Taiwan's Climate Crisis and Net-Zero Transition: The Day the Nuclear Three Referendum Failed Was Just the Beginning of Choosing Physical Limits

30-Second Overview: On the evening of August 23, 2025, the results of the Nuclear Three extension referendum were announced: 4.34 million yes votes, 74% approval, and a turnout of 29.53%, missing the threshold by 650,000 votes. The next day, President Lai Ching-te announced three principles: "nuclear safety assured, nuclear waste solved, social consensus." Seven months later, on March 27, 2026, Taipower submitted the Nuclear Three re-operation application to the Nuclear and Material Safety Administration (NAC), with the earliest restart expected in 2028¹². The referendum failed, yet Taipower is walking back towards nuclear power. This is the deepest contradiction of an island that relies on imports for 98% of its energy and has pledged to invest 9 trillion NTD to achieve net-zero by 2050³. The government's geothermal target is 200 MW by 2030, but actual commercial operation by the end of 2025 is only 7.4 MW, a 27-fold gap; the Lanyu storage facility opened in 1982, holding 97,672 barrels of nuclear waste, with relocation deadlines missed four times⁴⁵. Energy is a question of physical limits.

The Nuclear Three Plant (Maanshan Nuclear Power Plant) in Hengchun, Pingtung, located on the Nanwan coastline. Unit 1 shut down on January 1, 2025, and Unit 2 on May 17, 2025. Image: M. Weitzel, CC BY-SA 3.0, via Wikimedia Commons

The Day of the Nuclear Three Referendum

On the evening of August 23, 2025, vote counting was completed across all 22 counties and cities in Taiwan, and the results of the Nuclear Three extension referendum were announced: 4,342,206 yes votes, 1,511,693 no votes, with 74.17% approval. However, the turnout was only 29.53%. According to the Referendum Act, a threshold of one-quarter of the total number of voters (5,005,230 votes) was required. The yes votes fell short by 658,317 votes¹. Those who voted yes were more than twice those who voted no, yet the referendum failed.

📝 Curator's Note: The common interpretation is "74% approval = clear public support for nuclear energy," but this reverses causality. The design of the Referendum Act was never about who had more votes; it required a mobilization threshold to prove that "enough people cared." A 29.53% turnout meant that over 70% of voters chose not to go out. This is an even more awkward third signal: many people do not have strong enough feelings about energy issues to enter a polling station.

Two days later, on August 25, President Lai Ching-te held a press conference and gave a response: to restart nuclear power, three gates must be passed: "nuclear safety assured, nuclear waste solved, social consensus"². It sounds reasonable, but each is a problem unsolved for 50 years.

Then came March 27, 2026. Taipower submitted the Nuclear Three re-operation application plan to the NAC, initiating the safety inspection procedures for Unit 1 after shutdown. The safety inspection schedule is expected to take about 18 months, with the earliest restart in 2028¹. This is a step backward.

From the referendum results to the submission for review, exactly seven months passed. Nothing changed in between: nuclear waste is still in Lanyu, the ultimate repository site is not yet selected, and social consensus remains divided. But administrative procedures moved. This is the question this article seeks to answer: when democratic voting rejects something, but the executive branch advances it simultaneously, who is actually deciding Taiwan's energy policy?

Lanyu: 1982 to 2057

To understand the story of Nuclear Three, one must first understand the story of Lanyu.

In 1982, Taipower opened a low-level radioactive waste storage facility on the coastline south of Longmen in Lanyu. At the time, the Dyak (Yatu) residents were told it was a "fish cannery," a claim that later became the most frequently cited case of deception in Taiwan's environmental justice history⁶. In 1988, Dyak residents launched their first large-scale protest, using the traditional ritual of "exorcising evil spirits" to express their rejection of nuclear waste, marking the starting point of Taiwan's indigenous environmental movement.

Over the next 38 years, relocation promises were missed four times: the government promised relocation by 2002 in 1996, missing it in 2002; then missed again in 2016, 2019, and 2023. As of 2024, the Lanyu storage facility had accumulated 97,672 barrels of low-level radioactive waste. The Nuclear Regulatory Commission (NRC) requires Taipower to complete relocation by 2029, but the destination remains undefined⁴.

If 2029 is missed again (widely expected in the industry), Lanyu's nuclear waste will be stored from 1982 to 2057, a total of 75 years. An island with a population of 4,000 bears the byproduct time of four nuclear power plants operating for the entire country, longer than the lifespan of most people in Taiwan.

⚠️ Controversial Viewpoint: Pro-nuclear advocates often say, "Nuclear waste is technically solvable; it's just political resistance." But the problem of nuclear waste has always been about time. Lanyu has been storing waste since 1982, 44 years ago, and promises have never been fulfilled. In the most optimistic scenario, relocation happens in 2029; but what happens after "relocation from Lanyu"? The site selection for the ultimate repository is still stalled, and local resistance in Daren Township, Taitung County, remains unresolved. Technically feasible ≠ Politically feasible ≠ Ethically feasible. Lanyu is where these three gaps are concretized.

PanSci reports that spent nuclear fuel rods remain hot and highly radioactive after reactor decommissioning and must be cooled in the plant's fuel pool for at least 5 years before they can be moved. The dry storage facility land issue for Nuclear One and Nuclear Two has been stalled for over 11 years, with the New Taipei City Government refusing to approve the dry storage facility, causing spent fuel rods to remain in the plant's fuel pool, exceeding the original design capacity⁷⁸. "The biggest obstacle to nuclear power extension is the destination of spent nuclear fuel," citing industry consensus from PanSci, is the most awkward background noise for the Nuclear Three re-operation application⁸.

The Physical Limits of Nuclear Waste

Shift the lens to Olkiluoto Island in southern Finland.

500 meters underground, in the granite layer, a 5-kilometer-long tunnel is dug. At the end of the tunnel lies Onkalo, humanity's first high-level radioactive waste ultimate repository to officially receive an operating license. In August 2024, the Finnish Nuclear and Waste Safety Authority (STUK) issued the license. This project, planned since the 1970s, took nearly half a century⁹.

The entrance to Finland's Onkalo high-level radioactive waste ultimate repository, in the granite layer 500 meters underground, received an operating license in 2024. Image: kallerna, CC BY-SA 4.0, via Wikimedia Commons

Onkalo's design goal is to isolate nuclear waste for over 100,000 years. How extreme is this timescale? Human civilization is about 10,000 years old, the oldest pyramids are 4,500 years old, and 100,000 years ago our ancestors had not yet left Africa¹⁰.

💡 Did You Know? The timescale for isolating nuclear waste in an ultimate repository is so long that it reaches the "end of human memory." The Onkalo design team spent years discussing a question: how to make humans 100,000 years from now know "do not dig here"? Because by then, no existing language, symbol, government, or religion will remain. The final solution uses nuclear waste warning symbols with multilingual warnings, but designers admit this is only a "message for the next 1,000 years." What happens after 1,000 years, no one knows the answer.

What about Taiwan's ultimate repository? The candidate site for the low-level radioactive waste ultimate repository is Daren Township, Taitung County, but the site selection process is stuck under local political resistance¹¹. The site selection process for the high-level radioactive waste ultimate repository has not even started. Finland took 50 years to reach trial operation; Taiwan has zero years.

PanSci also mentions another "physical limit solution" once seriously discussed: space disposal of nuclear waste. "The idea of space disposal of nuclear waste is physically feasible, but requires very stable and reliable rockets; otherwise, if a launch fails, the radiation pollution caused to Earth would be incalculable"¹². SpaceX's Falcon 9 failure rate is about 1%, meaning 1 in 100 launches would explode high-level radioactive nuclear waste into the atmosphere. Physically feasible, but also physically unfeasible.

This is the physical limit behind the four words "nuclear waste solved." Its timescale is longer than the entire human civilization.

Hydrogen Rainbow: Green, Blue, Gray, White, Gold

If nuclear power is too heavy, can we bypass it?

In the past five years, hydrogen has been seen as the next wave of energy transition. The problem is that hydrogen itself is an energy carrier, not an energy source: other energy must first be used to "produce" hydrogen, which is then used for power generation or as fuel. Where it comes from determines if it is truly "clean."

PanSci categorizes hydrogen by color based on production methods: "Hydrogen color codes correspond to different production methods: Gray Hydrogen (Natural Gas SMR, emits CO₂), Blue Hydrogen (Gray Hydrogen + CCS), Green Hydrogen (Renewable Electricity Electrolysis), Blue-Green Hydrogen (Methane Pyrolysis, solid carbon capture, no CO₂ emission). From a carbon emission perspective, Green Hydrogen is ideal but has the highest cost"¹³.

Hydrogen Color	Production Method	Carbon Emission	Cost	Taiwan Status
Gray Hydrogen	Natural Gas Steam Methane Reforming (SMR)	High (emits CO₂)	Low	Most used in industry
Blue Hydrogen	Gray Hydrogen + Carbon Capture and Storage (CCS)	Medium (reduced after CCS)	Medium-High	No commercial operation
Green Hydrogen	Renewable Electricity Electrolysis	Zero	High	CPC Planning
Blue-Green Hydrogen	Methane Pyrolysis (Decarbonization Burning Hydrogen)	Zero (produces solid carbon)	Medium	Hsingda Power Plant Trial
White/Gold Hydrogen	Naturally Formed Underground	Zero (no production needed)	To Be Explored	None

Taiwan's hydrogen trials are at Hsingda Power Plant in Kaohsiung. Taipower collaborates with the Academia Sinica to test "decarbonization burning hydrogen" technology: decomposing natural gas (methane) into hydrogen and solid carbon at high temperatures, producing no carbon dioxide, with solid carbon usable as an industrial raw material¹³. The attraction of this technology is that it can use existing natural gas infrastructure, avoiding the need to rebuild the entire energy system.

But hydrogen has its own physical limits. "Although hydrogen is a clean energy, its effect as a greenhouse gas is 11.6 times that of CO₂ (GWP100); if it leaks during production, transportation, or use, it will exacerbate warming"¹⁴. The hydrogen molecule is the smallest in the universe, and leakage rates are inherently high. This is a physical limit of materials science, not something that can be fully overcome by engineering efforts.

There is also a newer role: White/Gold Hydrogen. A 2023 study by the U.S. Geological Survey (USGS) estimated that underground hydrogen reserves formed naturally by tectonic movements may reach "tens of billions of tons," sufficient to provide energy for human needs for hundreds of years¹⁴¹⁵. France and Mali already have commercial exploration. Taiwan's plate boundaries are active, theoretically holding potential, but currently there are no exploration plans, making it the energy option furthest from theory.

📝 Curator's Note: The core takeaway for readers from the hydrogen rainbow classification is to always ask "where does the energy come from" behind the term "clean energy." Green hydrogen is only cost-effective when renewable electricity is in surplus with nowhere to sell; this scenario has not yet arrived in Taiwan. Until then, hydrogen is actually a showroom for another fossil fuel.

Geothermal Taiwan: 33 GW Potential vs. 7.4 MW Reality

If hydrogen is a "carrier debate," geothermal is a "depth debate."

Taiwan should naturally be a geothermal powerhouse. Located at the junction of the Eurasian Plate and the Philippine Sea Plate, volcanoes, hot springs, and earthquake zones form a natural geothermal resource bank. In 1981, the Yilan Qingshui Geothermal Power Plant built a 3 MW experimental unit, Taiwan's first geothermal power plant. However, due to technical issues such as wellbore scaling and acid corrosion, it closed in 1993.

For the next 30 years, geothermal in Taiwan fell into silence. Until 2020, the privately invested Qingshui Geothermal 4.2 MW unit restarted commercial operation, bringing geothermal back into public discussion. In 2024, the Yilan Tufang Geothermal 5.4 MW project broke ground, expected to start in early 2026. By the end of 2025, Taiwan's total geothermal commercial capacity will be 7.4 MW¹⁶.

What is the government's official target? 200 MW by 2030, 6 GW (6,000 MW) by 2050. From 7.4 MW to 200 MW, a 27-fold gap; to 6 GW, an 810-fold gap. This is a 5-year and 25-year timeline.

PanSci cites National Taiwan University research stating, "Taiwan's geothermal resources are widely distributed. According to NTU research, deep geothermal (depth below 5 km) potential power generation reaches 33,640 MW, equivalent to about 12 Nuclear Four plants"¹⁷. But this is only a theoretical value. Developing deep geothermal requires EGS (Enhanced Geothermal System) technology, drilling wells several kilometers underground and injecting water artificially to create heat exchange layers. Currently, only a few demonstration projects exist globally, and the technology is not yet commercialized.

And "Taiwan's shallow geothermal (depth within 3 km) development potential is estimated at no more than 1,000 MW, with several experimental projects currently underway in Yilan Qingshui and Taipei Datun Mountain areas"¹⁷. Even if shallow geothermal is fully opened, it would only provide about 3% of Taiwan's total electricity demand.

Geothermal's advantage is stability. "Geothermal's advantage is that it is not affected by weather like wind or solar, providing stable 24-hour baseload power generation, giving it unique value in the energy mix"¹⁸. There are few renewable energies that can replace nuclear baseload functions; geothermal is one of them, provided it can actually be built.

⚠️ Controversial Viewpoint: Geothermal development in Taiwan is often attributed to "immature technology." But PanSci's industry interviews reached a different conclusion: the real bottleneck is underground uncertainty + difficulty in loans. Before drilling a geothermal well, no one can guarantee water output, temperature, or duration. Banks do not lend; operators dare not invest. Japan and New Zealand face similar dilemmas, but both have government-led funds sharing risks. Taiwan's geothermal developers currently use solar PV financing models: solar generates power once installed; geothermal does not. Copying the financing structure guarantees stagnation.

Public Television's Our Island produced a two-part series "Using Heat for Power" in March 2023, following the national team into Datun Mountain, Yilan, and Taitung tribes, comprehensively filming "geothermal exploration observation" and "tribal geothermal power generation":

Public Television's Our Island Official Channel: Episode 1195 "Geothermal Exploration National Team Mobilized" (2023-03-06). In Datun Mountain and Yilan Jiaoxi, follow the Ministry of Economic Affairs' Geological Survey Bureau, Taipower's Research Institute, and ITRI exploration teams to see how "geothermal mushrooms" are pieced together from three clues: seismic waves, rock samples, and well temperature gradients. How much heat is in Taiwan's underground only becomes concrete from here.

Public Television's Our Island Official Channel: Episode 1196 "Geothermal Power Generation in Tribes" (2023-03-13). Simultaneously presenting two levels: "energy transition vs. local justice"—Yilan Lize, Taitung Hongye, Chiayi Zhonglun. Geothermal development reaches which tribal negotiation line? Technical issues are one thing; sociologically, whether it can enter is a completely different matter.

Ocean Energy: The Trial Stage of Black Current's 9.4 GW

After the ground, there is the sea.

The Black Current (Kuroshio) off eastern Taiwan is one of the world's strongest currents. Flow speed 1.5-2.5 meters/second, width about 100 km, flowing north year-round. Theoretically, this is a continuous river of energy. The Academia Sinica completed offshore testing of a 100 kW experimental unit in 2021, a milestone for Taiwan's ocean energy development¹⁹.

PanSci cites Academia Sinica estimates: "Taiwan's surrounding waters have huge renewable energy potential. Ocean energy (including current energy, wave energy, thermal energy) theoretical potential is estimated at 9.4 GW. The Black Current flowing past Taiwan's east coast is the most promising current energy source"¹⁹.

Another direction is OTEC (Ocean Thermal Energy Conversion): using the temperature difference between surface warm water (25-28°C) and deep cold water (5°C) to drive generators. Taiwan's eastern waters have large depth differences, considered ideal for OTEC. "Taiwan's eastern waters have large depth differences, theoretically ideal for OTEC development, but currently still in the experimental stage"²⁰.

But ocean energy's physical limit is encountered even earlier than geothermal: marine engineering durability. Typhoons, salt corrosion, biofouling, deep-water maintenance—each is a century-level engineering challenge. Internationally, OTEC has no commercial power plants operating; the global leading case for Black Current power generation is Japan's Okinawa 100 kW demonstration. Taiwan's 100 kW trial is just the starting point; from here to commercialization, international experience shows it takes 15-20 years.

Fourth-Generation Nuclear Power SMR: Bill Gates' Bet

If we go back to the path of nuclear power, is fourth-generation nuclear power the answer?

PanSci reports: "The biggest difference between the Natrium reactor and traditional nuclear power plants is its coolant. Traditional nuclear reactors use water as a coolant, while Natrium uses liquid metal sodium. Sodium has a high boiling point, allowing operation at higher temperatures to improve reaction efficiency; sodium's thermal conductivity is 100 times that of water"²¹.

This is the sodium-cooled fast neutron reactor promoted by TerraPower, founded by Bill Gates. In April 2026, TerraPower's Natrium project officially broke ground in Kemmerer, Wyoming, expected to be completed in 2030²², one year later than originally scheduled, but still a key milestone for fourth-generation nuclear power commercialization.

The selling point of fourth-generation nuclear power is "Small Modular Reactors" (SMR): generating capacity drops from traditional 1000 MW class to 100-300 MW, factory-prefabricated and site-assembled, theoretically reducing costs and shortening construction time.

But physical limits remain. PanSci points out two key risks:

"Fast neutron reactors require high-concentration uranium fuel, and breeding reactions produce Plutonium-239, an important raw material for nuclear weapons. Therefore, how to manage nuclear materials and prevent nuclear proliferation becomes a challenge fast neutron reactors must face"²³.

"The construction of the Natrium reactor marks a major advance in fourth-generation nuclear power plant technology, but its development also comes with significant challenges"²⁴. Liquid sodium reacts violently and is flammable when it contacts water; reactor operation and maintenance have extremely high requirements for materials science, and there is currently no large-scale commercial operation safety data.

Does Taiwan have an SMR plan? Currently no official planning. Even if evaluation starts now, from site selection, environmental impact assessment, safety review to commercial operation, international experience shows it takes 15-20 years. That is, fourth-generation nuclear power is not the answer for 2050 net-zero; the most optimistic scenario is online by 2045-2050.

📝 Curator's Note: Fourth-generation nuclear power is often treated in international discourse as "future nuclear power," thus becoming "a good excuse to delay current energy transition": since better technology will be available in 15 years, why rush now? This is the most dangerous confusion in physical limit questions. Renewable energy's engineering bottleneck is "not built enough yet"; fourth-generation nuclear power's bottleneck is "safety and non-proliferation data for commercial operation not yet accumulated." The two timelines cannot replace each other. Missing the 2030 renewable energy construction window, 2045 SMR cannot save the climate bill.

Offshore Wind: The Leading Piece in Asia

Shift the lens back to what has already happened.

Hai Long Offshore Wind Farm (Formosa 1) off Miaoli, commercial operation started in 2019, Taiwan's first large-scale offshore wind farm. Image: Ministry of Economic Affairs, Attribution, via Wikimedia Commons

The Taiwan Strait is one of the world's best wind farms. "Due to topographical factors, the Taiwan Strait forms a 'pipe effect,' making wind speeds in the strait far higher than surrounding seas, making Taiwan one of the most potential locations for offshore wind development globally"²⁵. Winter Northeast Monsoons are squeezed into the strait by the Central Mountain Range and Fujian hills, with average wind speeds of 10-12 meters/second. This geographical fact makes offshore wind the core bet for energy transition.

From almost zero in 2016 to a cumulative installed capacity of about 4.5 GW by March 2026³, Taiwan's offshore wind expansion speed ranks among the highest in Asia. Ørsted from Denmark completed the construction of 920 MW for the Da Changhua Southwest Phase II and Northwest wind farms off Changhua²⁶. The third phase block development starting in 2026 has an allocated capacity of 3.6 GW, targeting completion and grid connection by 2030-2031³.

The government's blueprint is larger: 13 GW offshore wind by 2030, challenging 55 GW by 2050.

But offshore wind turbines bring not only power but also conflict. In February 2022, over a hundred Changhua fishermen went north to the Executive Yuan to protest, accusing the government of "eliminating fishermen" for wind power²⁷. Navigation zone bans for offshore wind farms blocked seas where they had operated for generations. In May 2025, a court ruled navigation restrictions illegal, the first time a court challenged offshore wind spatial governance in Taiwan²⁸.

Solar PV takes another path. In 2024, solar PV installed capacity reached 14,281 MW, accounting for 68% of total renewable energy, with 14.9 billion kWh generated²⁹. Rooftop, ground-mounted, floating, agrivoltaics—diverse installation modes make solar the mainstay of renewable energy. But agrivoltaic policies have sparked质疑 of "fake farming, real power generation," forcing the Ministry of Agriculture to strengthen inspections. On an island with only 790,000 hectares of arable land, every piece of land's use is a political issue.

Xihu Service Area rooftop solar panels. Taiwan's solar PV installed capacity reached 14,281 MW in 2024, accounting for 68% of renewables. Image: lienyuan lee, CC BY 3.0, via Wikimedia Commons

Wind and light are Taiwan's fastest piece of the energy transition puzzle, but inherently intermittent: no power when the sun sets, no power when the wind stops. This is also the most common argument used by pro-nuclear advocates in the Nuclear Three extension discussion: "Renewables are unstable, need baseload." The question returns to the geothermal section: baseload renewables built too slowly, a 27-fold gap in targets, is the political support for Nuclear Three's 2028 restart timeline.

The Afternoon of May 13

At 2:37 PM on May 13, 2021, at the Luzhou Extra-High Voltage Substation of Hsingda Power Plant in Kaohsiung, an operator opened switch 3541; he should have opened 3542³⁰.

This human error triggered a busbar grounding fault, tripping four units, instantly losing 2.2 GW of generation capacity. Starting at 3 PM, Taiwan implemented six rounds of regional power rationing, each 50 minutes, affecting about 4 million households. To make matters worse, solar power generation decreased with sunset, and drought reduced hydropower. By 7 PM, coal units came back online, and full power restoration was achieved at 8 PM.

Four days later, on May 17, Hsingda Unit 1 failed again, bringing a second round of blackouts. The two events combined affected over 5.62 million households³⁰.

May 13 and May 17 exposed how fragile a transforming power system is, far beyond just human error. The government's solution is energy storage: planning 1.5 GW battery storage by 2025, expanding to 8.6 GW by 2030. But storage costs remain high, and technology is still maturing.

This is the most honest side of energy transition: the old system is no longer sufficient, the new system is not yet ready. Whether the Nuclear Three extension referendum passes or not cannot change this reality, only delay or accelerate the time of facing it.

Giving Carbon a Price

On August 7, 2023, the Taiwan Carbon Exchange was listed in the Asia New Bay Area of Kaohsiung, with initial paid-in capital of 1 billion NTD, planned capital of 1.5 billion NTD; the Stock Exchange contributed 600 million NTD, and the National Development Fund contributed 400 million NTD³¹. On December 22, the same year, the first batch of international carbon credit transactions was completed: 45 enterprises purchased about 88,500 tons of CO₂ equivalent international carbon credits for over $800,000³¹.

In 2025, the domestic carbon fee system officially went online, and Taiwan entered the "Year of Carbon Pricing"³². Formosa Plastics' energy efficiency improvement project was listed at 3,000 NTD per ton; Hanbao Agriculture's biogas power generation project was priced at 3,000-4,000 NTD. But the market is still exploring: trading volume is low, and enterprises generally believe domestic carbon credit prices are too high.

Meanwhile, tech giants like TSMC and Foxconn are already seizing the beachhead in another battlefield. Under the RE100 initiative, these enterprises commit to using 100% renewable energy. TSMC plans to reach net-zero emissions by 2050. When international customers make green power a supply chain threshold, green power supply becomes an industrial survival issue, not just an environmental one.

The EU Carbon Border Adjustment Mechanism (CBAM) will enter formal system after 2026, increasing carbon costs and reporting pressure for high-carbon products like steel, cement, aluminum, fertilizers, electricity, and hydrogen exported to Europe³³. Taiwan's manufacturing industry is dominated by energy-intensive industries; steel, petrochemicals, cement, and papermaking account for 60% of industrial emissions. This is another physical limit, a timeline set by international trade structure for Taiwan.

In the 2024 National Day Speech, Lai Ching-te announced the launch of "Secondary Energy Transition," covering three directions: diverse green energy, deep energy saving, and advanced energy storage³⁴. But renewable energy share in 2025 remains significantly below the original 20% target; depending on statistical口径, it falls between 12.7% and 13.1%³⁵. The Ministry of Economic Affairs has changed its estimate to reach 20% from November 2026, and about 30% by 2030.

Algae Reef, Dyak, Meigong: Fault Lines of Environmental Justice

Every energy path has its opponents, and every opponent has their history.

Taoyuan Algae Reef. The 2021 "Save the Algae Reef Referendum" (Case 20) opposed Taipower building the Third Natural Gas Receiving Terminal on the Datan Coast, aiming to protect the world's largest columnar algae reef terrain. The referendum failed, and the three-terminal compromise plan went online: port area pushed out, avoiding high-density algae reef areas. Algae reef scholars still believe the EIA was insufficient, but the EIA Committee passed the review in 2023. The controversy has not subsided; this is the physical/ecological intersection of "building natural gas for carbon reduction, moving algae reefs for natural gas."

Lanyu Dyak. From 1982 to 2026, 44 years of nuclear waste storage history is Taiwan's longest environmental justice wound. Dyak residents continued to protest relocation delays in 2024; in May of that year, the NRC announced requiring Taipower to complete relocation by 2029. But where to relocate remains unanswered⁴.

Meigong Anti-Dam. The Meigong Anti-Dam Movement in the 1990s,凝聚ing resistance with "Meigong Yellow Butterfly Festival" and "Hakka Spirit," ultimately forced back the dam project, a classic of Taiwan's community-based environmental movement. Reading Meigong today, one finds its spirit still influencing other energy battlefields: every wind turbine, every solar panel, every transmission line entering a locality encounters the response "we oppose the transition costs being borne by us, not opposing energy transition itself."

📝 Curator's Note: Common environmental justice discussions often stop at "balancing development and environmental protection," but this framing flattens the problem. The true commonality of Lanyu, Algae Reef, and Meigong is: they are all post-1980s decision legacies, paid for by 1990-2020s social movements. Energy transition will produce many new "Lanyus" and "Algae Reefs" before 2050 (Changhua fishermen for offshore wind, Yilan indigenous for geothermal, Tainan salt fields for solar). The real question is "can we avoid repeating the 1982 decision model."

For detailed environmental justice historical context, see Taiwan Environmental Movement History and Taiwan Marine Pollution Control and Conservation Challenges.

9 Trillion NTD and Physical Limits

Putting all energy on the same table, the gap in physical limits becomes visible.

Energy	Taiwan Theoretical Potential	2025 Status	Government Target / Timeline	Main Physical Limit
Offshore Wind	One of the best globally	4.5 GW	13 GW by 2030, 55 GW by 2050	Marine Engineering / Fishery Conflict
Solar PV	Rooftop + Agrivoltaics	14.3 GW	31 GW by 2030	Land Acquisition / Intermittency
Geothermal (Shallow)	≤ 1,000 MW	7.4 MW	200 MW by 2030, 6 GW by 2050	Underground Uncertainty / Financing
Geothermal (Deep EGS)	33,640 MW (Theoretical)	Lab Stage	2040+	EGS Technology Not Commercialized
Ocean Energy	9.4 GW (Theoretical)	100 kW Trial	2030+	Marine Engineering Durability
Hydrogen (Green)	Requires Large Renewable Power	Hsingda Trial	2030+	Electrolysis Cost / Leakage GWP
Nuclear Three Extension	1,902 MW	2025 Shutdown	Earliest 2028 Restart	Nuclear Waste / Nuclear Safety Review
Fourth-Gen Nuclear SMR	No Local Plan	US 2030 Trial	2045+	Sodium Safety / Nuclear Proliferation

This table answers one question: Without nuclear power, can Taiwan achieve 2050 net-zero?

Technically, yes. The NDF's roadmap lists 12 key strategies, estimating 9 trillion NTD investment³⁶. But it requires offshore wind, solar, geothermal, ocean energy, hydrogen, and storage to simultaneously achieve their respective targets. And currently, geothermal is 27 times short, ocean energy is at kW level, hydrogen is in trial, storage costs are high.

Every physical limit is time.

In Hsu Hsiung-hsiung's model, Taiwan after 2060 has no winter³⁷. In coastal risk assessments, western low-lying areas face higher sea-level rise and storm surge pressure³⁸. From 1911 to 2020, Taiwan's annual average temperature has risen 1.6°C, almost one and a half times the global average (1.1°C) for the same period³⁷.

An Island Warmed by 1.5 Degrees

In the summer of 2017, Hsu Hsiung-hsiung from the Academia Sinica's Center for Ecology and Evolutionary Biology stared at data on his screen and made a prediction colleagues were afraid to say publicly: if emission trends continue, Taiwan's winter may disappear completely after 2060³⁷. Winter days drop to zero, summer extends to seven months.

This is not science fiction. Days above 35°C in Taipei, from 3 days a year in the 1960s, surged to 15 days in the last decade³⁹. The south is more severe; Tainan and Kaohsiung annual high-temperature days have exceeded 30 days.

In the same building, Wang Chung-ho from the Institute of Earth Sciences calculated another set of numbers. His conclusion was equally unsettling: Taiwan's surrounding sea-level rise speed is twice the global average³⁸. Multiple simulations point out that sea-level rise and storm surges will increase flooding risks for Taiwan's western low-lying coasts; among the six municipalities, New Taipei, Tainan, and Kaohsiung's exposed population and land area are particularly concerning.

Rain's temperament has also changed. Taiwan's total rainfall has not significantly decreased, but it doesn't fall when it should, and falls heavily when it does. Spring rainfall sharply decreased, dry seasons drier. In 2021, Taiwan suffered its most severe drought in 56 years, reservoir storage rates hit historical lows, and TSMC once sent water trucks to factories for water⁴⁰. In May of the same year, two major blackouts struck the island consecutively.

Days with over 200 mm of rainfall increased from an annual average of 5 days in the 1960s to 8 days recently. Typhoon Morakot in 2009 set a cumulative rainfall record of 2,884 mm at Alishan⁴¹, rain falling in three days equivalent to Taipei's annual rainfall. In that typhoon, Lin Village in Jiaxian, Kaohsiung was buried by collapsing earth from Xiandu Mountain at dawn, killing 491 people⁴².

"Every chair represents one family member." Survivor Wang Minliang later said to visitors in the Lin Memorial Park like this. He established the Sunlight Lin Community, touring Taiwan with the tribe's Grand Manzu Dance Troupe. (Quote from Public Television's Our Island)

Hsu Hsiung-hsiung's主持 (hosted) 2024 National Climate Change Science Report points out: extreme rainfall events once occurring once every 50 years may come once every 10 years in the future⁴³. Yunlin, Tainan, and Keelung are the areas with the highest coastal flood risk.

For an island of 23 million people, Taiwan's carbon emissions are disproportionately large: measured by fossil fuel CO₂ emissions, annual emissions are about 280 million tons, per capita about 11.7 tons, ranking in the global top tier; depending on different databases and statistical口径, ranking is about 20th globally⁴⁴. Emissions are highly concentrated in energy use and power supply, with the energy sector having the highest proportion; the power generation structure remains the core of carbon reduction pressure. The root of the problem lies in the power generation structure: in 2024 Taiwan's power generation mix, natural gas was about 42.4%, coal about 39.3%, natural gas exceeding coal for the first time; renewables about 11.6%, nuclear about 4.2%³⁵. This is an energy system still highly dependent on fossil fuels, and Taiwan relies on imports for 98% of its energy. Energy security and climate crisis are the same problem.

Parallelism of Democracy and Physics

The evening of August 23, 2025, the Nuclear Three referendum pushed all contradictions of this problem onto the ballot screen.

74% approval, 29.53% turnout, missed threshold, Taipower submission in March 2026, earliest restart 2028. Simultaneously: Lanyu 97,672 barrels, Finland Onkalo took 50 years, geothermal 27 times short, ocean energy still at 100 kW, fourth-generation nuclear power to 2045. Every number asks: Can the speed of democracy keep up with the speed of physics?

Democratic Timeline	Physical Timeline
Aug 23, 2025 Referendum Result	Lanyu opened 1982, still there by 2057
Aug 25, 2025 Three Principles Press Conference	Nuclear Waste Isolation for 100,000 Years
Mar 27, 2026 Taipower Submission	Ultimate Repository Finland Took 50 Years
Earliest 2028 Restart	Geothermal 27 Times Short
2050 Net-Zero Target	Ocean Energy Still at 100 kW Trial

Whether 9 trillion NTD can buy a different future, no one knows. But the consequences of not spending this money, we are already seeing: Hsu Hsiung-hsiung's 2060 no winter, Morakot's 2,884 mm, May 13 rolling blackouts, Algae Reef Referendum撕裂 (tearing), Lanyu's 44-year wait.

PanSci reports citing industry consensus, "The fastest progress globally for ultimate repositories is Finland's Onkalo project, receiving an operating license in August 2024. This project planned since the 1970s took nearly half a century to reach trial operation"⁹. Taiwan's ultimate repository site selection is not even done. Even if Nuclear Three restarts in 2028, every new fuel rod produced during the restart period must find a place to be stored.

Lanyu's 97,672 barrels will not disappear because the referendum passed or failed. They are there now, likely still there in 2029, still there in 2057 (if relocation is missed again).

✦ On August 23, 2025, the referendum failed. On March 27, 2026, Taipower still submitted. Between these two dates, physical limits did not change once. What changed is whether we are willing to admit that this island, relying on imports for 98% of its energy, is queuing to face all physical limits no one wants to face.

---

Further Reading:

• Taiwan and Nuclear Energy Discussion — This article writes about energy and physical limits; that article writes about the nuclear energy debate itself: 40 years of anti-nuclear/pro-nuclear, three referendums, and social struggle over Lanyu nuclear waste.
• Taiwan Environmental Movement History — From anti-nuclear to anti-air pollution, how Lanyu Dyak, Meigong Anti-Dam, and Algae Reef Referendum shaped today's energy politics.
• Taiwan Marine Pollution Control and Conservation Challenges — Nuclear Three plant outflow 80% coral bleaching, marine waste, and ecological intersection of offshore wind.
• Taiwan Hot Springs and Geothermal — From Qingshui Geothermal 1981 failure to 2024 restart, how 30 years of geothermal silence formed.
• Taiwan Environmental Justice and NIMBY Controversies — Lanyu, Algae Reef, Meigong: distribution politics of energy transition costs.
• Taiwan Industrial Transformation and Upgrading — From energy-intensive manufacturing to green energy industries, TSMC RE100, CBAM, and the energy bill of the "Guardian Mountain."
• Taiwan Agricultural Modernization Development — Agricultural transformation pressure and land use conflicts behind agrivoltaics.
• Plum Rain — Climate change local observations of "Spring rain not coming, Plum Rain concentrating."

References

Image Sources

• Nuclear Three Plant Appearance (Pingtung Hengchun Maanshan, hero): Maanshan Nuclear Power Plant, Nan Wan (Photography: M. Weitzel, Wikimedia Commons, CC BY-SA 3.0)
• Onkalo Underground Repository (Finland Olkiluoto): Onkalo spent nuclear fuel repository entrance (Photography: Posiva Oy, Wikimedia Commons, CC BY-SA 4.0)
• Hai Long Offshore Wind Farm off Miaoli: Hai Long offshore wind farm (Wikimedia Commons, CC BY-SA 4.0)
• Highway Service Area Rooftop Solar Panels: Xihu Service Area solar panels (Wikimedia Commons, CC BY-SA 3.0)

1. Central Election Commission: August 23, 2025 National Referendum Results Announcement; CNA: Nuclear Three Extension Referendum 4.34 Million Yes Votes Miss 1/4 Threshold, Failed; Taipower: Nuclear Three Re-Operation Plan Submitted to NAC for Review Explanation (2026/03/27) — Aug 23, 2025 Nuclear Three Extension Referendum: Yes votes 4,342,206 (74.17%), No votes 1,511,693, turnout 29.53%, missed Referendum Act 1/4 threshold of total voters (5,005,230 votes), referendum failed. Taipower submitted Nuclear Three re-operation plan application to NAC on March 27, 2026, expected safety inspection schedule about 18 months, earliest restart 2028.↩
2. CNA: Lai Ching-te Post-Nuclear Three Referendum Talk Proposes Nuclear Safety, Nuclear Waste, Social Consensus Three Principles — Aug 25, 2025 President Lai Ching-te formally responded to Nuclear Three extension referendum results, proposing future nuclear power restart "Three Principles": nuclear safety assured, nuclear waste solved, social consensus, and instructed Ministry of Economic Affairs and NAC to start safety inspection procedures for evaluation.↩
3. Ministry of Economic Affairs Energy Bureau: Offshore Wind Block Development Phase 3 Selection Mechanism Announcement — MOEA March 27, 2026 announcement, stating cumulative offshore wind installed capacity as of March 26, 2026 is about 4.5GW, Phase 3 allocated capacity is 3.6GW, targeting completion and grid connection by 2030-2031.↩
4. Nuclear Regulatory Commission: Lanyu Storage Facility Storage Volume Announcement (2024) — NRC official announcement, as of 2024 Lanyu low-level radioactive waste storage facility accumulated storage 97,672 barrels, since 1982 opening experienced 1996, 2002, 2016, 2019, 2023 multiple relocation promise delays, NRC requires Taipower to complete relocation by 2029. Content Curation Partner per MOU 2026-05-05.↩
5. Ministry of Economic Affairs Energy Bureau: Geothermal Power Generation Targets and Commercial Capacity (2025) — Government geothermal power generation policy targets: 200 MW by 2030, 6 GW (6,000 MW) by 2050; as of end of 2025 Taiwan total geothermal commercial capacity about 7.4 MW, mainly Yilan Qingshui Geothermal 4.2 MW and some small units, 27 times short of 2030 target.↩
6. Wikipedia: Lanyu Storage Facility — Before Lanyu Storage Facility opened in 1982, Taipower told Dyak residents it was building a "fish cannery," failing to fully inform nuclear waste storage nature; 1988 Dyak launched first "exorcise evil spirits" protest action, marking starting point of Taiwan indigenous environmental movement.↩
7. PanSci: Nuclear Two Exit, Nuclear Waste Still to Be Stored for 20 Years — Content Curation Partner per MOU 2026-05-05. Nuclear Two plant officially decommissioned end of 2023, but nuclear fuel rods remain hot and highly radioactive after reactor decommissioning, must be cooled in plant fuel pool for at least 5 years before movable; Daren Township, Taitung County is low-level nuclear waste ultimate repository candidate site, site selection process stuck under local political resistance.↩
8. PanSci: What Is the Real Problem of Nuclear Power Extension — Content Curation Partner per MOU 2026-05-05. Nuclear power plant dry storage facility land issue stalled for over 11 years, New Taipei City Government refused to approve dry storage facility, causing Nuclear One and Two spent fuel rods to remain in plant fuel pool and exceed original design capacity; nuclear power extension biggest obstacle is spent nuclear fuel destination.↩
9. PanSci: Since Nuclear Waste Has Nowhere to Go, Are There Other Methods — Content Curation Partner per MOU 2026-05-05. Fastest progress globally for ultimate repositories is Finland Onkalo project, receiving operating license August 2024, planned since 1970s took nearly half century to reach trial operation; ultimate repository needs to isolate waste for over 100,000 years, this timescale far exceeds human civilization existence time.↩
10. Posiva Oy: Onkalo Ultimate Repository Design Introduction — Finland Posiva Company operated Onkalo repository official explanation, design goal to isolate high-level radioactive nuclear waste for at least 100,000 years, including multi-barrier system (copper shell + bentonite + granite layer) and long-term memory warning system design.↩
11. Taitung County Daren Township Office: Low-Level Radioactive Waste Ultimate Repository Issue — Daren Township, Taitung County is one of two candidate sites for low-level nuclear waste ultimate repository (other is Kinmen Wuchiu), local public opinion divided, indigenous tribe opposition voices strong, site selection referendum not yet successfully held.↩
12. PanSci: Nuclear Waste Space Disposal Feasibility Analysis — Content Curation Partner per MOU 2026-05-05. Nuclear waste space disposal physically feasible but requires very stable and reliable rockets, if launch fails radiation pollution to Earth incalculable; with current rocket failure rate, 1 in 100 launches risk, not meeting engineering practice needs.↩
13. PanSci: Improved Natural Gas Power Generation Technology Does Not Produce CO₂? Gray Hydrogen, Blue Hydrogen, Green Hydrogen — Content Curation Partner per MOU 2026-05-05. Hydrogen color codes correspond to different production methods: Gray Hydrogen (Natural Gas SMR emits CO₂), Blue Hydrogen (Gray Hydrogen + CCS), Green Hydrogen (Renewable Electricity Electrolysis), Blue-Green Hydrogen (Methane Pyrolysis solid carbon capture no CO₂ emission); Taipower and Academia Sinica already collaborated at Hsingda Power Plant for decarbonization burning hydrogen technology test.↩
14. PanSci: Musk Disdains; Bill Gates Treasures Like Treasure! Hydrogen Energy — Content Curation Partner per MOU 2026-05-05. Besides Green Hydrogen, emerging White/Gold Hydrogen is naturally formed underground hydrogen, USGS estimates reserves may reach tens of billions of tons; but hydrogen GWP100 is 11.6 times CO₂, leakage will exacerbate warming (academic community still has 7-37 range controversy on GWP numbers).↩
15. USGS: Geological Hydrogen — A New Energy Frontier (2023) — U.S. Geological Survey 2023 geological hydrogen research report, estimating global underground natural hydrogen reserves may reach tens of billions of tons, sufficient to provide human energy needs for hundreds of years; France, Mali already have commercial exploration cases, Taiwan plate boundaries active but currently no exploration plans.↩
16. CNA: Yilan Tufang Geothermal Power Plant Breaks Ground, Expected 2026 Start — 2024 Yilan County Tufang Geothermal Power Plant 5.4 MW unit broke ground, expected start early 2026, Taiwan second MW-scale commercial geothermal power plant; as of end 2025 Taiwan total geothermal commercial about 7.4 MW, including Qingshui Geothermal 4.2 MW and other small units.↩
17. PanSci: Is Taiwan Developing Geothermal Power Generation Feasible (Part 1) — Content Curation Partner per MOU 2026-05-05. According to Taiwan University research, deep geothermal (depth below 5 km) potential power generation reaches 33,640 MW, equivalent to about 12 Nuclear Four plants, but development requires EGS enhanced geothermal system technology still in R&D stage; shallow geothermal (depth within 3 km) potential estimated no more than 1,000 MW.↩
18. PanSci: Geothermal Advantages and Taiwan Field Application — Content Curation Partner per MOU 2026-05-05. Geothermal not affected by weather, 24-hour stable baseload power generation, unique value in energy mix; but underground uncertainty leads to financing difficulty, fundamental bottleneck for Taiwan geothermal development delay.↩
19. PanSci: "Guardian Mountain" Higher, Power Pressure Greater: Taiwan Ocean Energy Is Solution — Content Curation Partner per MOU 2026-05-05. Taiwan surrounding waters ocean energy (current, wave, thermal) theoretical potential reaches 9.4 GW; Black Current flowing past Taiwan east coast most promising current energy source, Academia Sinica completed 100 kW trial unit test in 2021.↩
20. PanSci: Ocean Thermal Energy Conversion (OTEC) Taiwan Possibility — Content Curation Partner per MOU 2026-05-05. OTEC uses surface warm water (25-28°C) and deep cold water (5°C) temperature difference for power generation, Taiwan eastern waters large depth difference, theoretically ideal location, but currently still experimental stage, no commercial operation plants globally.↩
21. PanSci: Bill Gates' Fourth-Generation Nuclear Power Plant Finally Starts Construction — Content Curation Partner per MOU 2026-05-05. Natrium reactor biggest difference from traditional nuclear power plants in coolant: traditional uses water, Natrium uses liquid metal sodium; sodium high boiling point, can operate at higher temperatures to improve reaction efficiency, thermal conductivity 100 times water.↩
22. TechOrange: TerraPower Natrium Project 2026 Breaks Ground in Wyoming — TerraPower's Natrium fourth-generation nuclear power plant project officially breaks ground in Kemmerer, Wyoming in April 2026, slightly delayed from original plan, expected completion 2030, global fourth-generation sodium-cooled fast neutron reactor commercialization key milestone.↩
23. PanSci: Fourth-Generation Nuclear Power Nuclear Proliferation Risk — Content Curation Partner per MOU 2026-05-05. Fast neutron reactors require high-concentration uranium fuel, breeding reactions produce Plutonium-239, important raw material for nuclear weapons; how to manage nuclear materials, prevent nuclear proliferation, challenge fast neutron reactors must face.↩
24. PanSci: Natrium Reactor Safety Challenges — Content Curation Partner per MOU 2026-05-05. Natrium reactor construction marks fourth-generation nuclear power plant technology progress, but development comes with major challenges; liquid sodium reacts violently and flammable when contacting water, reactor operation and maintenance extremely high materials science requirements, currently lacks large-scale commercial operation safety data.↩
25. PanSci: Offshore Wind Turbine Construction Expensive and Troublesome, Why Taiwan Still Develops Heavily — Content Curation Partner per MOU 2026-05-05. Taiwan Strait due to topographical factors forms "pipe effect," making strait wind speeds far higher than surrounding seas, making Taiwan one of most potential offshore wind development locations globally.↩
26. PV Magazine: Taiwan solar and offshore wind targets — Reporting Ørsted completed 920MW construction for Da Changhua Southwest Phase II and Northwest wind farms, and Taiwan's plan to add 8.2GW solar and offshore wind by end of 2026.↩
27. Environmental Information Center: Changhua Fishermen Protest Offshore Wind (2022) — Reporting over a hundred fishermen going to Executive Yuan to protest offshore wind farm navigation zone bans blocking generational operating seas, shouting "Eliminate Fishermen" slogan.↩
28. Environmental Information Center: Court Rules Offshore Wind Navigation Restrictions Illegal (2025) — Taiwan's first court challenge to offshore wind spatial governance ruling, finding navigation restrictions infringe fishermen's rights, causing shock in energy circle.↩
29. Taipower: Renewable Energy Generation Statistics — Taiwan Power Company official statistics, recording annual data for various renewable energy installed capacity and generation; 2024 solar PV installed capacity 14,281 MW, generation 14.9 billion kWh.↩
30. Taipower: May 13 Blackout Accident Preliminary Investigation Out; Ministry of Economic Affairs: May 13 and May 17 Blackout Accident Review Report — Official data explains May 13 accident misoperation of switch 3541, causing instant reduction of about 2.2GW supply capacity and affecting about 4 million households, and organizing May 17 event and subsequent review.↩
31. Presidential Office Press Release: Taiwan Carbon Exchange Unveiling; Taiwan Stock Exchange 2023 Annual Report; Cnyes: Taiwan Carbon Exchange Expected to Go Online End of July, North-South Division Operation — Carbon Exchange unveiled August 7, 2023; planned capital 1.5 billion NTD, initial paid-in capital 1 billion NTD, Stock Exchange contributed 600 million NTD, National Development Fund contributed 400 million NTD. Stock Exchange annual report also records first batch international carbon credit transaction total 88,520 tons CO2e, 27 participating enterprises (45 including financial holding subsidiaries).↩
32. KPMG Taiwan: Carbon Pricing Trend Analysis (2025) — Analyzing Taiwan 2025 carbon fee system market dynamics post-launch, including domestic carbon pricing (Formosa 3,000 NTD/ton, Hanbao Agriculture 3,000-4,000 NTD/ton) and challenges of low trading volume.↩
33. EU Carbon Border Adjustment Mechanism Official Page — CBAM entered transition period in October 2023, fully implemented in 2026, covering six product categories: steel, cement, aluminum, fertilizers, electricity, and hydrogen.↩
34. Reccessary: Taiwan Energy Policy 2025 Outlook — Reporting Lai Ching-te 2024 National Day Speech announced "Secondary Energy Transition" policy direction: diverse green energy, deep energy saving, advanced energy storage.↩
35. Ministry of Economic Affairs Energy Bureau Statistics: Generation Mix by Fuel Type; Ministry of Environment Energy Information Platform: Power Structure; Economic Daily: MOEA Says Renewable Energy Share Can Reach 20% from Nov 2026 — According to MOEA Energy Bureau statistics, 2024 natural gas about 42.4%, coal about 39.3%, renewables about 11.5%-11.6%, nuclear about 4.2%; Ministry of Environment Energy Information Platform shows 2025 national total generation renewables share is 13.1%, while Taipower system purchase-generation structure common口径 is about 12.7%, two different口径. MOEA May 2025 stated, estimating 20% from Nov 2026, about 30% by 2030.↩
36. Presidential Office Press Release: Tsai Ing-wen 2021 Earth Day Talk — Tsai Ing-wen first declared as President "2050 Net-Zero Transition is the world's goal, and also Taiwan's goal," laying policy tone for subsequent NDF net-zero roadmap.↩
37. United Daily News Vision Project: Hsu Hsiung-hsiung Interview — Academia Sinica Center for Ecology and Evolutionary Biology Distinguished Researcher Hsu Hsiung-hsiung's team analyzed Taiwan temperature data from 1911-2020, pointing out Taiwan's century warming of 1.6°C, winter shortened by nearly half, worst case scenario winter days may reach zero after 2060.↩
38. CSRone Sustainable Think Tank: Wang Chung-ho Interview — Academia Sinica Institute of Earth Sciences Adjunct Researcher Wang Chung-ho has long tracked Taiwan sea-level changes, pointing out Taiwan's surrounding sea-level rise speed is higher than global average; article uses conservative risk description to avoid simplifying different research scenarios into absolute conclusions.↩
39. Central Weather Bureau Climate Change Information Platform — Taiwan climate observation historical database, recording century temperature, rainfall, extreme weather event records for each station, including Taipei 35°C+ day increase trend statistics.↩
40. BBC Chinese: Taiwan's 56-Year Worst Drought (2021) — Reporting 2021 central-southern Taiwan drought, reservoir storage rates dropped below 10%, tech factories like TSMC initiated water truck emergency measures.↩
41. National Center for Disaster Prevention and Rescue Technology Center: Morakot Typhoon Disaster Record — Official disaster archive, recording Morakot Typhoon Alishan station cumulative rainfall 2,884 mm, highest record in Taiwan meteorological observation history.↩
42. The Reporter: Lin Village Extinction Investigation — In-depth investigation of Lin Village Xiandu Mountain collapse process and 491 deaths context, including geological causes and warning system failure analysis.↩
43. Environmental Information Center: 2024 National Climate Change Science Report — Reporting key findings of Hsu Hsiung-hsiung's latest scientific report: 50-year extreme rainfall may become 10-year, 36°C+ high-temperature days may increase by 75 days.↩
44. Ministry of Environment Greenhouse Gas Emission Statistics — Taiwan official greenhouse gas emission database, recording annual national emission inventory, sector emissions, and per capita emission data.↩