For a pathogen to survive it has to find new hosts to infect. This may sound simple, but if you consider the entangled mesh that is the biology of a host species you realize that there are plenty of ways that things can go wrong, stopping the chain of transmission. First of all, the harm the pathogen incurs on its host – the virulence – needs to be balanced between being too low – the infection will be cleared before any transmission opportunities have occurred – or too high, so that it causes the demise of the host before transmission can take place. Second, the pathogens must overcome the hurdles of moving from one host to the next, be it in water, air or through the bite of an arthropod vector. And third, it has to overcome the fact that most hosts are not sedentary, but move varying distances in response to changes in the environment they inhabit. Finally, it needs to be able infect the new host and evade the immune system to establish infection. Not an easy feat, but something that is happening all the time in the world of viruses, bacteria, fungi, parasites and their hosts.
In the avian influenza field, the realization of the importance of bird migration in the epidemiology has a long history but we haven’t really been able to address it in the required detail. Most studies have addressed the process at the population level, inferring movements either from ring recoveries or from virus phylogenetic perspectives. If you have followed what we do, it will not come as a surprise that we are interested in both influenza viruses and bird migration. A longstanding goal for us has been to integrate virology and movement ecology to better understand the epidemiology of avian pathogens. This is where it gets exciting, as the technology needed for these types of studies are available. Last year we deployed loggers on migrating mallards at our main study site, the Ottenby Bird Observatory on the island of Öland in SE Sweden, and followed them during migration as a part of the H2020 program DELTA-FLU.
We programmed the loggers to record GPS-positions in bursts, hoping to retrieve as much data as possible during active flight. From the flight data we extracted the metrics of flight: how does a mallard migrate – how fast, how high, and in which direction? And how do these parameters change during the flight? These metrics formed the basis for a Mallard Migration Simulator with which we could simulate different types of migrations, based on the normal flight behaviors of mallards.
The next step was to use the ring recoveries retrieved from the study site over the last 50 years to get realistic headings of migratory flights. Finally, we introduced individual-level epidemiological parameters from our study populations and built classical SIR-models. Combined, this allowed us to look at the likelihood that a bird that migrated was infected with low-pathogenic avian influenza, and that it maintained infection during migration, controlled by season, age of birds and other factors that could contribute. The resulting data can be transformed into a risk map for transmission.
I am very pleased with this approach, and think it is a novel way of analyzing this type of data. The next step, of course, is to consider such models for highly-pathogenic avian influenza viruses on larger spatial scales. We are collecting tracks of four species of ducks in different parts of Eurasia and hopefully we will be able to make realistic models of virus dissemination among migratory ducks in a flyway perspective.
Zoonotic Ecology and Epidemiology 