Tree Canopy Monitoring:

Protocol and Monitoring from 2000-2020

February 2022
Tree Canopy Monitoring: Protocol and Monitoring from 2000-2020
February 2022

Prepared by Portland Parks & Recreation Urban Forestry Staff

Project Staff
Angie DiSalvo, Science, Outreach, and Planting Manager
Jeff Ramsey, Science and Policy Specialist
Nick Rossmiller, Botanic Specialist II - Forestry

Cover Photo: A tulip tree shades Alberta Park on a summer afternoon

Portland Parks & Recreation

1120 SW Fifth Avenue, Suite 1302
Portland, Oregon 97204
(503) 823-PLAY Carmen Rubio, Commissioner
www.PortlandParks.org Adena Long, Director
 Contents

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Monitoring Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Appendix A: Canopy Monitoring Protocol . . . . . . . . . . . . . . . . . . 17

 Summary

To monitor trends in Portland’s urban forest canopy cover, Portland

Parks & Recreation established a protocol for measuring canopy
change using point interpretation of aerial photos. Tree canopy cover
was measured in 2000, 2005, 2010, 2015, and 2020 citywide and in
commercial, industrial, open space and residential zoning classes.

20-year findings

• From 2000 to 2020, statistically significant increases in tree canopy

cover were found citywide and in commercial, industrial, and
residential zones

• Increases over the 20-year period represent an addition of 2,289 acres

of tree canopy

2015-2020 reporting period

• Tree canopy covers 29.8% of Portland’s area in 2020, down from

30.7% in 2015

• Tree canopy loss occurred in all zoning classes between 2015–2020,

totaling 823 acres. Losses during this period are as follows:

○ Commercial: 13.3% to 13.0% (-15 acres)

○ Industrial: 9.5% to 8.7% (-164 acres)
○ Open space: 54.9% to 54.2% (-121 acres)
○ Residential: 34.0% to 32.9% (-523 acres)

• Tree canopy losses found across all zoning classes and citywide during
this study period are within the margin of error, and not statistically
significant

• Future updates of this study will show whether losses found during
this period continue, representing a reversal in the long-term trend
of tree canopy expansion in Portland. The next measurement will be
taken in 2025.

Portland Parks & Recreation 1

2 Tree Canopy Monitoring: Protocol and Monitoring from 2000-2020
 Introduction

Tree canopy cover is identified as an important measure of urban forest

health by the City of Portland. Tree canopy cover is a measure of
Portland Parks & Recreation bureau-wide performance and is also cited
as an important indicator in the Portland Urban Forest Management Plan
(2004), Urban Forest Action Plan (2007), the Climate Action Plan (2015), and
the 2035 Comprehensive Plan (2016). Monitoring Portland’s tree canopy is
important in order to understand how canopy coverage may be changing,
and understanding these trends will allow managers to make important
decisions regarding management strategies.

Tree canopy cover has been measured in a variety of ways within the city
of Portland. Past studies have varied in methodology and time frame, and
citywide canopy estimates from 1972 to 2014 range from 25% – 31%
(Metro 2008 and 2016, Nowak & Greenfield 2012, Poracsky & Lackner
2004, PP&R 2007). These studies have provided important estimates
of tree canopy cover, but differences in methodology preclude direct
comparison of results for the purpose of detecting change.

Accurately detecting change requires establishing and using a replicable

protocol with a low error rate. Canopy change, especially growth, occurs
slowly and in order to detect a change, the same method must be used
over a period of time long enough for change to be evident. A successful
monitoring protocol will use the same type and resolution of imagery,
minimize and measure error, set thresholds for determining whether
or not change has occurred, define a statistical method for comparing
results, and be repeated at a regular time interval. This is vital to ensure
that change reported is due to actual change, and is not a result of
measurements being taken using different methods. If weighing tree
canopy measurements against targets, progress towards targets will be
measured using the same protocol.

To monitor trends in Portland’s urban forest canopy, PP&R established

a protocol for measuring tree canopy change according to the guidelines
above, using point interpretation of aerial photos across four zoning
classes and citywide, over five-year time increments. This report
documents the adopted protocol and reports results for the study period
from 2000-2020.

Portland Parks & Recreation 3

4 Tree Canopy Monitoring: Protocol and Monitoring from 2000-2020
 Monitoring Protocol

CHOOSING A METHODOLOGY

The goal of this canopy monitoring protocol is to determine how canopy

is distributed among land use classes and citywide, and to determine
how canopy is changing over time. Available methods for quantifying
canopy were evaluated for their ability to answer these questions,
including classification of remotely sensed data, ground sampling, and
point interpretation of aerial photos. The benefits and drawbacks of each
method were carefully weighed using the guidelines below.
• Canopy change methodology requirements:
• Low error rate
• Use imagery and technology that will continue to be available in
future years
• Cost effective
• Replicable
• Peer reviewed with a recognized protocol
• Ability to subject results to quality assurance testing
• Ability to determine canopy cover for pre-defined strata and citywide
• Produce results that can be statistically compared for significance

Point interpretation of aerial imagery was selected, as it best met the

above requirements. The primary drawback of point interpretation is
the inability to produce cover maps. Point interpretation also cannot
analyze canopy by categories not established at the beginning of the
study (for example, neighborhood boundaries), as each strata requires a
large number of sample points. However, the key goal of this project was
to monitor canopy in predetermined strata and citywide, and cover maps
and additional analysis are not required for this effort.

DEFINING STRATA

Recognizing that the city has different land use areas with varying
characteristics and goals, strata were determined according to zoning
classifications. Zoning classes are good proxies for the city’s different land
use types and best represent development intensity. Zoning classes also
have some connection to the Urban Land Environments outlined in the
2004 Portland Urban Forest Management Plan. Four strata were established
corresponding to zoning code: commercial, industrial, open space, and
residential (Table 1). All areas within the city’s boundary were assigned to
one of the zoning classes.

Portland Parks & Recreation 5

Monitoring Protocol

Table 1: Zoning Class Descriptions

Zoning Class Zoning Code Zoning Class Description Acres % of City

CO1, CN1, CO2, Storefronts, neighborhood and

Commercial CN2, CG, CS, CM, office commercial areas, and mixed 6,237 6.7%
CX residential commercial areas

Manufacturing and warehousing

EG1, EG2, IG2,
Industrial areas, industrial and wholesales 21,507 23.2%
IG1, IH, EX
sales, and industrial parks

Natural areas, developed parks, and

Open Space OS 16,819 18.1%
schools

RF, R20, R10, R7,

Single and multifamily residential
Residential R5, R3, R2.5, R2, 48,149 51.9%
homes
R1, RH, RX, IR
92,712 100.0%

Note that some zoning boundaries and designations were updated in 2018 (Bureau of Planning and
Sustainability 2018). Table 1 reflects zoning classes prior to this update. For the purposes of the 2015 –
2020 reporting period, the prior zoning designations were kept in place. Future updates of this report
will reflect the new zoning map.

APPLYING THE MONITORING PROTOCOL

PP&R contracted with Davey Resource Group, an experienced urban forestry consultant agency, to
assist in establishing a protocol in 2012. The complete monitoring protocol is described in Appendix A.

Point interpretation was conducted by first establishing randomly located points across each zoning
class. To keep standard error low, a minimum of 1,000 points were used for each zoning class for a total
of 4,521 points. High resolution imagery was available back to 2000, thus years 2000, 2005, and 2010
became the first study years, and Davey Resource Group conducted point interpretation for these years.
Subsequent point interpretation was conducted by Urban Forestry staff for years 2015 and 2020.

For each study year, points were laid in the same geographic location on aerial images and a trained
photo interpreter examined the points to determine whether the points coincided with tree canopy or
not. To ensure that the photo interpretation process was completed with the highest degree of accuracy,
a second photo interpreter performed quality assurance inspections on 10% of the work performed to
verify the interpretations, with a 95% agreement threshold.

A percent tree canopy cover was determined for each zoning class, and the number of acres of tree
canopy was calculated by multiplying the percentage of tree canopy by the total acres within the zoning
class. Citywide tree canopy levels and acreages were calculated as weighted averages of the zoning
classes. Standard error and 95% confidence intervals were calculated, and change over time was tested
for significant difference using a Chi-squared test (McNemar’s test) and significant differences were
found if p < 0.05.

6 Tree Canopy Monitoring: Protocol and Monitoring from 2000-2020

 Findings

TREE CANOPY COVER AND ACRES OF CANOPY

Overall tree canopy cover ranged from 27.3% in 2000 to 30.7% in 2015,
falling to 29.8% in 2020 (Table 2). Total tree canopy acres found in the
city ranged from 25,348 in 2000 to 28,460 in 2015, falling to 27,637 in
2020. Tree canopy cover was found to be unevenly distributed among the
four zoning classes (Figure 1).

Table 2: Percent tree canopy cover and acres of tree canopy from 2000–2020

2000 2005 2010 2015 2020

Tree Canopy Tree Canopy Tree Canopy Tree Canopy Tree Canopy
Cover Cover Cover Cover Cover

Zoning Class Percent Acres Percent Acres Percent Acres Percent Acres Percent Acres

Commercial 9.1 571 11.0 687 12.5 781 13.3 827 13.0 812
Industrial 6.4 1,374 6.8 1,467 7.9 1,690 9.5 2,043 8.7 1,879
Open Space 53.9 9,057 54.6 9,182 55.3 9,306 54.9 9,239 54.2 9,118
Residential 29.8 14,345 30.4 14,629 33.1 15,955 34.0 16,351 32.9 15,827

City Total 27.3 25,348 28.0 25,965 29.9 27,732 30.7 28,460 29.8 27,637

Portland Parks & Recreation 7

Findings

Figure 1: Acres of canopy in zone classes from 2000–2020. Error bars represent 95% confidence
intervals.

20,000 

18,000 

16,000 

14,000 

12,000 
ACRES

10,000 

8,000 

6,000 

4,000 

2,000 

‐
Commercial Industrial Open Space Residential
Year 2000 571 1,374 9,057 14,345
Year 2005 681 1,467 9,182 14,629
Year 2010 775 1,690 9,306 15,955
Year 2015 827 2,043 9,239 16,351
Year 2020 812 1,879 9,118 15,827

8 Tree Canopy Monitoring: Protocol and Monitoring from 2000-2020

 Findings

In commercial zones, tree canopy rose from 571 acres to 827 acres
between 2000-2015, falling to 812 acres in 2020. This represents total
tree canopy cover ranging from 9.1% in 2000 to 13.3% in 2015, falling to
13.0% in 2020. Commercially zoned lands contain approximately 3% of
the city’s total tree canopy.

In industrial zones, tree canopy rose from 1,374 acres to 2,043 acres
between 2000-2015, falling to 1,879 acres in 2020. This represents total
canopy cover ranging from 6.4% in 2000 to 9.5% in 2015, falling to
8.7% in 2020 – the lowest of any zoning class. Industrial zoned lands
contain approximately 7% of the city’s total tree canopy.

In the open space zone, tree canopy rose from 9,057 acres to 9,306 acres
between 2000-2010, falling between 2010-2020, to 9,118 acres. Open
space zones have the highest rate of canopy cover found in any zoning
class, at 54.2% in 2020, down from 55.3% in 2010. Land zoned open
space contains approximately 33% of the city’s tree canopy.

In residential zones, which make up the largest portion of the city’s land
base, tree canopy rose from 14,345 acres to 16,351 acres between 2000-
2015, falling to 15,827 acres in 2020. This represents total tree canopy
cover ranging from 29.8% in 2000 to 34.0% in 2015, falling to 32.9%
in 2020. Residential zones contain the majority (57%) of the city’s tree
canopy.

CHANGE OVER TIME

From 2000 to 2005, increases in tree canopy cover were found citywide
and in all zoning classes (Table 3), however only changes in commercial
zones were statistically significant (McNemar’s test, p < 0.05), where tree
canopy cover rose from 9.1% to 11.0% during the time period.

Table 3: Change in tree canopy cover in five-year intervals, 2000 – 2020

2000-2005 2005-2010 2010-2015 2015-2020

Percent Change Percent Change Percent Change Percent Change
Zoning Class
Change in Acres Change in Acres Change in Acres Change in Acres

Commercial +1.9* +116* +1.5* +94* +0.7 +46 -0.3 -15

Industrial +0.4 +93 +1.0* +223* +1.6* +353* -0.8 -164
Open Space +0.7 +124 +0.7 +124 -0.4 -66 -0.7 -121
Residential +0.6 +284 +2.8* +1,326* +0.8 +395 -1.1 -523

City Total +0.7 +617 +1.9* +1,767* +0.8 +728 -0.9 -823
* Change significantly different with p < 0.05 (McNemar’s test)

Portland Parks & Recreation 9

Findings

From 2005 to 2010, significant increases in tree canopy cover were found citywide and in all zoning
classes except open space. Citywide tree canopy cover increased by 1.9%, commercial by 1.5%,
industrial by 1.0%, and residential by 2.8%.
From 2010 to 2015, increases in tree canopy were found in all zoning classes except open space, however
the only significant change was seen in industrial zones, which increased by 1.6%, from 7.9% to 9.5%
tree canopy cover.
In the latest reporting period, from 2015-2020, decreases in tree canopy cover were found citywide, and
across all zoning classes. None, however, were statistically significant.
Over the twenty-year period, from 2000 to 2020, tree canopy cover increased significantly citywide and
in all zoning classes with the exception of open space (Table 4). Citywide tree canopy cover increased
by 2.5%, commercial by 3.9%, industrial by 2.3%, and residential by 3.1%. Citywide, these increases
represent an estimated addition of 2,289 acres of tree canopy.

Table 4: Change in canopy cover from 2000 to 2020

Zoning Class Percent Change Change in Acres

Commercial +3.9* +241*

Industrial +2.3* +505*
Open Space +0.3 +61
Residential +3.1* +1,482*

City Total +2.5* +2,289*

* Change significantly different with p < 0.05 (McNemar’s test)

10 Tree Canopy Monitoring: Protocol and Monitoring from 2000-2020

 Discussion

CANOPY DISTRIBUTION AND TRENDS

From 2000-2020, tree canopy cover significantly increased citywide and

across commercial, industrial, and residential zones. For the first 10 years
of the study period, tree canopy cover increased significantly citywide
and across all zones, with tree canopy in commercial, industrial, and
residential zones continuing to grow between 2010-2015. In the latest
reporting period, 2015-2020, some of these gains were lost as tree canopy
cover declined across all zones, although not at statistically significant
levels.

Twenty years is a relatively short period of time, but clear trends are
revealing changes in the urban forest. While Portland’s tree canopy cover
is greater today than 20 years ago, canopy growth has slowed or reversed
in the past 5 years. Interestingly, this time period coincides with the 2015
adoption of improved regulations for tree preservation and removal with
Portland’s tree code, Title 11. Despite these new rules resulting in the
preservation of thousands more trees than would have been preserved
in previous years, the city was still not able to realize the gains in tree
canopy cover seen in the early periods of this study.

Tree canopy cover varies greatly between zoning classes, reflecting land
use and development intensity. Land use also explains the likelihood for
changes in tree canopy cover during each reporting period. For example,
the open space zone has the highest level of tree canopy cover, but
changes over the study period were statistically insignificant. Open space
lands include natural areas and developed parks, where land use patterns
are largely set. In contrast, residential, commercial, and industrial zones
are more likely to undergo development and, with less overall tree
canopy, may have more opportunities for planting new trees, making
them more likely to experience changes in canopy cover over short 5-year
increments.

COMPARISON TO CANOPY COVER TARGETS

PP&R’s 2004 Portland Urban Forest Management Plan (UFMP) set

aspirational tree canopy cover targets for Urban Land Environments
(ULEs) (Table 5). Targets were established by reviewing
recommendations for tree canopy cover in scientific literature. ULEs
were derived from Metro’s Regional Land Information System, and have
some connection to the zoning code categories used in this study. Note
that ULEs are now outdated and may include up to 20% classification
error (PP&R 2009). The two ULEs that correspond best with zoning
categories are the residential ULE and the commercial/ industrial/
institutional ULE. The UFMP recommends targets of 35-40% tree

Portland Parks & Recreation 11

 Table 5: Existing canopy cover targets within the City of Portland
Canopy goals in PP&R
Canopy cover targets in UFMP
Category Canopy Report (2007) and
(2004)
Climate Action Plan (2015)
Residential ULE 35-40% n/a
Commercial/Industrial/Institutional
15% n/a
ULE
Natural Areas and Stream
Targets set by City Framework Plan n/a
Corridors ULE

Transportation Corridors and

35% n/a
Rights of Way ULE
Developed Parks and Open
30% n/a
Spaces ULE
Citywide No target set 33.3%

canopy cover for the residential ULE and 15% for the commercial/industrial/institutional ULE. In
2020, tree canopy cover did not meet these goals, having declined across every zone: in the residential
zone tree canopy cover is 32.9%, the commercial zone is 13.1%, and the industrial zone is 8.7%.

PP&R’s Canopy Report (2007) and City of Portland’s Climate Action Plan (2009 and 2015) set a goal
of expanding urban forest canopy to cover one-third of the city’s area. The 33.3% citywide goal was
established from tree canopy cover data produced using a different method than that used in this report,
and direct comparison of results is not recommended.

This report provides baseline data that can be used to establish and refine canopy targets. The Portland
Urban Forest Management Plan is scheduled to be updated in coming years, providing an excellent
opportunity for revising tree canopy goals, using new information on areas of potential growth to set
realistic targets (PP&R 2018). Well-developed tree canopy targets will provide the opportunity to make
deliberate and clear decisions for planning and goal setting for the future of the urban forest.

In addition to a citywide tree canopy goal, goals for each zoning class are recommended due to the
fundamental differences between zones in land use characteristics, existing tree canopy, and capacity to
accommodate tree canopy in the future. Zone class targets will assist managers in developing effective
strategies for increasing tree canopy, and may also assist the City in reaching its other tree goals, such as
more equitable distribution of trees.

OPPORTUNITIES FOR FUTURE STUDY

Establishing and applying a monitoring protocol has been an important step in a long-term commitment
to tracking tree canopy trends. The protocol outlined in this study will continue to serve as guide for
PP&R in future years and the next tree canopy measurement will occur using 2025 aerial images.

This monitoring study reports trends in tree canopy cover, but does not provide information on why
changes are occurring. Canopy increases may be attributed to growth of existing trees and planting
of new trees. Tree removal for development, tree loss from pests and diseases, natural mortality, and
weather events may negatively affect tree canopy cover. Examination of the reasons behind tree canopy

12 Tree Canopy Monitoring: Protocol and Monitoring from 2000-2020

trends requires additional study and would allow for more informed
strategies for meeting tree canopy goals.

Additionally, this study does not provide information on tree canopy

cover levels or change in areal units other than zoning classes or citywide.
Other boundaries of interest may be useful, such as at the neighborhood
level or across private versus public property. Maps of Portland’s tree
canopy have been developed for the years 2014 and 2019 (Metro 2016
and 2022) providing an opportunity to pursue these questions. While
not strictly comparable to data presented in this report, these maps of
Portland’s tree canopy are complementary and will aid understanding of
recent changes in the distribution of tree canopy in Portland.

Portland Parks & Recreation 13

14 Tree Canopy Monitoring: Protocol and Monitoring from 2000-2020
 References

Bureau of Planning and Sustainability. 2018. City of Portland Zoning

App. https://www.portlandmaps.com/bps/zoning/#/map/

Bureau of Planning and Sustainability. 2016. 2035 Comprehensive Plan.

https://www.portland.gov/bps/comp-plan/2035-comprehensive-plan-
and-supporting-documents

Bureau of Planning and Sustainability. 2015. Climate Action Plan.

https://www.portland.gov/bps/climate-action/history-and-key-documents

Lindren, B.W. and G.W. McElrath. 1969. Introduction to Probability

and Statistics. Macmillan, London.

Metro. 2008. State of the Watersheds Monitoring Report. http://www.

oregonmetro.gov/index.cfm/go/by.web/id=27579

Metro. 2016. Canopy 2014. http://rlisdiscovery.oregonmetro.gov/?action

=viewDetail&layerID=3552

Metro. 2022. Canopy 2019. Still pending at time of publishing.

Nowak, D. J. and E.J. Greenfield. 2012. Tree and impervious cover

change in US cities. Urban Forestry & Urban Greening. 11: 21-30.

Poracsky, J. and M. Lackner. 2004. Urban Forest Canopy Cover in

Portland, Oregon: 1972 - 2002. Final Project Report prepared for
Portland General Electric and City of Portland Urban Forestry
Commission.

Portland Parks & Recreation. 2004. Portland Urban Forestry

Management Plan. https://www.portland.gov/trees/urban-forest-
management-plan

Portland Parks & Recreation. 2007. Portland’s Urban Forest Canopy:

Assessment and Public Tree Evaluation. https://www.portland.gov/sites/
default/files/2020-08/urban-canopy-report-10-07.pdf

Portland Parks & Recreation. 2007. Urban Forest Action Plan. https://
https://www.portland.gov/sites/default/files/2020/urban-forest-action-
plan-02-07.pdf

Portland Parks & Recreation. 2009. A Standardized Methodology to

Track Urban Forestry Canopy Cover Change. White paper

Portland Parks & Recreation. 2018. Tree Canopy and Potential in

Portland, Oregon. https://www.portland.gov/sites/default/files/2020-08/
tree-canopy-and-potential-2018.pdf

Portland Parks & Recreation 15

 Sokol, R.R. and F.J. Rohlf. 2003. Biometry: The Principles and Practices of Statistics in Biological
Research. WH Freeman and Company, New York, NY.

Thompson, S. K. 2002. Sampling, second edition. John Wiley and Sons, Inc., New York, New York.

16 Tree Canopy Monitoring: Protocol and Monitoring from 2000-2020

 Appendix A: Canopy Monitoring Protocol

Method: Point interpretation of aerial photos

Measurement frequency: 5 years
Image standards: Color digital orthorectified photos at 6” resolution
taken during leaf on season
Strata: Commercial, industrial, open space, and residential according to
zoning code
Points: A minimum of 1,000 randomly selected points are established
within each zoning class. The high sample number is needed to
minimize standard error. A standard error threshold of 2% is established.
If standard error for any zone exceeds 2%, additional sample points
should be established until the standard error threshold is reached.
Future analyses use the same established points.
Interpreting points: Points are interpreted as tree, non-tree, or
unreadable. Unreadable points are removed from the sample.
Photo interpretation guidelines:
• Photo interpreters should have extensive experience interpreting
aerial photography and relating photos to locations on the ground.
Interpreters should have a high degree of confidence that they can
differentiate between trees, lawn, buildings, roads, and other ground
surfaces. This is a strategy to reduce errors that would occur when
the interpreter records a tree when there is no tree, or fails to see the
tree as occupying the point.
• The same photo interpreter should be used throughout the study,
except for quality assurance testing.
• A second photo interpreter performs quality assurance testing on
10% of the data points. A 95% agreement must be reached for the
data interpretation to be considered valid.
• Dead trees are considered “not tree.” Because photos are analyzed
in leaf-on season, trees devoid of leaves are considered dead or “not
tree.”
• Non-tree vegetation (e.g., hedges, low shrubs, green roofs, lawn) is
considered “not tree.”
• Points falling on water are included and are recorded as “not tree.”
• In cases where the point falls on the edge of a tree, the interpreter
will need to zoom in and carefully consider the image. Changes over
time may be due to canopies growing into the location of the point,
and it is important to spend the time to carefully analyze and capture
these borderline changes.

Portland Parks & Recreation 17

Appendix A

• Images that are too difficult to interpret due to large dark shadows from buildings or very large trees
are considered “unreadable” and are excluded from the study.
• Due to the nature of aerial photography, minor displacement occurs due to horizontal and parallax
variation from year to year. To minimize bias, these changes are ignored and each photo is assumed
to be correct. Although this may introduced error in some borderline cases, it is assumed that error
is equally randomly distributed between tree and non-tree points.

DATA ANALYSIS
Zoning class tree canopy cover percentage (p): The number of sample points (N) interpreted as “tree”
divided by the total number of sample points (n) within the zone (p=N/n).
Zoning class tree canopy acres: The percentage of tree canopy cover (p) multiplied by the total acres of
land within that zone.
Citywide tree canopy acreage: The sum of tree canopy acreages in each zone.
Citywide tree canopy cover percentage: The total acres of tree canopy divided by the total acres of land
in the city.
Standard error (SE): √ ((p x (1-p)/n) (Lindren and McElrath 1969)
Confidence interval: A 95% confidence interval is set and is calculated as: SE x 1.96 (Thompson 2002).
Significance testing: For each zone and citywide, McNemar’s test is used to determine whether changes
observed in canopy coverage are statistically significant (Sokal and Rohlf, 2003). McNemar’s test is a
non-parametric method used on nominal data. The test provides a chi-squared value, which is compared
against a p-value for statistical significance. Canopy cover between years and across zones is considered
significantly different if p < 0.05. A weighted total is used to calculate citywide chi-squared using
McNemar’s test. Each number of sample points (N) was multiplied by the portion of the city covered by
each zone to calculate the total.

18 Tree Canopy Monitoring: Protocol and Monitoring from 2000-2020