Impact of Urban Growth & Development on Vegetation in San Antonio. by Eric Bowman and Kari Papelbon

Transcription

1 Impact of Urban Growth & Development on Vegetation in San Antonio by Eric Bowman and Kari Papelbon

2 Source: Introduction

3 San Antonio 1909 Source:

4 Growth Patterns Migration out from center of city More affluent Higher elevation What stayed downtown? Incorporated cities within San Antonio Annexation Growth North of downtown, now extending West and East

5 Study Area & Data Hwy 281, North of 1604 Chosen because of San Antonio s annexation history Tremendous growth in the 281 corridor Continued development North to City Limits Landsat 7 ETM+ images (NLAPs) from Texasview.org, Path 2739, Bands 7, 4, , day , day , day 16

6 ,000 50,000 40,000 34,000 30,000 20,000 10,000

7 ,000 50,000 40,000 37,000 30,000 20,000 10,000

8 ,000 50,000 40,500 40,000 30,000 20,000 10,000

9 ,000 50,000 44,500 40,000 30,000 20,000 10,000

10 ,000 51,000 50,000 40,000 30,000 20,000 10,000

11 ,000 56,400 50,000 40,000 30,000 20,000 10,000

12 Purpose To detect changes and the extent of changes in vegetation when compared to urban growth and development in an area known for increasing growth.

13 Methods Data from Data Archives Download nlaps images for Landsat 7 ETM + Visually inspected 3 images Clipped 1999 image using Resize Data Used 1999 clipped image to clip 2002 and 2003 images

14 1999 Clipped Image 1604

15 2002 Clipped Image 1604

16 2003 Clipped Image 1604

17 Methods Chose ROIs for each image Fallow: Red Trees: Green Old residential: Blue New residential: Yellow Roads: Cyan Fallow #2: Magenta Grass: Sea Green Quarry: Maroon Applied Supervised Spectral Angle Mapper - SAM

18 1999 SAM 0.10

19 1999 SAM, 0.90 Trees Refined

20 2002 SAM 0.10

21 2002 SAM, 0.50 Residential Refined

22 2003 SAM, 0.10

23 Methods Compute Difference Map Initial: 1999 Final: 2002 # of Classes: 7 based upon ROI s Change Detection Statistics Initial: 1999 Final: 2002 Set up class pairs Trees Residential new Residential old Fallow Grass Roads Quarry

24 Change Map Results

25 Results: Statistics Table Pixel Counts Pixel Counts Trees [Green] 112 points Residential_new [Blue] 128 points Fallow [Red] 82 points Residential_old [Yellow] 293 points Roads [Cyan] 33 points Quarry [Maroon] 33 points Row Total Class Total Unclassified Trees [Green] 56 points Residential_new [Blue] 62 points Residential_old [Yellow] 121 points Fallow #1 [Red] 84 points Grass [Sea Green] 74 points Roads [Cyan] 27 points Quarry [Maroon] 53 points Class Total Class Changes Image Difference

26 Results: Statistics Table Percentages Trees [Green] 112 points Residential_new [Blue] 128 points Fallow [Red] 82 points Residential_old [Yellow] 293 points Roads [Cyan] 33 points Quarry [Maroon] 33 points Row Total Class Total Unclassified Trees [Green] 56 points Residential_new [Blue] 62 points Residential_old [Yellow] 121 points Fallow #1 [Red] 84 points Grass [Sea Green] 74 points Roads [Cyan] 27 points Quarry [Maroon] 53 points Class Total Class Changes Image Difference

27 Methods Compute Difference Map Initial: 1999 Final: 2003 # of Classes: 7 Change Detection Statistics Initial: 1999 Final: 2003 Set up class pairs Trees Residential new Residential old Fallow Roads Quarry

28 Change Map Results

29 Results: Statistics Table Pixel Counts Pixel Counts Fallow [Red] 82 points Trees [Green] 112 points Quarry [Maroon] 33 points Residential new [Blue] 128 points Residential old [Yellow] 293 points Roads [Cyan] 33 points Row Total Class Total Unclassified Fallow [Red] 112 points Trees [Green] 55 points Residential new [Blue] 56 points Residential old [Yellow] 152 points Roads [Cyan] 27 points Grass [Magenta] 27 points Quarry [Maroon] 60 points Class Total Class Changes Image Difference

30 Results: Statistics Table Percentages Percentages Fallow [Red] 82 points Trees [Green] 112 points Quarry [Maroon] 33 points Residential_new [Blue] 128 points Residential old [Yellow] 293 points Roads [Cyan] 33 points Row Total Class Total Unclassified Fallow [Red] 112 points Trees [Green] 55 points Residential new [Blue] 56 points Residential old [Yellow] 152 points Roads [Cyan] 27 points Grass [Magenta] 27 points Quarry [Maroon] 60 points Class Total Class Changes Image Difference

31 Conclusions 31.6% of trees did not change from ; 31.7% of trees did not change from % of trees changed to grass from ; 7.65% of trees changed to grass from % of fallow lands changed to grass from ; 23.4% of fallow lands changed to grass from Expected results differed from outcomes Perhaps need to look at pre-1999 data Perhaps ROIs or classifications were off Better methods exists to do object-based classification to deal with errors