US EPA Toxic Release Inventory (TRI)

This article demonstrates a compact workflow for EPA TRI facility emissions data.

This vignette runs its live workflow when rendered locally. The heavy download, processing, extraction, and plotting chunks are skipped automatically on CI, CRAN checks, and pkgdown builds; set AMADEUS_RUN_VIGNETTES=true to force live execution in those environments.

Available inputs and data availability

download_tri() works with EPA’s annual TRI basic data files and exposes a small set of high-value selectors:

jurisdiction supports the nationwide file ("US"), any two-letter state or territory code such as "NC", or the tribal file ("tbl").
year accepts a single year or a start/end pair, so multi-year requests download one annual TRI file per year.
TRI downloads are delivered directly as CSV files rather than zip archives.
Output names reflect the jurisdiction requested: U.S.-wide files keep the historical tri_raw_<year>.csv pattern, while state and tribal requests append a jurisdiction suffix such as _NC or _tbl.
TRI does not require authentication. Because these are annual facility-reported releases and waste-management totals, temporal resolution is yearly.

Download representative requests

directory_to_save <- file.path(tempdir(), "tri_workflow")
download_data(
  dataset_name = "tri",
  year = 2023L,
  jurisdiction = "US",
  directory_to_save = directory_to_save,
  acknowledgement = TRUE
)

Demonstate processing and covariate calculation for a single chemical

The helper function get_tri_info() lists the available chemicals in the downloaded files, which can be used to filter the processing and covariate calculation steps. Here we demonstrate with Polychlorinated biphenyls (PCBs), a group of persistent organic pollutants that were widely used in industrial applications.


chems <- get_tri_info(path = directory_to_save, type = "chemicals")

processed_pcb <- process_covariates(
  covariate = "tri",
  path = directory_to_save,
  chemical = c("Polychlorinated biphenyls"),
  year = 2023
)
# Note that extent is an option in process_covariates() to limit the domain

Calculate covariates at points

domain_x <- c(terra::xmin(processed_pcb), terra::xmax(processed_pcb))
domain_y <- c(terra::ymin(processed_pcb), terra::ymax(processed_pcb))
domain_dx <- diff(domain_x)
domain_dy <- diff(domain_y)

candidate_xy <- expand.grid(
  lon = seq(domain_x[1] + 0.12 * domain_dx, domain_x[2] - 0.12 * domain_dx, length.out = 5),
  lat = seq(domain_y[1] + 0.12 * domain_dy, domain_y[2] - 0.12 * domain_dy, length.out = 5)
)
example_points_sf <- sf::st_as_sf(
  candidate_xy,
  coords = c("lon", "lat"),
  crs = 4326
)
example_points_sf$site_id <- paste0("site_", seq_len(nrow(example_points_sf)))


point_values_pcb <- calculate_covariates(
  covariate = "tri",
  from = processed_pcb,
  locs = example_points_sf,
  locs_id = "site_id",
  decay_range = 5000, # 5 km decay range for illustrative purposes
  use_threshold = FALSE,
  geom = "sf"
)

Plot the covariates at points along with the facility locations

pcb_sf <- sf::st_as_sf(processed_pcb)
point_basemap <-
  sf::st_as_sf(maps::map("usa", plot = FALSE, fill = FALSE))

ggplot() + 
  geom_sf(data = point_basemap, fill = "gray80", color = "white") +
  geom_sf(data = pcb_sf, color = "red", size = 2) + 
  geom_sf(data = point_values_pcb, aes(color = STACK_AIR_0001336363_05000), size = 3)

Demonstate with multiple chemicals and a total emissions



processed_chems <- process_covariates(
  covariate = "tri",
  path = directory_to_save,
  chemical = c("Polychlorinated biphenyls", "Vinyl Chloride", "Tetrachloroethylene"),
  year = 2023,
  variables = "ON-SITE RELEASE TOTAL"
)

point_values_chems <- calculate_covariates(
  covariate = "tri",
  from = processed_chems,
  locs = example_points_sf,
  locs_id = "site_id",
  decay_range = 5000, # 5 km decay range for illustrative purposes
  use_threshold = FALSE,
  geom = "sf"
)