Nepal-Tibet floods: when disaster strikes before communities recover from previous climate extremes, crisis escalates

Major flash floods on the Nepal-Tibet border have destroyed homes, roads, bridges and local hydropower projects. The flooding of the Bhote Koshi river is affecting 50,000 residents. More than 380 tourists are missing. At least 95 people have already died. This region of the Himalayas had not yet recovered from the Bhote Koshi flooding in 2025 that killed nine people as it washed away a “friendship bridge” linking Nepal and China, disrupting trade and transport for months. In Sri Lanka, recent floods and landslides struck affected communities that were still recovering from Cyclone Ditwah, which had caused widespread damage only nine months earlier. With little time to recover from the cyclone, communities had no protection from further flooding. More than 15,000 people have been directly affected and 6,800 people were displaced this August. Disaster risk usually considers how severe the next flood, wildfire, cyclone or drought will be. But the Nepal and Sri Lanka examples show another dimension that is often overlooked: how quickly can a community recover and will it do so before the next disaster arrives? A disaster rarely leaves a place exactly as it found it. Roads and bridges may remain damaged. Families may have spent their savings. Businesses may still be struggling. Governments can still be financing reconstruction, while hospitals and other public services remain under pressure. If another hazard arrives before these systems have recovered, it does not encounter the same community that existed before the first disaster. It encounters a place with less resilience and a weakened infrastructure. A useful way to understand this is through the recovery gap: the relationship between how long recovery takes and how much time is available before another damaging disaster occurs. Imagine two communities experiencing similar floods. In one, the previous major disaster occurred ten years earlier. Homes have been rebuilt, infrastructure repaired and emergency resources replenished. In the other, another disaster happened only six months ago. Roads remain under repair, families are still displaced or in debt, and public authorities are still paying for reconstruction. The new flood may be physically similar in both places. Its consequences may not be. Incomplete recovery can substantially increase long-term losses. A 2025 study of recurrent flooding in the Philippines that losses increased by 40% between 2000 and 2018 because of incomplete recover. This suggests that repeated disaster losses are not always simply additive. The first disaster can change the conditions under which the next one occurs and the effects of subsequent disasters can be much more severe. Recovery has many clocks Recovery is often described as if a community is either recovered or not, a simplistic binary. It’s more of a sliding scale. Electricity and freshwater supplies may return within days. A damaged road may take months to repair. Rebuilding homes can take years. Businesses may reopen while still carrying substantial debt. Household finances, ecosystems and mental wellbeing can recover on entirely different timescales. A place can therefore look recovered according to one indicator while remaining highly vulnerable according to another. It took more than a decade for New Orleans in the US to achieve substantial physical and structural recovery following Hurricane Katrina in 2005. While basic emergency operations and power restoration took weeks to months, full neighbourhood rebuilding, population stabilisation and major flood protection system upgrades spanned 15 years or more. And this was in one of the richest countries in the world. Recovery is also unequal. Wealthier households may have insurance, savings and access to credit. Poorer households often have fewer resources to repair homes, replace belongings or absorb lost income. In the case of Hurricane Katrina, many poorer households who had hurricane insurance were refused payouts as the damage was deemed to be flood-related because of the burst levees and not by the hurricane. So, it is not only the time between disasters that matters. It is also how much recovery can happen during that time. This issue becomes increasingly important where climate change alters the frequency or severity of weather extremes. If reconstruction requires five years but another damaging event occurs after two, the interval between disasters starts competing with the time needed to recover. That should change how we manage disaster risk. Measure recovery, not only damage Governments are generally good at recording immediate losses: damaged buildings, displaced people, destroyed infrastructure and financial costs. But they also need to know what remains unrepaired months or years later. Recovery monitoring needs to include housing, infrastructure functionality, business activity, household finances, public services and displacement. The question should not stop at “how much did we lose?” It should continue with “how much vulnerability remains?” Disasters are not isolated occurrences. Floods, wildfires, droughts and storms are often assessed separately. These assessments are essential, but communities experience these hazards through the same infrastructure, economy and population. The road damaged by one event may be needed for evacuation during the next. An electricity system weakened today may be supporting hospitals or water services tomorrow. Risk assessments should therefore consider plausible hazard sequences and the condition critical systems may be in when another disaster hits. This does not mean predicting the exact order of future disasters. It means abandoning the assumption that every new disaster meets a fully recovered system. Recovery should not be seen only as the final stage of the disaster that just happened. It is also preparation for whatever comes next. Restoring infrastructure, supporting households and reopening essential services quickly can reduce the vulnerability carried into the next event. But simply rebuilding the same vulnerable systems recreates the same risks. Recovery should therefore combine speed with adaptation: stronger infrastructure, better redundancy, improved emergency planning and greater support for communities that would otherwise take the longest to recover. Recent research shows that the government plays a critical role in ensuring rapid, efficient and effective recovery. While asking “how severe could the next disaster be?”, we should also ask “how recovered will we be when it arrives?” Because the next disaster does not start from zero. It inherits what the last one left behind.