The complex network of tidal waterways, mudflats and small islands that make up the Sundarbans mangrove forest of South Asia is in constant flux. And so are its people.
The Sundarbans stretch across the border between India and Bangladesh, but many of the largely poor families who live there treat it as one ecosystem. When their homes come under threat from erosion and rising seas, or are flooded by storm surges, they often move to another island and sometimes, in doing so, to another country.
On the Indian side, there are two main types of migrant, writes Sahana Bose, assistant professor at Karnataka's Manipal University, in the latest issue of the journal Forced Migration Review (FMR). You'll find Indian Sundarbans dwellers shifting from one island to another, alongside rural Bangladeshis infiltrating the porous border, who are recognised neither by their own government as Bangladeshi citizens nor by India as "climate refugees".
Dhaka doesn't try to stop them, and neither does New Delhi, happy to have a supply of cheap labour. But Bangladesh won't take back those identified as illegal migrants, and people smuggling networks are "flourishing" on both sides of the border, Bose says.
How do you put a name to this kind of migration, and what do you do about it? These are the kind of questions the FMR issue on "crisis migration" tries to address – and finds no easy answers.
A key argument made by the authors is that it is often counter-productive to try to put migrants in separate boxes. When a boatload of exhausted Africans, from different countries, reaches the shores of Europe, they are likely there for a whole bunch of reasons - escaping war, political persecution, drought, hunger, grinding poverty, or a combination.
One thing they do have in common is that they are in crisis and doing what they can to survive, just like the people of the Sundarbans who island-hop in a desperate effort to stay afloat both literally and metaphorically.
WHO ARE CRISIS MIGRANTS?
The FMR defines "crisis migrant" as "a descriptive term for all those who move, including those who require relocation in the context of humanitarian crises". The concept, it adds, "reflects the endless historical reality and significance of movement as a crucial response to crises."
Crisis migration covers people displaced by events beyond their control, such as conflicts or earthquakes, people who leave in anticipation of disasters or threats, and people who are relocated because they might otherwise be trapped.
The reasons why people migrate should not of course be disregarded, says Susan Martin, director of the Crisis Migration Project at the Georgetown University Institute for the Study of International Migration, not least because they help determine whether the move will be permanent or temporary, and what the options are for assistance. But causes should only be one part of the decision-making process, she argues.
"What we're suggesting is that there's an umbrella...under which a lot of these causes can be placed, and which can be a starting point for thinking through what are the questions you ask," she tells me in an interview.
It is Martin who coined the term "crisis migration" a few years ago, when the Intergovernmental Consultations on Migration, Asylum and Refugees (IGC) decided to look at the migration consequences of humanitarian crises. It caught on as a more useful lens for looking at a growing problem than common epithets like "refugees" and "internally displaced persons".
CRISIS CRYSTAL BALL
One very strong argument for adopting a broader approach is the emergence of new types of migrant in a fast-evolving world - whether due to climate change, nuclear accidents like Fukushima after the Japanese tsunami of 2011, conflicts and disasters that affect illegal or poor foreign workers as in the Libya uprising, or the explosion of gang violence in Central America.
"If you try to think too far ahead with regard to what are the causes of the crises, then you are going to continue to have gaps because some unanticipated crisis happens, and then what do we do? Whereas if you have a framework that can cross crises, you've got a better shot at responding effectively when something new emerges," explains Martin.
The big question is what that framework should be. Recognising the gaps in existing legislation and policy is a first step to finding the answer, and there's been progress here.
Processes have been launched to work out how to plug holes in the system. For example, the Nansen Initiative, launched in October 2012, aims to support policy makers in deciding how to respond to the protection needs of people displaced across borders by disasters, including the effects of climate change.
The United States and the Philippines have set up a working group, involving a handful of other governments, to explore and produce guidance on how best to help migrants who find themselves in a foreign country in crisis.
Peter D. Sutherland, the U.N. Secretary-General's Special Representative for International Migration and Development, notes in a foreword to the FMR that despite shortcomings in laws, coherence and resources, "when it comes to protecting migrants’ well-being and rights, smart practices abound".
The FMR includes articles on an Eskimo community in western Alaska and Carteret islanders in Papua New Guinea who, at risk from erosion, have organised their own relocation to islands in the area. It also describes how the International Organization for Migration has trained tens of thousands of immigration and border officials around the world in human rights and refugee law, trafficking in persons and freedom of movement.
Then there's the example of how a 6,000-strong community of Vietnamese residents in eastern New Orleans moved back in and rebuilt their own community after Hurricane Katrina, ignoring city council orders that their parish be turned into a non-residential green zone.
At the same time, the FMR documents some of the many cases where people who have moved or cannot move due to crisis have fallen through the protection cracks - from floods in Pakistan and Colombia, to West African adolescents forced by food crises to seek work away from their families, and the growing number of Mexicans who are leaving home out of fear of violence linked to organised crime.
NEW CONVENTIONS NOT NEEDED?
Martin says there is "a fair amount of good practice" in dealing with crisis migration, but it is mainly ad hoc, and there has been relatively little exchange so far. But it's premature to try to come up with a new international legal treaty like the 1951 Refugee Convention, she emphasises.
"You can't just do a convention and hope governments will come and adopt it. You really have to build up a body of practice, so that once you go more formalised, governments are comfortable with the responsibilities," she says. "It's a process of consensus-building, and if you are really successful at that, you may not even need a convention, because it becomes second nature."
So we're likely to see more non-binding codes like the "Guiding Principles on Internal Displacement", which were endorsed by 193 heads of state in 2005 and have filled a major gap in the international protection system for people uprooted inside their own countries.
And perhaps the rise in extreme weather events hitting richer nations, like the floods swamping southwest England and the destruction wrought by Superstorm Sandy in New York City, will serve as a wake-up call that they urgently need to get to grips with crisis migration.
"If New York is going to be affected in this way, what is Dhaka going to do?" asks Martin.
A book edited by the Georgetown University Crisis Migration Project, "Humanitarian Crises and Migration: Causes, Consequences and Responses", will be published by Routledge in March.