Sitemap
Print
RSS

Putting dynamic models of migration to the test

Workshop Dynashop report June 2001   

By Marcel Klaassen (Netherlands Institute of Ecology)
Bruno J. Ens (Alterra, Centre for the Green Environment)


Is it possible to give an advance prediction of the optimum route along which birds must migrate? And is it then also possible to predict the effects of climate change on that route and the consequences for the survival of the migratory birds? Theoretical biologists, glued to their computer screens, believe that it is. Field biologists, hardened by the rain and wind, believe it is not.

Top people from both camps met on the island of Texel from Thursday 27 February to Saturday 1 March 1997 during 'DYNASHOP': an international workshop on 'Putting dynamic models of migration to the test'. Support by the National Research Programme on Global Air Pollution and Climate Change made it possible to organise the workshop.  

Input parameters

By making use of the DYNAMIG computer package the field biologists were able to do their homework in advance trying to apply the stochastic dynamic program (see box) for the bird migratory systems that they themselves investigated. The field biologists that were invited had all been intensively researching the migration of a particular species of bird for many years. The species studied, in order of decreasing size, are given in Table 1. The theoretical biologists are pioneers in stochastic dynamic programming (SDP), the mathematical technique used to solve the complex optimisation problem of how the birds should migrate.

Satisfactory climate scenarios

For those species for which the model did appear to be applicable, one could investigate where the most important bottlenecks lie during the migration. Furthermore, the possible consequences of climate change could also be explored. For a number of species it appeared that the wind conditions during the migration were very important. A little extra headwind sometimes makes it impossible for the relevant species to reach the breeding grounds in time. This was a surprising discovery and it emphasises the importance of satisfactory climate scenarios that can make accurate predictions of wind conditions.

Table 1: Bird species (in descending order of size) investigated
by DYNASHOP participants

English name

Latin name

Bewick's Swan

Cygnus bewickii

Greylag Goose

Anser anser

Pinkfooted Goose

Anser brachyrhynchus

Barnacle Goose

Branta leucopsis

Brent Goose

Branta bernicla

Avocet

Recurvirostra avosetta

Knot

Calidris canutus

Pectoral Sandpiper

Calidris melanotus

Sanderling

Calidris alba

Western Sandpiper

Calidris alaskensis

Other simulations investigated the consequences of the unavailability of specific stopover locations, or the consequences of a shift in the times at which stopover locations become available, both as a result of climate change. Consider the growth of food plants and the melting of ice. This was investigated for the barnacle goose as an example. The model predicted that an earlier growth of food plants would allow the birds to arrive at the breeding grounds earlier, and with an increased body mass, the result being more successful breeding. The computations make it clear that satisfactory, precise climate scenarios are required, which can predict changes of temperature and other factors, like wind velocity, at all locations.

 

Stochastic Dynamic Programming (SDP)

In SDP you can distinguish three distinct procedures.

  1. Definition of a terminal reward. In the case of most DYNAMIG applications in this report this is the number of offspring you may expect a bird to produce in relation to its body condition and date of arrival at the breeding grounds.
  1. Calculation of decision matrices (also called policy diagrams) starting from the final time step and looking backward in time. In the case of most DYNAMIG applications two-dimensional matrices are constructed for each staging site during migration. Each site-specific decision matrix provides optimal decisions in relation to date and the bird's condition.
  1. Forward simulation. A number of birds is sent on migration with these decision matrices as a travel guide. In this step the full impact of the various parameter settings and their stochasticity becomes evident.

References

Ens, B. 1996. Climate change: consequences for migratory birds. Change 32: 12-14.

 

Downloads 

You are welcome to use our computer package now DYNAMIG computer package (Dnmgarch.ex 91KB)

The computer package is compressed in a selfextracting archive file.

Save the archive file in a appropriate directory and start Dnmgarch.exe in Windows

 

The DYNAMIG REPORT 'Linking dynamic migration models to the real world'  (report_dynashop.pdf 4,479KB) is attached.

 

AttachmentSize
report_dynashop.pdf4.37 MB
Dnmgarch.exe90.93 KB