The duck, the seed, the distance – mallards as seed dispersers

A seed disperser in action (Flickr, USFSW Mountain Praire CC BY 2.0)

(This post is by Mariëlle van Toor, on Twitter as @mlvantoor)

Most of the stories on this blog somehow relate to birds (particularly ducks) and pathogens (most often avian influenza viruses, or AIV). Throughout the years, we have established that ducks are quite good at transporting AIV during their daily movements, and even migration. But viruses are not the only thing that ducks are able transport! Specialist wetland plants, which inhabit discrete habitats that are often not connected by waterways, have usually very little opportunity for (long-distance) dispersal that would allow them to colonise new wetlands. Dabbling ducks such as the mallards forage on all kinds of seeds, which they crush in a specialised organ called gizzard to make them easier to digest. However, seeds often escape from the gizzard unharmed. These seeds will be excreted by the ducks rather than digested, and can grow into mature plants. By transporting the seeds of wetland plants in this way, ducks could fill an ecologically important role.

But just how much could migratory ducks such as mallards contribute to seed dispersal? This was the question that our collaborator Erik Kleyheeg, then a PostDoc at the Max Planck Institute for Ornithology, approached us with. Erik (now also known as the “Teal man”) had already accumulated a wealth of knowledge on mallard seed dispersal during his PhD, and had collected spring migratory tracks of mallards wintering at Lake Constance. Combined with the extensive ring recovery data set from the same population of mallards, the duration of seed passage through the gut, and the models we previously applied for our virus dispersal paper (see post here), we developed our “comprehensive mallard seed dispersal model” that is now published in Frontiers in Ecology and Evolution.

But what does comprehensive mean in terms of mallard seed dispersal? For that we need to know what previous attempts at modelling seed dispersal by mallards during migration looked like. Often, previous studies have estimated seed dispersal distances during bird migration by multiplying distributions of gut retention time with the flight speed of birds. But neither do ducks fly in a straight line, nor do they fly uninterruptedly between their wintering and breeding areas. We thus need to account for staging behaviour and specifics of the migratory flight, and consider dispersal during the different stages of migration.

The black line shows the curve describing the retention of seeds in a mallard’s gut. Most seeds will be excreted soon after ingestion (depending on seed size), but the long tail could enable ducks to transport seeds over long distances both during migratory legs and staging periods.

Our model includes both seed dispersal during the actual migratory flight, and during the time that ducks are staying at stopover sites.

Another important aspect is that we need to account for when individuals stop foraging. Do they eat seeds right up until the start of a migratory flight, or do individuals fast prior to migration to avoid carrying around any extra weight? As the answer to this question is not satisfactorily resolved yet, we decided that our model should be able to account for fasting time. For our paper, we considered three scenarios – no fasting, short fasting (1 hour prior to migration), and long fasting (12 hours prior to migration) – to understand how fasting would affect seed dispersal. As soon as we better understand the behaviour of pre-migratory fasting in ducks, however, we can feed any number into the model for more specific (and realistic) predictions.

The duration of pre-migratory fasting affects potential dispersal distances quite substantially. Above you can see how short and long retention times (for small and large seeds, respectively) are affected by fasting duration.

Finally, all seed dispersal is to no end if seeds are not transported to habitat for suitable for germination. For wetland plants, that obviously means transport to other wetlands. While all other components of the seed dispersal kernel derived from our model should be transferable to other mallard populations, the availability of wetlands along the migratory corridor were specific to this population. We calculated the probability of dispersed seeds to end up in wetland habitat for both staging mallards, which can be found in wetlands most of the time, and for migratory individuals, for which we used the Global Lakes and Wetlands Database (GLWD). Our summary map for wetland availability already shows that it is not very likely for a seed dispersed by an actively migrating mallard to be deposited in a wetland:

Most migratory mallards wintering at Lake Constance migrate along a N-E corridor. The highest concentration of wetland area lie towards the East and North East.

All this taken together, the question is: How good are mallards as seed dispersers? You won’t like the answer: it depends! Theoretically, mallards are amazing at transporting seeds – if they don’t fast before starting migration. If mallards foraged right up until migration, they could transport a large part of the swallowed seeds over hundreds of kilometers, and many of them would end up in suitable habitat during the first stopover period. But the duration of pre-migratory fasting hugely influences dispersal distances, and in the 12 hour scenario, many seeds were already excreted prior to migration. But even then, some seeds can be transported over exceptional distances.

These are the final seed dispersal kernels predicted from our model, shown for short and long retention times, and three different scenarios of pre-migratory fasting. The upper row shows the general probability of seeds being dispersed over distances up to 950 km (the general prediction), whereas the lower row shows the probability of seeds being dispersed into suitable habitat (the prediction specific to Lake Constance mallards).

In conclusion, migratory ducks such as mallards likely play an important role for the short- and long-distance dispersal of wetland plants, both during periods of migration and residency. But, as always, there are more questions to be answered before we can say for sure.

Link to the paper:

Kleyheeg, E., Fiedler, W., Safi, K., Waldenström, J. & Wikelski, M. & van Toor, L. M. 2019. A comprehensive model for the quantitative estimation of seed dispersal by migratory mallards. 2019. Frontiers in Ecology and Evolution 7:40 [10.3389/fevo.2019.00040]

As the duck flies: Avian influenza virus and migratory mallards

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.

Link to the paper:

van Toor, M.L., Avril, A., Wu, G., Holan, S.H. & Waldenström, J. 2018. As the duck flies – estimating the dispersal of low-pathogenic avian influenza viruses by migrating mallards. Frontiers in Ecology and Evolution 6:208. doi: 10.3389/fevo.2018.00208

Perdeck revisited – or how well does a Mallard know its way?

By Jonas Waldenström

At this time of the year the air is full of migrating birds. Some, as cranes or geese with their conspicuous formations are easily spotted with the naked eye, while other birds, including most smaller songbirds, fly at altitudes where you need a scope to see them. But you can often hear them; each species has its own tune, and an experienced ear can tell them apart on call alone.

The question “how do they find their way” is as old as the field of ornithology itself. Generally, migration wouldn’t be possible without some sort of compass; a way of telling the bird in which direction to move. It has been shown that birds may use the sun, the stars, and the earth’s magnetic field for assessing their heading. And in some species also visible cues, a sort of map sense from previous travels, or even olfactory cues (a posh word for smelling where home is). As the vast majority of birds migrate without the guidance of their parents (which seems reserved to some flock-living species), a juvenile bird must be born with not only the tools to assess where it is, but also a sense of where it should go.

One of the pioneering fathers of ornithology was the Dutch professor Albert Christiaan Perdeck. He made one of the first real tests on how birds can sense where they are going, and how they can adjust the course if they get out of track. In order to test this he wanted to do a displacement study, where birds should be experimentally transported to a novel site, far from the catching site. As this study was conducted in the 1950s, in the pre-gadget era of ornithology, he needed a species that he could catch in large quantities, and where ring recovery data could be collected. His choice of study animal was the European Starling Sturnus vulgaris, a common farmland bird in most of Northern Europe. Starlings in autumn can aggregate in huge flocks, sometimes consisting of several thousand individuals, and was thus a good target species for Perdeck.

With a remarkable enthusiasm, the team caught and ringed thousands of starlings. Some were released at the ringing site in the Hague, while the other half were transported with airplanes to Switzerland and released. After some time the ring recoveries started to come in, and the results were extremely interesting. It seemed as the young starlings had a vector compass, as the birds that were transported south stayed on the same heading as they had when they were caught. But instead of ending up in Holland, the young starlings ended up way south, sometimes even on the Iberian peninsula. I wrote ‘young’ deliberately, as there was a clear age effect. Where the juvenile birds continued on the same vector, the adult starlings compensated for the displacement, changed course and headed to the original winter quarters. Adult birds are more experienced, and in the starling case they were able to adjust to the circumstances and get back on the right track. A quite remarkable feat – some of my colleagues cant find their way to the university canteen without a helper…

Spurred by the old studies (classics, you could say) and the advancement of new tracking tools we conducted a similar experiment with Mallards. The study was a collaborative effort with scientists from Sweden, Germany, the UK and Denmark (with the lead from Professor Martin Wikelskii at the Max Plank Institute for Ornithology, in Constance, Germany). Today’s gadgets can do stuff Perdeck could only dream about. During two autumn seasons, we caught juvenile Mallard females at Ottenby – our beloved duck field site – and equipped a total of 76 birds with satellite GPS transmitters. Half of the ducks were released at Ottenby, and the other half were transported in a private airplane to Lake Constance in southern Germany and released there. The tags had solar panels and, in the best of circumstances, had the potential to send data for at least two years; providing highly accurate GPS fixes at several times a day.journal.pone.0072629.g002

However, the best of circumstances is not often met in nature. The tags on the birds in Ottenby had problems with the lack of sunshine during Swedish late autumn and winter, and many of them just went offline. But a fair number of tags delivered data on movements both in autumn/winter and in spring, when birds headed to their breeding grounds. Contrary to the Perdeck’s starlings, our displaced Mallards did not continue migration in autumn; they stayed in the Lake Constance region. Of the Mallards released at Ottenby, some continued migration to the general wintering area of our study population, that is Denmark and Germany, south to The Netherlands.

After the winter: “most of the translocated ducks headed straight north-north-east, as if heading towards Ottenby, with one duck going as far as northern Sweden. Three of the transported ducks, however, first headed in a more easterly direction and turned northwards when reaching the longitudes of the area the control birds migrated to. It is unclear how these birds decided when to turn north, but the movement trajectories could be interpreted as if individuals had noticed that they were in the wrong place and then corrected for the southward translocation. Based on the observation that this second group of transported ducks ended up in their potential natural breeding grounds, and the first group had a more northerly heading than the control group, we conclude that mallards, just like the starlings from Perdeck’s original experiment, can correct for translocation during the spring season following the experiment.journal.pone.0072629.g004

Thus, there was quite large differences between individuals in the translocated group, from those that seemed to take the shortest route north to Ottenby in spring, to those that followed a eastern direction (and then going north), more in the direction of what they should have had if the stayed in the normal wintering grounds: a flexibility in continental navigation and migration.

The article is open access and can be found here.