338 Walton et al.: Assessing Urban Forest Canopy Cover Table 2. Urban forest canopy cover comparison for Syracuse, New York.z Method Imagery Photograph interpretation Supervised classification Subpixel–Cubist (specific) NLCD subpixel (regional) USFS Forest Density USFS RPA (water removed) UFORE ground sample Photographic prints Emerge 2 ft (60 cm) (digital) NYS Orthos 1 ft (30 cm) (digital) Emerge 2 ft (60 cm) Landsat 30 m Landsat 30 m AVHRR 1.1 km AVHRR 1.1 km Year 1984 1999 2003 1999 ca. 2001 ca. 2001 1991 1991 1999 2001 2004 zResults in shaded rows were generated using digital classification techniques. yARowntree (1984); BZhu (1994); CDwyer et al. (2000); DMyeong et al. (2001); EUSGS National Land Cover Database (2001); Homer et al. (2007); F USDA Forest Service, Northern Research Station, Syracuse, NY, unpublished data. NLCDNational Land Cover Database; USFS RPAU.S. Forest Service’s Resources Planning Act; UFOREUrban Forest Effects; AVHRRAdvanced Very-High Resolution Radiometer; MAE mean absolute error; SE standard error. used as the basis for change analyses (Walton 2008); specific local analyses can produce more acceptable results. Comparison of the imagery-based estimates with field-based estimates shows good agreement. The three ground-sampled estimates do not indicate a definitive change in the urban forest canopy cover between 1999 and 2004. However, because tree cover is esti- mated to the nearest 5% on each plot, detection of canopy change from ground-based methods should focus on changes at the in- dividual tree level (which trees are declining or growing), be- cause individual tree field measures are more precise. Accurate analysis of cover change through time is relatively difficult depending on the scale of change detection. To detect changes in overall tree cover using digital image processing, the actual change in cover through time must be greater than the uncertainty of the overall cover estimates to reasonably state that change has actually occurred. Thus, increased classification ac- curacy will allow for detection of smaller changes in overall canopy cover. Mapping actual changes in canopy cover offer additional chal- lenges. At the individual pixel level (mapping unit), uncertainty of the estimate increases relative to the uncertainty for the over- all tree cover estimate. With relatively high uncertainty at the pixel level for both the base year and future year of the change analysis, mapping actual changes in canopy location is difficult. Maps of changes in tree canopy cover should include estimates of probability of change (such as in Morisette et al. 1999) to help determine the certainty of specific canopy cover change across a landscape. Methods to improve the accuracy of tree canopy cover maps (e.g., using LIDAR data to differentiate between trees and grass) could help improve the ability to detect changes in tree canopy locations. To help determine changes in canopy cover, digital photo- graph interpretation can also be used. By sampling urban areas using photograph interpretation, specific locations can be fol- lowed through time using random sampling to statistically esti- mate changes and probability of changes in cover types for vari- ous land uses. The advantage of estimating change with photo- graph interpretation is that it is relatively inexpensive and accurate but lacks the capability to map specific location changes ©2008 International Society of Arboriculture in canopy cover. Rather, canopy changes are estimated for geo- graphic areas or land use classes. CONCLUSION Urban tree canopy cover analyses provide useful data on the extent and distribution of the urban forest resource as well as for estimating various ecosystem services. The use of airborne or satellite imagery to assess urban forest canopy cover can yield reliable and repeatable results if the limitations and accuracy of the imagery and classification processes are understood. High- resolution imagery, photograph interpretation, and ground-based cover analyses can produce reasonable estimates of tree canopy at the citywide scale. Medium-resolution satellite imagery is useful for city to regional analyses and can provide accurate urban forest cover and distribution information when analyzed using subpixel techniques derived from local urban training data. In mapping or determining overall canopy cover change between two digital cover classifications, it is important to understand the accuracy of the cover data and maps so that actual changes in canopy cover can be quantified. Acknowledgments. Funding for this project was provided, in part, by the USDA Forest Service’s RPA Assessment Staff and State and Private Forestry’s Urban and Community Forestry Program. The use of trade, firm, or corporation names in this article is for the information and convenience of the reader. Such use does not constitute an official endorsement or approval by the U.S. Department of Agriculture Forest Service of any product or service to the exclusion of others that may be suitable. LITERATURE CITED American Forests. 2007. Urban Ecosystem Analysis. www.americanfor- ests.org/resources/urbanforests/analysis.php (accessed 12/20/07). Campbell, J.B. 2007. Introduction to Remote Sensing. 4th Ed. The Guil- ford Press, New York, NY. 626 pp. Dwyer, J.F., E.G. McPherson, H.W. Schroeder, and R.A. Rowntree. 1992. Assessing the benefits and costs of the urban forest. Journal of Arboriculture 18:227–234. Dwyer, J.F., D.J. Nowak, M.H. Noble, and S.M. Sisinni. 2000. Con- necting People with Ecosystems in the 21st Century: An Assessment Percent canopy cover 24% 25.7% 24.7% 26.6% 26.6% 12.7% 21.6% 23.0% 24.4% 23.1% 21.4% SE: 2.5% SE: 2.5% Acc: 82% MAE: 11.7% Standard error or accuracy Sourcey A F F SE: 1.97% SE: 1.82% SE: 1.79% D F E B C F F F
November 2008
| Title Name |
Pages |
Delete |
Url |
| Empty |
Search Text Block
Page #page_num
#doc_title
Hi $receivername|$receiveremail,
$sendername|$senderemail wrote these comments for you:
$message
$sendername|$senderemail would like for you to view the following digital edition.
Please click on the page below to be directed to the digital edition:
$thumbnail$pagenum
$link$pagenum
Your form submission was a success. You will be contacted by Washington Gas with follow-up information regarding your request.
This process might take longer please wait
