# How River Floods Work > Since mid-September 2026 floods have reached 51 Thai provinces. This film explains the fundamentals behind every river flood, using the Chao Phraya basin that covers a third of Thailand: where monsoon rain comes from, why soaked soil sends water to the river, how four rivers become one funnel, why dams cannot both store and protect, why a sinking delta city fears the tide, and what a “100-year flood” really means. Soak the soil, send a storm down the basin, run a dam through a wet season and roll thirty years of odds yourself. Interactive film: https://www.planetpast.com/stories/how-river-floods-work/ Publisher: PlanetPast (https://www.planetpast.com/) · d => T('readPublished', l, { d: dateIn(d, l) }) 2026-10-16 · Earth science, Society, History ## Prologue ### The funnel to the sea *A flood is a chain, not an event* Follow a raindrop from the mountains to the Gulf of Thailand. Water that fell on mountains weeks ago is now standing in rice fields and city streets. Since mid-September 2026, floods have reached fifty-one Thai provinces, and by the twelfth of October seventy-one people had died. Floods like this are not bad luck. They follow rules: where rain falls, what the ground does with it, and how a whole country's water finds its way to the sea. This year's floods spread across several river basins. This film follows the largest of them, the Chao Phraya, link by link, from the clouds to the tide. ![Low streets fill first; in 2026 some areas of the city stayed knee-deep for days. AI-generated still.](https://assets.planetpast.com/a/41/41ef5ea7e2501a90.jpg) - **71**: deaths - **51**: provinces flooded - **~908,000**: households affected - Sky: monsoon rain - Soil: soak in or run off - River: the funnel - Dam, delta, tide Prologue · The news #### Floods across Thailand According to the Interior Ministry, floods between 16 September and 12 October 2026 affected 51 provinces and about 908,000 households and killed 71 people. Bangkok and about twenty other provinces still had flooded areas on 12 October. AFP reported on 7 October that 19 people had died in Prachinburi, the highest provincial toll it listed, and that Bangkok received about a month of rain in a few days in late September. This film is not a report on 2026: it explains the physics and history behind every river flood in the Chao Phraya basin. How to use: pick a chapter or press ← →; drag to rotate. Eight steps have buttons you can press. Sources: [The Star / Interior Ministry: flood toll since 16 September rises to 71 (12 Oct 2026)](https://www.thestar.com.my/aseanplus/aseanplus-news/2026/10/12/thailand-flood-death-toll-since-september-16-rises-to-71-interior-ministry); [Malay Mail / AFP: floods kill 60 across Thailand (7 Oct 2026)](https://www.malaymail.com/news/world/2026/10/07/floods-kill-60-across-thailand-as-bangkok-races-to-clean-up-before-imf-meetings/238107) ## Chapter I · The rain ### The wind that brings the water *Chapter I · The rain* Half a year of rain, arriving from the ocean. It starts over the Indian Ocean. Each year from about mid-May, the south-west monsoon carries warm, wet sea air over mainland Southeast Asia. When that air rises over hills and heats over land, its water vapour condenses into cloud and falls as rain. Thailand's rainy season runs until about mid-October. The rain peaks in August and September. By then, the land has been drinking for three months. ![Moist air from the Indian Ocean and the Bay of Bengal. Schematic illustration, AI-generated.](https://assets.planetpast.com/a/f2/f2ec4e47f1b7d608.jpg) - mid-Feb – mid-May: Hot season; April is hottest - mid-May – mid-Oct: Rainy season, south-west monsoon - mid-Oct – mid-Feb: Cool season, north-east monsoon - **Aug–Sep**: nationally - **3 months**: of rain already soaked up by August Chapter I · The rain #### The monsoon Thailand’s Land Development Department describes three seasons. The rainy season, driven by the south-west monsoon, runs from mid-May to mid-October and peaks in August and September. The southern east coast follows a different calendar: rain stays high through the end of the year, with November the wettest month. Sources: [Land Development Department, Thailand: the seasons](https://webeng.ldd.go.th/pv/?cd=229) ## Chapter II · The ground ### Soak in or run off *Chapter II · The ground* Every raindrop has two roads. A raindrop that lands on soil has two roads. It can soak in, filling the gaps between grains, or it can run across the surface toward the nearest stream. Soil can only take in water at a certain rate. In a drizzle, almost everything soaks in. In a downpour, rain arrives faster than the ground can swallow it, and the extra runs off. Try it. Switch between drizzle and downpour and watch where the drops go. $$\text{runoff} = r - f \quad (\text{if } r > f)$$ - $r$: rain intensity (mm per hour) - $f$: infiltration capacity of the soil ![Water pools on the surface only when the soil below cannot take it fast enough. AI-generated still.](https://assets.planetpast.com/a/b4/b4c0f50b24e3ae1c.jpg) - soil already wet from earlier rain - steep slopes with thin soil - bare or compacted ground Chapter II · The ground #### Infiltration and runoff Hydrologists call the first road infiltration. When rain falls faster than the soil’s infiltration capacity, the excess flows over the surface — “Hortonian” or infiltration-excess overland flow. The numbers in the interactive model are illustrative, not measurements of Thai soils. Sources: [University of Regina, Geography 327: runoff processes](https://uregina.ca/~sauchyn/geog327/runoff.html) ### A full sponge *Chapter II · The ground* Why late-season rain does the most damage. There is a second way to make runoff. Below every field lies the water table, the level under which the ground is already full of water. After weeks of rain it rises until the soil is saturated from below, and then even gentle rain has nowhere to go. Press a week of rain and watch the blue layer climb. As it does, the share of water running straight to the river grows. That is why the same storm is harmless in June and dangerous in September. And where the ground is paved over, almost nothing soaks in, wet season or dry. - water table: the top of the ground that is already full - happens first in hollows and at the foot of slopes - the wet area grows as the basin gets wetter - field studies find it more common than runoff from heavy rain alone - **June**: dry soil: most of it soaks in - **September**: soaked soil: most of it runs off - dry soil, drizzle: 0 % - dry soil, downpour: 58 % - soaked soil, downpour: 93 % - pavement: 98 % Chapter II · The ground #### Saturation from below In saturation-excess overland flow the soil is already full, so rain — and groundwater seeping back up — runs over the surface. The area that produces runoff grows as the whole watershed gets wetter, which hydrologists call the variable source area. The bar chart comes from the simplified model on screen, not from field data. Sources: [University of Regina, Geography 327: runoff processes](https://uregina.ca/~sauchyn/geog327/runoff.html) ### Higher and sooner *Chapter II · The ground* What a city does to a river. Hydrologists draw a flood as a curve: how much water passes a point in the river, hour by hour, after a storm. Rain falls at the start; the river rises later. In a forested valley, much of the rain soaks in and seeps out slowly, so the peak comes late and stays low. Now switch to a paved city. Roofs, roads and drains send water to the river at once. In one US Geological Survey study around Washington, D.C., urban development raised flood peaks by about two to nearly eight times. - peak discharge: the most water at once - lag time: from the rain to the peak - recession: how fast it drains away - **slow**: water held in the soil seeps out to the river - **low**: a late, gentle peak - **2 – 8×**: higher flood peaks after urban development - **down to ⅛**: of the natural delay before the peak, with piped drainage Chapter II · The ground #### The storm hydrograph Paving replaces ground that would soak up rain, and drains carry runoff straight to streams, so peaks are higher and arrive sooner. A USGS study of 81 sites, most around Washington, D.C., found development could raise peaks by a factor of about two to nearly eight, with drainage that could cut lag time to as little as one-eighth of natural channels. The curves on screen are schematic: their shape, not their height, is the point. Sources: [USGS Water Science School: impervious surfaces and flooding](https://www.usgs.gov/water-science-school/science/impervious-surfaces-and-flooding); [USGS Water-Supply Paper 2001-C: effects of urban development on floods](https://www.usgs.gov/publications/effects-urban-development-floods-northern-virginia) ## Chapter III · The funnel ### Four rivers become one *Chapter III · The funnel* A third of Thailand drains through Nakhon Sawan. Now zoom out. In the northern mountains, four rivers collect the runoff: the Ping and the Wang, the Yom and the Nan. The Wang joins the Ping, the Yom joins the Nan, and at Nakhon Sawan the two meet to become the Chao Phraya, which runs about three hundred and seventy kilometres to the sea. Altogether the basin covers almost a hundred and sixty thousand square kilometres, about a third of Thailand. Press Send a storm: rain from four valleys pours into one river, and it swells far more than any of them alone. ![Every slope tilts toward one outlet, like a funnel. Schematic relief, AI-generated.](https://assets.planetpast.com/a/82/82dc531ed4f7b2d7.jpg) - **372 km**: Nakhon Sawan to the gulf - **Chai Nat**: where the Tha Chin branches off: a second, smaller way to the sea - **157,924 km²**: drains to the Chao Phraya - **35%**: of Thailand’s land - the four valleys’ flows add up - below Nakhon Sawan the land is almost flat - on flat land, spilled water spreads wide Chapter III · The funnel #### The Chao Phraya basin The Chao Phraya begins at Nakhon Sawan (Pak Nam Pho), where the Ping — with its tributary the Wang — meets the Nan, which has already taken in the Yom. It runs 372 km to the Gulf of Thailand; at Chai Nat the Tha Chin branches off and reaches the sea separately, west of Bangkok. The whole watershed draining into it and its tributaries covers 157,924 km², about 35% of Thailand. The 3D relief is schematic and its height is exaggerated. Sources: [Wikipedia: River systems of Thailand](https://en.wikipedia.org/wiki/River_systems_of_Thailand) ## Chapter IV · The dams ### Two jobs, one bucket *Chapter IV · The dams* A reservoir cannot be full and empty at once. Two giant dams sit high on the funnel's branches. Bhumibol, finished in 1964 on the Ping, can hold about thirteen and a half billion cubic metres. Sirikit, finished in 1972 on the Nan, holds about nine and a half billion. Each has two jobs that pull against each other. Farmers need water stored for the dry season. Flood control needs empty space for the storms to come. Try the rainy season both ways. Keep the dam full, and once it is full it must pass on everything that flows in, all at once. Release early, and the flood downstream stays small, but the dry season starts with less water. - Bhumibol (Ping, 1964): 13.46 - Sirikit (Nan, 1972): 9.51 ![Stored water feeds irrigation canals long after the rains stop. Bhumibol stands 154 m high, the tallest dam in Thailand. AI-generated still.](https://assets.planetpast.com/a/86/868335dfb92b1965.jpg) - irrigation and power want it full - flood control wants room to spare - nobody knows the rain in advance - **9.76 bn m³**: usable storage out of 13.46 billion Chapter IV · The dams #### Bhumibol and Sirikit Bhumibol Dam on the Ping (built 1958–1964, 154 m) has a total capacity of about 13.46 billion m³, of which about 9.76 billion is usable storage; Sirikit Dam on the Nan (completed 1972, 113.6 m) about 9.51 billion m³. Both serve irrigation, hydropower and flood control. The rainy-season game is schematic, with invented inflows for one very wet year. It shows the trade-off, not how either dam is actually operated. Sources: [Wikipedia: Bhumibol Dam](https://en.wikipedia.org/wiki/Bhumibol_Dam); [Wikipedia: Sirikit Dam](https://en.wikipedia.org/wiki/Sirikit_Dam) ### When both dams were full *Chapter IV · The dams* How a drought set up a flood. In 2010 rain was scarce, and by June the reservoirs were at record lows. Keeping water became the priority. Then 2011 turned. In March, rain over northern Thailand was more than four times the normal amount. In three months, Bhumibol collected over eight billion cubic metres and filled. Storms kept coming. By early October most dams were near or over capacity and had to release more, adding water to a river that was already overflowing. ![Dam levels hit record lows after a dry 2010, so keeping water became the priority. AI-generated still.](https://assets.planetpast.com/a/c9/c98dbaa2699fa941.jpg) - **+344%**: above the March mean: about 4.4 times normal - **242.8 mm**: at Bhumibol Dam, against a normal 25.2 mm - Jun 2010: Dams at record lows - Mar 2011: Freak dry-season rain in the north - 31 Jul 2011: Tropical Storm Nock-ten brings floods - Early Oct 2011: Dams near or over capacity, releases rise Chapter IV · The dams #### The set-up for 2011 Low rainfall in 2010 left dam levels at record lows in June 2010. In March 2011 rainfall over northern Thailand was 344% above the mean; Bhumibol received 242.8 mm against a normal 25.2 mm. Bhumibol collected over 8 billion m³ in three months, and by early October most dams were near or over capacity and forced to increase discharge. The larger releases drew accusations of mismanagement at the time; how much they added to the flood is debated, and this film does not try to settle it. Sources: [Wikipedia: 2011 Thailand floods](https://en.wikipedia.org/wiki/2011_Thailand_floods) ## Chapter V · The delta ### A city at sea level *Chapter V · The delta* The last link in the chain pushes back. At the bottom of the funnel lies Bangkok, built on the flat mud of the Chao Phraya delta. Most of the city stands only half a metre to two metres above the sea. Here the river is tidal. Once or twice a day the gulf rises and pushes back up the channel, so floodwater coming down can only drain when the tide goes out. Studies of Bangkok's historic floods found that the highest monthly tides, and possibly storm surges, can make a river flood worse, just as the peak flow from the north arrives. ![At high tide the river can rise to the edge of riverside walkways even without rain. AI-generated still.](https://assets.planetpast.com/a/14/14d6679dc620eb88.jpg) - **0.5–2 m**: above sea level - **1–2 a day**: high tides pushing up the river - spring tides, the highest of the month - wind-driven meteorological tides - storm waves, and possibly storm surge Chapter V · The delta #### River meets sea Bangkok sits only about 50 cm to 2 m above sea level. A 2020 study of the Bangkok floods of 1785 and 1983 found that spring tides, wind-driven tides, storm waves and possibly surge can aggravate seasonal peak river flows in the tidal lower Chao Phraya. The 3D model exaggerates heights so the tide can be seen. Sources: [Wikipedia: 2011 Thailand floods](https://en.wikipedia.org/wiki/2011_Thailand_floods); [Terry, Winspear & Goff (2020), Natural Hazards: Bangkok floods of 1785 and 1983](https://zuscholars.zu.ac.ae/works/2159) ### The ground goes down *Chapter V · The delta* Pumping water out of the ground lowered the city. The delta has another problem: it sinks. Its soft clay is squeezed when water is pumped out of the layers below, and the ground above settles. In the 1980s, with factories and homes pumping groundwater freely, parts of Bangkok were sinking about ten centimetres a year. Press the pump and watch the street drop toward the tide. Laws, groundwater charges and close monitoring slowed it. By 2020 most of the city was sinking about one centimetre a year or less, but the lost height does not come back. ![Pumping drains water from the clay layers, which compact and do not spring back. AI-generated still.](https://assets.planetpast.com/a/59/59b895b68c326fb3.jpg) - 1980s: ~10 - most of the city, by 2020: 0–1 - legislation on wells - a price on groundwater - monitoring wells and extraction Chapter V · The delta #### Land subsidence Rapid development and uncontrolled groundwater pumping made Bangkok sink about 10 cm a year in the 1980s. A hydrogeologist at Thailand’s Ministry of Natural Resources and Environment credits legislation, groundwater pricing and close monitoring of wells for slowing it to about 1 cm a year or less in most of the city. The pump button lowers the model city one metre per press, a deliberately exaggerated scale. Sources: [Mekong-U.S. Partnership: how Bangkok stopped sinking (2020)](https://mekonguspartnership.org/2020/10/02/how-bangkok-stopped-sinking-and-what-the-mekong-delta-could-learn-october-2-2020/) ## Chapter VI · 2011 ### Every link failed at once *Chapter VI · 2011* The flood that stopped factories around the world. In 2011 every link gave way together. Soaked soil, a full funnel, brimming dams and a low delta. The water spread south for months, from late July until the middle of January. Eight hundred and fifteen people died, and thirteen point six million were affected. North of Bangkok, barriers around industrial estates in Ayutthaya and nearby provinces failed, and seven major estates went under, in places three metres deep. Thailand made about a quarter of the world's hard drives, so the shortage was felt worldwide. The World Bank put the damage at forty-six and a half billion dollars, then the fourth costliest disaster on record. ![In 2011 floodwater stood in some towns for weeks before it could drain south. AI-generated still.](https://assets.planetpast.com/a/5d/5dba82e20c1eed1c.jpg) ![The 10-metre Nikom Rojna barrier collapsed on 8 October 2011. AI-generated still, not a photograph of the event.](https://assets.planetpast.com/a/f7/f7cef4fdc628cc98.jpg) - **815**: deaths - **13.6 million**: people affected - **65**: provinces declared disaster zones ![About 25% of the world’s hard disk drives came from Thailand. AI-generated still.](https://assets.planetpast.com/a/a4/a4ef80fbc37a8600.jpg) - ahead: Tōhoku earthquake and tsunami, 2011 - ahead: Great Hanshin (Kobe) earthquake, 1995 - ahead: Hurricane Katrina, 2005 - No. 4: Thailand floods, 2011 Chapter VI · 2011 #### The great flood of 2011 Floods lasted from late July 2011 to mid-January 2012, killed 815 people and affected 13.6 million; 65 provinces were declared disaster zones. Seven major industrial estates were inundated, up to 3 m deep. The World Bank estimated damage at 1.425 trillion baht (US$46.5 billion), the fourth-costliest disaster as of 2011. Downtown Bangkok was largely defended by late November 2011. Sources: [Wikipedia: 2011 Thailand floods](https://en.wikipedia.org/wiki/2011_Thailand_floods); [World Bank / GFDRR: Thai Flood 2011 rapid assessment](https://preventionweb.net/quick/69697) ## Chapter VII · Living with water ### The hundred-year flood *Chapter VII · Living with water* A name that misleads almost everyone. Will a flood like 2011 come back? Engineers answer with odds. A “hundred-year flood” does not come once a century. It is a flood with a one per cent chance of happening in any given year. Every year starts with the same odds, so two can come back to back. Over thirty years, the chance of at least one is about one in four. Roll thirty years for twenty towns and count how many get hit. Then try a ten-year flood: almost every town meets one. - **1%**: chance in any single year - **not**: a schedule: two can come in a row $$P = 1-(1-p)^{n}$$ $$1-(0.99)^{30} \approx 0.26$$ - $p$: chance in one year (0.01 for a “100-year” flood) - $n$: number of years - “100-year” flood: 26% - “50-year” flood: 45% - “10-year” flood: 96% Chapter VII · Living with water #### Return periods The USGS explains that the recurrence interval is the average number of years between floods of a given size, and that agencies prefer the annual exceedance probability: a 100-year flood has a 1% chance in any year, and each year begins with the same chance. The 30-year figures come from P = 1 − (1 − p)ⁿ, which assumes years are independent. A changing climate or a changing basin moves p itself. Sources: [USGS Water Science School: the 100-year flood](https://www.usgs.gov/special-topic/water-science-school/science/100-year-flood) ### Monkey cheeks *Chapter VII · Living with water* Store the water, wait for the tide. After Bangkok flooded in 1995, King Bhumibol proposed an idea named after a monkey storing bananas in its cheeks: kaem ling. Ponds, lakes and canals take in floodwater when rain is heavy. Floodgates to the sea stay shut while the tide is high, and open when the sea drops below the water in the canals. Watch the gate follow the tide, then add the monkey cheeks and see how much lower the river runs. The same idea now shapes parks built to hold rain. ![Storage first, release later: the water is also kept for the dry season. AI-generated still.](https://assets.planetpast.com/a/4d/4d6ab793e87678bc.jpg) - heavy rain: canals and ponds fill - high tide: gates shut to keep the sea out - low tide: gates open, water drains - dry season: stored water is used ![Chulalongkorn University Centenary Park, by Kotchakorn Voraakhom, takes its idea from the monkey cheek. AI-generated illustration, not the real park.](https://assets.planetpast.com/a/15/15db6e254a206f61.jpg) Chapter VII · Living with water #### Kaem ling King Rama IX introduced the kaem ling (“monkey cheeks”) concept in 1995, when Bangkok faced a flooding crisis. Canals and ponds store excess water; floodgates open to release it to the sea when the sea is lower than the canals and close when the sea is higher. Stored water is used in the dry season. In the model the ponds lower the river by a made-up amount; real benefits depend on how much storage there is. Sources: [Bangkok Post: King Rama IX’s water legacy lives on](https://bangkokpost.com/news/general/1557358/king-rama-ixs-water-legacy-lives-on); [Wikipedia: Kaem Ling (monkey cheeks)](https://en.wikipedia.org/wiki/Kaem_Ling) ## Epilogue ### The funnel to the sea *Epilogue* Every link can be strengthened. None can be removed. The floods of 2026 came at the peak of the rainy season, after Bangkok had a month of rain in a few days. Every link in this film decides how bad a year like that becomes. None of these links can be removed. But each can be made stronger: by keeping room in reservoirs, leaving space for water, and building with the flood in mind. The water that reaches the Gulf of Thailand tomorrow fell on the mountains weeks ago. Understanding the chain is how a country gets ready for it. - 1964: Bhumibol Dam completed - 1972: Sirikit Dam completed - 1983: Floods in 42 provinces - 1995: Bangkok floods; kaem ling proposed - 2011: 815 dead, US$46.5 bn damage - 2026: 71 dead in 51 provinces by 12 Oct ![Forests and soil in the headwaters hold back part of every storm. All stills in this film are AI-generated.](https://assets.planetpast.com/a/29/29595bbf4da85fdf.jpg) > Rain in the north today is water in Bangkok in weeks. (The lesson of the funnel) Epilogue · Sources #### How river floods work A river flood links weather, soil, the shape of a basin, human structures and the sea. This film used the Chao Phraya basin as its example; the same chain runs through every flood-prone river delta. All interactive models are schematic, with illustrative numbers, and all stills are AI-generated. Figures for 2026 are provisional official counts as reported on 12 October 2026. This is general education, not flood-safety guidance: follow your local disaster agency. Sources: [The Star / Interior Ministry: flood toll since 16 September rises to 71 (12 Oct 2026)](https://www.thestar.com.my/aseanplus/aseanplus-news/2026/10/12/thailand-flood-death-toll-since-september-16-rises-to-71-interior-ministry); [Malay Mail / AFP: floods kill 60 across Thailand (7 Oct 2026)](https://www.malaymail.com/news/world/2026/10/07/floods-kill-60-across-thailand-as-bangkok-races-to-clean-up-before-imf-meetings/238107); [Land Development Department, Thailand: the seasons](https://webeng.ldd.go.th/pv/?cd=229); [University of Regina, Geography 327: runoff processes](https://uregina.ca/~sauchyn/geog327/runoff.html); [USGS Water Science School: impervious surfaces and flooding](https://www.usgs.gov/water-science-school/science/impervious-surfaces-and-flooding); [USGS Water-Supply Paper 2001-C: effects of urban development on floods](https://www.usgs.gov/publications/effects-urban-development-floods-northern-virginia); [Wikipedia: River systems of Thailand](https://en.wikipedia.org/wiki/River_systems_of_Thailand); [Wikipedia: Bhumibol Dam](https://en.wikipedia.org/wiki/Bhumibol_Dam); [Wikipedia: Sirikit Dam](https://en.wikipedia.org/wiki/Sirikit_Dam); [Wikipedia: 2011 Thailand floods](https://en.wikipedia.org/wiki/2011_Thailand_floods); [World Bank / GFDRR: Thai Flood 2011 rapid assessment](https://preventionweb.net/quick/69697); [Mekong-U.S. Partnership: how Bangkok stopped sinking (2020)](https://mekonguspartnership.org/2020/10/02/how-bangkok-stopped-sinking-and-what-the-mekong-delta-could-learn-october-2-2020/); [Terry, Winspear & Goff (2020), Natural Hazards: Bangkok floods of 1785 and 1983](https://zuscholars.zu.ac.ae/works/2159); [USGS Water Science School: the 100-year flood](https://www.usgs.gov/special-topic/water-science-school/science/100-year-flood); [Wikipedia: Kaem Ling (monkey cheeks)](https://en.wikipedia.org/wiki/Kaem_Ling); [Wikipedia: Floods in Thailand](https://en.wikipedia.org/wiki/Floods_in_Thailand); [Bangkok Post: King Rama IX’s water legacy lives on](https://bangkokpost.com/news/general/1557358/king-rama-ixs-water-legacy-lives-on)