In Metropolitan-Council/planting.shade: Tree planting and preservation prioritization tool

Growing Shade developed out of a collaboration between the Metropolitan Council, The Nature Conservancy, and Tree Trust. We thank members of the advisory group for initial consultations and thank all individuals who provided feedback during the development phase of this project.

Methods and data sources for the analyses presented within the Growing Shade are detailed below. Please contact us if you have questions or feedback.

Prioritization layer

Priority variables were sourced from several locations including:

• Demographic and socioeconomic information comes from the 2020 decennial census, American Community Survey 5-Year Summary File (2017-2021), and the Equity Considerations data published by the Metropolitan Council. Census block group data is used.
• Health information comes from the PLACES dataset published by the Center of Disease Control and Prevention (2023). PLACES data is reported by census tract.
• Tree canopy and green space information is obtained from the Copernicus Sentinel-2 satellite mission.
• Climate data including temperatures and flood risk was processed using the Climate Vulnerability Assessment published by the Metropolitan Council.

Priority variables were standardized and scaled so that the z-score was normally distributed on a 0-10 scale (by multiplying the normal distribution of the z-score for each variable by 10).

Based on user-defined selection of priority variables, standardized scores are averaged to create a single, integrated priority value.

Tree canopy

Growing Shade uses and shows a tree canopy layer from 2021. A machine learning method was created in Google Earth Engine and used to detect tree cover from other land cover types using Sentinel-2 satellite imagery. Any areas identified as open water or cultivated cropland were removed.

Next, the tree canopy as identified with Sentinel-2 data was calibrated to the tree canopy identified in 2015 using LiDAR data from 2011 (Twin Cities Metropolitan area 1-meter land cover classification). With 1000 equal-area regions across the 7-county area, a scaling factor of 0.885 was used to bring the Sentinel data in line with on-the-ground tree canopy. This scaling factor is appropriate for our methods of using 10 m x 10 m resolution data, which is often larger than tree canopies. This scaling factor makes our data align very closely with other reports (r^2 = 0.96) while still leveraging the scalability and temporal accuracy of our method.

# and processed in Google Earth Engine 
# - Census tract average greenness was calculated from Sentinel-2 satellite data processed on Google Earth Engine. Briefly, the Normalized Difference Vegetation Index (NDVI) was used as a measure of greenness. A composite image of the year 2020 was made where each pixel contained the maximum NDVI observed within the calendar year. Sentinel-2 collects measurements approximately 2-3 times a week, with a pixel resolution of 10 meters x 10 meters. Then, the tract-average NDVI value from this 'maximum NDVI' composite image was taken. NDVI over water bodies (rivers or lakes) was not included.



# Plant phenological patterns were leveraged in order to identify tree canopy. Five distinct phenological time periods were used. For each time period, a composite image was made showing the maximum NDVI observed. Then different thresholds of NDVI were used to separate trees from grasses and crops. 

# - Winter (1 January 2020 - 15 March 2020): pixel classified as a conifer tree if winter NDVI is above 0.3 (identify trees which are green in the winter) OR
# - Spring (15 March 2020 - 30 April 2020): pixel classified as a deciduous tree if spring NDVI is less than 0.5 (remove cool season grass) AND
# - Early summer (1 May 2020 - 15 June 2020): early summer NDVI is greater than 0.55 (remove warm season crops) AND
# - Summer (1 July 2020 - 15 September 2020): summer NDVI is greater than 0.55 (identify trees which are are green in the summer) AND
# - Fall (15 September 2020 - 30 October 2020): fall NDVI is greater than 0.4 (remove early senescing crops)
# 
# <br>
# 
# 
# 
# https://browser.creodias.eu/#lat=45.15999&lng=-92.79540&zoom=15&time=2020-07-05&preset=3_NDVI&datasource=Sentinel-2%20L1C
# <h2><span style='font-size:16pt'>Data dictionary</span></h2>
# 
# **Demographics**<br>
# *Age, % age 65 or older*<br>
# *Age, % under age 18*<br>
# *Age, % under age 18 or 65+*<br>
# *Disability, % any disability*<br>
# *Housing unit density (units / acre)*<br>
# *Population density (persons / acre)*<br>
# *Race, % Asian*<br>
# *Race, % Black or African American*<br>
# *Race, % Hispanic or Latino*<br>
# *Race, % Indigenous*<br>
# *Race, % Multiracial or other*<br>
# *Race, % people of color*<br>
# &nbsp; &nbsp; &bullet; <a href = "https://www.itreetools.org/tools/which-tool-should-i-use" target = "_blank">iTree tools</a> to quantify the benefits and values of trees, including tools for individuals such as homeowners concerned with individual or small amounts of trees<br>

Metropolitan-Council/planting.shade documentation built on Feb. 25, 2024, 3:15 a.m.