4 Cowett and Bassett: Street Tree Diversity in Three Northeastern U.S. States tistics for Simpson’s Diversity Index, the Shannon- Wiener Diversity Index, and distribution evenness (Buzas and Gibson 1969) at species and genus levels were calculated using PAST Paleontological Statis- tics soſtware Version 3.0 (Hammer et al. 2001). Sta- tistics for the inverse of Simpson’s Diversity Index were also determined. Some municipalities either collected data for genus and not species, or collected data at the species level for some but not all genera. Therefore, the number of municipalities for which relative abundance percentages and diversity- index statistics can be calculated at the genus level exceeds the number of municipalities for which relative abundance percentages and diversity-index statistics can be calculated at the species level. Because the relative abundance percentages and diversity-index statistics mentioned comprise a non-random sample, there is a potential for selec- tion bias and geographic variability to compromise the accuracy of further statistical analysis. Post- stratification of data and weighting with auxiliary information is a technique oſten used to correct for selection bias due to non-random sampling (Bethle- hem 2010). A New York State street tree assessment conducted previously by the authors (Cowett and Bassuk 2014), based on a non-random sample, strat- ified data by the 1990 USDA Plant Hardiness Zones (U.S. National Arboretum 1990) and then weighted the stratified data by a measure of street length con- tained within each zone. To correct for potential selection bias due to non-random sampling, a simi- lar technique was considered for this assessment. A Geographic Information System (GIS) shape- the 2012 Plant Hardiness Zones for the file of Mid-Atlantic region was purchased from Climate Source (Corvallis, Oregon, U.S.), the exclusive public distributor of the Plant Hardiness Zone GIS data sets. The 2012 Plant Hardiness Zones account for a general warming trend and changes in zone boundaries since the 1990 version (Daly et al. 2008; U.S. Department of Agriculture 2012). GIS soſt- ware was used to clip the zones to the boundaries of New Jersey, New York, and Pennsylvania (Figure 2). Each municipality was assigned to a zone based on the location of the municipality’s inner centroid (i.e., a geometrically calculated center point within a municipality’s boundaries). The relative abun- dance percentages for street tree species and genera found in inventoried municipalities were then aver- ©2017 International Society of Arboriculture aged. The means for prevalent species and genera were regressed on the 2012 USDA Plant Hardiness Zones in a one-way analysis of variance (ANOVA). Significant effects (α = 0.05) that satisfied sta- tistical assumptions for normality of residuals and homoscedasticity were found for many, but not all species and genera. Effects were generally greater for New York and Pennsylvania than for New Jersey. Based on these findings, it was decided to stratify data by the 2012 Plant Hardiness Zones. Auxiliary information used for weighting pur- poses in the 2014 New York statewide assessment was a measure of street length contained within each 1990 USDA Plant Hardiness Zone. This mea- sure reflected, first, obtaining a GIS shapefile of all street centerlines statewide from New York State; second, deleting street types, such as drive- ways, interstate highways, and divided highway segments, unlikely to contain street trees using New York Accident Location Information Sys- tem (ALIS) codes; third, selecting centerlines contained within cities, villages, and Census Des- ignated Places (CDPs) as well as Census Blocks with a population density of at least 500 persons per square mile (ppsm); and, fourth, calculating the percentage of selected street length contained within each Plant Hardiness Zone as a percentage of the selected statewide whole. Due to differences found in the coding and formatting of each state’s most current street centerline data, GIS shapefiles provided by the states could not be used as auxil- iary information for weighting purposes. Instead, U.S. Census TIGERLine All Roads GIS shapefiles (U.S. Census Bureau 2014) composed of data coded and formatted similarly for all three states, were used; first, to select street types by MTFCC (MAF/ TIGER Feature Class Codes) codes in a manner matching as closely as possible the 2014 New York statewide assessment, and second to select streets contained within all Census Places, Census Urban- ized Areas, and Census Blocks with a population density of at least 250 ppsm located in each state (Figure 3). The 250 ppsm threshold, a less stringent threshold than the 500 ppsm used in the 2014 New York statewide assessment, was deemed necessary to select streets contained within unincorporated communities and population concentrations in rural and suburban areas where inventories did not exist, but street trees could be expected to be found.
January 2017
| 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
