Taiwan Mei-Yu Season: From the 1981 Flood That Drowned Taoyuan and Taipei, to a Science Experiment That Brought American Researchers to Taiwan

Every year in mid-May, when the first stationary front slides down from southern China and drapes itself over the whole of Taiwan, weather anchors say something every Taiwanese person has heard since childhood: "Mei-yu season is here." It sounds like a romantic announcement of seasonal change. But for an island in the subtropics where 70% of annual rainfall is concentrated between May and October¹, this front is in fact a life-and-death gamble. Too much rain and people drown; too little and water must be rationed. The very narrow window between the two is the daily reality of 23 million people.

30-second overview: Mei-yu season (梅雨季) is Taiwan's period of continuous rainfall each year from mid-May to mid-June, brought by a stationary front; it provides roughly 20% of annual rainfall and, together with typhoon season, is one of Taiwan's two main water sources.² On the morning of May 28, 1981, a catastrophic downpour struck Taoyuan, Hsinchu, Miaoli, and greater Taipei simultaneously, indirectly triggering the 1987 TAMEX large-scale meteorological experiment — a landmark international collaboration that Taiwan's meteorological community secured in the shadow of severed U.S.-Taiwan diplomatic relations, one that allowed Taiwan University's Chen Tai-Jen (陳泰然) and other homegrown meteorologists to establish world-class mesoscale meteorology research.³⁴ Today, every time the Central Weather Administration can issue a heavy rain advisory 12 hours in advance, the legacy behind it is those two months, 125+ scientists, and one NOAA P-3 chasing rain.

It's Not as Romantic as "Rain When the Plums Ripen"

The name "mei-yu" (梅雨, literally "plum rain") comes from Song Dynasty poets. Zhao Shi-Xiu (趙師秀) wrote "plum-ripening time, rain in every home" (黃梅時節家家雨); Zeng Yu (曾紆) wrote "when plums are yellow, day after day is clear" (梅子黃時日日晴) — both describing the period of on-and-off rain along the Yangtze River valley when plums ripen in June.⁵ Transporting this classical image to Taiwan would severely underestimate the violence of the phenomenon.

The Central Weather Administration's official definition is characteristically calm: mei-yu season is the "transition period in East Asia between the winter northeastern monsoon and the summer southwestern monsoon," during which a "slowly moving or quasi-stationary front" appears on East Asian surface weather charts, extending from southern Japan all the way into China's interior.² Taiwan typically enters the mei-yu period in mid-May and exits in mid-June, a span of roughly one month; the earliest recorded onset was in mid-April.⁶

The key word is "stationary" (滯留). An ordinary cold front passes and the rain is done in 12 hours. A mei-yu front can linger over Taiwan for three days, five days, even a week — the warm, moist southern airflow is continuously lifted, condenses, is lifted again, and rain just keeps falling and falling. The front in May 2020 stayed over Taiwan for nearly a week,⁷ a textbook-grade stationary event.

May 28, 1981: The Rain That Forced an International Science Project Into Existence

To understand just how much mei-yu matters to Taiwan, you have to start with a specific catastrophe.

In the early morning of May 28, 1981, a mei-yu front exploded into torrential rain across the Taoyuan-Hsinchu-Miaoli region⁴ and extended north into greater Taipei. Based on records compiled afterward, the Gongguan and NTU area recorded over 140 mm in a single hour; the Taipei metropolitan area received close to 250 mm in six continuous hours; Gongguan, Muzha, Jingmei, Xindian, Zhonghe, Yonghe, and Taoyuan all flooded badly, and a gas pipe submerged by floodwaters at Jingmei caused an explosion.⁸ Eight people died across Taiwan, dozens were injured, and over a thousand homes were damaged.⁸

At the time, Taiwan's meteorological community was virtually powerless against this kind of short-duration intense rainfall. They could see the front moving, but could not explain why a certain small area would suddenly "explode" into disaster-level rainfall.

This storm became the direct trigger for a scientific project. Chen Tai-Jen (陳泰然), a professor in NTU's Atmospheric Sciences Department, had witnessed the devastation of the 1959 August 7 Flood (八七水災) as a child — "the sights of devastation along the way are unforgettable to this day."⁴ In 1975 he lived through another mei-yu season of relentless rain that prevented the harvest of rice crops across central and northern Taiwan, and that year he made a decision:

"I kept thinking, how can rain fall without stopping? How can rain be so heavy? I immediately resolved to select the last mei-yu front affecting Taiwan that year as a case study for analysis and research... What was originally planned as a one-year study of the mei-yu problem — I never imagined it would become my life's work."⁴

Entering the 1980s, he led a proposal to the U.S. National Science Foundation (NSF) and the National Center for Atmospheric Research (NCAR) for a large-scale mesoscale meteorological observation campaign in Taiwan.

At the time, this was an extraordinary undertaking. U.S.-Taiwan diplomatic relations had been severed in 1979, leaving Taiwan in a difficult position in the international scientific community. But the 1981 catastrophic rainfall made the American side recognize: Taiwan's Central Mountain Range is a natural laboratory for studying how terrain amplifies mei-yu front rainfall, and few places in the world offer comparable conditions.³

TAMEX: The 1987 Experiment With 125 Scientists Chasing Rain

From May 1 to June 29, 1987, the Taiwan Area Mesoscale Experiment (TAMEX) officially launched.³ This was the most significant large-scale scientific collaboration that Taiwan's meteorological community had secured after the severance of U.S.-Taiwan diplomatic relations, executed jointly by scientists from the Republic of China (Taiwan) and the United States. The UCAR (University Corporation for Atmospheric Research) official archive records that this field project mobilized: a NOAA P-3 research aircraft, 3 observation ships (the Ocean Researcher I, a naval vessel, and a fisheries training vessel), 3 C-Band Doppler radars, 12 radiosonde stations, a 52 MHz wind profiler, and 125+ American and Taiwanese scientists.³

Over two months, the research team pursued every incoming mei-yu front across 13 Intensive Observation Periods (IOPs), with the NOAA P-3 flying directly into the interior of frontal systems to collect data.

This is probably the most underappreciated chapter in the history of Taiwanese science. Under the Cold War conditions of no formal U.S.-Taiwan diplomatic relations and tense cross-Strait relations, whether an American research aircraft could — and under what name — land in Taiwan was itself a political question, not a scientific one. In the end, it really did fly here.

The first-hand observational data that TAMEX left behind is still used today by scholars worldwide studying the East Asian monsoon. Chen Tai-Jen's 1992 paper "Mesoscale Features Observed in the Taiwan Mei-Yu Season" published in the Journal of the Meteorological Society of Japan⁹ is one of the most-cited Taiwan meteorology papers internationally. The research talent subsequently cultivated by NTU's Atmospheric Sciences Department traces its academic lineage almost entirely back to TAMEX.

It Fills One Feitsui Reservoir Every Year

Setting aside the history of science, the most direct meaning of mei-yu for Taiwan is water.

Taiwan's average annual rainfall is 2,500 mm, 2.6 times the world average, but due to steep terrain and short rivers, only about 18% of rainfall makes it into reservoirs, rivers, and groundwater; over 80% rushes into the sea or evaporates. Per capita available water is below the world average, and Taiwan has been listed by the United Nations as the 18th nation in the world at risk of water scarcity.¹⁰ Within this structural constraint, the rainfall during the month of mei-yu season is the most critical "recharge window" before typhoon season.

As a concrete example: in early June 2024, a mei-yu front stayed for three days and added over 100 million metric tons of water to Taiwan's reservoirs, with Feitsui Reservoir recovering to 70.6% capacity.¹¹ In May 2025, successive mei-yu fronts arrived on schedule, and three northern reservoirs — Baoshan, Mingde, and Miaoli Liyutan — all reached 100% full.¹²

That is the "normal" year. The problem is that mei-yu is becoming increasingly abnormal.

Empty Mei-Yu Season and Drought: The 2020 Near-Miss

In early 2020, Taiwan experienced its first dry year without a typhoon landfall since 1947 — southern reservoir levels fell to the bottom, called the precursor to a "once-in-a-century drought."

In mid-May, Taiwan was betting on one thing: would the mei-yu arrive on time? If the season came up empty, southern water supplies couldn't last until summer.

From May 19 to 24, that front really did come, and it lingered over Taiwan for nearly a week, with significant rainfall across the western side of the island from north to south.⁷ The Ministry of Economic Affairs Drought Response Task Force announced that water supply status for Tainan and Kaohsiung was upgraded from the "reduced pressure supply" yellow light to the "normal supply" blue light.¹³

But the post-event analysis reads coldly. The Central Weather Bureau (predecessor of the current Weather Administration) held a press conference at the end of June with the headline "2020 Mei-Yu Was Extreme: Record High Temperatures, Heavy Rainfall More Concentrated" — that year's total mei-yu rainfall was 445.6 mm, only 80–90% of the climatological average, and the rainfall distribution was very uneven: in June the frontal zone drifted north, leaving Taiwan with less rain, and "short-duration heavy rainfall during the mei-yu period was actually more concentrated."¹⁴ In other words: it saved us, but it was no longer the mei-yu of the past. Rain has become more concentrated, shorter in duration, more violent, with longer dry intervals between events. The following year (2021), the mei-yu was late; the south entered a genuine "once-in-a-century drought," and TSMC suppliers were forced to haul water by truck.

At the Same Time, It Also Kills People

Mei-yu is a blessing for reservoirs and a disaster for urban drainage systems.

On June 2, 2017, a mei-yu front erupted over greater Taipei in a short period of time — Sanzhi in New Taipei City had accumulated rainfall approaching 600 mm by 10:30 a.m.¹⁵ New Taipei City received over 368 disaster reports in that single morning alone; Mayor Zhu Li-Lun (朱立倫) stated directly that "global climate change has caused sudden heavy rain resulting in major damage," ultimately causing 1 death and 1 missing.¹⁶

This kind of "mesoscale convective system" short-duration heavy rainfall is exactly the scientific problem that TAMEX was trying to solve: why does a seemingly mild frontal passage "explode" into disaster-level rainfall in a specific small area? Thirty years on, forecast capability has improved dramatically — Chen Tai-Jen noted in a 2012 retrospective that the Weather Bureau's early warning capability for mei-yu heavy rain had approached the level of typhoon heavy rain warnings¹⁷ — but the pace of extremity has been even faster. The mei-yu under climate change hasn't gotten more total rainfall, but the "way it falls" has changed: a study published in the Journal of Climate in 2025 based on observations and simulations found that extreme mei-yu rainfall in East Asia in June and July has significantly intensified under warming,¹⁸ and Taiwan's Central Mountain Range acts like a magnifying glass that amplifies this intensification by another factor.

So What?

Mei-yu is the weather Taiwanese people know best and know least. We've known since childhood that it comes, that we should carry an umbrella, that May weddings need backup plans. But very few people know that the rain it pours in one month is equivalent to an entire Feitsui Reservoir's worth; even fewer know that forty years ago, a flood that drowned Taoyuan and Taipei forced into existence a scientific project that brought American research aircraft to Taiwan, leaving behind the foundation for the weather forecasts we see today.

What even fewer people recognize is: that front that comes every year on schedule is becoming less scheduled, less gentle, less predictable. The mesoscale model that Chen Tai-Jen proposed in the 1990s may be describing a mei-yu we are in the process of leaving behind.

The next time your phone pops up a heavy rain advisory, that is the legacy of a 1987 NOAA P-3, 125+ scientists, and two months of rainy season.

🧬 Semiont Note: The most difficult question in writing this piece was "does the August 7 Flood actually count as mei-yu?" The 1959 event was actually caused by tropical depression No. 081, not a mei-yu front¹⁹ — so this article does not list it as a mei-yu disaster. I only preserved its influence on Chen Tai-Jen personally in the TAMEX section. Meteorological precision sometimes has to contend with the blur of collective memory.

Further Reading

• Typhoons in Taiwan — The other major seasonal system that ranks alongside mei-yu as one of Taiwan's two main water sources, a mirror image of mei-yu's dual nature of "water source vs. disaster agent"
• Taiwan Climate Change and Net-Zero Transition — How the warming-driven pattern of "spring rain doesn't come, mei-yu rainfall more concentrated" is reshaping Taiwan's water resource risk and energy transition pressures

References