TreePoint, a database of urban trees in the Eastern and Midwestern United States

Abstract Many municipalities maintain urban tree databases that describe individual tree identity, size, and coordinates but these data lack interoperability and are not conducive to national analyses. These tree inventories could be combined with high spatial resolution remote sensing data or other spatial datasets to answer research questions in several fields. However, these urban tree censuses are scattered, have disjointed collection protocols and file structures, and are not always easily available to the public. We systematically searched for and compiled municipal tree inventories into a single tree database with consistent formatting and quality control measures for 643 cities in the eastern and midwestern United States. The resulting database contains 7,125,351 individual trees of 1,665 species and 515 genera. This publicly available database can be used to answer questions relevant to tree management, ecology, and evolution across broad spatial and climatic scales. Files Description Main database of street trees: treepoint.csv The treepoint.csv file contains information for all the street trees in the database. Each record is an individual tree. At minimum, the record contains the tree's unique identifier (TREE_ID), botanical name (GENUS and SPECIES), diameter at breast height measured in centimeters (DBH_CM), location in decimal degrees aligned to the the WGS 84 datum, EPSG: 4326 (LON and LAT), and source municipality (CITY and STATE). Optional fields that are included when available are tree height measured in meters (HEIGHT_M), condition on a five-point scale (CONDITION), and inventory year (INV_YR). treepoint.parquet The same database is also available as a parquet file. treepoint_sample.csv The treepoint_sample.csv is a lightweight file that contains up to 25 trees from each city. Each record is an individual tree and contains the same information as the full database. Individual tree inventories: preprocess.zip The preprocess.zip folder is a collection of each muncipality's raw inventory files before processing, cleaning, and any quality control measures. Each file in the folder is the preprocessed inventory of a single municipality named by the following convention [city]_[state_abbreviation]_pre.csv. (For example: the Ann Arbor, Michigan file is named annarbor_mi_pre.csv). Users may attempt to link the cleaned files with the preprocessed files by matching the LOCAL_ID column with the corresponding local id column in individual file. Column names for each municipality inventory can be found in the inventoryCols.csv supplemental file. postprocess.zip The postprocess.zip folder is a collection of the cleaned tree inventory from each municipality, making it identical to the records in the main TreePoint database. Each file in the folder is the cleaned inventory of a single municipality named by the following convention [city]_[state_abbrevation]_post.csv. (For example: the Ann Arbor, Michigan file is named annarbor_mi_post.csv). Users interested in analyzing select cities may index files of interest through the municpality.csv metadata file instead of downloading the entire TreePoint database. Key metadata files: municipality.csv The municipality.csv file contains information for individual tree inventories that were compiled to create TreePoint. Each record reports a tree inventory for a municipality. The record contains the name of the muncipality (CITY and STATE), the filename of the inventory which can be matched with the pre-processed (FILENAME_PRE) and cleaned files (FILENAME_POST), the type of inventory (INVTYPE), a Census Bureau unique geographic identifier (UCGID), the number of trees in the municipality (TREES), the year the inventory was last updated (UPDATE_YR), and the year the inventory was accessed (DOWNLOAD_YR). Cities with less than 100 trees were combined to a post processed file named [state]_[state_abbreviation]_post.csv. When the update date is not available, the year of download was noted. The file also notes summary statistics for location validation. The (NDVI_01 and NDVI_00) columns note the number of trees located at an NDVI less than 0.1 and 0.0, respectively, when compared to regional NAIP imagery (2019-2025). The (TTEST) column reports test statistic from the Welch's t-test when comparing the mean NDVI of reported tree points to a random offset upload to 10 meters away. Location validation is not available for cities with less than 1,000 trees or more than 196,000 trees. Finally, the source for the tree inventories is reported in (SOURCE). treepoint_dataDictionary.csv The treepoint_dataDictionary.csv file contains information regarding each column in the TreePoint database. The data dictionary includes a description of each column and the type of data found in each one. municipality_dataDictionary.csv The municipality_dataDictionary.csv file contains informaiton regarding each column in the municipality.csv file. The data dictionary includes a description of each column and the type of data found in each one. Supplemental information: inventoryCols.csv The inventoryCols.csv file contains information about the file structure for individual tree inventories within TreePoint. Each record is a municipality's tree inventory. The record contains the names of each relevant column for processing and quality control. Column names in individual inventories may be similar to those in the TreePoint database but users should take caution as postprocessed files were significantly cleaned and processed compared to raw inventory files. The column name indicated by ID (LOCAL_ID in TreePoint) is the only column not altered significantly in the processing workflow. Column names with the _rep suffix indicate added columns during processing which do not exist in the raw inventory. Processing R scripts are available at the following Github repository: https://github.com/russellkwong/street-tree-database