Disaster Risk Reduction Day: A Supplier Continuity Questionnaire for Climate Shocks

Honda's plant at Rojana industrial estate stopped on 4 October 2011. By the end of January 2012 the company was still only forecasting a restart, telling the market it expected to resume production by the end of March and to normalise output across its operations during April.

Close to six months, from one flood. Guy Carpenter's retrospective a year on records that more than 684 factories, over 80 per cent, across the seven Thai industrial estates that went under that October had fully or partially reopened by late September 2012, which is another way of saying that nearly a fifth had not managed it inside a year. Smaller electronics makers were running at roughly 40 per cent of capacity twelve months on, with half their factories still idle.

Nissan's Thai plant was never directly hit. It halted anyway, on parts shortages, and put its losses at an estimated 40,000 vehicles.

If your supplier questionnaire asks whether a vendor has a business continuity plan, and the vendor ticks yes, you have learned nothing that would have helped you in October 2011.

The Day And Its 2026 Theme

The International Day for Disaster Risk Reduction began in 1989, after a UN General Assembly call, and is held on 13 October. UNDRR announced the 2026 theme on 21 June: "Resilience Starts at Home."

The campaign is about housing. It argues that homes are where disasters are felt most acutely and where resilience has to begin, rooted in the people-centred focus of the Sendai Framework for Disaster Risk Reduction 2015-2030, and it divides the work between governments setting land-use planning and building standards, the private sector in how homes are built, financed and insured, and households knowing their local hazards and what to do when a warning arrives.

That sounds adjacent to procurement rather than central to it. It is not. One of the more reliable predictors of how fast a flooded supplier resumes production is whether its workforce still has somewhere to sleep within commuting distance. A plant can be dry, powered and staffed by nobody. Worker housing belongs in a continuity questionnaire for the same reason it belongs in a human rights audit, and almost nobody puts it in either.

Why Closed Questions Fail

The standard supplier resilience questionnaire asks closed questions about documents. Do you have a plan, are you certified, do you have backup power. Every answer is yes, and none of them is evidence.

Certification is worth distinguishing from capability here. ISO 22301 sets out requirements for a business continuity management system and ISO/TS 22318 gives guidelines specifically for supply chain continuity. A valid certificate tells you a supplier has a documented system that an auditor sampled. It does not tell you the recovery time that system actually achieved the last time it was exercised, which is the number you need.

So ask for artefacts instead of assurances. Five requests cover most of the real exposure, and each one has a document behind it that either exists or does not.

Request One: Floor Elevations And Freeboard

Most suppliers who answer a flood question send a screenshot of a hazard map. That is close to useless, because the free global screening layers behind many of those maps work at kilometre scale, and flood depth is decided by metres of elevation.

Ask for four things instead. The finished floor level of each production and utility building, as a figure relative to a stated datum. The design flood level used when the site was built or last modified, and the freeboard between the two. The depth and date of any previous inundation, including events that reached the yard but not the building. And the resolution of whatever model produced their flood figure, because a supplier quoting a 10 kilometre grid cell as a site assessment has told you something useful about their risk function.

A site that cannot produce a finished floor level has not done the analysis. The number is on the as-built drawings.

Request Two: Heat Records At The Work Face

Heat questions get answered with ambient temperature, which is the wrong metric. The international standard for occupational heat stress is wet bulb globe temperature under ISO 7243, combining temperature, humidity, radiant heat and air movement. Inside a shed under a metal roof, WBGT and the weather forecast diverge sharply.

Ask for the WBGT measurement record at the work face, not at the gate, and the count of days in the last two years that exceeded the relevant action level. Those levels, derived from NIOSH and reproduced in OSHA guidance, scale with workload and acclimatisation: roughly 30 degrees Celsius for acclimatised workers on light work against about 28 for the unacclimatised, about 28 and 25 for moderate work, about 26 and 23 for heavy work.

Then ask two questions nobody asks. What is the design ambient temperature of the site's cooling and electrical plant, and what is the current summer average against it? Equipment derates above its design basis. Guidance built on IEEE C57.91 puts a self-cooled oil-filled transformer at roughly 1.5 per cent of capacity lost per degree Celsius above a 30 degree base design ambient. A supplier running three degrees hot has several per cent less electrical headroom than its nameplate claims, permanently, with no incident in any log.

And ask for the shift curtailment history. If production has been cut on hot days, somebody wrote it down.

Request Three: Generator Class And Test Records

This is the request that separates serious answers from reassuring ones, because NFPA 110 already provides the vocabulary and most buyers never use it.

Under that standard, an emergency power system carries a Level, a Type and a Class. Level 1 applies where failure could cost a life. Type is the time to accept load, so Type 10 means ten seconds. Class is the minimum run time at full load, and the scale runs from Class 0.083, which is five minutes, through Class 2 at two hours, to Class 48 and beyond.

Ask for the Class. A Class 2 installation is designed to run for two hours. In a regional flood where the grid is down for five days and the roads are closed, that set is a formality unless there is fuel behind it, which is why the second question is whether the fuel supply contract gives the site priority or shares a tanker fleet with every other customer in the region.

Then ask for test records. Under clauses 8.4.1 and 8.4.2, NFPA 110 expects a monthly run of at least 30 minutes, counted as successful only if the set reaches the minimum exhaust gas temperature in the owner's manual or runs at operating temperature at no less than 30 per cent of nameplate kW. Warm-up and cool-down time does not count toward the 30 minutes, which is worth checking in the log, because a 30-minute entry that includes both is not a 30-minute test. Where an oversized diesel set cannot meet that load monthly, clause 8.4.2.3 provides for an annual supplemental load sequence of two continuous hours: 25 per cent of nameplate for 30 minutes, 50 per cent for 30 minutes, then 75 per cent for an hour. For Level 1 systems there is also a three-year test under clause 8.4.9, running up to four hours, and that report is the single most useful document in this request. Records of inspection, testing and repair are themselves a requirement.

A supplier who exercises the generator unloaded every month has a machine that starts, not one that carries the plant.

The reason to go this deep is that elevating the generator is not sufficient, and two well-prepared hospitals demonstrated it during Hurricane Sandy. At Bellevue, the generators sat on high floors but the fuel pumps were at basement level, and reporting at the time put roughly 17 million gallons of water through the loading docks into a million square feet of basement. The fuel feed stopped, and Bellevue's generators were in any case designed to carry only about 30 per cent of normal electrical load. At NYU Langone, generators sat on upper floors, the fuel tank was in a watertight vault, and new pumps had been installed that year in a hardened pump house. The pump house stayed dry. Water reached the fuel vault anyway and sensors shut the generators down. Close to a thousand patients were evacuated between the two sites.

So the questions are about components, not the set: the elevation of the generator, the day tank, the bulk tank, the transfer switch, the pumps and the switchgear, each one separately, plus the fraction of normal site load the system is designed to carry.

Request Four: Alternate Routes And Shared Floodplains

Suppliers answer the routing question by naming a second road. Ask three follow-ups.

Which river, bridge or culvert does the alternate cross, and does it share a floodplain with the primary route? Two roads out of an estate that both cross the same watercourse are one road. What is the alternate's transit time and vehicle capacity, because a route that exists but adds nine hours and takes half the trucks is a partial answer. And what is the plan when the constraint is not the road but the port, the rail head or the single substation feeding the estate?

The Thai timeline is the argument for taking this seriously. Dewatering and cleanup at the affected estates did not begin until mid-November at the earliest and late December at the slowest. Access was not restored in days. It was restored in months, and the alternate route question was really a question about whether the supplier could ship from somewhere else entirely.

Request Five: Tested Recovery Times

The last request is the one most likely to be answered with a number that means something different from what you assume.

ISO 22301 separates concepts that get used interchangeably. Maximum tolerable period of disruption is the point at which the interruption becomes unacceptable. Minimum business continuity objective is the reduced level of output an organisation will accept while recovering. Recovery point objective is how much data or work in progress can be lost. And recovery time objective is the period following an incident within which a product, service or activity is resumed.

Note what that last definition leaves out. It fixes a time and says nothing about the level of output restored within it, which makes any quoted recovery time ambiguous by construction. A supplier offering a 72-hour objective may mean 72 hours to a trickle or 72 hours to your full order, and the standard does not settle it. Some national standards close the gap by tying the objective to a minimum operational level; ISO does not. So ask the two questions the number cannot answer: what output level that objective restores, as a percentage of normal volume, and how long to full capacity.

Then ask for the exercise. The standard expects a documented exercise programme with after-action records and achieved metrics against targets. Request the most recent exercise report, the recovery time actually achieved against the objective, the corrective actions raised, and whether those actions are closed. Ask specifically whether the exercise assumed that staff could reach the site, because that assumption is where the theme of this particular awareness day meets the spreadsheet.

Honda's near six months at Rojana is the benchmark to hold any answer against. So is the observation that a year after the water went down, half of the smaller electronics factories were still idle.

Using The Answers

These five requests feed a site-level risk register rather than a supplier scorecard, and the distinction matters. A score of 7 out of 10 cannot be insured, financed or escalated. A row that reads "single source, finished floor level 300 mm below design flood level, Class 2 generator, no priority fuel contract, untested 72-hour objective" can be taken to a category manager with a decision attached.

Two practical notes. Check that your contingent business interruption cover actually names the suppliers your register flags, because the usual discovery is that the policy schedule and the risk register disagree. And treat an honest "we do not know" as a better answer than a confident one. A supplier who admits it has never load-tested the generator has told you where to spend the audit day. A supplier who claims a tested 24-hour recovery and cannot produce the exercise report has told you something less comfortable.

Resilience starting at home is a useful frame for a procurement team, not a slogan to borrow. The supplier whose workers live in flood-exposed housing has a recovery time nobody has measured, and it is probably the longest number in your supply chain.

General guidance only. NFPA 110 Level, Type and Class designations, test intervals, load fractions and clause references cited here reflect published summaries of the standard, including guidance referencing its 2010 edition, and should be confirmed against the current edition and against the authority having jurisdiction before use in a contract or specification. ISO 22301 and ISO/TS 22318 definitions are summarised from secondary sources and should be read in the published standards; definitions have also changed between editions. Occupational heat exposure limits differ by jurisdiction, workload definition and measurement protocol, and the WBGT values cited are indicative action levels rather than legal limits in most countries. Equipment derating depends on the specific unit, cooling class and loading profile, and should be confirmed with the manufacturer. Accounts of the 2011 Thailand floods and of backup power failures during Hurricane Sandy follow contemporaneous reporting and retrospective analysis, not primary investigation. Take professional engineering, legal and insurance advice on any specific supplier, contract or site.

Sources

United Nations Office for Disaster Risk Reduction, International Day for Disaster Risk Reduction, Sendai Framework for Disaster Risk Reduction 2015-2030, ISO 22301, ISO/TS 22318, NFPA 110, ISO 7243, National Institute for Occupational Safety and Health, United States Occupational Safety and Health Administration, IEEE C57.91, Minnesota Department of Health, Guy Carpenter, Honda Motor Co., Agence France-Presse and IndustryWeek, CBS News, Fox News, World Resources Institute

This article is intended for general professional information and does not constitute legal, financial, or investment advice.