先把问题问对，再谈转型。
A right question carries no assumption. It stays open, goes to the real data, and keeps splitting into the next question — until one of them lands on an action. Then the action is built: capability, pilot, adoption, deployment, integration — until it is simply how we decide. By default.
对的问题不带预设。它保持开放，回到真实数据，不断裂变出下一个问题——直到其中一个能落成行动。 然后把行动做实：能力建设、试点、推广、部署、融入系统与行为，直到它成为常识与默认判断。
Transparent by default · 透明为默认从一个问题，到默认动作
Six stages. Nothing is skipped, and nothing is declared “done” before it is boring.
六个阶段。不跳步；在它变得平常无聊之前，不算完成。
Write the question so that a “no” is still a useful answer. No conclusion smuggled in.
把问题写成——即使答案是「否」也依然有用。不预埋结论。
Primary sources, field measurement, the farm gate. Then visit. Numbers you cannot trace are opinions.
原始数据、田间实测、走到源头去看。追溯不到出处的数字，只是意见。
Each answer must open the next question. Stop only when a question names an owner, a place and a change.
每个答案都要打开下一个问题。直到某个问题能指名责任人、地点与要改的动作。
Skills, tools, measurement, incentives. Most pilots fail on readiness, not on technology.
先建能力：技能、工具、计量、激励。多数试点败在准备度，而非技术。
One district, one mill, one season. Publish the result including what failed. Then multiply.
一个片区、一家米厂、一个季。公开结果，包括失败之处，然后复制放大。
Contracts, specs, pricing, audit, job descriptions. When it needs no champion, it is done.
写进合同、规格、定价、审核与岗位职责。当它不再需要推动者，才算完成。
执行摘要：高品质、低排放稻米
Rice feeds more than half of humanity and is the largest single source of methane in the food system — roughly 8–10% of global anthropogenic methane, because flooded paddies are engineered anaerobic reactors. Around 90% of the world’s rice is grown in Asia, mostly by smallholders on fields smaller than a hectare.
“High-quality, low-emission rice” is therefore not a variety. It is a water-management decision, a straw decision, a nutrient decision and a record-keeping decision — verified per field, per season. The best-documented lever, alternate wetting and drying (AWD), cuts paddy methane by roughly 30–70% and irrigation water by about 25%, usually without yield loss, and often with lower arsenic in grain.
The constraint is rarely the agronomy. It is water control (levelled fields, a bund, a gate, a perforated water tube), trust (a farmer will not drain a field on a promise), and MRV — the ability to prove what happened without an auditor standing in every field.
稻米养活全球一半以上人口，同时是食物系统中最大的单一甲烷来源——约占全球人为甲烷排放的 8%–10%，因为长期淹水的稻田本质上是一个厌氧反应器。全球约 90% 的稻米产自亚洲，多为不足一公顷的小农户。
因此，「高品质、低排放稻米」不是一个品种，而是一组可核查的决策：水管理、秸秆处理、养分管理与记录留痕——按田块、按季核查。证据最充分的措施是干湿交替灌溉（AWD）：稻田甲烷可降低约 30%–70%，灌溉用水约省 25%，通常不减产，并常伴随籽粒砷含量下降。
瓶颈很少在农艺，而在：控水条件（平整田面、田埂、闸口、观察水管）、信任（农户不会为一句承诺去排水）、以及 MRV 计量核证——如何在没有审核员站在每块田里时，仍能证明真实发生了什么。
If it cannot be measured at the field, it cannot be claimed at the pack. 在田里量不出来的，就不该印在包装上。
什么算，什么不算
There is no single global “low-methane rice” label. There is a stack: a farm-level sustainability standard, a GHG quantification methodology, an accounting framework, and an assurance scheme. A credible claim names all four.
目前没有统一的全球「低甲烷稻米」标签，而是一个叠加结构：农场层面的可持续标准 + 温室气体量化方法学 + 核算框架 + 保证/审核机制。可信的声明必须四者具名。
SRP Standard for Sustainable Rice Cultivation (Sustainable Rice Platform, hosted with UNEP/IRRI): requirements across water, nutrients, pest management, harvest, labour and safety, with performance indicators including a GHG indicator. Verified through the SRP Assurance Scheme at three levels — self-assessment, internal (second-party), and third-party verified.
SRP 可持续稻米种植标准（可持续稻米平台，与 UNEP／IRRI 合作）：涵盖水、养分、病虫害、收获、劳工与安全等要求，并含温室气体等绩效指标。通过 SRP 保证体系分三级核查：自评、内部（第二方）、第三方核查。
方法学：IPCC 2019 修订版第四卷第 5.5 章（稻田排放因子与水分／有机物料／土壤调整系数，Tier 1–3）；CDM AMS-III.AU；Gold Standard 稻田水管理方法学；Verra VM0042；美国 Climate Action Reserve 稻作协议；日本 J-Credit 中期延长排水方法学。
GHG Protocol Land Sector & Removals Guidance for Scope 3 category 1 inventories; SBTi FLAG for target setting; ISO 14064-2 (project-level quantification), ISO 14067 (product carbon footprint), ISO 14068-1 (carbon-neutrality claims); ISO 14016 / 14065 for validation and verification bodies.
核算与声明：GHG Protocol 土地部门与移除指南（范围三第 1 类清单）；SBTi FLAG 目标设定；ISO 14064-2（项目层量化）、ISO 14067（产品碳足迹）、ISO 14068-1（碳中和声明）；ISO 14065／14016（核查机构与验证）。
验证流程：界定边界（田块 GPS 边界、季次、加工厂、生育期）→ 设定基线（历史水分制度、秸秆处理、产量；记录不足时取保守默认值）→ 监测计划（灌排日期、水位、秸秆去向、氮投入、产量；观察水管或传感器＋农户记录；Sentinel-1 雷达遥感判识淹水状态）→ 第三方事前审定（Validation）→ 季后核查（Verification：文件审阅、抽样、田间检查、农户访谈，出具带不确定度说明的核查声明）。核查机构须依 ISO 14065 获认可或经机制（Verra／Gold Standard／SRP）批准，如 Control Union、SGS、Bureau Veritas、TÜV、DNV 等。
不算的情形：只有品种名称、无田间水分数据；用区域平均值套到具体批次（无田块边界即无声明）；仅有模型推算、无实测排水记录；以氧化亚氮升高、籽粒镉升高或减产（导致扩张耕地）为暗中代价换取甲烷下降；重复计算（同一减排既卖碳信用又计入范围三）；自我宣称「已认证」却说不出认可机构、标准版本与核查日期。
全球现状与路径：2025 → 2050
Read the verdict as: is this geography measuring, moving and scaling on paddy methane — not whether it grows good rice. Figures are order-of-magnitude and must be checked against the primary sources listed at the end.
此处的判定是：该地区在稻田甲烷上是否在计量、在行动、在放大——而非稻米品质本身。数字为量级参考，请以文末原始来源核对。
这件事值不值得做？
Illustrative arithmetic with the assumptions written out, so anyone can disagree with a number instead of with a feeling.
以下是把假设明写出来的示意算术——好让人能对某个数字提出异议，而不是对一种感觉。
of global rice area adopting AWD or extended mid-season drainage by 2030 — an ambitious but not fantastical target where water control exists.
情景假设：到 2030 年，全球约 30% 稻作面积采用干湿交替或延长中期排水。
average methane reduction on those hectares — the mid-point of the 30–70% range reported for AWD.
这些面积上的甲烷平均下降约 40%（干湿交替 30%–70% 区间的中值）。
the resulting annual reduction — comparable to the total emissions of a mid-sized industrial country, from a practice whose main cost is timing.
由此产生的年减排量约 1 亿吨二氧化碳当量——相当于一个中等工业国的全年排放，而其主要成本只是「时机」。
How to scale it to default: put the practice in the purchase specification, not in the CSR report. Price the water, not the promise. Fund the three cheap enablers (levelling, a gate, a water tube). Pay the service provider per verified hectare. Then let the mill’s intake sheet, not a campaign, carry it.
如何放大到「默认」：把措施写进采购规格，而不是写进社会责任报告；为水定价，而非为承诺定价；资助三项廉价使能条件（平地、闸口、观察水管）；按已核查公顷数支付服务商报酬；最后让米厂的收粮单——而非一场宣传活动——承载它。
田间培训：做什么，为什么
Written to be taught standing in a field, in one morning, with a stick and a plastic tube. Every “what” comes with a “why” — a farmer who understands the reason will adapt it correctly when the season is odd.
这套内容是为「一个上午、站在田里、只带一根竹竿和一截塑料管」而写的。每一个「做什么」都配一个「为什么」——理解原理的农户，才能在异常年份做出正确调整。
甲烷由只能在无氧条件下工作的微生物产生。长期淹水使土壤缺氧；让水位下落使空气回到土中，这几天基本不再产甲烷。
埋设观察管（长 30 cm，入土 20 cm，土面以下打孔）。返青立苗后停灌；管内水位降至土面下 15 cm 时复灌至5 cm，循环往复。开花期严禁落干：幼穗分化至开花保持 5 cm 水层。
许多农户本就在中期排水一次以促根。只需将该次排水延长约一周，即可在既有习惯之上获得大部分甲烷减排效益。
记录开闸与复灌日期。若土壤未深裂、且未进入幼穗分化期，延长至 7–10 天。若中午叶片卷曲，提前复灌。
新鲜秸秆在淹水条件下翻埋，等于给产甲烷菌「投料」，排放可成倍上升。同样的秸秆在有氧条件下堆肥、做食用菌基质、饲料或垫料，排放低得多且可增收。焚烧最差：污染空气，碳白白损失。
收获后移出或打捆；若必须还田，应在干旱休闲期、淹水前至少 30 天翻入，或先行有氧堆沤。切勿在插秧前直接翻埋鲜秸秆。
留不住、放不走的水，无法管理。田面不平时，一角已干、另一角尚有 10 cm 水，农户只能为保高处而多灌，落干便无从谈起。
用激光平地机或拖板平地，控制在 ±3 cm；季前修补田埂与鼠洞；设置一进一出、真正能关严的闸口。
淹水天数越少，产甲烷天数越少。旱直播还可省去育秧、插秧用工与整田泡水的耗水。
必须具备除草能力与平整田面；使用合格种子并制定苗前除草方案。第一年先在 10%–20% 面积试行。
土壤落干可能增加氧化亚氮——其增温潜势约为二氧化碳的 270 倍。氮肥分次施用、按需匹配、深施，可避免甲烷减排被抵消。
用叶色卡或土壤检测决定用量；氮肥分 2–3 次施用；避免在排水前将尿素撒施于水面；全部记录留痕。
减排只有能被展示，才算存在。一行带日期的记录——「7 月 12 日开闸，7 月 21 日复灌，管内水位 −15 cm」——正是把好农艺变成可核查声明、并在有机制时变成收入的关键。
每田每季一页：田块编号与 GPS 点位、播种／插秧日期、每次灌排日期、水位读数、肥药名称与日期、秸秆去向、收获日期、产量、水分。
现场审核／田间检查清单
Sixteen checkpoints an auditor can complete in one field visit. Tick them here as a rehearsal — nothing is submitted anywhere.
十六个检查点，可在一次田间走访中完成。此处可勾选作为预演——不会提交到任何地方。
议题 0 是稻米。以下是排队中的十个问题。
Priority order = (emission or risk size) × (our influence) × (readiness of a credible standard). Each one gets the same treatment rice got: define, verify, map, train, audit, quantify, default.
优先级 =（排放或风险体量）×（我们的影响力）×（可信标准的成熟度）。每一项都走稻米走过的同一套流程：定义 → 验证 → 摸底 → 培训 → 审核 → 量化 → 默认化。
The worked example above. Standard stack exists; the gap is field-level MRV and water control.
Selected next — the largest agricultural methane pool. Brief opened below.
Soy, palm, beef, cocoa, coffee, rubber. Plot-level geolocation, legality, and a due-diligence statement.
Both a field-practice question and a supplier-process question. Also the main trade-off risk for topic 00.
Small volumes, extreme potency. Leak-rate data, retrofit plans, end-of-life recovery.
Where extended producer responsibility fees and consumer-visible claims collide.
Volume is the wrong metric alone; context of the basin is the question. Directly coupled to rice.
Boiler fuel switching, heat pumps, real hourly-matched renewable electricity rather than paper claims.
Modal shift, load factor, ocean and road fuels, and what a book-and-claim unit really represents.
Post-harvest drying and storage first — the cheapest emission reduction is the crop you do not lose.
No environmental practice survives on a farm that cannot pay for the next season.
The right question: for the milk and meat actually in our products, what is the measured methane intensity per kilogram of fat-and-protein-corrected milk or per kilogram of liveweight, by supplying region — and which of feed quality, herd productivity, manure management or additives changes it most, at what cost, verified how?
对的问题：就我们产品中实际使用的乳与肉而言，按供应区域计，每公斤脂蛋白校正乳或每公斤活重的实测甲烷强度是多少？在饲料品质、畜群生产效率、粪污管理与饲料添加剂之中，哪一项改变最大、成本多少、如何核查？
First data pulls: herd productivity by region, manure storage type (open lagoon vs. covered vs. daily spread), feed digestibility, additive regulatory approval status by market.
首批取数：各区域畜群生产效率、粪污储存方式（敞开式厌氧塘／覆盖／日清日运）、饲料消化率、各市场对添加剂的监管批准状态。
Methane first, because it is short-lived: reducing it changes temperature within our own planning horizon rather than a century from now. Then land conversion, because it is irreversible. Then nitrogen, energy and freight, because they are engineering problems with known answers. Then loss, packaging and people — because a supply chain that does not hold together socially will not hold together environmentally.
先甲烷，因为它寿命短——减下来能在我们自己的规划期内影响温度，而非一百年后。其次是土地转换，因为不可逆。再是氮、能源与运输，因为它们是答案已知的工程问题。最后是损耗、包装与人——一条在社会层面撑不住的供应链，在环境层面也撑不住。
这值得进入议程吗？值得——前提是它最终变成默认动作。
Baseline every sourcing region. Choose the standard stack and stick to it. Train trainers, not slides. Fix levelling, gates, tubes, records.
能力建设期：为每个采购区域建立基线；确定并固守标准组合；培训「培训者」而非做课件；解决平地、闸口、水管与记录。
Verified hectares in the purchase spec. Pay per verified hectare. Publish results including failures. Target: 30% of sourced volume on documented low-methane practice.
试点与推广期：把「已核查公顷」写进采购规格，按核查结果付费，公开成果与失败。目标：采购量的 30% 采用有据可查的低甲烷措施。
Contracts, supplier scorecards, product specifications and audit programmes carry it. Digital MRV replaces most manual sampling.
全面部署期：由合同、供应商评分卡、产品规格与审核方案承载；数字化 MRV 取代大部分人工抽样。
Low-emission is the base specification; conventional becomes the exception that needs justification. The topic leaves the agenda — because it has become common sense.
默认化：低排放成为基准规格，常规做法反而需要说明理由。议题因此从议程上消失——因为它已成为常识。
What’s next is not another pledge. It is the first field with a tube in it, and a dated line in a notebook. 下一步不是又一个承诺，而是第一块插了观察管的田，和笔记本上那一行带日期的记录。
数据与内容来源