Why Disaster Loss Data Matters More Than Ever for Climate Adaptation

In Cox's Bazar, Bangladesh, I watched a reforestation programme collide with a community that was already losing ground, literally. The sudden influx of approximately 700,000 Rohingya refugees in 2017 had caused immense environmental strain, stripping hillsides of forest at a rate estimated at roughly four football fields every single day. IOM's reforestation effort, part of the broader Safe Access to Fuel and Energy Plus (SAFE+) programme, was replanting over 778 hectares with more than 775,000 trees. The goal was to stabilise soil against landslide and flood risk in and around the camps.

The host communities pushed back, and hard. Not because they opposed reforestation. They were dealing simultaneously with rising sea levels eating into their own agricultural land, increasingly severe cyclone seasons battering their livelihoods, and the social and economic pressure of hosting one of the world's largest displaced populations on land they were already losing to the Bay of Bengal. IOM adopted a cash-for-work approach to the reforestation, hiring host community members for planting, site preparation, and nursery management. It was a proven method for creating livelihood opportunities and reducing the refugee-host community tensions that were building across the district. The approach was smart. It turned an environmental intervention into an economic one and gave host communities a material stake in the programme's success.

But when we sat with local disaster management authorities to discuss reforestation priorities, the conversation still cut deeper than livelihoods. It was about competing vulnerabilities, compounding risks, and a community whose own climate losses felt invisible next to the scale and resourcing of the refugee response. The land they were losing to the sea. The embankments failing each monsoon. The crops destroyed by cyclone flooding. Cash-for-work addressed the economic tension. It could not address the evidentiary one: that the host community's disaster losses were undocumented, unquantified, and therefore unfundable.

That experience crystallised something I had been observing across multiple crisis contexts. Disaster loss data is no longer a back-office record-keeping exercise. It is the evidentiary backbone of the entire global climate adaptation architecture. And the communities that need it most are usually the ones whose losses are least documented.

The host communities around Cox's Bazar had a legitimate grievance. Their flood losses, their coastal erosion, their cyclone damage. None of it was systematically recorded in a format that could compete for adaptation funding against the well-documented refugee response. The refugee operation had registration data, displacement tracking, cluster-level needs assessments, and donor reporting pipelines. The host community's climate losses had fragmented local records and anecdotal evidence.

This asymmetry is not unique to Bangladesh. Of 193 UN Member States, only 153 report to some degree on the Sendai Framework targets, and significant data quality gaps persist even among those that do. Many countries, particularly in the Global South, still rely on paper-based disaster records or fragmented spreadsheets that cannot be aggregated, compared, or verified. The Global Assessment Report 2025 estimated the true cost of disasters at $2.3 trillion globally. Yet in many of the most disaster-affected countries, the loss data that would justify DRR investment simply doesn't exist in a usable form.

The Cox's Bazar experience laid bare an uncomfortable truth. The responsibility for collecting disaster loss and damage data was never in the hands of humanitarian organisations in the first place.

IOM, UNHCR, WFP, and the cluster system collect operational data (displacement figures, needs assessments, response coverage) because they need it to run emergency programmes. That data is designed for coordination, not for national statistical accounting. It tracks what humanitarian agencies are doing. It does not systematically track what disasters are costing a country's population, infrastructure, agriculture, and ecosystems over time. Yet in the absence of functioning national systems, humanitarian data has become a proxy for loss data, and a poor one at that, because it captures response activity rather than comprehensive impact.

The responsibility for systematic, complete, and quality disaster loss and damage data sits with two national institutions. National Disaster Management Agencies (NDMAs) collect operational impact data from the ground. National Statistical Offices (NSOs) certify that data as official statistics meeting international standards. This is where the Sendai Framework places the mandate. This is also where the G-DRSF, endorsed by the UN Statistical Commission in March 2026, assigns institutional roles.

But we have to be honest about why these institutions have not always fulfilled this mandate. NDMAs in many countries operate with skeleton staff, outdated tools, and budgets that prioritise emergency response over data management. NSOs rarely have disaster statistics units. When they do, those units compete for resources against census operations, economic surveys, and demographic monitoring. UNDRR's own capacity assessments have consistently found gaps in governance, technical infrastructure, data quality, and human capacity across the countries they support. The UNDRR Strategic Framework 2026-2030 identifies risk knowledge as a critical gap requiring systematic institutionalisation and resourcing. An acknowledgement, basically, that the mandate exists but the means to fulfil it often do not.

The result is predictable. One of the most common challenges among reporting countries is the reliability of data and the systematisation of datasets from different sources generated by different institutions. Data completeness, consistency, and disaggregation remain uneven. And communities like those in Cox's Bazar, whose losses fall outside both the humanitarian data pipeline and the capacity of under-resourced national systems, end up in an accountability void where nobody is counting what they have lost.

Meanwhile, humanitarian data capacity itself is shrinking. The State of Open Humanitarian Data 2026 revealed that crisis data availability has fallen from 74% to 68% across 22 humanitarian operations. OCHA, UNHCR, and IOM have all experienced significant reductions in data staff. The proxy system is degrading at the same time as the demand for the real thing has never been higher.

The importance of disaster loss data has not changed. It has always mattered. What has changed is that two global policy processes now simultaneously demand it, and the consequences of not having it are financial.

The Loss and Damage Fund has $768 million in pledges against $580 billion in estimated need. Its first COP30 call for proposals made one thing clear: evidence-based loss data is the prerequisite for accessing finance. Communities like those around Cox's Bazar cannot access this funding without structured proof of what they have lost. The Sendai Framework Endgame enters its final five-year implementation window in 2026, with the "Beyond the Numbers" acceleration strategy demanding disaggregated, validated, internationally comparable data. The 38 Sendai indicators feed directly into 12 SDG indicators across targets 1.5, 11.5, 11.b, and 13.1.

Each process independently requires granular disaster loss data. Together, they create an unprecedented demand signal, and an unprecedented penalty for countries that cannot respond.

Back in Cox's Bazar, what would structured disaster loss data have changed? Almost everything about how that conversation with local disaster management authorities unfolded.

Imagine the district had maintained disaggregated records: hectares of agricultural land lost to coastal erosion by year, number of households displaced by cyclone flooding by union, damage to embankments and infrastructure by monsoon season. The host community's climate vulnerability would have been quantifiable, comparable, and fundable. The reforestation programme would not have needed cash-for-work alone to earn community consent. It would have been framed from the outset as what it actually was, a dual-benefit climate adaptation measure where replanting stabilised hillsides for refugees at risk of landslide and restored watershed function for a host community losing agricultural land to erosion and flooding. The data would have made both vulnerabilities visible in the same frame, and made the case that investing in one community's resilience was inseparable from investing in the other's.

The consent problem we faced was, at its root, a data problem. The refugee response had data infrastructure (registration systems, displacement tracking, cluster-level needs assessments, donor reporting pipelines). The host community's climate losses had fragmented local records and anecdotal evidence. Cash-for-work could address the economic grievance. Only structured loss data could have addressed the deeper one: the feeling that your crisis does not count because nobody is counting it.

This is precisely the problem that DELTA Resilience is designed to solve. Co-developed by UNDRR, UNDP, and WMO to replace the legacy DesInventar platform, DELTA is a comprehensive system of tools, standards, and governance frameworks built to give NDMAs and NSOs the infrastructure they have lacked. Its Data Ecosystem Maturity Assessment diagnoses gaps in governance, infrastructure, data quality, and human capacity before any technology is deployed.

Critically, DELTA applies no minimum thresholds for recording. Localised, cascading, slow-onset, and rapid-onset events can all be documented consistently across sectors and scales. Legacy systems tend to capture headline disasters while the slow erosion of agricultural land, the seasonal flooding that displaces a few hundred families, and the localised landslide that destroys a school go unrecorded. For communities like those in Cox's Bazar, whose losses were incremental, compounding, and politically invisible, a no-threshold system means their crisis finally gets counted. DELTA uses universally unique identifiers (UUIDs) to systematically connect hazardous-event observations to their impacts (including cascading and compound effects), producing the granular, multi-hazard loss records that the Sendai Framework, the Loss and Damage Fund, and the Belém Indicators all require. Its "one-report-two-purposes" design means data entered once for the 38 Sendai indicators automatically feeds 12 SDG indicators, eliminating double-reporting. The Arab States regional rollout, launched in Doha with 18 Member States, demonstrated the model: country-specific roadmaps drafted around institutional capacity, not technology wish lists.

The investment case is direct. Disaster loss data is the input that makes every other DRR investment measurable. The countries that invest now will access the Loss and Damage Fund. Those that do not will find themselves locked out, not because their losses are less real, but because they cannot prove them.

The architecture is finally in place: DELTA Resilience, the G-DRSF, the Sendai Framework Monitor, the Loss and Damage Fund. What remains is the hardest part. Building national capacity to collect, validate, analyse, and publish disaster loss data that meets these standards. That means investing in data ecosystem maturity assessments before deploying technology, forging NDMA-NSO partnerships that outlast project cycles, and recognising that the data officer in a district disaster management office, the person who could have documented what Cox's Bazar's host communities were losing to the sea, is doing some of the most consequential climate work on the planet.

Every dollar of climate finance that flows to the wrong place because the loss data wasn't there is a dollar stolen from the communities who need it most. The data systems exist to prevent that. The question is whether we will build them fast enough.