Coastal Bangladesh’s low-lying topography is often depicted by media, donors and even scholars as particularly vulnerable to rising sea levels. In this paper, I problematise the current framing of changing rainfalls and sea level rise as the main causes of floods in the Bengal delta, especially the suggestion that large-scale flood-protection infrastructure constitutes climate adaptation. As Zaman (1993: 987) points out, the complete elimination of flooding in Bangladesh is undesirable ‘for flood is intricately linked with the very survival of the people in this delta country’. Drawing on environmental history and ethnography, I show how narratives of climate vulnerability ignores the fact that there are three types of floods in Bangladesh: borsha (annual monsoon rains), bonna (irregular destructive floods in the wake of cyclones and storms) and jalabaddho (waterlogging). In a delta known for its heavy sedimentation, attributing all floods to rising sea levels due to melting ice caps caused by climate change is problematic. I use archival research and oral histories to highlight the ways in which embankments changed from ‘salinity-protection’ infrastructure during the East India Company deforestation of the Sundarbans to ‘flood-protection’ infrastructure from the 1850s onwards, and the ways they have contributed to damaging floods in the coastal zone. The framing of higher and wider embankments to protect Bangladesh against floods is unsustainable as it ignores the way these very embankments exacerbate siltation and increase the risk of damaging flood.