ArthritisMap

How the ArthritisMap Climate Index works

One number from 0 to 100, and one to five stars, for how easy a city’s typical weather is on arthritic joints. This page shows the whole calculation, the research it leans on and what it cannot tell you.

The short version

  • We measure ten features of a city’s climate, using long-term averages from NOAA weather stations.
  • Each feature is turned into a comfort score from 0 to 100, where 100 is the easiest condition found at major U.S. stations and 0 is the hardest.
  • The ten comfort scores are averaged using weights that reflect how well research ties each feature to joint pain. Humidity, pressure swings and cold make up 57% of the rating.
  • The result is rounded and turned into stars: 75 and up is five stars, below 30 is one.

Read this first. The index describes weather, not your body. People differ widely in how much weather affects their pain, and some are not affected at all. Nothing here is medical advice; see the medical disclaimer.

The formula

FactorWhat we measureWeight100 points at0 points at
HumidityAverage relative humidity (mean of morning and afternoon readings)24%30 %82 %
Pressure swingsAverage day-to-day change in sea-level barometric pressure18%1 hPa6.5 hPa
ColdHeating degree days per year (base 65°F)15%0 degree days10,000 degree days
Cold, damp monthsMonths that are both cold and humid, counted in month-equivalents10%0 months6 months
WindAverage wind speed7%5 mph13 mph
Rainy daysDays per year with measurable rain or snow (0.01 inch or more)6%30 days180 days
Temperature swingsAverage day-to-day change in daily mean temperature6%1 °F6.5 °F
Cloud coverCloudy days per year (sky at least 8/10 covered)6%60 days230 days
Snow and iceAverage snowfall per year4%0 inches100 inches
Extreme heatDays per year reaching 90°F or hotter4%10 days150 days

For each factor:

comfort score = 100 × (worst − measured value) ÷ (worst − best), held between 0 and 100

Then:

Climate Index = sum of (weight × comfort score) ÷ 100

The “cold, damp months” factor needs one more step. Each month gets a coldness value (0 at an average of 55°F or warmer, rising to 1 at 30°F or colder) and a dampness value (0 at 55% average humidity or less, rising to 1 at 80% or more). The two are multiplied, and the twelve results are added. A city with four months that are fully cold and fully damp scores 4.

Stars

RatingIndexIn a wordCities
75 to 100Excellent55
60 to 74Good91
45 to 59Mixed148
30 to 44Tough111
0 to 29Harsh11

Worked example: Phoenix, AZ

FactorMeasuredComfort scoreWeightPoints
Humidity35%9124%21.9
Pressure swings2.3 hPa7718%13.8
Cold9359115%13.6
Cold, damp months0.010010%10.0
Wind6.1 mph867%6.0
Rainy days35976%5.8
Temperature swings3.6°F526%3.1
Cloud cover70946%5.6
Snow and ice0.0 in1004%4.0
Extreme heat16904%0.0
Climate Index83.9

83.9 rounds to 84, which is 5 of 5 stars. See the Phoenix page.

Why these factors, and why these weights

Researchers have studied weather and joint pain for decades without settling it. The authors of the largest study to date wrote that there remains no scientific consensus1. So the weights below are our reading of the evidence, heaviest where several good studies point the same way and light where the support is thin or practical rather than clinical. They are a judgment, and we show them so you can disagree.

Humidity, 24%
The best-supported factor. In a U.K. study that tracked more than 2,600 people with long-term pain for up to 15 months, higher relative humidity was the weather condition most strongly tied to painful days1. A study of older adults with osteoarthritis in six European countries found the same link3, and a 2023 meta-analysis found humidity correlated with osteoarthritis pain2.
Pressure swings, 18%
Barometric pressure shows up in most positive studies, but they disagree about direction: the U.K. study found more pain when pressure was lower1, while a U.S. study of knee osteoarthritis found pain tracked changes in pressure4. Because the common thread is pressure that moves, we score how much it changes from one day to the next, not how high or low it sits.
Cold, 15%
Colder temperatures were independently linked to worse knee osteoarthritis pain in the U.S. study4, and the 2023 meta-analysis found pain rose as temperature fell2. Heating degree days capture both how cold a place gets and for how long.
Cold, damp months, 10%
The European study found humidity affected pain more in colder weather3, which matches what many people with arthritis report. This factor gives extra weight to months that are both.
Wind, 7%
Stronger wind was linked to more painful days in the U.K. study, though less strongly than humidity or pressure1.
Rainy days, 6%
A deliberately small weight. An analysis of more than 11 million Medicare outpatient visits found no relationship between rainy days and visits for joint or back pain5. We keep rain in the index because wet days tend to arrive with humidity and shifting pressure and make it harder to stay active.
Temperature swings, 6%
Often reported by people with arthritis, but thinly supported: the European study found that changes between consecutive days were not significantly tied to pain3. It counts for little.
Cloud cover, 6%
None of the studies we rely on ties cloud itself to joint pain. We include it at a low weight as a practical measure of long gray stretches, which discourage activity.
Snow and ice, 4%
A practical factor, not a pain trigger: snow and ice raise the risk of falls and make it harder to get out and move.
Extreme heat, 4%
Also practical. Very hot days limit outdoor activity. Heat counts lightly, so some desert cities score well in summer months that many people would still find hard going.

Month-by-month scores

City pages show a score for each month. It uses the same weights and anchors, applied to that month’s normal weather. Counts such as rainy days are multiplied by twelve so the yearly anchors apply, and cold is measured by how far the month’s average temperature falls below 65°F.

Where the numbers come from

  • Humidity, wind, cloud cover, wet days, snowfall and hot days are long-term station averages from NOAA’s Comparative Climatic Data tables, 2018 edition7. Records at most stations run for several decades. One exception: at Alaska stations NOAA tabulates days of 70°F or warmer in place of 90°F, so hot days there are counted as zero.
  • Temperature and heating degree days are NOAA’s 1981 to 2010 climate normals, from the same publication.
  • Pressure swings and temperature swings are calculated from the NCEP/NCAR Reanalysis, a gridded reconstruction of past weather from NOAA’s Physical Sciences Laboratory, and interpolated to each station. Pressure uses one year of daily sea-level pressure fields (1963) and temperature uses two years (2013 and 2014), on a grid of 2.5 degrees. Storm tracks change little from year to year, so the pattern across the country is stable, but this is the coarsest input to the index and the first we plan to upgrade to 30-year station records.
  • City populations are U.S. Census Bureau estimates for July 1, 20258.

The data set behind this version was built in November 2026.

Which cities are rated

We rate 416 cities using 207 weather stations: every incorporated place of 100,000 or more people with a station within 60 miles, plus smaller cities that host one of NOAA’s first-order stations, so that every state is covered.

  • Each city uses its nearest station. A suburb usually shares its metro area’s airport station, so neighboring cities often have identical scores.
  • When the station is more than 35 miles away, the city page says so and labels the rating a regional estimate. Hills, elevation and distance from the coast can change conditions over that distance.
  • Nine cities of 100,000 or more are not rated because no suitable station lies within 60 miles: Bend, College Station, Hesperia, Laredo, Salinas, St. George, Tyler, Victorville and Yuma.
  • At a few stations NOAA’s tables lack one measure, usually cloud cover. That one value is taken from the nearest station that has it, and the city page says which.

What the index does not measure

  • You. Sensitivity to weather varies from person to person and by type of arthritis. A five-star city will not help everyone, and plenty of people do well in one-star cities.
  • Air quality. A 2026 study linked fine-particle pollution to higher disease activity and flares in rheumatoid arthritis6. Air quality is not in this version; we plan to add it.
  • Health care. How many rheumatologists practice nearby, and whether they take your insurance, may matter more than the weather.
  • Indoor life. Heating, air conditioning and dehumidifiers change the air you actually spend most of your time in.
  • Trends. Climate normals look backward. They do not capture how a place is changing.

Sources

  1. Dixon WG, et al. How the weather affects the pain of citizen scientists using a smartphone app. npj Digital Medicine, 2019. Link
  2. Wang L, Xu Q, Chen Y, Zhu Z, Cao Y. Associations between weather conditions and osteoarthritis pain: a systematic review and meta-analysis. Annals of Medicine, 2023. Link
  3. Timmermans EJ, Schaap LA, Herbolsheimer F, et al. The influence of weather conditions on joint pain in older people with osteoarthritis: results from the European Project on OSteoArthritis. The Journal of Rheumatology, 2015. Link
  4. McAlindon T, Formica M, Schmid CH, Fletcher J. Changes in barometric pressure and ambient temperature influence osteoarthritis pain. The American Journal of Medicine, 2007. Link
  5. Jena AB, Olenski AR, et al. Association between rainfall and diagnoses of joint or back pain: retrospective claims analysis. BMJ, 2017. Link
  6. Air pollution linked to spikes in rheumatoid arthritis activity and flare risk (study in Annals of the Rheumatic Diseases, 2026). Elsevier press release. Link
  7. NOAA National Centers for Environmental Information. Comparative Climatic Data (CCD-2018), including 1981 to 2010 climate normals. Link
  8. U.S. Census Bureau. City and Town Population Totals (July 1, 2025 estimates, taken from the table in Wikipedia's "List of United States cities by population"). Link

Spotted an error, or have a study we should weigh? Tell us.