! ! !!!! !!!! ! !! ! !!!! !! !!! !! !! !! !!! !! !!! !!!! !! !! !! ! !! !!!! ! ! !! !!! !!!! !! !! ! !!!!! !! ! !! !! !! ! !!!! !!! !! !!! !! !! ! !! !! ! ! !!!!! !!!! ! !! !!! !! !! ! !!! !! !! !!! ! !! ! !! ! !! !!! ! !! !! !!! !! !! !!! !! !!!! !! !! !!!!!! !!!!! !! !! ! !!!!! !!!!!! !! !! ! !! ! !! !! !! !!! !!!!! ! !! !!! !! !! !!!!! !!!!!! !!! !! !! !!! !!!!!! !!!!! !!! !!! !!!! !! !!! !!!!! !!!!! !!!!!!!!!! !!! !!!
!!!!!!! !!! !! !! !!!! !! !! !! ! !!! !! !! !! !!! !! !! !!! !!! !!!!! !! !! !!! !!!!! ! !! !! !!!! !! !! !!!!!! !!! !!! !!! !!! !! ! !!! ! ! !!!! !! ! !!!! !! !!! !! ! !!!!! !! !! !! !! !!! ! !! !!! !! ! !!! ! !!! ! !!! !!!!!! !! !! !! !! !!!! !! ! !!! !!!! ! !!! !!! ! !!!! !! ! !!! !!! ! !!! !!!! !!!!! !!!!! ! !!!! !! !!!!!! !! !! !!! ! ! !!! ! ! !! ! !!! !!! !! !!! ! !!! ! !! !!! !! !!! !! !!! !! !! ! !! !!! !! !!!!! ! !!! !! !! ! ! !! !! !!!!! !! !! ! !! ! !!! !!!! !!! !!! !!!! !!!! !!! !! !!!!!! !! !! ! !!! !!! !! ! !! !!! !!! !!! ! ! !!!!! !! !!!!!!!!! !!! ! !! !!! !! ! !!! !!!! ! !! !! !! !! !! !!!!! !! !! !!! !! !!!!! !! !!!! !! !!!!! !!! ! !! !! !! !!! !! !!! !!! ! !!!! !!! !! ! !!!!!!! ! ! !! !! !! ! !! ! !!! !! !!! !! !!! !!! !! !! ! !! !!!! ! !! ! !! !!! !!!! ! !! !! !! !!! !!! !! !!! ! !!!! !! !!! ! !!!! ! !! !!! !!! !!! !! !! ! ! !! !!!! !!! !! !!! !!!! ! !!!! !!
Page 21
Regional Total ~ 93,000
Page 22
IV. Current Status & Preliminary Results
Page 23
Edge-matching between three mapping areas
Page 24
Considered correctly classified if majority of pixels agree with sample polygon
Accuracy Assessment with 20% withheld data:
Page 25
Accuracy Assessment with 20% withheld data: Southern Wasatch Range
Page 26
Park ValleyPark ValleyPark Valley
AGRICULTURE
GRASSLAND
GREASEWOOD
LOWLAND RIPARIAN
PICKLEWEED BARRENS
SAGEBRUSH
SAGEBRUSH/PERENNIAL GRASS
SALT DESERT SCRUB
WATER
PERENNIAL GRASSLAND
ANNUAL GRASSLAND
PLAYA
BIG SAGEBRUSH SHRUBLAND
XERIC MIXED SAGEBRUSH
MIXED SALT DESERT SCRUB
SEMI-DESERT SHRUB STEPPE
SEMI-DESERT GRASSLAND
LOWER MONTANE RIPARIAN WOODLAND AND SHRUBLAND
GREASEWOOD FLAT COMPLEX
CULTIVATED CROPS/IRRIGATED AG
PASTURE/HAY/NON-IRRIGATED AG
ANNUAL FORBLAND
1995 GAP 30 M 2004 GAP 30 M1995 GAP Pub.1KM
Page 27
Summary
• Approximately 100 Ecological Systems and 10 NLCD Land Use classes
• Generalized to 1 acre MMU
• Delivered via NBII data node
• Anticipated completion: 1 September, 2004