Daily questions, in public.

The first day began with a deliberately plain question: what, exactly, is failing for people in a particular village? “Climate resilience” can sound like a programme title rather than a lived condition. The Kerala State Action Plan and disaster-management guidance establish that climate and disaster planning systems exist at state level. They do not tell us whether the recurring failure in a given ward is a drain that backs up, a road that disconnects a settlement, unsafe water after a flood, a slope that moves, a lost day of work, or the inability to reach help. A useful village question therefore begins with a specific failure, its place and the people carrying its cost—not with a ready-made intervention.

Evidence:KRL-01KRL-02

Hazard information is indispensable, but it answers only part of this question. Kerala’s hazard maps can help a panchayat or resident see where flooding, landslide or other risks may be concentrated. The IPCC synthesis and Asia assessment explain why climate hazards can interact with exposure and vulnerability. None of that turns a map into a household diagnosis. A mapped flood-prone area may contain families with very different housing, savings, work arrangements, disability access, land rights, social networks and routes to information. Conversely, a household outside a mapped zone can still be harmed when water, transport, health care or income systems fail. The map should open an investigation; it should never close one.

Evidence:KRL-03GLB-01GLB-02

That distinction changes the unit of research. Instead of asking whether a village is “resilient”, the working unit becomes a failure chain: heavy rain or another trigger; the physical or service breakdown; the people exposed; the decision that was missing, delayed or inaccessible; and the recovery burden. Research on farmers’ vulnerability to floods and landslides in Kerala is useful here because it directs attention to livelihood exposure. It is observational work, not a test of a particular solution. Its proper role is to sharpen questions about who loses income, mobility or productive capacity—not to certify a technology or project as effective.

Evidence:KRL-07

This also disciplines how precedent is used. The Resilient Kerala programme materials show that basin and resilience work has been planned and reported in Kerala. An approval or a partner progress brief is evidence of intended scope or recorded activity; it is not independent evidence that a named household, ward or livelihood is safer. That is not a minor caveat. It prevents a familiar policy error: importing a project label into a new place and treating the label as proof of fit. Before borrowing any practice, the research has to ask which failure it addressed, who operated it, what resources it needed and what outcome was actually measured.

Evidence:KRL-04KRL-05

The practical implication is a short local inquiry before a solution card is written. It should combine the available hazard layer with people’s account of the last bad day: what happened first, which service or asset failed, who could not act, what made recovery harder and which institution had authority over the next decision. This is not a request for communities to supply data for its own sake. It is a way of checking whether an intervention has an owner, an operating burden and an equitable path to benefit. It also makes room for residents to disagree with the official description of a problem.

Evidence:KRL-02KRL-03

The day’s conclusion is therefore modest but consequential. A hazard map, state plan or global climate assessment can make a local question more informed; none can answer it alone. The next test is not to select a branded solution. It is to trace one water, flood or slope-related failure through an actual village operating chain and identify the smallest decision that could reduce harm. Until that chain is visible, any proposed fit remains a hypothesis—worth examining, but not yet a recommendation.

Evidence:KRL-01GLB-01

Day 2 asked a governance question that can easily disappear under the language of “community participation”: when does participation change a decision? Kerala’s local-government material and the annual or disaster plans in the source register show that panchayat planning and disaster-management documents exist. That is an important institutional starting point. It is not, by itself, evidence that residents can set an agenda, alter a priority, contest a budget or obtain a response when an intervention harms them. A meeting, attendance list or consultation photograph records contact. Decision power is a different thing and needs different evidence.

Evidence:KRL-08KRL-09KRL-10

The distinction becomes concrete when a plan reaches a hard choice. Who decides whether scarce funds go first to drainage, housing repair, drinking water, a road, an early-warning process or livelihood recovery? Who sees the criteria? Who can point out that a seasonal worker, an older resident, a person with a disability or a family without secure tenure is being missed? And where does that objection go after the meeting ends? The Pangode and Pallippad documents are useful as examples of documented planning surfaces. Their existence does not establish that these questions were answered, nor can one panchayat document stand in for Kerala as a whole.

Evidence:KRL-09KRL-10

The evidence register offers a useful warning from both Kerala and a comparable region. The Janakiya Pankalithavum Punarnirmanavum programme report records a feedback-oriented initiative in seven Kerala districts, but the available report does not show whether feedback changed decisions or continues to operate. Research on local adaptation plans in Nepal offers a mechanism-level governance lesson: a local plan can exist while authority, resources and accountability remain uneven. Neither source proves what will happen in a Kerala village today. Together they make a sharper research demand: do not infer meaningful participation from the existence of a plan or channel.

Evidence:KRL-12INT-04

This matters most where the cost of being unheard is high. Adivasi settlements, coastal communities, women carrying unpaid care work, migrant or seasonal workers and people facing livelihood disruption should not be presented as a single “vulnerable group”. They have different rights, practical knowledge, mobility constraints and claims on public systems. The Scheduled Tribes development scheme page records a state programme surface, but it does not demonstrate rights protection or outcomes. The responsible next move is therefore not to make a claim about any community from a webpage. It is to build place-specific profiles with safeguards for consent, representation and disagreement.

Evidence:KRL-11KRL-12

For this project, a credible voice-to-decision route has five observable parts: people can see what is being decided; they can enter a concern before a decision is closed; the decision-maker and authority are named; the response, change or refusal is recorded; and there is a remedy route when harm continues. This is a design and accountability test, not a claim that any current panchayat already meets it. It can be used with an intervention card just as rigorously as a technical feasibility test. If the route is missing, a proposal may be technically plausible yet socially unsafe.

Evidence:KRL-08KRL-11

The finding is not that panchayats or public participation fail. The source set is too limited and too varied for that conclusion. The finding is narrower: participation has to be researched as an operating right, not displayed as an event. The next day’s work should select a real planning or risk decision and trace who is included, what evidence they can challenge, where the money and authority sit, and how an unresolved concern travels. Only then can the project say whether a proposed climate action has a legitimate route from local knowledge to public decision.

Evidence:KRL-09KRL-10INT-04

Kerala’s public climate record gives the day a firm starting point. The Institute for Climate Change Studies reports that the state-averaged annual mean temperature in 2024 was 0.99°C above the 1991–2020 average and the highest in its record beginning in 1901. It also reports rising annual maximum and mean-temperature trends across all fourteen districts. These are observed state and district signals, not predictions for a particular home, field, school or worksite. The record can justify asking where heat exposure is already becoming harder to manage. It cannot tell a panchayat whose room is unsafe at night, whose work has no shade or who misses income when hours are changed.

Evidence:KRL-13

Rainfall needs even more careful language. The same climate statement records long-run decreases in state-average southwest and northeast monsoon rainfall, district variation and, at the same time, very heavy one-day rain at several stations in 2024. Kerala’s climate plan also discusses future change and variability. “Rainfall is decreasing” therefore cannot be turned into “this village will face drought”, just as a heavy-rain statistic cannot by itself predict which lane will flood. Season, intensity, slope or catchment, blocked drainage, storage, soil condition and coastal influence all matter. Annual rainfall is a broad signal. A useful local question is where water arrives, where it pauses, what it contaminates and which route or livelihood fails first.

Evidence:KRL-01KRL-13

Heat shows why projections, actions and outcomes must remain separate. A Kerala heat analysis using station records and downscaled models projects higher February-to-May heat, while warning that its modelled historical absolute temperatures are underestimated. KSDMA’s public timeline shows that worker-hour changes, school water breaks, local heat-plan guidance and funding routes have been introduced. That proves that public action has begun; it does not show how consistently the measures operate or whether illness, lost wages or unequal exposure fell. A panchayat should not copy a heat-plan label and declare success. It should ask whether water, rest, shade, health support and altered schedules actually reach informal workers, migrant workers, students, older people and people with disabilities.

Evidence:KRL-14KRL-15

Agriculture presents a similar trap. A Kerala Agricultural University-led model study reports rice-yield reductions under simulated warming across five agroclimatic zones. That is credible evidence that temperature can stress the crop under the model’s conditions. It is not a forecast of one farmer’s harvest and it does not identify a winning seed, irrigation system or insurance product. Observational research on Kerala farmers also shows that environmental and social vulnerability varies across panchayats. Farm risk therefore includes the crop and much more: land and lease arrangements, debt, labour availability, water access, market conditions, care work and the ability to absorb a failed season. Crop sensitivity should open a farm-system inquiry, not close a purchasing decision.

Evidence:KRL-07KRL-16

The working response is a small seasonal profile, built with residents and kept privacy-respecting. For heat, it would connect work and school hours, indoor conditions, shade, drinking water, reported illness and safe routes to help. For rain, it would connect water paths, low points, drainage condition, water quality, crop stage and roads that become unusable. For farming, it would connect the climate signal to soil, water, labour, tenure, credit and viable alternatives. The profile must show whose conditions are otherwise averaged away: outdoor workers, women carrying care work, students, small and tenant farmers, fishers, renters and people living in weak or poorly connected housing. Sensitive household information should not become a public map.

Evidence:KRL-13KRL-14KRL-15

This changes the role of climate evidence in the project. A state trend or scenario will no longer sit beside a proposed intervention as if one proves the other. It will trigger a named local question, a set of people whose exposure must be visible and a decision owner who can act on the answer. The next test is practical: can one panchayat assemble this profile without creating a surveillance burden, name who maintains each field and identify what evidence would reverse its first judgement? Until that test is done, heat, rain or crop responses remain possible lines of action rather than recommendations. The honest bridge from a state signal to a village decision is local verification.

Evidence:KRL-01KRL-13KRL-15

Kerala’s flood discussion often reaches quickly for the phrase “restore wetlands”. The state’s post-2018 technical committee gives that idea a credible mechanism: rice fields, ponds and lakes historically helped hold or move water, while deterioration, abandonment and reclamation weakened parts of that landscape. But the committee report is a diagnosis and recommendation, not a before-and-after evaluation of a restoration project. A wetland is not an empty plot waiting to become flood infrastructure, and every pond or field does not perform the same hydraulic job. The responsible proposition is narrower: connected storage and drainage landscapes may matter, but their function and downstream effects have to be tested at the scale where water actually moves.

Evidence:KRL-22

That water unit is also a social landscape. A 2024 study of the Kuttanad Wetland System examines local governments inside an ecological, multi-watershed boundary and finds that flood history or geographic position alone is insufficient for understanding population vulnerability. In that study, housing quality, density and a working population including fishers, agricultural labourers and women also shape vulnerability. The lesson is not a new statewide ranking. It is that two maps must be read together: how water moves and ecosystems function, and who has fragile housing, exposed work, care obligations, transport constraints or little power over land. A project that improves storage while cutting off a poorer settlement has not solved the flood problem fairly.

Evidence:KRL-23

Kerala already has an institutional route that could hold part of this decision. The 2008 Paddy Land and Wetland Act treats paddy land as including associated bunds, drainage channels, ponds and canals. It also establishes a Local Level Monitoring Committee in each panchayat or municipality, with monitoring and data-bank responsibilities. This matters because it offers a local public body around which land records, cultivation knowledge and water questions can meet. Yet a statute is not a performance report. The source does not tell us whether a committee is active, its data are current, tenant cultivators are heard, enforcement is fair or recurring maintenance is funded. Any real proposal would also need current legal advice on amendments and case law.

Evidence:KRL-24

Rights become unavoidable when a field is expected to hold water or a crop is exposed to loss. A peer-reviewed case study documents how Kerala’s Thayillam collective argued after the 2018 disaster that compensation should reach lease farmers rather than only landowners. It records a specific justice outcome in one collective; it does not establish statewide entitlement or show that lease farming reduces floods. Its value here is a failure test. A water project must identify owner, cultivator, agricultural worker and water-dependent user separately. It must say who agrees to a changed water regime, who loses a crop or workday, who receives compensation and where a grievance goes. Ecological benefit cannot be financed by making an already weak right-holder absorb the loss.

Evidence:KRL-25

Maps help only when their authority is bounded. A 2026 Pozhuthana case combines spatial analysis with surveys from 1,120 households and checks its mapping against reported 2018 impacts. The study is one hilly panchayat and one assessment method, not a universal risk score or a wetland-restoration result. It nevertheless reinforces an important operating rule: panchayat-scale analysis needs field validation. Residents can identify a culvert that blocks, a footpath that disappears, a house cut off before another, or a water path changed by recent construction. Technical expertise is still necessary; local memory is not a substitute for hydrology. The decision becomes stronger when model, event history, field observation and disagreement are recorded together.

Evidence:KRL-26

The resulting candidate is a water-commons and cultivator-rights compact, not a construction prescription. Before monsoon, a panchayat and its monitoring committee could walk one small connected unit with cultivators, workers, fishers, residents and technical reviewers. They would mark water paths, culverts, storage areas, blockages, water-quality concerns, ownership and lease arrangements, recent cut-off points and maintenance responsibility. Then they would choose one reversible, rights-cleared action, publish its owner and operating cost, and keep a complaint route open. Success after ninety days would mean a better shared record, a named maintenance budget and no hidden downstream or livelihood harm—not a sweeping claim that one cleared canal or restored pond has made the village flood-safe.

Evidence:KRL-22KRL-23KRL-24KRL-25KRL-26

The first finding is simple but easy to miss after a flood: water safety must be checked source by source. Kerala’s post-2018 damage and needs assessment reported that shallow wells supplied drinking water to 62% of people and described flood damage to water sources and sanitation facilities. A separate peer-reviewed study followed 26 wells in the lower Periyar Basin before, during and after the flood. Their water-quality scores did not move in one direction: 69% improved after the flood, 19% were poorer during flooding than before it and 12% were unchanged. This does not make flooded wells generally safe or lost. It shows why a flood label cannot replace an appropriate test and local follow-up.

Evidence:KRL-27KRL-28

Kerala also has public routes that a local plan can use. The Kerala Water Authority lists household water-quality testing through district quality-control laboratories and publishes contact points. Its 2024 drought directory recorded drinking-water vending locations across all fourteen districts. These sources establish that testing and emergency-supply routes have been implemented. They do not tell us whether a sample can be collected safely during a flood, how long a result takes, whether a fee excludes a low-income household, how far somebody must travel, how much water is available or whether a listed point is operating today. A service can exist on paper while access fails in practice, so availability and reach must be measured separately.

Evidence:KRL-29KRL-30

That distinction moves the problem away from a household product. The state assessment recommends water-quality surveillance with feedback and follow-up, improvements to shallow wells, flood-resilient latrines and a WASH preparedness plan with defined roles. It also records both KWA-managed and panchayat-managed supply arrangements. A village water decision should therefore name the source, the nearby sanitation or drainage risk, who decides temporary closure, where testing happens, how a safe alternative is supplied and who authorises reopening. A filter, pump, tanker or chlorination step may be useful in a particular chain. None is a complete answer by itself, and the available sources do not show that one such measure reduced illness or household time burden.

Evidence:KRL-27KRL-29

Access changes who experiences a water interruption as an inconvenience and who experiences it as a crisis. A public notice is of little use if it is only online, in one language or posted after a household has already used an uncertain source. A vending point is not accessible if the route floods, the queue is unsafe, identification is demanded or the person collecting water cannot carry it home. A ward-level plan must ask who usually collects and carries water, who cannot travel or wait, who uses a shared or rented source and who loses wages or care time when that source closes. Women and girls, older people, people with disabilities, tenants, migrant workers and remote settlements cannot be collapsed into one access category.

Evidence:KRL-27KRL-29KRL-30

The candidate response is a small water-continuity and safe-reopening compact. Before monsoon or dry-season stress, one ward would keep a limited, non-public working record of up to five primary or shared drinking-water sources and two reachable backups. For each source, the record would identify likely contamination pathways, physical access, users needing assisted access, the testing contact, a closure trigger, an understandable status notice, estimated backup volume and cost, and the public owner responsible for follow-up. The purpose is not to build another permanent household database. It is to make the operating chain visible before pressure arrives, while keeping household risk and sensitive source information out of the public site.

Evidence:KRL-27KRL-29KRL-30

The next test can remain deliberately small. The ward can rehearse one imagined flood or drought interruption and measure the time from trigger to status notice and from closure to a safe alternative. It can record whether priority users were reached without unsafe travel or queuing, whether a result led to repair or reopening, and whether water collection or care work increased. It must also watch for ecological harm: a backup pump or tanker plan should not hide falling groundwater yield, sanitation failure or transferred contamination. These are proposed operating indicators, not proven outcomes. The evidence supports testing the chain; it does not yet support claiming that the compact prevents illness or guarantees equal water access.

Evidence:KRL-27KRL-28KRL-29KRL-30

Kerala’s groundwater story cannot be reduced to “more recharge”. The Central Ground Water Board’s 2024 state assessment shows why. In its April 2023 monitoring observations, depth to water ranged from 0.07 metres below ground level in Alappuzha to 53.45 metres in Thiruvananthapuram. Eighty-three per cent of the analysed wells were within ten metres, while seventeen per cent were deeper. These are measured conditions in a monitoring network, not a village water balance. A shallow reading does not tell us that water is safe, reliably available to everyone or replenished faster than it is taken. The figures warn against carrying one engineering prescription from one Kerala landscape into another.

Evidence:KRL-31

The coastal dry season makes that caution concrete. A peer-reviewed study used monthly observations from 2015 to 2021 along the Tavanur–Ponnani stretch of the Bharathapuzha. It reported salt-water intrusion at specified locations and poorer pre-monsoon salinity and electrical-conductivity conditions in Purathur and Ponnani. The study’s suggestion to restrict groundwater use near this coast belongs to that setting; it is not a rule for every coastal panchayat. Its wider lesson is procedural. Before trying to “raise the water table”, a local decision must identify the aquifer, test the seasonal quality risk and make sure that people have a reachable alternative if an unsafe or saline source is restricted.

Evidence:KRL-33

Kerala does not lack public machinery. The Ground Water Department describes a conservation and recharge scheme using pits, dug-well and borewell recharge, small check dams and subsurface dykes. Its 2021–22 table records 514 proposals and 244 completed works. The department’s National Hydrology Project page also describes observation wells, digital water-level recorders, a hydrological information system and an intended groundwater-budgeting function. This is evidence that monitoring and construction routes have been implemented. It is not evidence that each completed work recharged the intended aquifer, survived maintenance, improved dry-season access, avoided contamination or salinity, protected nearby ecosystems, or shared benefits fairly. A completion count cannot stand in for a water-security result.

Evidence:KRL-32KRL-34

Technical advice is available too, but access to it must be examined rather than assumed. The department lists hydrogeological investigation, site-specific advice for artificial recharge, water-quality and yield testing, data services and permits. Published services include fees and, for some applications, a tax receipt or possession certificate. This does not prove that tenants, low-income households, migrant workers or people without formal tenure are excluded in practice. It does show what a panchayat must check before calling the route accessible. A shared source may matter to people who do not own the land under it. The right question is not only who can submit a form, but whose water dependence and carrying burden enter the decision.

Evidence:KRL-35

The proposed response is a monitored groundwater commons compact for one small, hydrologically coherent source-use area. With consent, a panchayat would follow three to five sentinel sources for 90 days, recording only the minimum needed: monthly level or availability, recent rainfall and extraction context, one quality screen matched to the suspected risk, the users and access barriers, sanitation or drainage warnings, and a reachable backup. Household coordinates and individual risk would stay out of the public record. The panchayat would name the budget and coordination owner; the technical authority would review siting and interpretation; local custodians would not be expected to provide unpaid labour. Construction would remain optional, not the assumed endpoint.

Evidence:KRL-31KRL-34KRL-35

The compact is still a hypothesis, not a proven intervention. Its first test is whether better observation changes a real decision. After 90 days, the panchayat should be able to explain why it chose one reversible source, demand or recharge action—or why it deferred construction. It should publish the decision, responsible owner, cost and review date, but not sensitive source locations. Success at this stage means that seasonal quality, access, maintenance, ecological warnings and alternatives were considered, and that the decision can be cancelled when its mechanism is unsupported. Whether the approach ultimately improves dry-season yield, reduces the time and cost of collecting water, or prevents salinity and contamination remains unknown and must be measured rather than promised.

Evidence:KRL-31KRL-32KRL-33KRL-34KRL-35

A flood rarely respects the neat boxes of local administration. Rain falls higher up, water gathers through fields and drains, a culvert narrows the flow, a crossing goes under, and a ward that did not receive the heaviest rain can still lose its road to a clinic, school, workplace or water point. Kerala’s post-2018 needs assessment describes this interdependence: upper-catchment soil and water conservation, erosion, dam and bund coordination, and damaged, silted or polluted drainage all sit inside the same river-basin problem. The assessment recommends integrated planning. That is a serious diagnosis and a policy route; it is not proof that clearing a channel, changing a bund or building a check dam will reduce harm in a particular reach.

Evidence:KRL-27

The Pamba basin offers a useful lesson about what a map can and cannot do. A Government of Kerala and Haritha Keralam technical report compared river form, land use and groundwater observations in selected highland, midland and lowland watersheds before and after the 2018 flood. It documented changes in channel bars and banks, while stating that plot-level land use, protection walls and riparian vegetation need field knowledge alongside remote sensing. In plain terms, the satellite view is incomplete without a walk and the people who know the reach. The report documents a mapping practice; it does not establish what caused every change, measure present carrying capacity or certify a safe engineering response.

Evidence:KRL-36

A 2024 model study of the Greater Pamba basin widens the question from water depth to access. It combines projected land use, rainfall, population, infrastructure and terrain. Under its 100-year-return-period scenario, the model estimates high-risk exposure for about 36 per cent of villages and municipalities and 70–80 per cent of selected critical facilities. Those numbers are a prompt to ask which clinic, school, relief point, road or water facility may become unreachable. They are not a present-day ward forecast, an evacuation instruction or a promise that rebuilding one crossing will prevent loss. The result depends on its input data, assumptions, return period and spatial scale.

Evidence:KRL-37

Kerala also has planning routes a panchayat can use without pretending it controls the whole river. State hazard maps, the recorded Pamba-basin resilience programme and local annual-plan examples can help officials ask for evidence, align budgets and state which authority owns a road, drain, irrigation structure or water asset. Programme reporting says risk-informed support and basin planning are under way. It does not show that coordination works during a flood, that local priorities shape every decision or that a panchayat can authorise an upstream or downstream physical change on its own. The honest use of these routes is to connect a small local decision to the larger system and record where local authority stops.

Evidence:KRL-03KRL-05KRL-09

The proposed compact starts with one repeatedly interrupted crossing, road segment or river reach. Residents from upstream and downstream, people who use the route and the relevant asset authorities would walk it together before monsoon. They would mark the water path, recent overtopping or closure, ecological and land-use features, a safe alternative, people who may need assisted travel and the public owner responsible for inspection and upkeep. The record must show who is cut off, not only what structure is damaged. It should include older and disabled people, pregnant people, students, renters, migrant households, water carriers, farm, fishing and plantation workers, and Adivasi or forest-edge users where relevant—without publishing sensitive household or settlement locations.

Evidence:KRL-27KRL-36KRL-37

The first action should be reversible and useful even if no construction follows: a route-status sign, a contact chain that works without smartphones, an assisted-access trigger, or a written inspection and maintenance agreement. The group can rehearse one closure at a table and measure how quickly an accessible warning and alternative route reach people. A physical change to a culvert, bank, channel or crossing comes later, and only after hydraulic, ecological, land-rights, fishing-access and downstream review identifies a safe option with a funded owner. The compact remains a mechanism-supported proposal; its effect on route closures, flood depth, erosion, livelihoods and unequal access is unverified. A sound first result may be the decision not to build.

Evidence:KRL-27KRL-36KRL-37

The first discipline is to say exactly what was measured. After Kerala’s severe 2018 flood, a peer-reviewed environmental microbiology study examined samples from flood-affected Kuttanad and reported faecal indicator organisms together with bacterial profiles related to virulence and antimicrobial resistance. That is evidence of a hazard in the sampled environments at that time. It is not evidence that every well, canal or household in Kuttanad was exposed, that a particular person became infected, or that the same organisms are present today. The study should sharpen a village question—where could contaminated water meet a drinking source or worker—not become a permanent label attached to a community.

Evidence:KRL-39

A second Kerala study sampled selected wells and waterbodies connected with the Pampa, Periyar and Vembanad–Kuttanad systems after the same flood. It found faecal indicators in some samples and discussed sanitation conditions, well cleaning and chlorination. Its sample was small and uneven across source types. The work therefore supports source-specific caution and appropriate testing; it does not establish a statewide contamination rate, clinical disease burden or the effectiveness of cleaning or chlorination in a particular household. A test result is one link in a decision chain. The useful question is what status follows the result, who communicates it, what safe water is available meanwhile and who checks whether the source can reopen.

Evidence:KRL-40KRL-27

Kerala’s 2025 faecal-sludge policy describes the wider pathway behind this risk. It says the state depends mainly on onsite sanitation while sewered coverage is minimal, links unmanaged septage and inadequate treatment to contamination of waterbodies, and estimates that less than ten per cent of generated faecal sludge was being treated or safely managed. This is a consequential government self-assessment, not an independent census of every tank, drain or water source. It tells us that contamination cannot be treated only as bad behaviour by one household. Containment, emptying, transport, treatment, drainage, water-source protection and public oversight all shape whether waste reaches floodwater.

Evidence:KRL-38

How results are communicated matters as much as the sample. Publishing the coordinates of a precarious household source can expose a family without making anyone safer. Naming a settlement as “dirty” can deepen stigma while hiding a failed public service. A village response should keep household and worker details confidential, publish only an understandable source status and safe alternative, and explain what the test did not measure. People who collect water, clean wells, empty tanks or handle waste need protective equipment, authority to refuse unsafe work and a route to report hazards. Environmental evidence should direct protection and accountability, not blame.

Evidence:KRL-27KRL-38KRL-39

The working response is a small, non-public water-and-sanitation pathway record for one flood-prone reach or ward. With consent and technical review, it would connect a limited set of drinking sources to nearby containment, drains, flood paths, safe sampling points, closure triggers, laboratories, backup water and a named follow-up owner. It would not map every toilet or household. Each unsafe or uncertain result would produce a plain-language status, a reachable alternative and a review date. The record would also identify who is expected to sample, clean or transport waste and whether that work can be done safely during waterlogging or road closure.

Evidence:KRL-27KRL-38KRL-40

The next test should stay modest. Trace one pathway before monsoon, rehearse a flooded-source scenario and check the time from warning to source status, safe alternative and follow-up. Use only an agreed sampling protocol and qualified interpretation. Record whether people without smartphones, tenants, migrant households, older and disabled residents, and people who cannot carry water can act on the notice. These are proposed operating checks, not proven health outcomes. The available evidence supports investigating the pathway and protecting people while uncertainty is resolved. It does not yet show that this compact prevents infection, restores every source or makes sanitation services equitable.

Evidence:KRL-38KRL-39KRL-40

Kerala’s new faecal-sludge policy begins with an uncomfortable admission: most sanitation is onsite, sewered coverage is minimal, and the state estimates that less than ten per cent of generated faecal sludge was being treated or safely managed. The policy links unmanaged septage and inadequate treatment with contamination of waterbodies. This is a government account of the problem and a defined policy route, not an independent audit of every panchayat. Even with that limit, it changes the unit of attention. A household having a toilet does not tell us whether waste stays safely contained, can be emptied without exposure, reaches a functioning facility or leaves that facility without harming another waterbody or settlement.

Evidence:KRL-38

The policy assigns a long list of responsibilities to local governments: funding, infrastructure ownership, operator licensing and regulation, fees, public toilets, local byelaws, operation-and-maintenance payments, cluster coordination and first response to grievances. It expects emptiers to use protective equipment and keep collection and manifest records. These provisions show that an accountable service chain has been designed on paper. They do not show whether a licensed service is reachable in each panchayat, whether a manifest matches a real receiving facility, whether treatment works during monsoon, whether workers can refuse unsafe tasks, or whether a tenant can complain without stigma or retaliation. Each of those is a delivery question that must stay visible.

Evidence:KRL-38

Place changes the engineering question. The policy says conventional septic tanks may not work efficiently in high-groundwater Alappuzha and coastal zones and calls for contextual containment. Suchitwa Mission guidance describes cluster planning for treatment plants across districts and records land availability as a constraint. Together, these sources support local hydrogeological and service review before a technology is chosen. They do not endorse one above-ground tank, plant design or cluster size for every landscape, and they do not prove that proposed facilities were built, accept every load, meet discharge standards or remain accessible when roads flood. A copied design can move contamination rather than resolve it.

Evidence:KRL-38KRL-41

Workers and residents reveal breaks that an asset register misses. The people who empty a tank, drive a vehicle, receive a load, clean after a leak or respond to a complaint need equipment, training, paid time and authority to stop dangerous work. Women’s groups or community volunteers cannot become unpaid sanitation logistics. Tenants, people using shared toilets, migrant households, older and disabled residents, coastal and backwater communities, and Adivasi or forest-edge settlements may face different access barriers. A complaint route should work in appropriate languages and formats, without a smartphone and without forcing a household to publish its toilet, land or water-source details. Policy promises are the start of this check, not evidence that these protections operate.

Evidence:KRL-38KRL-41

The proposed compact makes every link visible: containment condition, regulated emptying, a collection record, documented transport, a verified receiving and treatment route, effluent and residual management, nearby water-source protection, and a grievance with corrective action. The panchayat or cluster names the service-chain owner, budget holder and complaint lead. Operators record collection and destination; the facility records what it receives, how it operates and where residuals go. Only aggregate service information belongs in public. Household conditions, worker reports and precise risk locations remain confidential. A toilet, vehicle or treatment plant counts as an asset; safe sanitation depends on whether the whole chain works together.

Evidence:KRL-38KRL-41

A practical first test is a confidential 90-day audit in one flood- or waterlogging-prone panchayat or service cluster. Residents, sanitation workers, health and technical staff, and the local government would trace requests from containment to a verified destination, rehearse a monsoon road or power disruption and identify the first real break. The first repair may be a complaint route, destination verification, safe pickup access or a funded maintenance responsibility rather than a new plant. Track response time, documented destinations, worker-reported safety, ability to stop unsafe work and unresolved water or drainage warnings. These are operating indicators. Whether the compact improves water quality, health, worker safety or affordability remains unverified and must be measured.

Evidence:KRL-38KRL-39KRL-40KRL-41

Kerala already recognises coastal erosion as a disaster, but recognition does not choose a remedy. A Kerala State Disaster Management Authority training module, drawing on older state-profile data, describes 322 kilometres of the coast as prone to severe erosion and links erosion, waves, surge and Kallakkadal with beach loss and damaged fisher housing. This is public hazard framing, not a current survey of every coastal segment or household. It cannot tell one panchayat whether tomorrow’s failure will be overtopping, lost beach access, saline water, an unsafe home or a blocked landing site. Nor does an erosion map establish that a seawall, nourishment project, planting line or relocation plan will reduce that failure.

Evidence:KRL-42

A study of the Kayamkulam–Arattupuzha and Thottappally sector explains why the decision area must extend beyond the structure’s frontage. Using field information and numerical modelling, the researchers attributed observed morphological change around the Kayamkulam inlet to fishing-harbour breakwaters built between 2000 and 2007 and argued for integrated shoreline management. This is unusually useful place-specific causal evidence. It is still one coastal segment, one historical period and one modelling context. It does not prove that every seawall, groyne, harbour or breakwater in Kerala has the same effect, and it does not identify the right alternative. Its practical lesson is an alongshore test: who receives changed sand movement, erosion, access or maintenance burden beyond the project boundary?

Evidence:KRL-43

Warning is another part of protection, but an alert is not the outcome. The Indian National Centre for Ocean Information Services describes SIVAS, a Kerala coastal inundation forecasting capability for swell surge and Kallakkadal, with up to three days’ lead time and English, Hindi and Malayalam outputs. The page documents that such a forecasting route has been developed. Its wording does not establish where the service is currently disseminated, how well forecasts perform at every location, whether fishers, renters or shore workers receive and trust them, or whether people with disabilities can act. A warning becomes a resilience asset only when it reaches a named action: where people, boats, gear, water and care responsibilities can go safely.

Evidence:KRL-44

Valiyaparamba offers a different Kerala precedent. KSDMA case material describes a panchayat and MGNREGA-supported nursery, Casuarina planting, geotextile bunds, maintenance work and mangrove activity in response to erosion. That documents local mobilisation and an implemented maintenance route. It does not establish protective performance. The case provides no baseline or comparison shore, measured shoreline change, long-term cost and upkeep record, biodiversity assessment, distributional analysis or independent livelihood result. Its own statement that the work helped “to an extent” cannot carry a general recommendation. Green belts and geotextiles therefore remain place-specific options whose wave, sediment, habitat, space, access and maintenance fit must be tested—not a softer-looking substitute for evidence.

Evidence:KRL-45

The proposed compact treats one connected coastal segment as the decision area, including neighbouring and downdrift settlements, landing places, backwater links, habitats, water points and existing structures. A compensated shoreline walk would bring together small-scale fishers, women involved in processing, selling and care work, shore and harbour workers, tenants, displaced or informal households, older and disabled residents, students and boat or gear users. Their observations of waves, beach change, routes and work would sit beside official hazard and forecast information, not below it. The public record would show options, responsible authorities, costs, objections and review dates while keeping household risk, tenure details and precise vulnerable locations confidential.

Evidence:KRL-42KRL-43KRL-44KRL-45

The first 90-day test does not need an irreversible project. Select one erosion- or swell-surge-prone segment with the directly connected settlements and rehearse a three-day warning: who receives it, what changes, where people, boats and gear go, who needs help and which authority responds. Publish a non-sensitive action protocol. At the same time, log alongshore, habitat, landing, fishing, water, tenure and relocation concerns. Physical protection or relocation should be deferred until independent sediment, engineering, ecological and rights review is complete and a funded maintenance and harm-response owner is named. This compact is mechanism-supported; whether it reduces loss, preserves livelihoods or improves warning outcomes remains unverified.

Evidence:KRL-42KRL-43KRL-44KRL-45

On paper, Kerala has added an important public-response route. The Fisheries Department government-order register lists an August 2026 administrative sanction for sea-rescue and disaster-management coordination and information centres, harbour-based rescue squads, and a separate order concerning marine-ambulance operation and staffing. That is evidence that government has authorised a response structure. It is not evidence that every centre or squad is open, staffed, trained, reachable and equipped today. The register does not report communication coverage, response times, interoperability, maintenance, completed rescues, fisher oversight or avoided loss. Before a panchayat tells people to rely on the route, it must verify the specific contacts, shifts, assets and escalation responsibilities that are actually available.

Evidence:KRL-46

The warning chain also begins before an official message arrives. Participatory research across 20 Kerala marine fishing villages, involving 400 fishing households, documented how fishworkers read biotic, oceanic, atmospheric and celestial changes associated with local coastal hazards. The study makes local observation a legitimate input to warning design rather than a ceremonial consultation after decisions are made. It does not prove that any single sign predicts a hazard accurately, that every community reads conditions in the same way, or that local observation should replace scientific forecasting. The operational lesson is more modest: a protocol should state how an unusual local observation is reported, checked and joined with official information, and it should recognise knowledge-holders without transferring public safety liability to them.

Evidence:KRL-47

Ockhi shows why reception cannot be treated as a technical footnote. A qualitative case study in Vizhinjam examined the 2017 cyclone through interviews, focus groups, key informants and official-warning material. It records delayed or ineffective communication, coordination conflict and community perceptions of exclusion. These findings describe a historical, place-specific failure chain; they do not establish the present performance of Kerala warning or rescue services, represent every fisher community, or isolate one cause for every loss. They also do not license this project to reuse a participant’s grief as dramatic copy. The finding relevant to today is institutional: record who issued a message, who received and acknowledged it, what decision followed, and which authority took responsibility when the chain failed.

Evidence:KRL-48

The national workshop convened after Ockhi adds operational detail without becoming an outcome evaluation. Its proceedings bring together fishers, fisherwomen, officials and civil-society participants discussing warning reception, communication equipment, changing fleet patterns, search and rescue, and recovery. The record helps expose failures that a device inventory can hide: a message may not reach the boat, may arrive in an unusable form, may conflict with local conditions, or may demand a stop-work decision whose cost falls on crews and women’s shore-based work. Workshop testimony is not a representative survey, and a participant’s account cannot be generalised to all coasts. It should be used to design questions and failure tests, not to claim that a preferred technology would have prevented Ockhi losses.

Evidence:KRL-49

This leads to a four-link public test. First is the forecast: INCOIS documents SIVAS as a Kerala swell-surge and Kallakkadal forecasting capability with up to three days of lead time and outputs in Malayalam, Hindi and English, although current coverage, verification and operational dissemination still need checking. Second is a trusted relay, using redundant voice, text and in-person routes rather than assuming a smartphone. Third is a feasible action for different boats, shore workers, older or disabled residents, renters and people with care duties; “do not go” is not feasible if it simply moves debt or danger elsewhere. Fourth is accountable follow-up or search when acknowledgement is missing or a crew is overdue. The Fisheries order may support that final link, but current readiness remains unverified.

Evidence:KRL-44KRL-46

A useful first trial is a 90-day tabletop exercise in one willing harbour or landing-centre catchment, not a public tracking system. Test two cases: a SIVAS-style swell-surge affecting shore activity and a rapidly worsening offshore report affecting a crew. Map message origin, trusted relay, confirmed receipt, feasible action, overdue check and search escalation. Include at least one crew type, women in vending or processing, an older or disabled resident, and households without smartphones; pay people for advisory and drill time. Publish only the accessible protocol, verified public contacts, named institutional owners and corrected failure points. Keep individual boat, crew, household, debt, tenure and location information confidential and consent-based. This design is mechanism-supported. Whether it improves receipt, safer decisions, search speed or livelihood protection remains unverified until the chain is tested and its failures are independently reviewed.

Evidence:KRL-44KRL-46KRL-47KRL-48KRL-49

Kerala has an official route for households facing land certified as unsafe. The State Disaster Management Authority describes vulnerability-linked assistance for coastal families moving beyond 50 metres from the high-tide line: up to ₹6 lakh to buy three to five cents of land and ₹4 lakh for construction. For other certified non-liveable settings, the page says the legal owner is eligible. This documents a scheme, funding ceiling and eligibility route. It does not show whether the amounts meet present land and construction costs, whether safe sites are available, or whether applicants obtain secure occupancy, services and livelihood access. The legal-owner wording also raises a question rather than answering it: what route exists for renters, shared households and people with undocumented or contested tenure?

Evidence:KRL-50

Punargeham shows substantial reported administrative activity, but activity is not durable recovery. A Fisheries Department success-story page describes several routes, including household purchase of land and a house, purchase of an existing land-and-house package, and group land or flats. It reports 1,682 completed houses and 8,675 families expressing willingness to move. Those are departmental self-reported figures whose headline date and denominator definitions are not clear on the page. The record does not independently disaggregate informed consent, tenure, destination safety, maintenance, debt, services, fishing or shore-work access, or outcomes after handover. A completed unit can therefore be counted without answering whether the household can safely occupy it, afford daily travel, continue work and care, or obtain a correction when the arrangement fails.

Evidence:KRL-51

Historical Kerala research explains what a building count can miss. A Kerala Institute of Labour and Employment-sponsored mixed-method report used interviews, focus groups, community and official consultations, and local economic assessment across selected coastal cases in Kollam, Alappuzha and Thiruvananthapuram. It records participant accounts of post-tsunami rehabilitation-colony deficits, stigma, disrupted relationships with livelihood resources and difficulty continuing work. These are serious lived-experience warnings, but they remain selected historical cases with contested causal interpretations. The report is not an audit of Punargeham, a representative survey of coastal households, or proof that every rehabilitation colony fails. Nor does public availability give this project permission to reuse people’s identities, trauma or images. The responsible use is to turn the documented risks into questions that every new option must answer.

Evidence:KRL-52

Kerala’s State Planning Board supplies a useful design standard, though not proof of performance. Its working group on fishworker social protection recommends that coastal housing give weight to residents’ preference for independent houses, access to fishing livelihood assets and activities, and associated community amenities. This moves work, choice and everyday infrastructure into the housing decision instead of treating them as extras after construction. The report is policy advice: it does not establish a legal entitlement, confirm that a current programme follows the recommendation, or show that a particular housing form improves safety or wellbeing. The practical implication is still clear. Each proposed destination should be compared for travel time and recurring cost to landing sites, markets, shore work, schools, health care, water and care networks, with residents defining which changes are tolerable.

Evidence:KRL-53

The resulting compact begins before a beneficiary list or building plan is fixed. For every realistic option—including safer work on the present site, nearby movement, a purchased house or a group site—it should record the household’s preference and reasons; owner, tenant, shared or disputed occupancy; destination rights; hazard evidence; accessibility; construction and maintenance responsibility; and distance, time and cost to work, care and public services. The record must also show transition income and debt risk, ecological and common-access effects, and an appeal route with authority to change the plan. This information is necessary for the decision but dangerous as a public register. Household risk, documents, preferences, disability, debt and precise locations should remain confidential, consent-based and separated from the aggregate public accountability record.

Evidence:KRL-50KRL-51KRL-52KRL-53

The first test should be a 90-day, no-public-list audit in one proposed coastal cluster before site approval or beneficiary selection. Use anonymised household and worker profiles to compare risk, secure occupancy, work and service routes, accessibility, recurring costs, destination ecology, maintenance and appeal. Compensate women in fish vending or processing, renters and occupancy advocates, disabled and older residents, workers and local knowledge-holders for review time. Publish only aggregate gaps, named public responsibilities, corrected assumptions and unresolved appeals. Do not close the record at handover: require follow-up through at least one monsoon and work season. This compact is mechanism-supported, not outcome-proven. Whether it produces safer occupancy, meaningful choice, lower secondary loss or durable livelihood continuity remains unverified until programmes disclose comparable, independently reviewable results.

Evidence:KRL-50KRL-51KRL-52KRL-53

Kerala’s new river-basin framework starts with an important admission: salinity is not a problem with one owner or one consequence. The Water Resources Department says intrusion can affect fisheries, agriculture and the abstraction of water for supply. It calls for coordination across sectors, regular sustainability review, stakeholder participation and a joint calendar covering crops, barrage operation, drainage and weed control. That is an approved direction for government, not evidence of delivery. The same framework said an Integrated River Basin Management Plan had not yet been developed when it was issued. We therefore know that the conflict is officially recognised; we do not yet know whether a backwater community can see the relevant data, influence the seasonal decision or obtain a correction when the decision causes harm.

Evidence:KRL-54

Kuttanad makes the trade-off concrete. The Irrigation Department’s current responsibilities include salinity control, flood management and modernisation of the Thanneermukkom Bund to manage salinity while minimising ecological decay. Paddy-supporting water works and flood-water regulation sit in the same mandate. This is useful evidence that the State does not describe the structure as serving only one objective. But a responsibilities page cannot tell a resident whether gates operated as intended last season, whether water was safe at an intake, whether a landing or canoe route remained usable, or whether fish, shell work and wetland exchange suffered. It records an institutional route. It does not establish completion, operating quality, accountability or an equitable result for the people living around the connected water system.

Evidence:KRL-55

The ecological record warns against treating Vembanad as a tank that should simply be made fresher. A peer-reviewed synthesis describes fishing, farming, transport, aquaculture, shell work, tourism, canals and wetlands as one coupled system. It reviews evidence that the Thanneermukkom barrier altered hydrology and connectivity alongside reclamation, pollution, dredging and other pressures. That supports a connected-wetland view: changing water at one point can alter access or ecology elsewhere. It does not isolate the effect of a particular gate schedule, report today’s salinity at a village tap or landing, or identify an ideal operating regime. Historical and mixed studies can reveal mechanisms and risks, but a present seasonal decision still needs current, place-specific observations and a public account of the trade-offs.

Evidence:KRL-56

Drinking-water security needs its own test inside that wider system. Kerala Water Authority’s planning proposal identifies coastal supply exposure when summer flow falls and seawater rises. It proposes water-security and safety planning, surveillance and more resilient sources, while also noting reliability and maintenance constraints. These are a service provider’s diagnosis and proposals, not proof that a named household now receives affordable, continuous and salinity-safe water. Earlier measurements along the Ponnani–Tavanur stretch of the Bharathapuzha also show why a single statewide label is weak: salinity and water levels change by place and season. Even a salinity reading does not by itself certify potability. A useful public record has to distinguish source condition, treatment, service interruption, household access and the response offered when the normal source fails.

Evidence:KRL-33KRL-57

The working model is a seasonal backwater water-regime and access compact. Before and during an operating period, the authorised water operator would place its water-level and gate decisions beside a shared ledger: quality-assured salinity and basic water observations at agreed public points; drinking-water intake and service status; crop and drainage timing; fishing, shell-work, landing and navigation access; flood, weed and ecological observations; and barriers reported by users. A cross-sector table would publish, in accessible Malayalam and English, the reason for a decision and the authority responsible for any correction. Only the competent statutory authority could alter a regulated structure. The panchayat’s role would be local coordination and accessible issue reporting within its mandate—not unpaid gate operation, laboratory work or responsibility for another authority’s decision.

Evidence:KRL-54KRL-55KRL-56KRL-57

The first test should change no structure. For 90 days in one connected backwater reach, map existing official monitoring, intakes, drainage, farms, landing and navigation dependencies; agree safe public observation points; and rehearse a scenario in which salinity rises while a drinking-water source is disrupted. Fishers, shell and boat workers, farmers and lease cultivators, women carrying water and care work, older and disabled residents, students, renters, migrants and island or canal-edge households should help define what loss and feasible action mean, with payment for advisory time. Publish system conditions, decisions, response times, responsible authorities and unresolved trade-offs—not household locations, tenure, income or livelihood-sensitive records. The compact is a plausible, mechanism-supported fit. Whether it improves water, access, livelihoods or ecology remains unverified until an authorised trial is independently reviewed.

Evidence:KRL-33KRL-54KRL-55KRL-56KRL-57

Kerala is not starting from an empty page. The State Disaster Management Authority’s 2020 Heat Action Plan sets out warning, health, labour, education, water and local-government roles. It also reports 1,671 heat-related healthcare presentations through the 2019 IDSP reporting system, including recorded cases of sunstroke and sunburn. Those numbers establish that heat had entered routine public-health reporting and that a statewide response plan existed. They do not establish the completeness or accuracy of diagnosis, today’s incidence, which workers or settlements were missed, or how much illness or lost income any plan action prevented. A plan can name every department and still fail at the moment when one person must decide whether it is safe—or financially possible—to stop working.

Evidence:KRL-59

The local planning route is more recent. KSDMA’s 2025 guideline asks local bodies to assess place-specific exposure, name warning and coordination roles, link health surveillance to action, identify finance and plan both immediate and longer-term measures. It recognises that coastal humidity, hill settings, built form, public places, water and shade can alter what heat means locally. KSDMA’s public Heat Actions timeline also records training and workshop activity for local plans during 2026. Together these sources document guidance and capacity-building activity. They do not show that a particular panchayat has finished a plan, funded it, tested its warning chain, maintained water and shade, or reduced illness. A training event is a preparation record; the public still needs to know who acts, with what budget, by what deadline, when a warning arrives.

Evidence:KRL-58KRL-60

The official warning itself names why self-protection advice is not enough. It addresses construction and farm workers, vendors and people doing hard outdoor work, alongside students, Anganwadi children, older people, pregnant people, people with illness or disability, delivery riders, drivers, police and other public-facing workers. It recommends changes to schedules, rest, water, ventilation and institutional protection. That is an important public route, but advice is not evidence that it can be followed. A contract worker may lose pay, a delivery rider may keep the same target, a school may lack a cool room, or a household may have neither reliable water nor electricity. The unanswered question is not only whether the message arrived. It is whether employers, programmes and public services removed the penalty for acting on it.

Evidence:KRL-61

A Kottayam study brings workplace exposure into view without giving us a statewide answer. Researchers assessed 82 construction workers, 62% of whom were migrants, and reported 28% in the yellow or alarm zone and 2.5% in the red or danger zone of their heat-strain score. This is a measured result inside a small, purposively selected and mixed-risk sample. It is not a clinical diagnosis, a Kerala prevalence estimate or proof that migration caused greater risk. It also does not test whether an alert, rest break, water point, shade or changed work schedule reduced strain. Its value is narrower and still useful: any local heat plan that speaks about outdoor work should test real working conditions and include migrant and contract workers rather than assuming that public advice reaches a uniform workforce.

Evidence:KRL-62

The working model is a heat-to-action and income-protection compact. Each official warning level would trigger a short public action table: the trusted source and lead time; exposed work, school, care, market and transport settings; the schedule, shade, safe-water, rest, ventilation and referral changes required; and the named organisation responsible. Any work adjustment must also say what happens to attendance, targets, piece rates and pay. The panchayat can coordinate its local plan, public sites and accessible notices within its mandate. Health services own clinical readiness; employers, contractors, platforms and public-work programmes own the conditions of their workers. ASHAs, Anganwadi staff, students, women’s groups and workers must not become unpaid substitutes for those duties. Complaints and health signals should be reviewed without publishing personal medical or employment information.

Evidence:KRL-58KRL-59KRL-60KRL-61

The first test is a 90-day readiness audit in one willing panchayat before the next hot season. Follow one official alert from receipt to an accessible local notice, then to a safe change in work, school, care or a public service, and finally to referral, complaint and after-action review. Test separately with an outdoor worker, a gig or contract worker, a market vendor, a school or Anganwadi representative, an older or disabled resident and a household without dependable smartphone or electricity access. Verify water quality and refill, shade and rest-space maintenance, transport and health contacts, pay or target protection, and a non-retaliation route. Publish only aggregate gaps, responsible organisations, funding needs and correction dates. Whether this chain reduces illness, lost income or unequal exposure remains unverified until the trial is independently reviewed.

Evidence:KRL-58KRL-59KRL-60KRL-61KRL-62

Kerala now has a large institutional route for agricultural climate resilience, but the latest official account reviewed describes a project moving from preparation toward delivery. The State Planning Board’s Economic Review says the Kerala Climate Resilient Agri-Value Chain Modernization Project was launched in 2024–25 with climate-resilience, smallholder-commercialisation and food-system components. It records regional and project implementation units, budget allocation, preparation spending and consultancy studies, while saying implementation would follow. This is evidence that a programme, offices and preparatory work exist. It is not evidence that a farmer or worker received a useful service, made an informed choice, improved income, protected soil or water, accessed a fair market or resolved a grievance. Public reporting should keep launch, field delivery and measured outcome as three different states.

Evidence:KRL-63

A Bharathapuzha Basin study shows why a statewide climate label should not become advice for one plot. Researchers used hydrology, crop and CMIP6 climate scenarios to model streamflow and paddy-yield risk, including Palakkad and Alathur blocks. The work points to differentiated local sensitivities and discusses crop timing, water management, varieties, diversification and basin coordination as possible responses. The authors also state important limits: default rice parameters, constant land-use and management assumptions, outlet-based calibration and no direct modelling of adaptation measures or household behaviour. The result is useful for asking what conditions must be checked. It is not an observed loss, a farm forecast or proof that a named variety, irrigation schedule or diversification choice will protect yield, income or downstream water access.

Evidence:KRL-65

Decision power belongs inside the climate analysis. A Kerala coastal-agriculture study reports associations between its adaptive-capacity measure and ownership, cultivated area, freedom in decision-making and collective farming. That does not prove that ownership or collective farming causes resilience, and the restricted evidence cannot be generalised across Kerala. It does show why adaptation access cannot be inferred from location or programme enrolment alone. A landowner, lease cultivator, woman managing cultivation, landless worker and producer relying on collective or customary access may face the same season with different authority, credit, information and ability to absorb failure. Any option record must therefore ask who can consent, who bears debt and unpaid work, who receives the benefit and who has a remedy—without turning land title into the only passport to support.

Evidence:KRL-66

Labour outcomes can be measured separately from crop output. A 2026 study compared 80 Karshika Karma Sena or Labour Bank participants with 96 non-participant agricultural labourers in Palakkad and Thrissur. Using regression adjustment, it estimated that participation was associated with 24.74 more work days a year and ₹65,117 higher average annual household income. This is a bounded, outcome-supported livelihood finding—not merely a programme description. It is also not a universal causal result: the sample was non-random, covered two paddy districts and used 2024 reference conditions and a model-based counterfactual. The study does not resolve worker safety, payment timing, gender or migrant exclusion, crop and water outcomes, or performance in another climate season. It supports evaluating farm-labour institutions; it does not certify every labour-bank design.

Evidence:KRL-64

The working model is a seasonal farm-risk and livelihood-continuity compact. Before a season, a representative local table would combine verified weather, water, soil and hazard information with farmer and worker knowledge. For each realistic option it would record tenure and decision authority; crop, livestock or fishery timing; seed, input, credit and labour needs; heat and flood safety; water and ecological effects; market, storage and transport routes; household food and care implications; and the named service owner or remedy if the option fails. The panchayat can convene and record gaps within its capacity, while Agriculture, irrigation, labour, extension, markets and cooperatives own only their actual mandates and commitments. Kerala’s earlier lease-cultivator recovery case is a reminder that cultivators without title cannot disappear from the decision.

Evidence:KRL-25KRL-63KRL-64KRL-65KRL-66

The first test should precede a new input or capital work. In one willing agro-ecological cluster, run a 90-day pre-season and in-season audit using anonymous farmer, worker and lease-cultivator profiles. Map verified water, soil and hazard constraints; work and heat risks; market, transport and storage routes; debt and insurance conditions; and care burdens. Rehearse a dry-spell or badly timed flood, then publish an aggregate option matrix with named service owners, stop conditions and escalation points. Include smallholders, women farmers and workers, landless and migrant workers, Adivasi or forest-edge producers where relevant, older or disabled farmers, youth and households considering migration. Keep holdings, debts, tenancy, crop plans and Indigenous or local knowledge private unless people explicitly permit disclosure. Farm income, food security, soil, water and equity value remain unverified until complete seasons are independently reviewed.

Evidence:KRL-25KRL-63KRL-64KRL-65KRL-66

Kerala’s climate-and-health plan gives us the first connection, but not a measure of hunger. The 2022 State Action Plan on Climate Change and Human Health treats sufficient food as a health determinant and describes climate effects that can run through crop failure, displacement, contaminated water, food-borne disease and malnutrition risk. It also names variation in food production and barriers facing marginalised people among the state’s climate-health sensitivities. This is useful planning and mechanism evidence because it prevents food, water and health from being handled as separate emergencies. It does not tell us whether food became unaffordable during a particular flood or heat period, whether diets worsened, which households were missed, or whether a programme improved nutrition. Those outcomes remain unmeasured in this source.

Evidence:KRL-67

The public-distribution system is a large and legitimate route that any local continuity plan must understand. Kerala’s 2025 Economic Review reports 94.92 lakh valid ration cards as of March 2025 and 13,858 Fair Price Shops as of October 2025, alongside state responsibilities for identifying eligible households, supervising outlets and distributing food grains. These figures document administrative reach and physical infrastructure. They are not evidence that a person can use the route when a road is flooded, wages stop, electricity fails, care duties expand, a document is unavailable or the outlet is physically inaccessible. Nor do card and shop counts establish food quality, dietary adequacy, cultural fit, safety or grievance resolution. The useful question is whether the route remains usable under the precise constraints created by a climate shock.

Evidence:KRL-68

Kerala’s Orange Book supplies a starting division of responsibility during disaster response. Its 2020 framework assigns Food and Civil Supplies, Revenue, Health, local governments and other agencies roles in assessing food needs, checking quality and quantity, making emergency food and packaged water available, and arranging specific nutrients or supplementary diets for pregnant and lactating people, infants, children and disabled people. It also links equitable relief to safe water, cooking fuel or equipment, transport and shelter. This is a documented operating framework, not a current readiness audit. It does not show which staff, stocks, vehicles, water sources, accessible collection routes or local budgets are available today, or whether isolated households and relief settings received timely, safe and appropriate food in a later event.

Evidence:KRL-69

These sources change the unit of inquiry from a packet or stock figure to a continuity chain. A household needs understandable information, a safe route or alternative, physical and language access, food that fits medical, cultural and age-related needs, safe water, fuel or utensils, time to collect and cook, and a way to correct failure without publicly proving hardship. The same shock may also remove income from a fisher, farm worker, vendor or migrant household and weaken the local market that relief depends on. None of the sources compares food packages, vouchers, home delivery, community kitchens or market support. The evidence therefore supports mapping dependencies and responsibilities; it does not justify selecting a preferred delivery model or claiming that food security improved.

Evidence:KRL-67KRL-68KRL-69

The working model is a climate-shock food-continuity and dignified-access compact. Before a predictable risk season, authorised public agencies and the local body would map the PDS, Anganwadi or nutrition, health, market, water, transport and cooking dependencies for one ward. They would agree a hazard trigger, named decisions, accessible Malayalam and multilingual notices, alternate collection or support routes, and a confidential barrier-and-grievance channel. During an event, the record would follow entitlement or supply through information, safe access, water, fuel, utensils, food quality and dietary fit to actual dignified use. Statutory agencies retain their duties. Paid and trained staff—not women, students, ASHAs, Anganwadi workers or volunteers by default—need clear shifts, safety, transport, reimbursement and handover rules.

Evidence:KRL-68KRL-69

The first test should use no resident-identifying data. With the responsible agencies’ approval, run one ward-scale tabletop for a heavy-rain access disruption and another for heat-linked income loss. Trace two representative but fictional households—such as an older disabled resident and a migrant worker household with care duties—through only official routes. Test message comprehension, document and language barriers, an accessible alternate collection route, safe water and cooking needs, dietary fit, escalation and appeal. Record whether every barrier has a named owner, budget route and correction date. Publish only aggregate gaps and unresolved responsibilities, never names, addresses, ration details, health conditions or inferred vulnerability. These process results can show readiness defects; they cannot be presented as reduced hunger, better nutrition or a proven continuity system.

Evidence:KRL-67KRL-68KRL-69

Kerala now has a documented relief route for recurrent coastal high-tide flooding and sea incursion. A February 2026 government notice says these hazards were declared state-specific disasters and that eligible losses can receive State Disaster Response Fund assistance under applicable norms. The notice recognises harm affecting fishers, coastal farmers, settlement residents and small traders. That is important because it makes recurrent coastal livelihood loss visible inside a public relief route. It is still evidence of authorisation, not evidence of recovery. The notice does not show that a household received money, how quickly a decision was made, whether eligibility and appeals were fair, or whether support protected food access and livelihoods. It is also a post-loss route: the source does not document a forecast trigger or support arriving before the harm.

Evidence:KRL-70

A randomized comparison in Bangladesh and Nepal gives a bounded reason to test timing. A 2025 World Bank working paper studied 300 flood-exposed villages in Bangladesh and 138 in Nepal that received the same one-off assistance at different times. In Bangladesh, the anticipatory group received its transfer two days before the flood peak, while the comparison group received it roughly a month later. The pooled short-run analysis reported a higher food-consumption score and a lower coping-strategies index for households receiving earlier support. After the later transfers arrived, the results converged, and the paper reported no long-run difference. Random assignment makes this meaningful causal evidence about relative timing in those programmes. It does not establish cash versus no support, and it does not prove that mobile money, a forecast trigger, the transfer amount or the result will travel to Kerala. The paper is also a working paper, not a final peer-reviewed article.

Evidence:KRL-71

Household-level cash does not automatically create individual choice or safety. A peer-reviewed qualitative study in northern Bangladesh reports that women described forecast-based cash as helping with household food purchases, boat evacuation and reduced borrowing. The same accounts said money was not spent on women’s own utility and safety needs, while hygiene and service barriers continued. This study cannot estimate a general treatment effect, represent all women or tell us what will happen in Kerala. It does directly challenge an easy assumption: that an unconditional transfer is equally controlled and equally useful to everyone inside a household. Any Kerala test must therefore ask who receives information, who controls the support, who undertakes risky travel and care work, which personal or disability needs remain unmet, and whether a confidential remedy exists. A household record cannot substitute for individual access, consent or safety.

Evidence:KRL-72

The transfer lesson for Kerala is a decision architecture, not a cash recipe. Kerala already has documented public-distribution infrastructure and disaster roles for food, water and supplementary nutrition. The new coastal notification adds a potential post-loss relief route. None of those sources establishes authority for forecast-triggered livelihood payments, or shows that roads, markets, water, power, networks and banking will remain usable during a particular hazard. Before choosing cash, food, home delivery or another bridge, public authorities would need a verified trigger and stop rule; a legal and fiscal decision; continuation of PDS, nutrition, compensation and labour rights; current market and service checks; non-digital and non-bank alternatives; accessible Malayalam and other required languages; minimum private data; and a workable complaint and appeal route. The comparator supports testing whether timing changes choices. It does not justify replacing entitlements or transferring a foreign delivery system into Kerala.

Evidence:KRL-68KRL-69KRL-70KRL-71

The working model is a forecast-to-rights livelihood and food-support compact. Subject to verified legal authority, a named district and local coordination table would agree a hazard trigger, a public decision record, a funding boundary, safe delivery choices and an appeal before the season begins. When the trigger is met, it would first protect existing food, nutrition, water, compensation and labour routes. A financial or in-kind bridge would be considered only after checking that markets, transport, water, power and communication work and that people without a phone, bank account, biometric match, stable address, recognised card, mobility or secure tenure still have two safe options. State and district authorities retain trigger, fund and legal responsibility. Panchayats communicate verified local conditions only where authorised. Targeting, translation, delivery, care work and complaints must not be pushed without pay onto women, community workers, students or volunteers.

Evidence:KRL-68KRL-69KRL-70KRL-71KRL-72

The first test should disburse nothing and collect no resident-identifying data. With written public-authority approval, run a tabletop exercise in one hazard-exposed ward using fictional profiles: a coastal fisher household, an agricultural worker or tenant household, a migrant worker without a local address, an Adivasi or rightsholder household where relevant, and a disabled or older-person household. Trace alert, legal decision, entitlement and market checks, two accessible delivery routes, food-water-fuel continuity, complaint, appeal and an aggregate after-action record. Stop if legal authority, a safe alternative, privacy or a remedy cannot be shown. Record readiness measures such as decision time, exclusion errors corrected, accessible routes and named owners. Do not report these as reduced hunger, protected income, avoided harm or a proven anticipatory-support system. Those outcomes would require authorised real-world implementation and independent evaluation.

Evidence:KRL-68KRL-69KRL-70KRL-71KRL-72

Kerala has a visible heat-action route, but the route is not yet evidence of protection. KSDMA’s current Heat Actions timeline records daily alerts since 2016, a history of worker-hour rescheduling, the 2020 state Heat Action Plan, local-plan guidance issued in 2025, a reported first grama-panchayat heat plan at Moodadi, school water-bell activity and higher annual DDMA allocations for heat action in 2026. These records show that heat has entered state and local administrative work, with a planning and funding path that a panchayat can examine. The same timeline is an administrative account, not an independent audit. It does not show whether Moodadi’s plan was completed, funded, rehearsed or used; whether workers could change hours without losing pay; whether water, shade and referral services functioned; or whether illness, deaths, lost income or inequality fell in any locality.

Evidence:KRL-58KRL-59KRL-60KRL-61KRL-73

Ahmedabad supplies outcome-supported evidence for a bundled city plan, with important causal limits. A 2018 peer-reviewed study compared the city’s temperature–all-cause-mortality relationship in 2014–15, after full Heat Action Plan implementation, with a 2007–10 baseline. It reported lower post-plan heat-associated mortality rates and estimated 1,190 average annualised deaths avoided. That is more than a programme description: it is a measured before-and-after association. It is not a randomized test, and it cannot isolate a warning, outreach, clinical training or any other component. The comparison spans different heat seasons and a rapidly changing city, uses all-cause mortality, and cannot remove every concurrent change. It therefore does not prove causality, show effects for workers or low-income groups separately, or provide Kerala with a transferable mortality number, threshold or promise. Ahmedabad is a reason to test a complete action chain, not a benchmark Kerala can claim.

Evidence:KRL-74

The more transferable Ahmedabad lesson is how the plan was assembled and rehearsed. A 2014 peer-reviewed programme-development account describes analysis of local temperature, mortality and service data; surveys and focus groups with exposed communities and workers; named roles across agencies; public communication; health-worker capacity building; a table-top exercise; and changes to agency actions before the pilot. This supports a mechanism: a warning is more likely to become action when responsibilities, services and communication have been tested together before the heat arrives. The paper was produced by programme partners and is not an independent impact or equity evaluation. Ahmedabad’s dense urban services, municipal health structure and relatively low-humidity setting do not establish what a Kerala panchayat can fund, enforce or maintain. The design discipline can travel only after local authority, staffing, climate, work and care patterns, water conditions and resident knowledge are rebuilt into it.

Evidence:KRL-75

Kerala should not import Ahmedabad’s maximum-temperature thresholds or reduce a plan to alert colours. Ahmedabad’s early analysis was built around local mortality thresholds in a hot, relatively low-humidity city; its own development paper cautions that humid coastal cities can have different heat-health relationships. Kerala’s guidance already calls for local planning, while Kerala evidence also shows that warnings, worker guidance and a small Kottayam heat-strain study answer different questions. A threshold can start a public decision only if it fits local heat and humidity, night-time and indoor exposure, work and care schedules, water and electricity conditions, and health access. The operational test is whether a woman combining paid work and care, a migrant or contract worker, a fisher or farm worker, a student, or an older or disabled resident can understand the warning and take a safe action without losing income, water, dignity or a remedy.

Evidence:KRL-60KRL-61KRL-62KRL-74KRL-75

The working model is a locally grounded heat-action accountability compact. Under health and disaster authority, a local body would connect an official warning to a trusted and accessible message, a named action at public sites, worksites, schools and care settings, ready water and health referral, a no-retaliation complaint route and an after-action correction record. KSDMA and district authorities retain forecast and warning functions. Health agencies own clinical protocols and surveillance. Employers, contractors, platforms and public-work programmes own safe schedules, rest, water, shade, pay and non-retaliation within their duties. The panchayat coordinates its plan and public infrastructure only within its mandate. Budgets must cover staff, water safety and refill, shade and rest upkeep, schedule changes, transport and referral, translation, data protection and paid participation. Women, ASHAs, Anganwadi workers, students, informal workers and volunteers cannot be the unpaid enforcement layer.

Evidence:KRL-58KRL-59KRL-60KRL-61KRL-73KRL-75

The first test should collect no personal data and make no claim about lives saved. With public-authority approval, run a two-day pre-season tabletop in one willing panchayat: first, a humid night followed by a high-heat workday; second, a heat day combined with a water-service interruption. Include representative but fictional profiles for an agricultural worker, a woman with paid and care work, a construction or contract worker, a fisher or shore worker, a student, an older or disabled resident, and an Adivasi or rightsholder setting where relevant. Test warning comprehension, ability to change activity without an income penalty, safe water, rest and referral, employer and public-site action, non-digital access, complaint escalation and named correction dates. Publish only aggregate gaps and owners. These readiness results cannot be reported as reduced mortality, illness, lost income or inequality; those claims require independently reviewed real-world outcomes.

Evidence:KRL-58KRL-60KRL-61KRL-62KRL-73KRL-74KRL-75

Kerala already has a legitimate route for local disaster planning. The 2020 KILA and KSDMA framework, issued through the Local Self Government Department, sets a common template built around situation analysis, resource mapping and intervention reports. It directs discussion through disaster-management working groups and grama or ward sabhas, connects suitable projects to annual plans, and describes Local Resource Groups, ward teams and district scrutiny. That matters because it places preparedness inside public planning rather than treating it as an informal volunteer activity. But the document is a framework and order, not a current audit. It cannot tell a resident whether a named panchayat has an approved and updated plan, whether its contacts and resource lists are current, whether shelters, water, toilets, lighting and routes work, or whether participation was representative, accessible, safe and influential.

Evidence:KRL-76

Kerala also has a substantial climate-information input, but information is not capacity until a local institution can interpret and use it. KSDMA and KILA’s 2022 downscaling report describes precipitation and temperature analysis for all 1,034 local governments, alongside a Pamba Basin programme route involving training, climate-risk integration, investment planning, assessment and emergency-management exercises. The report itself records a baseline problem: local climate-risk information had not been integrated into preparedness. Its models and programme targets can inform questions about future conditions and planning needs. They do not establish that a particular panchayat received or understood the material, converted it into an authorised and funded action, rehearsed a service chain, or protected anyone. A downscaled projection is neither a ward-level observation nor an operational forecast, and a programme score is not evidence that a resident can use a safer route during a disruption.

Evidence:KRL-77

The research therefore shifts the unit of attention from a plan document to a maintained readiness loop. Once each season, and after a meaningful disruption, an authorised local process would verify one priority hazard, the service functions and groups affected, the minimum public resource and access map, named agency owners, operating funds and an accessible action chain. It would rehearse that chain, record only aggregate failures, assign corrections and return on a visible date. Existing Kerala material shows pieces of this architecture: local heat guidance asks local governments to assess risk and coordinate action; the disaster framework names food, water, health and relief responsibilities; and the Heat Actions timeline records plans, training and funding routes. These are documented routes, not proof that the links join up. The proposed refresh loop is a plausible mechanism whose value remains unverified until a named locality tests it under lawful authority and resident governance.

Evidence:KRL-60KRL-69KRL-73KRL-76KRL-77

Maintenance is mostly operating work, and hiding that work is a major failure mode. Contacts and routes need verification; meetings need translation and accessibility; shelters, pumps, warning equipment and public services need inspection; exercises need safe facilitation; complaints need response; and corrections need funding and follow-up. The competent agencies must retain their health, water, road, labour, shelter and warning duties, while the panchayat coordinates only what its mandate permits. None of this labour should be quietly shifted to women, ASHA or Anganwadi workers, Kudumbashree members, students, fishers, migrant workers or other volunteers without authority, choice, pay, insurance, transport and safety. Local knowledge must not become a free data source. Household routes, land or forest-right details and precise vulnerable locations should remain governed locally; a public dashboard should show aggregate gaps, responsible institutions and due dates, not create surveillance, land-value or exclusion risks.

Evidence:KRL-60KRL-69KRL-76KRL-77

The first test should be small, authorised and honest about what it can show. In one willing ward, use no personal data and choose one locally relevant chain: for example, heavy rain blocking the route to drinking water, or a heat day disrupting water and work. Compare the existing plan and resource list with a facilitated check by residents and the responsible service owners. Test one accessible notice, a non-digital alternative, an alternate route or service, escalation and a correction register. Record whether selected contacts, routes and resources were verified; whether language, disability and income barriers received an owner and due date; whether recurring costs and staff protections were named; and whether corrections were revisited within 90 days. Publish only aggregate gaps and authority or funding constraints. A cleaner list, a completed exercise or a faster message is evidence of process readiness only. It cannot be reported as reduced loss, illness, displacement or inequality without independently reviewed real-world outcomes.

Evidence:KRL-60KRL-69KRL-73KRL-76KRL-77

Kerala has moved beyond treating migrant or guest workers as an unnamed audience. KSDMA’s current guideline index records a 2025 first edition of a dedicated Guidelines and Response Plan for Migrant Labour Inclusive Disaster Risk Reduction. That is a meaningful public signal: the state recognises that inclusion needs its own disaster-planning route across preparedness, response and recovery. The admissible claim stops there. A guideline is not an implementation audit or an outcome evaluation. Its publication does not establish that a district, panchayat, employer, contractor, accommodation provider or worker organisation has adopted it, budgeted for it, translated it well, rehearsed it or corrected failures. The linked source is in Malayalam and has not received the fluent human language review required for detailed quotation or operational reliance in this project. This note therefore records the route and its limits; it does not represent any detailed provision as verified practice.

Evidence:KRL-78

Kerala’s 2018 post-disaster needs assessment shows why translation cannot carry the whole burden. The assessment reports that many migrant workers did not receive or understand Malayalam alerts. It also records reported shelter exclusion or early closure, unpaid wages, barriers connected to documents, address and bank access, weak access to water, electricity and health support, and unsafe cleaning work without protective equipment. These are coupled access failures: even a perfectly translated alert is unusable if leaving work risks wages or eviction, a shelter will not admit a person, food and water routes are uncertain, or a grievance creates more danger. The assessment is historical and post-event. Its migrant-worker account is not a representative prevalence survey, a causal analysis or evidence of current conditions. It cannot establish that every agency failed in 2018 or that later guidance has or has not improved practice. It does establish a serious question set that any current local test must answer.

Evidence:KRL-79

The defensible design is a warning-to-action chain, not a translated poster. Under authorised disaster, labour, health and local-government arrangements, an official trigger would move through a trusted multi-language and non-digital message to a worker-controlled safe action, verified shelter, transport, food, water and health routes, the employer or accommodation provider’s lawful duties, a confidential grievance hand-off and an aggregate correction record. Kerala’s heat advisory and the Kottayam construction-worker study add a relevant warning: workers can face exposure while having limited control over schedules, rest, water or income, but neither source measures a migrant-inclusive response. The disaster food and water framework names public roles, yet a role on paper is not service reach. The proposed chain is therefore mechanism-supported and plausible, not proven. It must be tested with worker-governed participation and responsible agencies before any claim of inclusion, safety or resilience.

Evidence:KRL-61KRL-62KRL-69KRL-78KRL-79

Operating ownership and data boundaries determine whether the chain protects or harms. Forecast and warning functions remain with competent public agencies. Local governments coordinate local access only within mandate. Labour authorities, employers and contractors retain workplace, wage and safety duties; health, civil-supplies, water and revenue agencies retain their services. Budgets must cover professional translation, interpreter access, printed and in-person communication, safe transport and shelter, protective equipment, complaint handling, staff time and corrective maintenance. A migrant “link worker” cannot become the unpaid, exposed substitute for these systems. Identity, accommodation and movement data are sensitive. Immediate safety should not depend on a registry, biometric match, immigration status, employer approval or police contact. Women workers, families, newly arrived people, workers without stable addresses or phones, disabled and older workers, isolated worksites and people facing caste, gender or employer power need distinct routes and the right to refuse unsafe data collection.

Evidence:KRL-69KRL-78KRL-79

A first test should be fictional, paid and privacy-preserving. With written agency permission, run a 90-minute tabletop for a heavy-rain disruption affecting a worksite and worker accommodation. Include paid worker representatives selected through safe channels, a woman worker or family perspective, disability-access input, the local government, Labour, health or water, civil supplies and a responsible employer, contractor or accommodation role. Test message comprehension in at least two locally relevant languages, a no-phone route, safe exit without pay penalty, shelter and service referral, confidential complaint hand-off and an aggregate correction record. Collect no names, addresses, employer identity, immigration data or precise worker locations. Stop if retaliation safeguards, service capacity or data governance cannot be demonstrated. Measure usable routes, named duty-holders, time to feasible action and corrections checked by the due date. These are process-readiness measures, not evidence of safety, income protection, health, inclusion or equity outcomes.

Evidence:KRL-61KRL-62KRL-69KRL-78KRL-79