Constanța · summers of 1961–2025
Rain after
a dry month
At Constanța, after 30 days without rain above 5 mm per day, the threshold was exceeded within the next 7 days in 12 of 64 summer episodes. Explore the record and compare stations.
Constanța · June–August · 1961–2025
A “dry” month can include light rain. Here we follow gaps between days with rain above 5 mm; agricultural drought also depends on water in the soil.
Historical observations. For the coming days, consult a weather forecast.
Historical calculator
When did rain return at your chosen station?
Choose a station and the months you want to examine. To compare earlier and more recent summers, change the observation period.
Constanța · 1961–2025 · Summer · June–August
18.8%
with precipitation above 5 mm in the next 7 days
12 out of 64 observed episodes, 51 years.
95% uncertainty interval: 10.5% – 27.8%
The same episodes, four waiting periods
Compare with usual rainfall frequency
Reference for the same calendar months 31.5%Result − reference-12.7 pp · 95%: -21.1 … -3.6 pp
Reference for the same month and year 14.3%Result − reference+4.5 pp · 95%: -2.8 … +12.3 pp
The first reference uses the same calendar months across all years in the period. The second also keeps each episode’s year. Frequencies are weighted to match the selected episodes, regardless of the preceding dry duration. The difference shows how the choice of comparison matters; neither reference measures an effect of prayer.
Full method, definitions and limitationsDoes a longer dry spell raise tomorrow’s chance?
Here the waiting period stays at one day. The chart compares episodes that reached different dry durations. Bars show uncertainty; gaps without points have too few observations.
Read the curve values and denominators
| Duration | Episodes | Probability | 95% uncertainty interval |
|---|---|---|---|
| 0 | 49 / 368 | 13.3% | 9.8% – 16.9% |
| 3 | 17 / 275 | 6.2% | 3.1% – 9.4% |
| 7 | 11 / 209 | 5.3% | 2.7% – 7.9% |
| 14 | 5 / 142 | 3.5% | 0.7% – 6.8% |
| 21 | 4 / 90 | 4.4% | 1.1% – 8.8% |
| 30 | 1 / 64 | 1.6% | 0.0% – 5.3% |
| 45 | 0 / 31 | 0.0% | not estimated |
| 60 | 0 / 11 | not estimated | not estimated |
| 90 | 0 / 3 | not estimated | not estimated |
A longer wait gives rain more chances to arrive.
In the Constanța example, rain above 5 mm arrived in 12 of 64 episodes within one week and in 28 of 64 within two weeks. This increase does not imply that each day’s chance rose as the dry spell lengthened. The chart above lets you examine that separate question.
In the Constanța record
The summers behind the percentage
Each row marks the day when a spell reached 30 consecutive days without precipitation above 5 mm. These are all 64 episodes with that date in June–August, during 1961–2025. The six most recent appear below; the complete list includes episodes with no day above the threshold in the following 28 days.
Across these 30-day windows, total precipitation ranged from 0 to 25.3 mm. Half the windows had at most 4.6 mm. Only one window had no precipitation at all.
| At 30 dry days | Next rain >5 mm | After another |
|---|---|---|
| 2025-09-139.2 mm | 27 days | |
| None within 28 days | — | |
| 2024-07-2616.9 mm | 20 days | |
| 2024-06-0618 mm | 4 days | |
| None within 28 days | — | |
| 2023-07-046 mm | 26 days |
See all 64 episodes in chronological order
| At 30 dry days | Previous rain >5 mm | Next rain >5 mm | After another |
|---|---|---|---|
| 1961-06-16 | None within 28 days | — | |
| 1962-05-12 | None within 28 days | — | |
| 1963-07-04 | 1963-08-1020.8 mm | 7 days | |
| 1965-05-15 | 1965-07-1144.4 mm | 27 days | |
| 1965-07-14 | 1965-08-166.7 mm | 3 days | |
| 1966-07-03 | 1966-08-127.8 mm | 10 days | |
| 1970-05-25 | 1970-07-188.6 mm | 24 days | |
| 1970-07-18 | 1970-08-235.5 mm | 6 days | |
| 1971-05-31 | 1971-07-0210.6 mm | 2 days | |
| 1971-07-02 | None within 28 days | — | |
| 1973-07-26 | 1973-09-115.1 mm | 17 days | |
| 1974-07-08 | 1974-08-1215.2 mm | 5 days | |
| 1975-07-09 | None within 28 days | — | |
| 1976-06-08 | 1976-07-1711.5 mm | 9 days | |
| 1977-07-04 | 1977-08-2527.8 mm | 22 days | |
| 1979-05-13 | 1979-06-246.9 mm | 12 days | |
| 1981-05-09 | 1981-06-2523.5 mm | 17 days | |
| 1981-07-01 | 1981-08-235.3 mm | 23 days | |
| 1984-05-14 | 1984-06-256.9 mm | 12 days | |
| 1985-06-30 | 1985-08-037 mm | 4 days | |
| 1986-07-18 | None within 28 days | — | |
| 1988-07-12 | 1988-09-078 mm | 27 days | |
| 1989-07-19 | 1989-09-0517.4 mm | 18 days | |
| 1990-06-25 | None within 28 days | — | |
| 1992-05-13 | 1992-07-019.9 mm | 19 days | |
| 1992-07-19 | None within 28 days | — | |
| 1993-07-13 | 1993-08-2913.6 mm | 17 days | |
| 1994-07-12 | 1994-08-2414.9 mm | 13 days | |
| 1995-05-21 | 1995-06-2620.8 mm | 6 days | |
| 1995-06-28 | None within 28 days | — | |
| 1996-05-23 | 1996-07-1719.3 mm | 25 days | |
| 1996-07-26 | 1996-09-076.3 mm | 13 days | |
| 1998-07-17 | 1998-08-236.5 mm | 7 days | |
| 1999-07-03 | 1999-08-0345.9 mm | 1 day | |
| 2000-06-27 | None within 28 days | — | |
| 2001-06-21 | None within 28 days | — | |
| 2002-06-04 | 2002-07-216.9 mm | 17 days | |
| 2003-07-31 | 2003-09-1128.8 mm | 12 days | |
| 2005-07-16 | None within 28 days | — | |
| 2006-07-01 | 2006-08-1115.8 mm | 11 days | |
| 2007-06-28 | 2007-08-0734.4 mm | 10 days | |
| 2008-05-12 | 2008-06-1933.2 mm | 8 days | |
| 2008-07-16 | 2008-09-0912 mm | 25 days | |
| 2009-05-31 | 2009-07-1028.8 mm | 10 days | |
| 2009-07-26 | 2009-09-0615.6 mm | 12 days | |
| 2010-05-21 | 2010-06-2213.6 mm | 2 days | |
| 2010-07-28 | None within 28 days | — | |
| 2011-07-21 | None within 28 days | — | |
| 2012-05-29 | None within 28 days | — | |
| 2014-06-26 | 2014-08-1796.4 mm | 22 days | |
| 2015-05-27 | 2015-07-0520 mm | 9 days | |
| 2016-06-14 | None within 28 days | — | |
| 2017-07-28 | 2017-08-2913.4 mm | 2 days | |
| 2018-07-29 | None within 28 days | — | |
| 2019-05-22 | 2019-07-0318.2 mm | 12 days | |
| 2020-07-08 | 2020-08-217.2 mm | 14 days | |
| 2021-07-07 | None within 28 days | — | |
| 2022-06-09 | 2022-07-1710.6 mm | 8 days | |
| 2023-05-09 | 2023-07-046 mm | 26 days | |
| 2023-07-28 | None within 28 days | — | |
| 2024-05-03 | 2024-06-0618 mm | 4 days | |
| 2024-06-06 | 2024-07-2616.9 mm | 20 days | |
| 2025-05-17 | None within 28 days | — | |
| 2025-07-18 | 2025-09-139.2 mm | 27 days |
This is the fixed Constanța example from the opening. Season applies to the date when the 30-day duration was reached; later rain may fall in September. Missing observations are never counted as dry days.
Download the episode list · JSON ↓Observations in Romania
Compare stations under the same conditions.
The question stays fixed: after 30 days with at most 5 mm per day, did a day above the threshold occur in the next 7 days? Starting dates are in June–August, during 1961–2025.
Natural Earth · 1:110m
The range above covers stations with enough episodes. Băneasa has only 28, so its percentage is withheld. Stations describe measurement sites; their values do not form a probability for the whole country.
See results and episode counts at all 23 stations
| Weather station | Probability | Episodes | 95% uncertainty interval |
|---|---|---|---|
| Arad | not estimated | 14 / 27 | not estimated |
| Bacău | not estimated | 9 / 18 | not estimated |
| Botoșani | not estimated | 4 / 10 | not estimated |
| București Băneasa | not estimated | 14 / 28 | not estimated |
| Buzău | not estimated | 10 / 21 | not estimated |
| Caransebeș | not estimated | 9 / 12 | not estimated |
| Ceahlău Toaca | not estimated | 1 / 2 | not estimated |
| Cluj-Napoca | not estimated | 6 / 13 | not estimated |
| Constanța | 18.8% | 12 / 64 | 10.5% – 27.8% |
| Craiova | 30.0% | 9 / 30 | 13.5% – 47.6% |
| Călărași | 32.4% | 11 / 34 | 14.3% – 50.0% |
| Deva | not estimated | 17 / 23 | not estimated |
| Drobeta-Turnu Severin | 33.3% | 13 / 39 | 18.2% – 48.7% |
| Galați | 28.6% | 12 / 42 | 15.9% – 43.2% |
| Iași | not estimated | 18 / 28 | not estimated |
| Miercurea Ciuc | not estimated | 3 / 6 | not estimated |
| Ocna Șugatag | not estimated | 4 / 7 | not estimated |
| Roșiori de Vede | 33.3% | 13 / 39 | 17.9% – 48.4% |
| Râmnicu Vâlcea | not estimated | 7 / 14 | not estimated |
| Sibiu | not estimated | 6 / 15 | not estimated |
| Sulina | 23.0% | 17 / 74 | 13.6% – 32.5% |
| Tulcea | 25.9% | 14 / 54 | 13.0% – 38.1% |
| Vârful Omu | not estimated | 1 / 2 | not estimated |
Which places are covered?
Of the 30 Romanian precipitation stations in NOAA’s inventory, 23 have sufficient coverage in 1961–2025. That does not guarantee enough episodes for every duration and season. The map shows stations, without delimiting drought-affected areas.
Inspect annual coverage and exclusionsRain and ritual
When weather can seem to answer a prayer
Anthropologist Radu Umbreș’s post about Praying for Rain prompted this analysis. The study asks when weather conditions can sustain belief in the effectiveness of rainmaking rituals. Read Radu Umbreș’s post ↗
Espín-Sánchez, Gil-Guirado and Ryan examine prayers in Murcia and practices recorded in ethnographic sources worldwide. In the raw comparison in Appendix D11, rainmaking is recorded for 44% of groups where estimated daily probability rises after very long dry intervals, compared with 30% in the comparison group.
That increase concerns the end of a fitted curve, near the local 99th percentile of intervals between wet days. It does not mean that every dry day brings a higher probability. The Romanian calculator shows observed frequencies and their uncertainty; it does not reproduce the paper’s model or cultural analysis.
What we can say about Romania
Romania’s national inventory of intangible cultural heritage documents Caloian and Paparuda. Dated Orthodox rain processions are also recorded, for example at Suceava on 9 August 2015. These sources attest to practices with different histories; weather records alone cannot explain why people took part or whether a ritual had any effect.
Heritage inventory, page 24The reported Suceava procession
A test of the social hypothesis would need dates and locations of rituals, including those not followed by rain, and a definition of “success” settled before making the comparison.
How the Romanian connection could be studied ↓Open method
How episodes are counted
We divide episodes followed by at least one day above the threshold by all eligible episodes. All four waiting periods use the same cases, with 28 complete days of subsequent observations. Each episode’s history starts at a known day above the threshold.
Uncertainty intervals come from resampling calendar years 1,000 times. They are approximate and do not include every source of error. Percentages require at least 30 episodes across 10 years; counts remain visible below that threshold.
Full method, definitions and limitations ↗
NOAA GHCN-DailySources and claim ledger
Download the data and code
Synthetic review and numerical checks ↗
NOAA data are public; retain source attribution. This project's calculations and text are available under CC BY 4.0.