Floods of large rivers
A river’s flow integrates rain, snow and the state of the soils of a whole basin. Marker colour places today’s flow among those of the same dates over the whole record: dark blue, among the highest 5%; red, among the lowest 5%.
- very low (≤ 5th percentile)
- low (5th–25th)
- normal (25th–75th)
- high (75th–95th)
- very high (≥ 95th)
USGS and Hub’Eau gauging stations; colour: percentile of today’s flow among the same calendar days (15-day window).
Garonne at Tonneins O900001002 · 1989-01-01 → 2026-10-24 · 37 complete water years
Show the data as a table
Show the data as a table
| # | Day of the maximum | m³/s | Estimated return (yr) |
|---|---|---|---|
| 1 | 2026-02-14 | 6,124 | 34 |
| 2 | 2021-02-03 | 6,034 | 29 |
| 3 | 2019-12-15 | 5,279 | 11 |
| 4 | 2009-01-25 | 5,175 | 10 |
| 5 | 2003-02-06 | 4,699 | 6 |
| 6 | 2014-01-27 | 4,676 | 6 |
| 7 | 1993-12-27 | 4,653 | 6 |
| 8 | 2022-01-12 | 4,627 | 6 |
| 9 | 2003-12-05 | 4,581 | 5 |
| 10 | 1995-02-27 | 4,545 | 5 |
This page follows the river floods of large basins: they rise over several days with lasting rain and snowmelt. Flash floods, on small fast basins, last a few hours and escape daily means. Coastal flooding comes from the sea (storm surge, tide).
Observed flow tells neither the flooded extent nor the forecast: for France, see Vigicrues; maps of flooded areas come from the activations of the Copernicus Emergency Management Service, and global forecasts from GloFAS (Copernicus key pending).
| River | Latest flow (m³/s) | Percentile | 10-year (m³/s) | 100-year (m³/s) | Record (m³/s) |
|---|---|---|---|---|---|
| Mississippi at St. Louis 1861– | 4,729 (10-24) | 78 | 21,751 | 28,998 | 29,733 (1993) |
| Missouri at Hermann 1928– | 1,733 (10-24) | 62 | 12,842 | 19,230 | 20,926 (1993) |
| Columbia at The Dalles 1878– | 2,291 (10-24) | 15 | 20,914 | 28,469 | 34,830 (1894) |
| Susquehanna at Harrisburg 1890– | 334 (10-24) | 53 | 11,504 | 19,758 | 27,014 (1972) |
| Rio Grande at Embudo 1889– | 13 (10-24) | 69 | 230 | 450 | 450 (1903) |
| Seine at Paris (Austerlitz) 2006– | 105 (10-24) | 4 | — | — | 1,746 (2016) |
| Loire at Montjean-sur-Loire 1900– | 134 (10-24) | 6 | 5,031 | 6,431 | 6,300 (1910) |
| Rhône at Tarascon 1900– | 1,101 (07-15) | 13 | 8,194 | 10,486 | 10,047 (1935) |
| Garonne at Tonneins 1989– | 208 (10-24) | 49 | 5,186 | 6,794 | 6,124 (2026) |
Daily mean flows from the USGS (Water Services, daily values, ft³/s converted) and Hub’Eau (French national hydrometric database, QmnJ elaborated observations, L/s converted), with their status: provisional or non-validated values flagged. Maximum of each water year (October to September, the USGS convention), a year being kept only with at least 330 days.
Generalised extreme value (GEV) distribution fitted to the annual maxima by L-moments (Hosking, 1990), with the shape approximation of Hosking, Wallis and Wood (1985), only from 30 years; return levels x(T) and return periods of the observed floods. Today’s flow is placed among the flows of the same calendar days (15-day window) over the whole record.
Dams and river works change floods: the Missouri and the Columbia are heavily regulated since the 1930s to 1970s, and the fitted distribution mixes two regimes. Old series (Loire, Rhône) rest on rating curves revised over time and contain values qualified as doubtful.
A “100-year” return period is not a flood every hundred years but a 1% annual probability, assuming an unchanged climate and basin; extrapolated beyond the length of the record, it stays very uncertain. This is not a warning service (France: Vigicrues).
- Daily mean discharge of four French rivers (Hub'Eau, hydrométrie)
-
Hub'Eau, API Hydrométrie (observations élaborées, débit moyen journalier QmnJ), données de la base hydrométrique nationale (HydroPortail), ministère de la Transition écologique.
· source page ↗
Licence : Licence Ouverte / Open Licence 2.0 (Etalab): free reuse with the source mentioned · retrieved on Oct 25, 2026 · SHA-256 6e69c4baeea0… - Hosking (1990)
-
Hosking, J. R. M. (1990). L-moments: analysis and estimation of distributions using linear combinations of order statistics. Journal of the Royal Statistical Society, Series B, 52(1), 105–124.
· doi:10.1111/j.2517-6161.1990.tb01775.x
Licence : Published literature; cited - Hosking et al. (1985)
-
Hosking, J. R. M., Wallis, J. R., & Wood, E. F. (1985). Estimation of the generalized extreme-value distribution by the method of probability-weighted moments. Technometrics, 27(3), 251–261.
· doi:10.1080/00401706.1985.10488049
Licence : Published literature; cited - Hosking et al. (1997)
-
Hosking, J. R. M., & Wallis, J. R. (1997). Regional Frequency Analysis: An Approach Based on L-Moments. Cambridge University Press.
· doi:10.1017/CBO9780511529443
Licence : Published literature; cited