Spatial Heterogeneity of Flood–Salinity WASH Vulnerability: An Integrated FS-WVI and CatBoost–SHAP Assessment across Eight Unions of Subarnachar Upazila, Noakhali, Bangladesh

In the low-lying coastal lands of Bangladesh, water is a constant companion, but it is not always a friend. For generations, communities here have navigated a delicate balance between the life-giving rains of the monsoon and the creeping threat of saltwater intrusion from the sea. When the two forces collide, the result is a complex crisis that goes beyond simple flooding. It creates a situation where the water that fills a home is also the water that poisons the soil and the drinking supply. Scientists call this a compound hazard, a term that simply means two or more dangerous conditions happening at once to make a situation worse than either one alone. Understanding how these combined forces affect the daily lives of people—specifically their access to clean water, safe toilets, and basic hygiene—is crucial for planning how to survive and adapt to a changing climate. Without this understanding, aid and infrastructure projects might miss the mark, helping the wrong people or solving the wrong problems.

A recent study in the Subarnachar Upazila, a coastal sub-district in Noakhali, Bangladesh, tackled this challenge by looking at the problem through two different lenses at the same time. The researchers wanted to know not just where the water was rising or where the salt was highest, but how those physical conditions actually translated into vulnerability for individual families. They focused on eight distinct communities, or unions, within the area, investigating why some households struggled immensely with water and sanitation while others, even in the same flood zone, managed better. By combining satellite images of the landscape with detailed interviews from hundreds of families, the team built a new way to measure risk that accounts for both the environment outside the door and the resources inside the home.

The investigation began with a massive flash flood in August 2024, which swept through eastern Bangladesh and brought unexpected water to areas that were already dealing with high salinity. To understand the scale of this event, the team used satellite technology to map exactly where the floodwaters spread and to measure the saltiness of the soil and surface water. They combined these two pieces of information into a single score that represented the intensity of the combined flood and salt pressure on the land. However, they knew that a map of the hazard was only half the story. A family living in a flooded house might be in a dangerous spot, but if they have a deep well, a strong roof, and enough money to buy clean water, they might survive the crisis. Conversely, a family in a less flooded area might be in deep trouble if they have no safe water source, poor housing, or no way to pay for medical care.

To capture this reality, the researchers surveyed 384 households across the eight unions. They asked specific questions about how long their homes stayed underwater, how deep the water rose, and whether they lost crops or livestock. They also asked about the quality of their drinking water, whether family members fell sick with diarrhea or skin infections, and what kind of toilets they used. Crucially, they gathered data on the families' ability to cope: their monthly income, the type of house they lived in, the depth of their water wells, and whether they received help from the government or aid organizations. By mixing these household answers with the satellite data, the team created a comprehensive vulnerability index. This index did not just measure how bad the flood was; it measured how exposed a family was, how sensitive they were to the damage, and how much capacity they had to bounce back.

The results revealed a landscape of vulnerability that was far more complex than a simple map of floodwaters would suggest. The study found that the eight unions did not all suffer in the same way. Three unions—Charaman Ullah, Mohammadpur, and Char Jubilee—showed the highest levels of overall vulnerability. In these places, the primary driver of the problem was exposure; the families were simply in the path of the worst flood and salt conditions. However, the story changed in other areas. In unions like Char Clark, the physical hazard was lower, but the families were still vulnerable because their water sources were naturally salty or their health conditions made them more susceptible to disease. In one specific union, Char Jabbar, the main issue was not the flood itself or the family's health, but a lack of resources; these families had the least ability to adapt or recover because they lacked money, strong housing, or institutional support.

Perhaps the most surprising finding was that the areas with the most severe physical hazards were not always the ones with the highest overall vulnerability. For instance, one union had very low flood and salt scores on the satellite maps but still ranked as highly vulnerable because its residents had very few resources to cope with even a moderate event. This mismatch proved that looking at the environment alone is not enough to understand human risk. The study confirmed that vulnerability is a product of the interaction between the physical world and the social world. A flood becomes a disaster not just because of the water level, but because of the depth of the water well, the strength of the house, and the availability of income to buy medicine or clean water.

To dig deeper into what specifically caused families to have salty drinking water, the researchers used a sophisticated computer model that could find patterns in the complex mix of environmental and social data. This model acted like a highly trained observer, analyzing how different factors worked together to predict whether a household's water would be fresh or saline. The analysis pointed to two main factors as the strongest predictors: the depth of the water well and the combined intensity of the flood and salt hazard. It showed that while the hazard on the land was important, the depth of the well a family used was even more critical in determining if their water was safe to drink. The model also highlighted that support from the government or non-governmental organizations played a significant role in keeping water safe.

The study did not stop at identifying the factors; it also explained how they worked together. The computer analysis showed that the impact of the flood and salt hazard was not the same for everyone. For a family with a very deep well, a high level of flood hazard might not result in salty drinking water. But for a family with a shallow well, that same level of hazard could mean their water became undrinkable. This interaction meant that the solution to the problem had to be tailored to the specific conditions of each community. In some places, the priority might be to build better flood defenses. In others, the focus might need to be on digging deeper wells or providing financial support to help families buy clean water during a crisis.

The researchers concluded that their new approach, which blends satellite data with household stories and advanced computer analysis, offers a powerful tool for planning. It moves beyond the old way of thinking that treated flooding and salinity as separate issues or assumed that all coastal communities were equally vulnerable. Instead, it provides a clear picture of where the risks are concentrated and, more importantly, why they exist. By understanding that vulnerability is shaped by a unique combination of exposure, sensitivity, and the ability to adapt in every single community, planners can design interventions that actually work. The study suggests that in the future, efforts to protect water and sanitation in coastal Bangladesh must be as diverse and nuanced as the communities themselves, targeting the specific weaknesses of each area rather than applying a one-size-fits-all solution.