Environment and hazards

What environmental observations and hazards are recorded?

Find monitoring records, climate summaries, environmental listings and historical disaster data. Read each table’s location, period and measurement method.

Choose a table or county below. View data opens the actual table on a place profile, with its dates, units and source notes.

Choose a data table

Open a county or state example to see the values. Table coverage counts supported profiles, not complete rows or every area. Original publisher links provide background documentation; they are separate from the data displayed here.

Air pollutant emissions inventory, 2023 NEI (EPA)

U.S. Environmental Protection Agency · 2023 National Emissions Inventory (NEI), Tier summaries of criteria air pollutants by state

Data period
2023 National Emissions Inventory, calendar year 2023 (EPA summary files dated 1 September 2026 on gaftp.epa.gov; retrieved 2026-11-03)
Geography vintage
States as coded in the NEI (the 50 states, the District of Columbia and Puerto Rico; other territories are not shown)
Units and scope
Tons per year of each pollutant (short tons, as EPA reports them). A figure below half a ton is shown as <1.
Pages with this table
52
Coverage and interpretation
  • These are inventory estimates of emissions for 2023, not measured air quality. The National Emissions Inventory is EPA’s compilation of how much of each pollutant sources are estimated to release in a year, built from state, local and tribal submissions, EPA models (for example the MOVES model for vehicles) and activity data. It says nothing about the concentration of any pollutant in the air a person breathes, which depends on weather, transport and chemistry, and it is not exposure, risk or a health effect. The air quality index shown elsewhere on county pages comes from monitors and is a different measure for a different year.
  • Emissions are in tons per year for calendar year 2023, the newest National Emissions Inventory EPA has published (summary files posted by EPA between May and September 2026). EPA states that the accuracy of individual estimates varies from county to county and that the inventory meets user needs largely in the aggregate. EPA also warns that methods change between inventory years, so these figures are not compared here with any earlier inventory year and should not be read as a trend.
  • The rows are EPA’s own Tier 1 source categories, shown as EPA names them; they are not the point, nonpoint, on-road and non-road data categories. Highway Vehicles are on-road mobile sources and Off-Highway are non-road mobile sources. Natural Resources is EPA’s biogenic emissions (vegetation and soils) and is large for volatile organic compounds. Miscellaneous holds agriculture and forestry, other combustion, fugitive dust and other sources as EPA groups them; the inventory counts fires as county totals, and which category holds them is EPA’s grouping.
  • The pollutants are carbon monoxide, nitrogen oxides, ammonia, sulfur dioxide, volatile organic compounds and EPA’s primary particulate matter for PM2.5 and PM10 (filterable plus condensable). Lead (reported in pounds), hazardous air pollutants and the elemental and organic carbon components in the same EPA files are not shown. “All categories” adds the Tier 1 rows the file reports for the pollutant, including Natural Resources.
  • A cell shows Not reported where the file has no row for that category and pollutant; that is the absence of a figure, not zero tons. A figure below half a ton is shown as <1 with its value kept in the source field. Whole tons are rounded half up from EPA’s values, which this site adds from EPA’s Tier 1 rows.
  • Emissions are placed in the county where EPA assigns the source or the model estimate; a county total does not say where the pollution is breathed or who is affected,. Emissions that tribes report for their own lands are filed by EPA under the tribe, not the county, and are not included here; emissions in federal waters are not shown.
  • State figures add every county row EPA files for the state, including rows EPA files under a state-wide county code (777) that belong to no county table, so a state is slightly more than its county tables add to. Where EPA’s own state sheet exists (all but Iowa, Kansas, Missouri and Nebraska, whose regional workbook has none) it agrees with this sum to within a millionth of a ton. Puerto Rico is shown as a separate state-level geography; Guam and the U.S. Virgin Islands are not shown.

Source periods, coverage and other tables · Original publisher: documentation and methods

Air pollutant emissions inventory, 2023 NEI (EPA)

U.S. Environmental Protection Agency · 2023 National Emissions Inventory (NEI), Tier summaries of criteria air pollutants by county

Data period
2023 National Emissions Inventory, calendar year 2023 (EPA summary files dated 1 September 2026 on gaftp.epa.gov; retrieved 2026-11-03)
Geography vintage
Counties as coded in the NEI (five-digit FIPS code); Connecticut by the nine Census planning regions as in the 2023 file
Units and scope
Tons per year of each pollutant (short tons, as EPA reports them). A figure below half a ton is shown as <1.
Pages with this table
3,222
Coverage and interpretation
  • These are inventory estimates of emissions for 2023, not measured air quality. The National Emissions Inventory is EPA’s compilation of how much of each pollutant sources are estimated to release in a year, built from state, local and tribal submissions, EPA models (for example the MOVES model for vehicles) and activity data. It says nothing about the concentration of any pollutant in the air a person breathes, which depends on weather, transport and chemistry, and it is not exposure, risk or a health effect. The air quality index shown elsewhere on county pages comes from monitors and is a different measure for a different year.
  • Emissions are in tons per year for calendar year 2023, the newest National Emissions Inventory EPA has published (summary files posted by EPA between May and September 2026). EPA states that the accuracy of individual estimates varies from county to county and that the inventory meets user needs largely in the aggregate. EPA also warns that methods change between inventory years, so these figures are not compared here with any earlier inventory year and should not be read as a trend.
  • The rows are EPA’s own Tier 1 source categories, shown as EPA names them; they are not the point, nonpoint, on-road and non-road data categories. Highway Vehicles are on-road mobile sources and Off-Highway are non-road mobile sources. Natural Resources is EPA’s biogenic emissions (vegetation and soils) and is large for volatile organic compounds. Miscellaneous holds agriculture and forestry, other combustion, fugitive dust and other sources as EPA groups them; the inventory counts fires as county totals, and which category holds them is EPA’s grouping.
  • The pollutants are carbon monoxide, nitrogen oxides, ammonia, sulfur dioxide, volatile organic compounds and EPA’s primary particulate matter for PM2.5 and PM10 (filterable plus condensable). Lead (reported in pounds), hazardous air pollutants and the elemental and organic carbon components in the same EPA files are not shown. “All categories” adds the Tier 1 rows the file reports for the pollutant, including Natural Resources.
  • A cell shows Not reported where the file has no row for that category and pollutant; that is the absence of a figure, not zero tons. A figure below half a ton is shown as <1 with its value kept in the source field. Whole tons are rounded half up from EPA’s values, which this site adds from EPA’s Tier 1 rows.
  • Emissions are placed in the county where EPA assigns the source or the model estimate; a county total does not say where the pollution is breathed or who is affected,. Emissions that tribes report for their own lands are filed by EPA under the tribe, not the county, and are not included here; emissions in federal waters are not shown.
  • A county has a table only if EPA’s file has emissions for it. Counties are matched to the file by the five-digit FIPS code, never by name. EPA’s 2023 inventory files Connecticut under the nine Census planning regions (09110 to 09190), the same geography this site uses, so Connecticut counties are matched directly; earlier inventory years used the former eight counties and are not mixed in.
  • These are county figures, not figures for any city or place within the county.

Source periods, coverage and other tables · Original publisher: documentation and methods

Air quality index county-days by year, 2022 to 2025 (EPA, state)

U.S. Environmental Protection Agency · Air Quality System, Annual AQI by County, 2022 to 2025

Data period
Calendar years 2022, 2023, 2024 and 2025, one EPA file per year
Geography vintage
States and the District of Columbia by the state name printed in the EPA files, matched exactly to the state list
Units and scope
County-days: one county on one day. The AQI is the EPA index from 0 to 500, where a higher number is more air pollution.
Pages with this table
52
Coverage and interpretation
  • These are monitor-derived summaries of reported days from the EPA Air Quality System, not a measure of the air everyone breathes. A county day takes the largest AQI value among the monitors in the county that reported that day, so the figure describes the monitored air at its most polluted, not the average air across the county, and a county without monitors has no row.
  • The AQI is the EPA index from 0 to 500 and its categories are Good (0 to 50), Moderate (51 to 100), Unhealthy for Sensitive Groups (101 to 150), Unhealthy (151 to 200), Very Unhealthy (201 to 300) and Hazardous (301 and above), as EPA publishes them. A day count is a number of calendar days, not hours, and a category is not a health finding for any person.
  • Days with an air quality reading can be fewer than 365 when monitors were off or reported too little, and the category counts and the main-pollutant counts then each add to that smaller total. Days without a reading are not counted as Good days or as zero. Some days are affected by exceptional events such as wildfire smoke; the files report them as EPA published them, and a daily AQI above 500, which the index scale does not define, is shown as published.
  • Each year is a separate EPA file (2022 to 2025). A county or state that is not in a year’s file shows Not reported for that year, which is not zero and does not mean the air was clean. The monitoring network and the number of monitors differ by year and by county, so years and counties are shown as published and are not combined into an index.
  • State figures are sums of county figures and add up county-days: the number of county days with a reading, in each category or with each main pollutant, summed over every county of the state that is in the EPA file (including counties that have no table on this site, such as Connecticut’s former counties). A county-day is one county on one day, so a state total counts a day once for each reporting county; it is not a number of days in the state. A median is not shown for states because county medians cannot be combined. The maximum is the largest of the county maximum values. Counties in Mexico are not shown, and the Virgin Islands are not on this site.
  • The EPA data are a U.S. government work (public domain, decision 0011); this site is not endorsed by EPA.

Source periods, coverage and other tables · Original publisher: documentation and methods

Air quality index days by year, 2022 to 2025 (EPA, county)

U.S. Environmental Protection Agency · Air Quality System, Annual AQI by County, 2022 to 2025

Data period
Calendar years 2022, 2023, 2024 and 2025, one EPA file per year
Geography vintage
Counties as named by EPA (state and county name), matched exactly to this site’s county list
Units and scope
Number of days in the year; the AQI is the EPA index from 0 to 500, where a higher number is more air pollution.
Pages with this table
1,002
Coverage and interpretation
  • These are monitor-derived summaries of reported days from the EPA Air Quality System, not a measure of the air everyone breathes. A county day takes the largest AQI value among the monitors in the county that reported that day, so the figure describes the monitored air at its most polluted, not the average air across the county, and a county without monitors has no row.
  • The AQI is the EPA index from 0 to 500 and its categories are Good (0 to 50), Moderate (51 to 100), Unhealthy for Sensitive Groups (101 to 150), Unhealthy (151 to 200), Very Unhealthy (201 to 300) and Hazardous (301 and above), as EPA publishes them. A day count is a number of calendar days, not hours, and a category is not a health finding for any person.
  • Days with an air quality reading can be fewer than 365 when monitors were off or reported too little, and the category counts and the main-pollutant counts then each add to that smaller total. Days without a reading are not counted as Good days or as zero. Some days are affected by exceptional events such as wildfire smoke; the files report them as EPA published them, and a daily AQI above 500, which the index scale does not define, is shown as published.
  • Each year is a separate EPA file (2022 to 2025). A county or state that is not in a year’s file shows Not reported for that year, which is not zero and does not mean the air was clean. The monitoring network and the number of monitors differ by year and by county, so years and counties are shown as published and are not combined into an index.
  • County figures, not figures for any city or place within the county. The EPA files carry state and county names and no FIPS codes, so each county is matched to this site’s county list only by identical state and normalized county name; a county whose name is ambiguous or absent has no table, and nothing is matched by guesswork. Connecticut files use the former eight counties, so Connecticut planning regions have no table.
  • The EPA data are a U.S. government work (public domain, decision 0011); this site is not endorsed by EPA.

Source periods, coverage and other tables · Original publisher: documentation and methods

Air quality standard designations: nonattainment and maintenance areas (EPA Green Book)

U.S. Environmental Protection Agency, Green Book · Green Book criteria pollutant nonattainment and maintenance areas, county-level export (NAYRO)

Data period
Designations as posted by EPA in the Green Book (county file exported 2026-10-31, EPA page last updated October 30, 2026; retrieved 2026-11-03)
Geography vintage
States as coded by FIPS in the Green Book (the 50 states, the District of Columbia and Puerto Rico; Guam is not shown)
Units and scope
Designation status, classification and county coverage; counts of areas and counties in the state table.
Pages with this table
52
Coverage and interpretation
  • EPA Green Book designations as posted, not a measure of current air quality. The Green Book records the area designations EPA has made under the Clean Air Act for each National Ambient Air Quality Standard, with the classification EPA assigned. A designation is a regulatory status. It does not report pollutant concentrations, monitor readings or exposure, and it says nothing about the air at any address.
  • Status is read from the Green Book county file: Nonattainment where the file marks the county as in a nonattainment area, and Maintenance where it does not. The file’s county rows carry no separate status column; the reading was confirmed against the Green Book’s area-level file, whose status for every area agrees (a maintenance area is one EPA redesignated to attainment with a maintenance plan). For a multi-state area that is split because some of its states are redesignated and others are not, the file marks the area as nonattainment and the area name carries the “portion” marker.
  • Only standards in effect are shown. The Green Book file also holds rows for two revoked standards (the 1997 8-hour ozone standard, revoked in 2015, and the 1-hour ozone standard of 1979, revoked in 2005); those rows are not shown and each table’s notes count them. A standard EPA lists on its Green Book page but whose designations are not in the county file is not covered.
  • A county with no row for any standard in effect is shown as None listed. The county file lists the nonattainment and maintenance counties for each standard on EPA’s Green Book page, so a county absent from the file is not designated nonattainment for any listed standard; it does not mean the county has no air quality problem under a standard EPA has not designated, and it is not a statement about monitor readings.
  • The file gives no original designation date. The date shown is the effective date of the area’s redesignation, which the file fills only for areas EPA has redesignated (maintenance areas and a few others); a blank is shown as Not reported.
  • Classification is as the file gives it, for example Marginal, Moderate, Serious, Severe or Extreme for ozone. For a maintenance area it is the classification at the time of redesignation. Some standards have no classifications and the cell says so.
  • State figures are counts derived from the county rows: designated areas with at least one county in the state (a multi-state area is counted in each state it touches, so state counts do not add to a national count of areas) and counties with a row for the standard and status. A county that is in nonattainment areas for one standard and a maintenance area for another is counted under each. Counts of counties are not populations and are not weighted.
  • Connecticut’s county counts are Not reported because the Green Book files it by former counties that this site does not use; its area counts are shown.

Source periods, coverage and other tables · Original publisher: documentation and methods

Air quality standard designations: nonattainment and maintenance areas (EPA Green Book)

U.S. Environmental Protection Agency, Green Book · Green Book criteria pollutant nonattainment and maintenance areas, county-level export (NAYRO)

Data period
Designations as posted by EPA in the Green Book (county file exported 2026-10-31, EPA page last updated October 30, 2026; retrieved 2026-11-03)
Geography vintage
Counties as coded by FIPS in the Green Book (Connecticut is filed by former counties and is Not reported)
Units and scope
Designation status, classification and county coverage; counts of areas and counties in the state table.
Pages with this table
3,222
Coverage and interpretation
  • EPA Green Book designations as posted, not a measure of current air quality. The Green Book records the area designations EPA has made under the Clean Air Act for each National Ambient Air Quality Standard, with the classification EPA assigned. A designation is a regulatory status. It does not report pollutant concentrations, monitor readings or exposure, and it says nothing about the air at any address.
  • Status is read from the Green Book county file: Nonattainment where the file marks the county as in a nonattainment area, and Maintenance where it does not. The file’s county rows carry no separate status column; the reading was confirmed against the Green Book’s area-level file, whose status for every area agrees (a maintenance area is one EPA redesignated to attainment with a maintenance plan). For a multi-state area that is split because some of its states are redesignated and others are not, the file marks the area as nonattainment and the area name carries the “portion” marker.
  • Only standards in effect are shown. The Green Book file also holds rows for two revoked standards (the 1997 8-hour ozone standard, revoked in 2015, and the 1-hour ozone standard of 1979, revoked in 2005); those rows are not shown and each table’s notes count them. A standard EPA lists on its Green Book page but whose designations are not in the county file is not covered.
  • A county with no row for any standard in effect is shown as None listed. The county file lists the nonattainment and maintenance counties for each standard on EPA’s Green Book page, so a county absent from the file is not designated nonattainment for any listed standard; it does not mean the county has no air quality problem under a standard EPA has not designated, and it is not a statement about monitor readings.
  • The file gives no original designation date. The date shown is the effective date of the area’s redesignation, which the file fills only for areas EPA has redesignated (maintenance areas and a few others); a blank is shown as Not reported.
  • Classification is as the file gives it, for example Marginal, Moderate, Serious, Severe or Extreme for ozone. For a maintenance area it is the classification at the time of redesignation. Some standards have no classifications and the cell says so.
  • County coverage is the file’s own flag: Whole county where the whole county is inside the area, Part of county where only part is. The Green Book does not say which part, and these are county figures, not figures for any city or place within the county.
  • A county in more than one area for the same standard has one row for each area. Counties are matched by FIPS code. The Green Book files Connecticut by its former eight counties; this site uses the nine Census planning regions, so Connecticut counties are Not reported rather than matched by name or guessed.

Source periods, coverage and other tables · Original publisher: documentation and methods

Brownfields properties recorded in EPA’s ACRES system (EPA)

U.S. Environmental Protection Agency, Brownfields Program / Facility Registry Service · EPA Facility Registry Service (FRS) Facility Interests, layer ACRES (Brownfields properties)

Data period
Layer contents as published by EPA (retrieved 2026-11-03); 44,615 ACRES property records
Geography vintage
County codes as recorded by EPA, joined by FIPS code to this site’s counties; Connecticut’s eight former counties are not matched to the nine planning regions
Units and scope
Number of ACRES property records.
Pages with this table
3,258
Coverage and interpretation
  • These are counts of property records in EPA’s ACRES system (Assessment, Cleanup and Redevelopment Exchange System) that carry this county’s code, as published in EPA’s Facility Registry Service. They are reported counts of records, not a measurement of contamination, exposure or health risk, and not a statement about any property’s present condition.
  • ACRES holds properties reported by EPA Brownfields grant recipients as assessed or cleaned up with grant funding, plus Targeted Brownfields Assessments performed by EPA regions. It is not a census of all brownfields: a county with few or no ACRES records may simply have had little EPA Brownfields grant activity, and a county with many records is not thereby more contaminated. The source layer carries no assessment, cleanup or redevelopment status, so none is shown, and no comparison of places is made.
  • The count is of ACRES property records, one per ACRES property id. One location can carry more than one record, so the count is not a count of distinct addresses.
  • A record is placed in a county only when EPA’s county code has five digits and its first two digits are the FIPS code of the state EPA names for the same record; records with a missing, short or conflicting code are not placed in any county. For state pages, these records are listed in a separate row under the state EPA names, where EPA names a state. Because of this, a county shown as 0 means no placed ACRES record, not that EPA has none for the county.
  • Connecticut is coded by EPA to its eight former counties (FIPS 09001 to 09015); this site’s Connecticut counties are the nine Census planning regions, and no official crosswalk exists for individual properties, so no Connecticut planning region has a table and nothing is allocated; Connecticut’s state figures are the sum of the eight former counties. Shannon County, South Dakota (46113) and Wade Hampton Census Area, Alaska (02270), which the Census Bureau renamed (same area, new code) in 2015, are mapped to Oglala Lakota County (46102) and Kusilvak Census Area (02158). Any other county code in the source that is not one of this site’s counties (for example Alaska’s retired 02201 and 02261) is counted in its state figures only, and counties in that state with no records of their own are then not shown rather than shown as zero.
  • The count is a snapshot of the layer on the retrieval date and changes as grantees report. These are county figures and state figures; nothing is allocated to cities or places. Export, API and commercial use have not been reviewed.

Source periods, coverage and other tables · Original publisher: documentation and methods

Community drinking water systems and violations (EPA)

U.S. Environmental Protection Agency, Enforcement and Compliance History Online (ECHO) · Safe Drinking Water Act (SDWIS) data download, 2026Q3 snapshot

Data period
SDWIS 2026Q3 snapshot published October 14, 2026; violations that began October 14, 2021 through October 14, 2026
Geography vintage
Counties served as coded by EPA (state and county number), matched to 2020 Census counties
Units and scope
Number of community water systems and people; population is a whole-system estimate.
Pages with this table
3,064
Coverage and interpretation
  • A community water system is a public water system that supplies water to the same people year-round. Only active community water systems are counted; transient and non-transient non-community systems (campgrounds, schools with their own wells) are not.
  • A system is counted in every county that EPA lists as served by it, so a county count includes systems that also serve other counties, and a system can appear on several county pages. The population is EPA’s estimate of the average daily population served by the whole system, summed over those systems. It is not a count of people in the county and can exceed the county’s population.
  • A health-based violation is a reported compliance status: a system exceeded a maximum contaminant level or maximum residual disinfectant level, or did not meet a treatment technique requirement. It is not a measurement of the water at any tap, does not say who was exposed, and does not say the water is unsafe today. The count covers violations whose noncompliance period began in the five years ending on the date shown; a system is counted once however many violations it has, and a violation may since have been resolved.
  • Monitoring and reporting violations, public notice violations and other violations that are not health based are not counted. Violations are reported by states and EPA regions to EPA, and states differ in how completely and how quickly they report, so a count of zero means none appears in the file, not that none occurred.
  • EPA codes counties served by state and county number. Counties were matched to Census counties on that code after checking it against EPA’s county reference table and the county name on the row. Where EPA still codes counties of an earlier vintage (Connecticut, which now has planning regions, and Alaska, where census areas changed), every county in that state is Not reported. A system with no validated county code is not counted anywhere, and a county with no matched system has no table.
  • These are county figures, not figures for any city or place within the county, and not a ranking or score of any water system.

Source periods, coverage and other tables · Original publisher: documentation and methods

Drought conditions, U.S. Drought Monitor weekly map of October 27, 2026 to November 2, 2026 (NDMC, USDA, NOAA)

National Drought Mitigation Center, University of Nebraska-Lincoln (with USDA, NOAA and NASA) · U.S. Drought Monitor statistics: percent of area by drought category, counties and states

Data period
Weekly map valid October 27, 2026 to November 2, 2026 (map date 2026-10-27); weeks counted: the 52 weekly maps ending with it (retrieved 2026-11-03)
Geography vintage
Counties and states by FIPS code as the service reports them; Connecticut as the nine Census planning regions
Units and scope
Percent of area (two decimals) in each drought category, and a count of weekly maps out of 52.
Pages with this table
3,274
Coverage and interpretation
  • The U.S. Drought Monitor is a weekly snapshot, not a forecast. Analysts at the National Drought Mitigation Center draw it each week from many indicators (rainfall, soil moisture, streamflow, vegetation and local reports) and the weekly map shown here is valid for the dates stated; conditions can change the next week. It does not measure water supply, crop loss or the effect on any person.
  • These are county figures and state figures: the percent of the county’s or state’s area that the map places in each category. A county with a small share in a category is not a county with no part in it, and the share says nothing about where inside the county the drought lies. Percents are rounded by the source to two decimals, so the six categories can miss 100.00% by a few hundredths.
  • The categories do not overlap: None, D0 (abnormally dry, which is not yet drought), D1 (moderate), D2 (severe), D3 (extreme) and D4 (exceptional). “D1 to D4 combined” is their sum, shown no higher than 100.00% where the source’s rounding makes the sum a hundredth over, and is the area in drought; D0 is excluded from it.
  • Weeks counted use the most recent 52 consecutive weekly maps ending with the map shown. A week counts as having any area in D1 or worse (or D3 or worse) when the source reports at least 0.01% of the area there; a smaller share rounds to 0.00% in the source and does not count. An area missing any of the 52 maps would be Not reported rather than counted short.
  • Connecticut is shown by the nine Census planning regions that replaced its eight counties. The Drought Monitor’s data service reports Connecticut under these planning-region FIPS codes in every one of the weeks used here (checked), so the regions join by code with no crosswalk and no week mixes the old counties with the new regions. Weeks before the service used planning-region areas are not read.
  • Areas join to this site’s counties and states by FIPS code only; no name matching is used. The source’s own area definitions apply (including how it treats water and Alaska’s and Puerto Rico’s areas), and these are not figures for any city or place within the county.
  • The weekly map is jointly produced by the National Drought Mitigation Center at the University of Nebraska-Lincoln, USDA, NOAA and NASA. Its published statistics are shown with the producers’ requested wording; the program’s permission page asks for that wording when its map is reproduced and states no other conditions on the statistics. Export, API and commercial use have not been reviewed.

Source periods, coverage and other tables · Original publisher: documentation and methods

EPA Map of Radon Zones: counties per zone (EPA and USGS, 1993, state)

U.S. Environmental Protection Agency with the U.S. Geological Survey · EPA Map of Radon Zones (1993), Text Version - Radon Zones Map (xls): county radon zones

Data period
1993 map (predicted screening level; not a current measurement)
Geography vintage
States and the District of Columbia by the state name printed in the EPA workbook, matched exactly to the state list
Units and scope
Zone: category 1, 2 or 3. Predicted average indoor radon screening level: picocuries per liter (pCi/L). Counts: number of list entries.
Pages with this table
51
Coverage and interpretation
  • The EPA Map of Radon Zones is a 1993 predictive screening classification, not a measurement of any home or of the current radon level. Each zone is a prediction of the average indoor screening level for a whole county as of the map’s development; it does not say what any one home, building or address contains.
  • EPA states that homes with elevated radon levels have been found in all three zones, that the map should not be used to decide whether an individual home should be tested, and that all homes should be tested. Radon varies from house to house and within a county.
  • Zone 1 is a predicted average above 4 pCi/L, Zone 2 from 2 to 4 pCi/L and Zone 3 below 2 pCi/L (picocuries per liter), as EPA defines them. A zone number is a category, not a measured concentration and not a health finding.
  • State counts are counts of county figures: numbers of entries in the EPA list under each state heading. Virginia’s 41 independent cities in the list are counted under Virginia, and Connecticut is counted by its former eight counties. Puerto Rico is not in the EPA list and has no table.
  • The EPA workbook carries the county list on two worksheets. They agree for all but 7 of the 3,144 entries; for those 7 counties (Anchorage Municipality, Alaska; La Plata County, Colorado; Somerset County, Maryland; and Macon, Madison, Marion and Maury Counties, Tennessee) no zone is shown, because the file does not say which worksheet is correct. Not reported is not zero.
  • Zones are shown as EPA published them; they are not combined into any index, and places are not compared or ordered. The map is an EPA and USGS work (public domain, decision 0011); this site is not endorsed by EPA or USGS.

Source periods, coverage and other tables · Original publisher: documentation and methods

EPA Map of Radon Zones: predicted indoor radon screening zone (EPA and USGS, 1993, county)

U.S. Environmental Protection Agency with the U.S. Geological Survey · EPA Map of Radon Zones (1993), Text Version - Radon Zones Map (xls): county radon zones

Data period
1993 map (predicted screening level; not a current measurement)
Geography vintage
Counties and county equivalents by the name printed in the EPA workbook, matched exactly to this site’s county list (Connecticut planning regions have no table)
Units and scope
Zone: category 1, 2 or 3. Predicted average indoor radon screening level: picocuries per liter (pCi/L). Counts: number of list entries.
Pages with this table
3,126
Coverage and interpretation
  • The EPA Map of Radon Zones is a 1993 predictive screening classification, not a measurement of any home or of the current radon level. Each zone is a prediction of the average indoor screening level for a whole county as of the map’s development; it does not say what any one home, building or address contains.
  • EPA states that homes with elevated radon levels have been found in all three zones, that the map should not be used to decide whether an individual home should be tested, and that all homes should be tested. Radon varies from house to house and within a county.
  • Zone 1 is a predicted average above 4 pCi/L, Zone 2 from 2 to 4 pCi/L and Zone 3 below 2 pCi/L (picocuries per liter), as EPA defines them. A zone number is a category, not a measured concentration and not a health finding.
  • County figures, not figures for any city or place within the county. The EPA workbook has county names and no FIPS codes, so each county is matched to this site’s county list only by identical state and normalized county name; a county with no exact match has no table, and nothing is matched by guesswork. Connecticut is listed by its former eight counties, which are not the nine planning regions used for Connecticut on this site, so Connecticut planning regions have no table. Alaska boroughs and census areas that have changed since 1993, Virginia independent cities that no longer exist, Yellowstone National Park and Puerto Rico also have no match or no row where the names differ or are absent.
  • The EPA workbook carries the county list on two worksheets. They agree for all but 7 of the 3,144 entries; for those 7 counties (Anchorage Municipality, Alaska; La Plata County, Colorado; Somerset County, Maryland; and Macon, Madison, Marion and Maury Counties, Tennessee) no zone is shown, because the file does not say which worksheet is correct. Not reported is not zero.
  • Zones are shown as EPA published them; they are not combined into any index, and places are not compared or ordered. The map is an EPA and USGS work (public domain, decision 0011); this site is not endorsed by EPA or USGS.

Source periods, coverage and other tables · Original publisher: documentation and methods

EPA-regulated facility records by program (ECHO, county)

U.S. Environmental Protection Agency, Enforcement and Compliance History Online (ECHO) · ECHO exporter download (echo_exporter.zip, ECHO_EXPORTER.csv)

Data period
Active registry records in the download of Sun, 01 Nov 2026 02:54:17 GMT (retrieved 2026-11-04); 1,689,073 active of 3,233,462 records
Geography vintage
County codes as recorded by EPA, joined by FIPS code to this site’s counties; Connecticut’s former counties are not matched to the planning regions
Units and scope
Number of active facility records.
Pages with this table
3,211
Coverage and interpretation
  • These are counts of registered facility records, not a measure of pollution, exposure, risk or compliance, and not a statement about any facility. Only the number of records is shown: no facility name, address, location, violation, compliance status, penalty or inspection result is published.
  • A record is counted when ECHO’s active-facility flag is Y. The count covers every kind of regulated place the registry holds, from large plants to small sites such as dry cleaners, gas stations and small water systems, so a larger count mostly reflects how many places a county has that fall under federal programs, not how much activity or harm there is.
  • Program flags overlap: one facility can carry several flags, so the flag rows do not add up to the total, and a record with no flag among the six shown is counted only in the total. A flag means the facility is linked to that program’s data system (Clean Air Act, Clean Water Act NPDES permits, RCRA hazardous waste, Safe Drinking Water Act, Toxics Release Inventory, Greenhouse Gas Reporting Program), not that it emits or discharges anything today.
  • This differs from the other EPA subsections on this site. The TRI flag counts active registry records linked to the Toxics Release Inventory in any year, whereas the Toxics Release Inventory subsection counts the facilities that filed a report for its single reporting year; likewise the GHG flag is not the Greenhouse Gas Reporting Program year-specific facility count, the drinking water flag counts all linked records rather than active community water systems, and Superfund National Priorities List sites and Brownfields (ACRES) records are separate programs counted separately. Counts from different subsections are not meant to agree.
  • A record is placed in a county only when ECHO’s county code has five digits and its first two digits are the FIPS code of the state ECHO names for the same record; records with a missing or conflicting code are in no county. State figures include those records in a separate row. A county shown as 0 therefore means no placed record, not that ECHO has none for the county.
  • Connecticut is coded by ECHO to its former counties; this site’s Connecticut counties are the Census planning regions and no official crosswalk exists for individual facilities, so no Connecticut planning region has a table and nothing is allocated. Shannon County, South Dakota (46113) and Wade Hampton Census Area, Alaska (02270), renamed by the Census Bureau in 2015, are mapped to Oglala Lakota County (46102) and Kusilvak Census Area (02158). Any other source county code that is not one of this site’s counties counts in its state figures only, and counties of that state with no record of their own are then not shown rather than shown as zero.
  • The registry changes continuously as facilities open, close and are reclassified; this is a snapshot of the monthly download on the retrieval date. These are county figures and state figures; nothing is allocated to cities or places. Export, API and commercial use have not been reviewed.

Source periods, coverage and other tables · Original publisher: documentation and methods

EPA-regulated facility records by program (ECHO, state)

U.S. Environmental Protection Agency, Enforcement and Compliance History Online (ECHO) · ECHO exporter download (echo_exporter.zip, ECHO_EXPORTER.csv)

Data period
Active registry records in the download of Sun, 01 Nov 2026 02:54:17 GMT (retrieved 2026-11-04); 1,689,073 active of 3,233,462 records
Geography vintage
County codes as recorded by EPA, joined by FIPS code to this site’s counties; Connecticut’s former counties are not matched to the planning regions
Units and scope
Number of active facility records.
Pages with this table
52
Coverage and interpretation
  • These are counts of registered facility records, not a measure of pollution, exposure, risk or compliance, and not a statement about any facility. Only the number of records is shown: no facility name, address, location, violation, compliance status, penalty or inspection result is published.
  • A record is counted when ECHO’s active-facility flag is Y. The count covers every kind of regulated place the registry holds, from large plants to small sites such as dry cleaners, gas stations and small water systems, so a larger count mostly reflects how many places a county has that fall under federal programs, not how much activity or harm there is.
  • Program flags overlap: one facility can carry several flags, so the flag rows do not add up to the total, and a record with no flag among the six shown is counted only in the total. A flag means the facility is linked to that program’s data system (Clean Air Act, Clean Water Act NPDES permits, RCRA hazardous waste, Safe Drinking Water Act, Toxics Release Inventory, Greenhouse Gas Reporting Program), not that it emits or discharges anything today.
  • This differs from the other EPA subsections on this site. The TRI flag counts active registry records linked to the Toxics Release Inventory in any year, whereas the Toxics Release Inventory subsection counts the facilities that filed a report for its single reporting year; likewise the GHG flag is not the Greenhouse Gas Reporting Program year-specific facility count, the drinking water flag counts all linked records rather than active community water systems, and Superfund National Priorities List sites and Brownfields (ACRES) records are separate programs counted separately. Counts from different subsections are not meant to agree.
  • A record is placed in a county only when ECHO’s county code has five digits and its first two digits are the FIPS code of the state ECHO names for the same record; records with a missing or conflicting code are in no county. State figures include those records in a separate row. A county shown as 0 therefore means no placed record, not that ECHO has none for the county.
  • Connecticut is coded by ECHO to its former counties; this site’s Connecticut counties are the Census planning regions and no official crosswalk exists for individual facilities, so no Connecticut planning region has a table and nothing is allocated. Shannon County, South Dakota (46113) and Wade Hampton Census Area, Alaska (02270), renamed by the Census Bureau in 2015, are mapped to Oglala Lakota County (46102) and Kusilvak Census Area (02158). Any other source county code that is not one of this site’s counties counts in its state figures only, and counties of that state with no record of their own are then not shown rather than shown as zero.
  • The registry changes continuously as facilities open, close and are reclassified; this is a snapshot of the monthly download on the retrieval date. These are county figures and state figures; nothing is allocated to cities or places. Export, API and commercial use have not been reviewed.

Source periods, coverage and other tables · Original publisher: documentation and methods

Fatal motor vehicle crashes and deaths (NHTSA)

U.S. Department of Transportation, National Highway Traffic Safety Administration · Fatality Analysis Reporting System (FARS) national CSV files, 2020 to 2023

Data period
Calendar years 2020 to 2023, final files (retrieved 2026-11-03)
Geography vintage
County FIPS codes as coded on the crash record, matched to Census counties by code
Units and scope
Fatal crashes and people killed (deaths within 30 days of the crash): counts, not rates.
Pages with this table
3,133
Coverage and interpretation
  • These are counts of fatal crashes on public roads and of the people who died in them within 30 days, as recorded by NHTSA from police reports, death certificates and other state sources. They are not counts of all crashes or injuries, and they say nothing about how safe a road is, how much anyone drives, or the cause of any crash. Counts are not adjusted for population or travel, so a county with more people or more driving can record more deaths for that reason alone.
  • FARS is a census of fatal crashes for the 50 states and the District of Columbia, so a county with no fatal crash in a year is shown as zero. Puerto Rico is published by NHTSA in separate files that this site does not use, so Puerto Rico counties are Not reported.
  • The county is the one coded on the crash record, the county where the crash happened, not where anyone lived. Counties are matched to Census counties by FIPS code only. A few crashes name no county (code not applicable, other, not reported or unknown) and appear on no county page, so county figures can sum to slightly less than a state total.
  • Where NHTSA used a county code that the Census list no longer has, the code was mapped only if the county was renamed with a new code and the same territory (Shannon County, South Dakota, now Oglala Lakota County; Wade Hampton Census Area, Alaska, now Kusilvak Census Area). The retired Valdez-Cordova Census Area, Alaska, was split into two areas, so crashes coded to it cannot be assigned and the two successor areas are Not reported. NHTSA still codes Connecticut by its former counties, which the Census planning regions on this site replaced, so Connecticut is Not reported.
  • Years shown are 2020 through 2023, all final files. The 2024 file is an Annual Report File that NHTSA says is subject to change and is not used. 2020 is the first year of the COVID-19 pandemic, when travel changed sharply, so years are shown separately rather than averaged.
  • These are county figures, not figures for any city or place within the county.

Source periods, coverage and other tables · Original publisher: documentation and methods

Federal disaster declarations by year and incident type (FEMA)

Federal Emergency Management Agency · OpenFEMA Disaster Declarations Summaries, version 2

Data period
Declared January 1, 2016 through October 27, 2026
Geography vintage
States, the District of Columbia and Puerto Rico; other territories are not shown
Units and scope
Number of federal declarations; one disaster number in one state is one declaration.
Pages with this table
52
Coverage and interpretation
  • These are counts of declarations, not damage, deaths, cost or aid. A declaration is a federal decision to make assistance available; a state without one in a year may still have been hit.
  • A major disaster, an emergency and a fire management assistance declaration are counted separately, so a single event can appear in more than one. Each disaster number in a state is counted once however many of its counties were designated.
  • Incident groups follow the incident type FEMA records: hurricane includes tropical storm, typhoon and tropical depression; severe storm includes tornado, straight-line winds, winter storm, severe ice storm, snowstorm and coastal storm; other is every remaining type, including biological (COVID-19), earthquake and landslide.
  • The year is the calendar year of the declaration date, not of the event, and 2026 is partial. Declarations are added continuously, so counts can rise.
  • A zero is a complete count with no declaration of that kind, because the source lists every declaration it has. Tribal declarations are not counted in a state.

Source periods, coverage and other tables · Original publisher: documentation and methods

Federal disaster declarations since 2016 (FEMA)

Federal Emergency Management Agency · OpenFEMA Disaster Declarations Summaries, version 2

Data period
Declared January 1, 2016 through October 27, 2026
Geography vintage
Counties named in each declaration
Units and scope
Number of federal disaster declarations that designated the county.
Pages with this table
3,210
Coverage and interpretation
  • A declaration is a federal decision to make assistance available. It is not a measure of damage, deaths or cost, and a county without one may still have been hit.
  • Each row counts declarations for one type of incident. A single event can bring several declarations, such as an emergency and then a major disaster.
  • Declarations for COVID-19 are recorded as Biological incidents and designated nearly every county in 2020.
  • Statewide and tribal designations without a county code are not counted. Declarations are added continuously, so the count can rise.
  • These are county figures, not figures for any city or place within the county.

Source periods, coverage and other tables · Original publisher: documentation and methods

Food, drug and device recall events by year, firms in this state (FDA enforcement reports)

U.S. Food and Drug Administration (openFDA) · openFDA food, drug and device enforcement reports (recall enforcement data)

Data period
Enforcement report dates 2021-01-01 to 2025-12-31 (five calendar years)
Geography vintage
The 50 States, the District of Columbia and Puerto Rico, by the state the source prints for the recalling firm; records for firms outside these (foreign firms, other territories, no state) are not on these pages
Units and scope
Count of distinct recall events per year and product type, and per year and FDA classification (all product types).
Pages with this table
52
Coverage and interpretation
  • These are counts of recall events reported in FDA enforcement reports and classified by the FDA, counted by the state of the firm that carried out the recall, not where the product was sold or where anyone was harmed, and not a measure of how safe a state's products are. A recall event is one FDA event identifier; one event can cover many products and is counted once.
  • The year is the date the recall appeared in the FDA Enforcement Report, which can be later than when the firm began the recall. Classification is the FDA's own: Class I is a reasonable probability of serious adverse health consequences or death, Class II a temporary or medically reversible health consequence, Class III unlikely to cause adverse health consequences. Recalls not yet classified are counted in the product totals only. An event with products in more than one class is counted in each class.
  • Zero is a count of zero events in the complete set of enforcement reports retrieved for that state, year and product type, not a missing value. Large firms and states with many manufacturers have more recalls; counts are not adjusted for the number of firms. Names of firms are not shown.

Source periods, coverage and other tables · Original publisher: documentation and methods

Forest land and timberland area (USDA Forest Service Forest Inventory and Analysis)

USDA Forest Service, Forest Inventory and Analysis Program · Forest Inventory and Analysis, EVALIDator forest land and timberland area

Data period
Latest whole-state FIA evaluation per state (evaluation years 2019 to 2026)
Geography vintage
States and Puerto Rico with a whole-state FIA evaluation; Alaska, the District of Columbia and the other territories are not shown
Units and scope
Thousand acres (rounded from acres); sampling error as a percent of the estimate.
Pages with this table
50
Coverage and interpretation
  • These are sample-based estimates of the land area in forest land (land at least 10 percent covered by trees of any size, or formerly so and to be regenerated, at least 1 acre and 120 feet wide) and in timberland (forest land capable of producing industrial wood and not legally withdrawn from timber use), from the latest FIA evaluation of the state, not a census or a satellite map. Timberland is a subset of forest land.
  • The evaluation year is the year of the FIA evaluation group; the plots behind it were measured over several earlier years. Texas shows the latest whole-state evaluation, which is older than the East Texas evaluation. Compare states with care where the evaluation years differ.
  • A value the source did not report is shown as Not reported, never as zero.

Source periods, coverage and other tables · Original publisher: documentation and methods

Greenhouse gas reporting facilities (EPA)

U.S. Environmental Protection Agency, Greenhouse Gas Reporting Program · Greenhouse Gas Reporting Program (GHGRP) facility and sector emission tables, Envirofacts

Data period
Reporting year 2023, the newest year EPA had published (retrieved 2026-11-03)
Geography vintage
County codes as assigned by EPA to each facility’s reported location, matched to Census counties by code
Units and scope
Facilities: count. Emissions: metric tons of carbon dioxide equivalent per year.
Pages with this table
3,183
Coverage and interpretation
  • These are facility-reported quantities, not exposure, risk or health effects. A facility reports greenhouse gases to EPA under a federal rule; the figure says nothing about what anyone in the county breathes, drinks or is exposed to.
  • Only facilities above the program’s reporting thresholds report (generally 25,000 metric tons CO2e a year from covered sources). “No facilities in the file” means no reporting facility above those thresholds was located in the county, not that the county has no greenhouse gas emissions. Farms, vehicles, homes and most small businesses are not covered.
  • Total reported direct emissions are the facility’s emissions of CO2, methane, nitrous oxide and fluorinated gases in metric tons of carbon dioxide equivalent (EPA’s AR4 global warming potentials), summed over the direct-emitter source categories. Biogenic CO2 is reported separately by EPA and is not included. Gas and product suppliers report quantities that are released elsewhere when the product is used; those are not direct emissions and are not included, and CO2 received for injection is not an emission.
  • The county is the one EPA assigns to the facility’s reported location. Basin-level oil and gas reporters and some pipeline systems report no single county and appear on no county page. Connecticut is omitted: EPA still uses the former Connecticut counties, which the Census planning regions replaced.
  • The sector shown is EPA’s sector for the source category (for example Power Plants or Petroleum and Natural Gas Systems) with the largest reported emissions in the county, and its share is that sector’s part of the county’s reported total, computed from the facility values by integer arithmetic. It describes which reporting sector emitted the most, not a comparison with other places.
  • These are county figures from the facilities’ latest submissions as published by EPA; facilities may revise earlier years. They are not figures for any city or place within the county.

Source periods, coverage and other tables · Original publisher: documentation and methods

Individuals and Households Program registrations and FEMA-approved amounts, disasters declared 2017 to date (FEMA)

Federal Emergency Management Agency · OpenFEMA Housing Assistance Program Data - Owners and Renters, version 2

Data period
Disasters declared 2017-01-01 through 2026-10-27 (ten calendar years, 2026 partial); amounts approved as of the data refresh; retrieved 2026-11-04
Geography vintage
County names as FEMA reports them with the state, matched exactly to Census counties by state and normalized name; no county code in the source; Connecticut’s former counties are not matched to planning regions
Units and scope
Number of disasters and valid registrations, and whole U.S. dollars approved by FEMA (not adjusted for inflation).
Pages with this table
2,384
Coverage and interpretation
  • These are amounts FEMA approved for households under its Individuals and Households Program, not damage and not total losses. They cover valid registrations for disasters declared from January 1, 2017 (ten calendar years, the last one partial) in counties FEMA designated for Individual Assistance. Housing assistance is repair, replacement and rental assistance; other needs assistance is personal property, transportation, medical, dental, funeral, moving and similar items. The program does not include Public Assistance to governments, flood insurance payments, Small Business Administration loans, insurance, other agencies’ programs or private losses.
  • A valid registration is a household application received in an Individual Assistance county during the registration period; it is not a household that received aid, and approved dollars are for the households FEMA approved. Owners and renters are separate FEMA tallies and are added here. Dollars are approved amounts as of the data’s refresh, not adjusted for inflation, and change as cases are amended. The source total can differ from the sum of its listed housing and other needs amounts, so the three amounts are shown as FEMA reports them and are not forced to add.
  • The period is set by the date each disaster was declared. A disaster counts for a county when the county has valid registrations for it. FEMA publishes these figures by disaster and ZIP code with a county name but no county code, so counties are matched to this site’s counties by exact state and county name (case, accents, punctuation, “Saint” and “St.” and the county type word only); rows whose county text is a city, tribe, island area or anything else not matched exactly, and Connecticut’s former counties, are in no table and are not guessed.
  • Privacy rule: FEMA’s figures are already counts and dollar totals, never applicants. A county-disaster combination with fewer than 10 valid registrations (owners plus renters) is small enough to be suppressed, and its registrations and dollars are not published or added into any figure on this site. A county’s figures are shown only when none of its disasters in the period was under 10 registrations; a county with any such disaster has no table, and Not reported is not zero.
  • These are county figures, not figures for any city, ZIP code or place within the county, and they say nothing about the severity of any disaster or the need of a place.

Source periods, coverage and other tables · Original publisher: documentation and methods

Modeled expected annual loss by natural hazard (FEMA, county)

Federal Emergency Management Agency (FEMA), National Risk Index · National Risk Index version 1.20 (December 2025), county table (NRI_Table_Counties.zip, NRI_Table_Counties.csv)

Data period
National Risk Index version 1.20, December 2025 (file last modified Thu, 18 Dec 2025 22:05:16 GMT, retrieved 2026-11-02); 3,232 county rows
Geography vintage
FEMA’s county list for this release, joined to this site’s counties by five-digit FIPS code (Connecticut planning regions as coded by FEMA)
Units and scope
Dollars of modeled average annual loss (Expected Annual Loss).
Pages with this table
3,222
Coverage and interpretation
  • These are modeled estimates of average annual loss from FEMA’s National Risk Index (version 1.20, December 2025), not a forecast of losses in any year, not a count of past events or past damage, and not a statement about any property. Expected Annual Loss is FEMA’s average yearly loss in dollars to buildings, population and agriculture from each natural hazard, built from hazard frequency, the value exposed and a historic loss ratio.
  • Population loss is monetized by FEMA at $13.7 million per fatality or per 10 injuries, so the population dollar figure is a statistical-life equivalent, not an economic cost that anyone pays. Combined figures are the sum across hazards, with the total equal to buildings plus the population dollar equivalent plus agriculture.
  • A county page shows each hazard’s total Expected Annual Loss and the combined total; the split into buildings, population and agriculture and the exposure are shown on the state page only, and annualized frequency and historic loss ratios are not shown, to keep the page compact. A hazard FEMA does not assess for a county, or does not report a value for, has no row; that is not zero. A zero shown is a zero in the source.
  • Some hazards are assessed for fewer consequence types: drought is assessed for agriculture only, and avalanche, coastal flooding, earthquake, ice storm, landslide, lightning, tsunami and volcanic activity have no agriculture component. State figures are the sum of the county values FEMA reports for the state’s counties, shown only where the quantity is additive (losses and exposure); a state cell sums the counties that have a value and a state with no county value for a quantity shows Not reported.
  • Exposure on the state page is the total value in dollars FEMA treats as exposed to a hazard (buildings, population equivalent and agriculture together). Figures are not adjusted for inflation or updated between FEMA releases. Counties are FEMA’s county list for this release (the nine Connecticut planning regions, Alaska boroughs and census areas, Virginia independent cities and Puerto Rico municipios); FEMA’s American Samoa, Guam, Northern Mariana Islands and U.S. Virgin Islands rows are not used. These are county figures and state figures; nothing is allocated to cities or places.
  • The National Risk Index draws on many third-party hazard, exposure and loss inputs, some licensed by FEMA; this site publishes only FEMA’s aggregate county quantities. Display only; export, API and commercial use have not been reviewed. FEMA does not endorse this site.

Source periods, coverage and other tables · Original publisher: documentation and methods

Modeled expected annual loss by natural hazard (FEMA, state)

Federal Emergency Management Agency (FEMA), National Risk Index · National Risk Index version 1.20 (December 2025), county table (NRI_Table_Counties.zip, NRI_Table_Counties.csv)

Data period
National Risk Index version 1.20, December 2025 (file last modified Thu, 18 Dec 2025 22:05:16 GMT, retrieved 2026-11-02); 3,232 county rows
Geography vintage
FEMA’s county list for this release, joined to this site’s counties by five-digit FIPS code (Connecticut planning regions as coded by FEMA)
Units and scope
Dollars of modeled average annual loss (Expected Annual Loss).
Pages with this table
52
Coverage and interpretation
  • These are modeled estimates of average annual loss from FEMA’s National Risk Index (version 1.20, December 2025), not a forecast of losses in any year, not a count of past events or past damage, and not a statement about any property. Expected Annual Loss is FEMA’s average yearly loss in dollars to buildings, population and agriculture from each natural hazard, built from hazard frequency, the value exposed and a historic loss ratio.
  • Population loss is monetized by FEMA at $13.7 million per fatality or per 10 injuries, so the population dollar figure is a statistical-life equivalent, not an economic cost that anyone pays. Combined figures are the sum across hazards, with the total equal to buildings plus the population dollar equivalent plus agriculture.
  • A county page shows each hazard’s total Expected Annual Loss and the combined total; the split into buildings, population and agriculture and the exposure are shown on the state page only, and annualized frequency and historic loss ratios are not shown, to keep the page compact. A hazard FEMA does not assess for a county, or does not report a value for, has no row; that is not zero. A zero shown is a zero in the source.
  • Some hazards are assessed for fewer consequence types: drought is assessed for agriculture only, and avalanche, coastal flooding, earthquake, ice storm, landslide, lightning, tsunami and volcanic activity have no agriculture component. State figures are the sum of the county values FEMA reports for the state’s counties, shown only where the quantity is additive (losses and exposure); a state cell sums the counties that have a value and a state with no county value for a quantity shows Not reported.
  • Exposure on the state page is the total value in dollars FEMA treats as exposed to a hazard (buildings, population equivalent and agriculture together). Figures are not adjusted for inflation or updated between FEMA releases. Counties are FEMA’s county list for this release (the nine Connecticut planning regions, Alaska boroughs and census areas, Virginia independent cities and Puerto Rico municipios); FEMA’s American Samoa, Guam, Northern Mariana Islands and U.S. Virgin Islands rows are not used. These are county figures and state figures; nothing is allocated to cities or places.
  • The National Risk Index draws on many third-party hazard, exposure and loss inputs, some licensed by FEMA; this site publishes only FEMA’s aggregate county quantities. Display only; export, API and commercial use have not been reviewed. FEMA does not endorse this site.

Source periods, coverage and other tables · Original publisher: documentation and methods

National Flood Insurance Program policies and paid claims (FEMA)

Federal Emergency Management Agency · OpenFEMA NFIP Redacted Policies v3 and NFIP Redacted Claims v3

Data period
Policies in force on September 8, 2026; claims for losses from January 1, 1978 through September 7, 2026
Geography vintage
County FIPS codes as reported by FEMA, matched to Census counties by code
Units and scope
Number of policies and claims, and whole U.S. dollars (not adjusted for inflation).
Pages with this table
3,213
Coverage and interpretation
  • These figures describe flood insurance bought through the National Flood Insurance Program (NFIP), not flood risk. A county with few policies is not a county with little flooding, because people choose whether to insure, lenders require it only in mapped high-risk areas, and some homes carry private flood insurance that is not counted here.
  • Policies in force are the policy records whose effective date is on or before, and whose termination date is after, the as-of date shown. FEMA counts one policy per insured unit, so a condominium association policy counts once for each unit it insures. Coverage in force is the total of the building and contents coverage limits on those policies. It is the maximum the policies could pay, not the value of the property or a prediction of loss.
  • Paid claims are NFIP claims whose building, contents and increased-cost-of-compliance payments add up to more than zero. The total paid is the sum of those payments as recorded, before recoveries such as salvage and subrogation, and includes the negative adjustments FEMA records when a check is reissued. Dollars are as paid in the year of the claim, not adjusted for inflation. The claims file begins with losses in 1978.
  • FEMA publishes a county code for each claim. The policies file has no separate county column, so the county of a policy is the first five digits (state and county FIPS code) of the Census block group identifier FEMA assigns to the insured address. Nothing else, such as a ZIP code or coordinates, is used to place a record in a county. Records without a usable code are not assigned to any county.
  • FEMA still reports Connecticut by its former counties, which the Census planning regions on this site replaced, so Connecticut is omitted. Where FEMA uses a county code that is not in this site’s county list, a county with no records in that state is Not reported rather than zero; other counties show a zero only when FEMA’s files carry no policy or claim for them.
  • This is FEMA’s public redacted data and it is revised monthly. No individual policy or claim is shown or kept; only county counts and sums are.
  • These are county figures, not figures for any city or place within the county.

Source periods, coverage and other tables · Original publisher: documentation and methods

Public Assistance projects and federal share obligated, disasters declared 2017 to date (FEMA)

Federal Emergency Management Agency · OpenFEMA Public Assistance Funded Projects Details, version 2

Data period
Projects of disasters declared 2017-01-01 through 2026-09-01 (ten calendar years, 2026 partial); federal share obligated as of the data refresh of 2026-11-02; retrieved 2026-11-03
Geography vintage
State FIPS codes as FEMA reports them (the 50 states, the District of Columbia and Puerto Rico; other territories are not shown)
Units and scope
Number of projects and whole U.S. dollars of federal share obligated (not adjusted for inflation).
Pages with this table
52
Coverage and interpretation
  • These are federal obligations for Public Assistance projects, not total damage and not all spending on a disaster. They are the federal share FEMA has obligated, as of the data’s refresh, for projects of disasters declared from January 1, 2017 (ten calendar years, the last one partial). Public Assistance reimburses governments and certain nonprofits for eligible emergency work and for repairing or replacing public facilities; it does not include FEMA aid to individuals and households, flood insurance payments, hazard mitigation grants, other federal agencies’ programs, insurance, state or local spending, or private losses.
  • The period is set by the year each disaster was declared, not the year the money was obligated or the work was done; a project can be obligated years after its declaration, so recent declarations are incomplete, and amounts change as projects are amended, closed out or de-obligated. The figures are obligations (money committed), not payments received. The newest disaster year in the file is partial.
  • A project is one Public Assistance project (project worksheet). Only projects whose eligibility status is Eligible are counted; Ineligible, Withdrawn and Pending Eligibility Review projects are set aside. “Projects with federal share obligated” counts the eligible projects with more than $0 obligated so far; an eligible project with nothing yet obligated counts in Projects only. One project can cover work in several places and is assigned to a single county.
  • The county is the applicant’s location as FEMA records it, not necessarily where the damage or the work occurred. Records whose county FEMA gives as Statewide, or leaves blank, are state-level applicants; they are included in the state figures and in no county, and a Statewide record is never given the county code FEMA attaches to it. Records with no state FIPS code (applicants such as some tribes, and territories outside this site’s state list) are in no table.
  • Categories are FEMA’s Public Assistance damage categories A to G, I and Z. Emergency protective measures (B) also includes donated-resource emergency work, and Management and administrative costs (Z) combines FEMA’s management cost, direct administrative cost and Section 324 management cost descriptions. Dollars are as obligated, not adjusted for inflation, and are shown rounded to whole dollars with the exact cents kept in the source field.
  • A place with no row in a category has no eligible project of that category in this file; a place with no table has no eligible project assigned to it. Neither is shown as zero. The figures describe federal funding for public projects and say nothing about the severity of any disaster or the need of a place.
  • State figures add every eligible project whose state FIPS code is the state’s, including state-level applicants and Connecticut’s former counties, so a state is more than its county tables add to. Puerto Rico is shown as a separate state-level geography; Guam, the U.S. Virgin Islands, American Samoa and the Northern Mariana Islands are not shown.

Source periods, coverage and other tables · Original publisher: documentation and methods

Public Assistance projects and federal share obligated, disasters declared 2017 to date (FEMA)

Federal Emergency Management Agency · OpenFEMA Public Assistance Funded Projects Details, version 2

Data period
Projects of disasters declared 2017-01-01 through 2026-09-01 (ten calendar years, 2026 partial); federal share obligated as of the data refresh of 2026-11-02; retrieved 2026-11-03
Geography vintage
Applicant county FIPS codes as FEMA reports them (state code plus county code), matched to Census counties by code; Connecticut’s former counties are not matched to planning regions
Units and scope
Number of projects and whole U.S. dollars of federal share obligated (not adjusted for inflation).
Pages with this table
2,867
Coverage and interpretation
  • These are federal obligations for Public Assistance projects, not total damage and not all spending on a disaster. They are the federal share FEMA has obligated, as of the data’s refresh, for projects of disasters declared from January 1, 2017 (ten calendar years, the last one partial). Public Assistance reimburses governments and certain nonprofits for eligible emergency work and for repairing or replacing public facilities; it does not include FEMA aid to individuals and households, flood insurance payments, hazard mitigation grants, other federal agencies’ programs, insurance, state or local spending, or private losses.
  • The period is set by the year each disaster was declared, not the year the money was obligated or the work was done; a project can be obligated years after its declaration, so recent declarations are incomplete, and amounts change as projects are amended, closed out or de-obligated. The figures are obligations (money committed), not payments received. The newest disaster year in the file is partial.
  • A project is one Public Assistance project (project worksheet). Only projects whose eligibility status is Eligible are counted; Ineligible, Withdrawn and Pending Eligibility Review projects are set aside. “Projects with federal share obligated” counts the eligible projects with more than $0 obligated so far; an eligible project with nothing yet obligated counts in Projects only. One project can cover work in several places and is assigned to a single county.
  • The county is the applicant’s location as FEMA records it, not necessarily where the damage or the work occurred. Records whose county FEMA gives as Statewide, or leaves blank, are state-level applicants; they are included in the state figures and in no county, and a Statewide record is never given the county code FEMA attaches to it. Records with no state FIPS code (applicants such as some tribes, and territories outside this site’s state list) are in no table.
  • Categories are FEMA’s Public Assistance damage categories A to G, I and Z. Emergency protective measures (B) also includes donated-resource emergency work, and Management and administrative costs (Z) combines FEMA’s management cost, direct administrative cost and Section 324 management cost descriptions. Dollars are as obligated, not adjusted for inflation, and are shown rounded to whole dollars with the exact cents kept in the source field.
  • A place with no row in a category has no eligible project of that category in this file; a place with no table has no eligible project assigned to it. Neither is shown as zero. The figures describe federal funding for public projects and say nothing about the severity of any disaster or the need of a place.
  • FEMA still codes Connecticut by its former eight counties (09001 to 09015), which the Census planning regions on this site replaced, and publishes no crosswalk, so no Connecticut county has a table and nothing is allocated to a planning region; the Connecticut state figures include every Connecticut record. County tables add to less than the state figure because state-level applicants belong to no county. Areas join to this site’s counties by FIPS code only (state code plus county code, zero-padded); no name matching is used.
  • These are county figures, not figures for any city or place within the county.

Source periods, coverage and other tables · Original publisher: documentation and methods

Public-use airports and passenger boardings (FAA)

U.S. Department of Transportation, Federal Aviation Administration and Bureau of Transportation Statistics · National Transportation Atlas Database Aviation Facilities (FAA NASR airport data) and FAA calendar year 2025 passenger boarding (enplanement) data, aggregated to state counts

Data period
FAA NASR airport records effective 2026/10/29 (retrieved 2026-11-04); passenger boardings (enplanements) for calendar year 2025
Geography vintage
States as recorded by the FAA (the 50 states, the District of Columbia and Puerto Rico; other territories are not shown)
Units and scope
Facilities: counts of public-use, operational FAA records. Passenger boardings: counts of enplaned passengers in the calendar year shown, as the FAA defines them. Counts, not rates.
Pages with this table
52
Coverage and interpretation
  • Counts of facilities and passenger boardings as the Federal Aviation Administration (FAA) records them, not a measure of air service, airport size or quality, and not rates. Facilities are the records in the FAA National Airspace System Resource (NASR) airport data, as published by the Bureau of Transportation Statistics in the National Transportation Atlas Database “Aviation Facilities” layer, and counted once each.
  • A facility is counted here when the FAA records it as open for public use (facility use code PU) and operational (status O). Private-use facilities and facilities the FAA lists as closed are not counted. Type and ownership are the FAA codes, named as the FAA names them; a public-use facility can be privately owned. Airport names, addresses and coordinates are not shown.
  • Passenger boardings are the FAA’s calendar-year passenger boarding (enplanement) counts, a calendar year being January 1 through December 31, taken from the FAA Air Carrier Activity database and published in “Enplanements at All Commercial Service Airports” (primary and nonprimary commercial service airports). A boarding counts at the airport where the passenger boards, not where the passenger lives or finishes the trip. The FAA publishes the preliminary figures first and the final figures about eight months after the year ends. Boardings appear only where the FAA list has an airport in the area; areas without one show no boardings row.
  • State figures use the state of the county the FAA records for each facility, so every record is counted whether or not its county could be matched. The FAA places the airports serving Washington, D.C. that stand in Virginia in Virginia. Territories other than Puerto Rico, and facilities outside the United States, are not shown.
  • Not reported is not zero: a value that cannot be placed is Not reported.

Source periods, coverage and other tables · Original publisher: documentation and methods

Public-use airports and passenger boardings (FAA)

U.S. Department of Transportation, Federal Aviation Administration and Bureau of Transportation Statistics · National Transportation Atlas Database Aviation Facilities (FAA NASR airport data) and FAA calendar year 2025 passenger boarding (enplanement) data, aggregated to county counts

Data period
FAA NASR airport records effective 2026/10/29 (retrieved 2026-11-04); passenger boardings (enplanements) for calendar year 2025
Geography vintage
Counties named by the FAA (state and county name, no code) and matched by exact name to the Census counties on this site; Connecticut cannot be matched
Units and scope
Facilities: counts of public-use, operational FAA records. Passenger boardings: counts of enplaned passengers in the calendar year shown, as the FAA defines them. Counts, not rates.
Pages with this table
3,099
Coverage and interpretation
  • Counts of facilities and passenger boardings as the Federal Aviation Administration (FAA) records them, not a measure of air service, airport size or quality, and not rates. Facilities are the records in the FAA National Airspace System Resource (NASR) airport data, as published by the Bureau of Transportation Statistics in the National Transportation Atlas Database “Aviation Facilities” layer, and counted once each.
  • A facility is counted here when the FAA records it as open for public use (facility use code PU) and operational (status O). Private-use facilities and facilities the FAA lists as closed are not counted. Type and ownership are the FAA codes, named as the FAA names them; a public-use facility can be privately owned. Airport names, addresses and coordinates are not shown.
  • Passenger boardings are the FAA’s calendar-year passenger boarding (enplanement) counts, a calendar year being January 1 through December 31, taken from the FAA Air Carrier Activity database and published in “Enplanements at All Commercial Service Airports” (primary and nonprimary commercial service airports). A boarding counts at the airport where the passenger boards, not where the passenger lives or finishes the trip. The FAA publishes the preliminary figures first and the final figures about eight months after the year ends. Boardings appear only where the FAA list has an airport in the area; areas without one show no boardings row.
  • The FAA records a county name and a state, not a county code. Each record was matched to a Census county by exact state and county name, after removing capitalization, punctuation and the words County, Parish, Borough, Census Area, Municipality and Municipio. A record whose name matches nothing in the current Census list, or matches both a county and an independent city of the same name (for example Fairfax, Franklin, Richmond or Roanoke in Virginia), is not assigned to any county and is never guessed. Where a state has such records, a county with none assigned shows no table, and a county that does have records carries a note that its count could be lower.
  • Connecticut is Not reported by county: the FAA still names Connecticut’s former counties, which the Census planning regions on this site replaced, and the two cannot be matched by name. Alaska’s FAA county names include boroughs and census areas that the Census Bureau has since renamed, divided or merged; those records are not assigned.
  • These are county figures, not figures for any city or place within the county.
  • Not reported is not zero: a value that cannot be placed is Not reported.

Source periods, coverage and other tables · Original publisher: documentation and methods

Railroad grade crossings and reported incidents (FRA)

U.S. Department of Transportation, Federal Railroad Administration · FRA Crossing Inventory Data (Form 71, current) and Highway-Rail Grade Crossing Incident Data (Form 57), 2021 to 2025, as published on data.transportation.gov, aggregated to state counts

Data period
Crossing inventory as filed and published by FRA (retrieved 2026-11-03); incident reports for calendar years 2021 to 2025 (retrieved 2026-11-03)
Geography vintage
States as coded by FRA (the 50 states, the District of Columbia and Puerto Rico; other territories are not shown)
Units and scope
Crossings and incident reports: counts. People killed and injured: counts as recorded on the reports. Counts of reports to FRA, not rates.
Pages with this table
52
Coverage and interpretation
  • State figures count every inventory record and incident report FRA filed under the state code, including reports with no county code. Puerto Rico is shown as a separate state-level geography; its inventory holds only crossings marked closed, so it shows no open crossing.
  • These are counts of what railroads and agencies reported to the Federal Railroad Administration (FRA), not measures of how safe a crossing, a road or a county is. The inventory is the crossing record railroads and state agencies file on FRA Form 71 and keep current; the incident counts are the highway-rail grade crossing incident reports railroads file on FRA Form 57 whenever a train and a highway user collide. Neither is adjusted for traffic, trains, population or the number of crossings, so a place with more crossings or more traffic can report more incidents for that reason alone.
  • Incidents are reports for calendar years 2021 to 2025, the latest five complete years in the file (2026 reports end 2026-08-31 and are not used). “People killed” and “People injured” are the totals FRA records on each report (highway users, train crew and passengers), as filed; a report can be amended later. A report counts once, at the county where the incident was reported to have happened.
  • Crossings are those the inventory does not mark closed, counted once each. Public and private are the crossing type FRA records; “Other” covers pedestrian pathway crossings and crossings whose type is blank. Inventory records are updated at different times by different filers, so some entries are old.
  • Counties are matched to Census counties by FIPS code only. Where FRA filed a code the Census list no longer has, it was mapped only for a documented rename or absorption (Dade County to Miami-Dade County, Florida; the former Virginia cities of Bedford, Clifton Forge and South Boston into their counties). Where a code was divided, its successors are Not reported. FRA still files Connecticut by its former counties, which the Census planning regions on this site replaced, so Connecticut is Not reported.

Source periods, coverage and other tables · Original publisher: documentation and methods

Railroad grade crossings and reported incidents (FRA)

U.S. Department of Transportation, Federal Railroad Administration · FRA Crossing Inventory Data (Form 71, current) and Highway-Rail Grade Crossing Incident Data (Form 57), 2021 to 2025, as published on data.transportation.gov, aggregated to county counts

Data period
Crossing inventory as filed and published by FRA (retrieved 2026-11-03); incident reports for calendar years 2021 to 2025 (retrieved 2026-11-03)
Geography vintage
Counties as coded by FRA (state FIPS code plus county FIPS code); Connecticut’s former counties cannot be mapped to planning regions
Units and scope
Crossings and incident reports: counts. People killed and injured: counts as recorded on the reports. Counts of reports to FRA, not rates.
Pages with this table
2,897
Coverage and interpretation
  • A county with no record at all in the inventory is Not reported: the inventory cannot tell a county with no crossings from a county whose crossings were not filed. A county with at least one record shows its counts as filed, and zero is a real zero. A county with no inventory record but at least one incident report shows the incident figures only. Puerto Rico’s inventory has no usable county codes, so its counties are Not reported.
  • The county of an incident report is the county on that report, which can differ from the county in the inventory for the same crossing. Reports with no county code are not assigned to any county (two in 2021 to 2025, both in Virginia, and the independent city each names is Not reported for incidents).
  • These are county figures, not figures for any city or place within the county.
  • These are counts of what railroads and agencies reported to the Federal Railroad Administration (FRA), not measures of how safe a crossing, a road or a county is. The inventory is the crossing record railroads and state agencies file on FRA Form 71 and keep current; the incident counts are the highway-rail grade crossing incident reports railroads file on FRA Form 57 whenever a train and a highway user collide. Neither is adjusted for traffic, trains, population or the number of crossings, so a place with more crossings or more traffic can report more incidents for that reason alone.
  • Incidents are reports for calendar years 2021 to 2025, the latest five complete years in the file (2026 reports end 2026-08-31 and are not used). “People killed” and “People injured” are the totals FRA records on each report (highway users, train crew and passengers), as filed; a report can be amended later. A report counts once, at the county where the incident was reported to have happened.
  • Crossings are those the inventory does not mark closed, counted once each. Public and private are the crossing type FRA records; “Other” covers pedestrian pathway crossings and crossings whose type is blank. Inventory records are updated at different times by different filers, so some entries are old.
  • Counties are matched to Census counties by FIPS code only. Where FRA filed a code the Census list no longer has, it was mapped only for a documented rename or absorption (Dade County to Miami-Dade County, Florida; the former Virginia cities of Bedford, Clifton Forge and South Boston into their counties). Where a code was divided, its successors are Not reported. FRA still files Connecticut by its former counties, which the Census planning regions on this site replaced, so Connecticut is Not reported.

Source periods, coverage and other tables · Original publisher: documentation and methods

Recall events by product type and FDA classification, 2021 to 2025, firms in this state

U.S. Food and Drug Administration (openFDA) · openFDA food, drug and device enforcement reports (recall enforcement data)

Data period
Enforcement report dates 2021-01-01 to 2025-12-31, five years combined
Geography vintage
The 50 States, the District of Columbia and Puerto Rico, by the state the source prints for the recalling firm; records for firms outside these (foreign firms, other territories, no state) are not on these pages
Units and scope
Count of distinct recall events over the five years by product type and classification as the FDA publishes it.
Pages with this table
52
Coverage and interpretation
  • These are counts of recall events reported in FDA enforcement reports and classified by the FDA, counted by the state of the firm that carried out the recall, not where the product was sold or where anyone was harmed, and not a measure of how safe a state's products are. A recall event is one FDA event identifier; one event can cover many products and is counted once.
  • The year is the date the recall appeared in the FDA Enforcement Report, which can be later than when the firm began the recall. Classification is the FDA's own: Class I is a reasonable probability of serious adverse health consequences or death, Class II a temporary or medically reversible health consequence, Class III unlikely to cause adverse health consequences. Recalls not yet classified are counted in the product totals only. An event with products in more than one class is counted in each class.
  • Zero is a count of zero events in the complete set of enforcement reports retrieved for that state, year and product type, not a missing value. Large firms and states with many manufacturers have more recalls; counts are not adjusted for the number of firms. Names of firms are not shown.

Source periods, coverage and other tables · Original publisher: documentation and methods

Recreational boating accidents, deaths, injuries and registered vessels (U.S. Coast Guard)

U.S. Coast Guard, Office of Auxiliary and Boating Safety · Recreational Boating Statistics, Table 30 (accident, casualty and damage data by state) and Table 38 (registration data by state)

Data period
Calendar years 2020 to 2024; each year is the figure printed in that year's report
Geography vintage
The 50 States, the District of Columbia and Puerto Rico; other territories and offshore waters are in the source but not on these pages
Units and scope
Counts of reports and persons, vessels registered by the state, and deaths per 100,000 registered vessels as the source prints it.
Pages with this table
52
Coverage and interpretation
  • These are counts of boating accidents reported to state authorities and compiled by the U.S. Coast Guard, not a measure of how safe boating is. Federal rules set minimum reporting requirements and states may set more stringent ones, for example a lower damage threshold, so reporting thresholds differ by state and counts are not strictly comparable between states.
  • Registered vessels are the vessels each state registers, and the scope of the registration system differs by state (some register all watercraft, others only motorized vessels or boats above a length), so registration counts and the published rate are not comparable between states. The source notes that a state's rate can include deaths on unregistered vessels or on vessels registered in another state.
  • For incidents on or after January 1, 2024 the Coast Guard changed some reporting thresholds and definitions (injury, damage and what counts as a vessel), so 2024 is not strictly comparable with earlier years. The 2020 report calls these events accidents; later reports call them incidents.
  • Each year is the figure printed in that year's report and is not revised to later reports. The deaths and injured columns count persons; reported accidents count events. A zero is a count the source prints as zero.

Source periods, coverage and other tables · Original publisher: documentation and methods

State temperature and precipitation, 2025 against 1991-2020 (NOAA)

NOAA National Centers for Environmental Information · nClimDiv statewide monthly temperature and precipitation (climdiv-tmpcst, climdiv-pcpnst)

Data period
Calendar year 2025; comparison is the average of calendar years 1991 to 2020
Geography vintage
The 48 contiguous states, Alaska and Hawaii as defined by NCEI; the source has no District of Columbia or Puerto Rico record
Units and scope
Degrees Fahrenheit and inches of precipitation.
Pages with this table
50
Coverage and interpretation
  • These are area-weighted averages of a 5 kilometre grid interpolated from weather station readings, not a reading at any one station or city. Alaska uses a different base climatology from the other states.
  • The annual temperature is the average of the twelve monthly average temperatures, and the precipitation is the sum of the twelve monthly totals. The comparison is the average of the thirty calendar years 1991 to 2020, NOAA's standard climate normal years. The difference is computed from the unrounded values, so it can differ by 0.1 from the difference of the two rounded figures shown.
  • NOAA revises recent months as late station data arrive. A state with any missing month in a year has no figure for that year, and none for the average if any of its thirty years is incomplete.
  • One year differing from the thirty-year average is weather, not climate change, and is not a forecast.
  • These are statewide figures, not figures for any city, place or county within the state.

Source periods, coverage and other tables · Original publisher: documentation and methods

Storm events, 2016 to 2025 (NOAA)

NOAA National Centers for Environmental Information · Storm Events Database, bulk CSV details files

Data period
Calendar years 2016 through 2025, by event beginning date
Geography vintage
Counties by state and county FIPS code as recorded by NOAA (county-based events only)
Units and scope
Number of events recorded for the county, deaths and injuries (direct plus indirect), and estimated damage in whole U.S. dollars of the year of the event.
Pages with this table
3,105
Coverage and interpretation
  • Storm events are reports, not measurements. NOAA’s National Weather Service offices compile them from spotters, emergency managers, the public, the media and other sources, so a count reflects who was watching and reporting as well as what happened. Reporting practices, the event types recorded and the way damage is estimated have changed over the years and differ between places, and the figures are not uniform enough to rank counties.
  • NOAA records some events by county and others by forecast zone, which is a group of counties or part of one. Only events recorded by county are shown here, and no event recorded by zone is assigned to a county. Of 653,920 events in the 10 years, 368,604 were recorded by county and 285,316 (44%) were recorded by zone, or labeled as county events whose code and name did not agree (4,376), and are not in any county figure. Zone events include many winter weather, heat, cold, wind and drought events, so a county's table is not a complete account of its weather; a county with no county-recorded event is shown with zeros but may have had zone events.
  • Deaths and injuries add NOAA’s direct and indirect counts. Property and crop damage are NOAA’s estimates, read from its dollar values with thousand, million and billion suffixes, in dollars of the year of the event, and are often rough or missing; a blank is counted as zero. When one event affected several counties NOAA records a row for each county and splits its damage and casualties between them.
  • NOAA’s event types are listed as NOAA names them; this site has not grouped or renamed them. The six types with the most events in the county are shown, and all remaining types are summed in one row.
  • The Storm Events Database for the most recent years is revised as NOAA completes quality control, and the latest year can still change.
  • NOAA and the National Weather Service identify counties by FIPS codes in these files. Connecticut’s former counties are not the planning regions on this site, so Connecticut is omitted. Where NOAA uses a county code that is not in this site’s county list, a county with no events in that state is Not reported rather than zero.
  • These are county figures, not figures for any city or place within the county.

Source periods, coverage and other tables · Original publisher: documentation and methods

Superfund National Priorities List sites (EPA)

U.S. Environmental Protection Agency, Superfund · Superfund National Priorities List (NPL) Sites with Status Information

Data period
Listing status as published by EPA (retrieved 2026-11-03)
Geography vintage
Counties as named by EPA, matched to Census counties by state and county name
Units and scope
Number of sites on the National Priorities List.
Pages with this table
3,179
Coverage and interpretation
  • The National Priorities List is EPA’s list of sites of national priority among known releases or threatened releases of hazardous substances. A listing is a regulatory status. It is not a statement about current exposure, health risk, water or air quality, or whether anyone is affected, and a county with no listed site is not certified free of contamination.
  • Final means the site is currently on the list, proposed means EPA has proposed adding it, and deleted means EPA removed it after determining that no further response is required. Deleted sites are shown because they were listed; they are not current priorities.
  • A site that spans more than one county is counted in each county EPA names for it, so county counts do not add up to the national total and one site can appear on several county pages.
  • EPA names counties in text without a county code. Counties were matched to Census counties by exact state and county name, ignoring case, accents, punctuation and spacing; Virginia’s independent cities and Puerto Rico’s municipios were matched to the name with “city” or “municipio” added. Where a county and an independent city share a name, EPA’s bare name is resolved only when EPA’s Facility Registry county code names one of the two; otherwise both are Not reported (Maryland and Missouri have such cases). Where a listed county could not be matched at all (Connecticut, which EPA still reports by its former counties, and Alaska), every county count in that state is Not reported rather than zero.
  • The list is complete as published, so a matched county with no listed site is shown as zero. Federal facility sites are included. Sites in territories outside the Census county list are not shown.
  • These are county figures, not figures for any city or place within the county.

Source periods, coverage and other tables · Original publisher: documentation and methods

Temperature and precipitation, 2025 against 1991-2020 (NOAA)

NOAA National Centers for Environmental Information · nClimDiv county monthly temperature and precipitation (climdiv-tmpccy, climdiv-pcpncy)

Data period
Calendar year 2025; comparison is the average of calendar years 1991 to 2020
Geography vintage
Counties of the 48 contiguous states, Alaska and Hawaii as defined by NCEI
Units and scope
Degrees Fahrenheit and inches of precipitation.
Pages with this table
3,133
Coverage and interpretation
  • These are area-weighted averages of a 5 kilometre grid interpolated from weather station readings, not a reading at any one station or city.
  • The annual temperature is the average of the twelve monthly average temperatures, and the precipitation is the sum of the twelve monthly totals. The comparison is the average of the thirty calendar years 1991 to 2020, which are NOAA's standard climate normal years.
  • NOAA revises recent months as late station data arrive. A county with any missing month in a year has no figure for that year.
  • One year differing from the thirty-year average is weather, not climate change, and is not a forecast.
  • These are county figures, not figures for any city or place within the county.

Source periods, coverage and other tables · Original publisher: documentation and methods

Toxics Release Inventory facilities and on-site releases (EPA)

U.S. Environmental Protection Agency, Toxics Release Inventory Program · Toxics Release Inventory (TRI) basic data file, 2024, U.S.

Data period
Data year 2024, the newest year in EPA’s basic data files (retrieved 2026-11-03)
Geography vintage
Counties as named in the TRI file, matched to Census counties by state and county name
Units and scope
Facilities: count. Releases: pounds reported released on site in the year (dioxin and dioxin-like compounds, reported in grams, excluded).
Pages with this table
3,099
Coverage and interpretation
  • These are facility-reported quantities, not exposure, risk or health effects. A pound released on site is a quantity a facility reported to EPA; it does not say how toxic the chemical is at the amount released, where it went, or whether anyone is exposed.
  • Only facilities that meet the program’s reporting thresholds report (covered industry sectors, ten or more full-time employees, and manufacture, process or use of a listed chemical above its threshold). “No facilities in the file” means no facility above those thresholds was found in the county, not that nothing is released there. Small facilities and sources outside the covered sectors do not report.
  • Facilities are distinct TRI facility identifiers that filed at least one chemical form (Form R or the short Form A) for the data year. Total on-site releases are the sum of the file’s “on-site release total” column over each facility’s chemical forms in pounds: releases to air (fugitive and stack), water, underground injection, landfills, land treatment, surface impoundments and other disposal at the facility. Transfers to other locations, to treatment plants and for recycling or energy recovery are not included. Form A filings report no release quantities.
  • Dioxin and dioxin-like compounds are reported in grams, not pounds; they are left out of the pound totals and the chemical shown, though their facilities are counted. The chemical shown is the one with the largest summed on-site pounds in the county (ties by name), which can be a very large release of a low-toxicity substance.
  • EPA names counties in text without a county code. Counties were matched to Census counties by exact state and county name, ignoring case, accents, punctuation and spacing, with “city” tried for Virginia and “municipio” for Puerto Rico. Where a bare name could be a county or a same-named independent city, both are Not reported. Where a name in a state could not be matched at all, every county in that state without a matched facility is Not reported rather than shown as having none. Connecticut is omitted: EPA files use the former Connecticut counties, which the Census planning regions replaced.
  • Facilities may revise earlier submissions, and the national file is updated over time. These are county figures, not figures for any city or place within the county.

Source periods, coverage and other tables · Original publisher: documentation and methods

Water withdrawals, estimate for 2015 (USGS; not current water use)

U.S. Geological Survey (Water Resources Mission Area) · Estimated use of water in the United States, county-level data for 2015 (ver. 2.0)

Data period
Calendar year 2015 (USGS five-year compilation, data release version 2.0 of June 2018; retrieved 2026-11-03)
Geography vintage
Counties and county equivalents as USGS coded them for 2015 (FIPS); Connecticut’s eight former counties and Alaska’s Valdez-Cordova Census Area are not matched to this site’s planning regions or successor areas
Units and scope
Million gallons per day (Mgal/d) withdrawn, averaged over the year, with the groundwater and surface-water split; fresh water only for domestic, irrigation and livestock.
Pages with this table
3,263
Coverage and interpretation
  • These figures are USGS estimates for 2015, not current water use. They come from the most recent USGS county-level compilation of all eight use categories (the data behind USGS Circular 1441, published in 2018); USGS has since released some 2020-and-later estimates by modeling and by single category, which are different products and are not combined here. Water use in a county has likely changed since 2015.
  • The figures are withdrawals: water taken from the ground or from a river, lake or other surface source, in million gallons per day averaged over the year. A withdrawal is not the amount delivered to a user and not the amount consumed. Public supply withdrawals are counted in the county where the supplier withdraws the water, which can differ from the county where people use it, and a utility’s water can cross county lines. Domestic is self-supplied water only (private wells and similar); water delivered by public suppliers to homes is in public supply.
  • Domestic, irrigation and livestock are fresh water only in the source. Public supply, industrial, aquaculture, mining, thermoelectric power and the all-categories row include saline water where the source reports it, so the groundwater and surface-water columns add to the total in every row. A 0.00 is a quantity USGS estimated as zero; a value the source does not report is shown as Not reported, never as zero. The source reports every figure shown here for every county, so none is Not reported in this release.
  • Connecticut is shown by the nine Census planning regions that replaced its eight counties. The USGS data are for the eight former counties (FIPS 09001 to 09015) and there is no published crosswalk for withdrawals, so no Connecticut planning region has a table and nothing is allocated to one; the Connecticut state figures are the sum of the eight former counties, which cover the whole state. Likewise Alaska’s Valdez-Cordova Census Area (02261), which was later split into two areas, is not allocated to them; those two areas have no table and Alaska’s state figures include it. The U.S. Virgin Islands are outside this site’s state list and are not shown.
  • State figures are the sum of that state’s county rows in the same USGS file. USGS rounds each county figure to two or three decimals, so a sum can differ from a state total USGS publishes elsewhere by a few thousandths. Areas join to this site’s counties and states by FIPS code only; no name matching is used. These are county figures and state figures; nothing is allocated to cities or places.
  • The estimates were compiled by USGS from state and local reports, permit data and its own estimation methods, which differ by state and category; they are estimates, not meter readings, and USGS notes that categories and methods changed between five-year compilations, so comparing this year with other years needs care. The data release is marked CC0 1.0 and USGS-authored data are in the U.S. public domain; it is shown with USGS’s citation. Export, API and commercial use have not been reviewed.

Source periods, coverage and other tables · Original publisher: documentation and methods

Explore environment and hazards in a county

Choose a county below to open a supported table on its profile. Dates and coverage vary by source.

Choose a state or territory

Keep this topic while choosing any available county or county equivalent. Counts below are profiles; each county's table availability is shown on the next page.

Reviewed county examples

Browse all supported county profiles. The broader directory opens a profile; the state selector and example links above preserve the selected topic.

Before comparing places

No monitor does not mean clean air. A listed site is not a measured personal exposure; a declaration is not a forecast. County summaries and modeled risk do not establish risk at an address.

Use the same definition, geography and period. Read uncertainty and missing-value notes beside the observation. A topic label alone does not establish that two figures are comparable.

Choose a state · Choose a county · All topics

These guides are under editorial review. Source display permission does not grant downloads, redistribution or commercial use. Public-source discovery is not approval for a new use.