Metapopulations EcoBeaker Lab
Sarah Cates and Scott Sylvester
A "metapopulation" has been defined as a group of populations which are connected to each other through dispersal. Dispersal within the smaller populations is more common than dispersal among them.1 In a broad sense, many animal populations, including most small-mammal populations, are probably structured this way.2 The number, size, and connectedness (amount of dispersal) of the small populations within the larger metapopulation are all at least partially dependent on features of the landscape. A "heterogenous" landscape is one with variation in features such as elevation and/or vegetation. Some parts of a heterogenous landscape provide better habitats than others for any given species, whose individuals tend to be concentrated in those "patches" of preferable habitat. Because a metapopulation is spread over many different patches, a catastrophic disturbance, such as a quickly-spreading disease, a fire, or a bulldozer, which kills many, or even all, of the individuals in one or a few patches, has less chance of permanently affecting the metapopulation as a whole. Accordingly, metapopulations can persist in the face of disturbances which might otherwise wipe out a less patchily-distributed population. However, if the frequency of disturbance becomes too high, it is possible for the metapopulation to experience large variations in size, or even become extinct. In this lab, you will look at how population size and persistence is affected by three aspects of metapopulations: the level of extinction of individual patch populations, the amount of dispersal among patches, and the number and spatial arrangement of the patches.
The lab consists of five different arrangements of habitat patches which are occupied by a species of small mammal. The populations in the individual patches are subject to periodic catastrophic disturbances which wipe out the entire population within those patches. You can control the number and frequency of those disturbances. You are also able to change the dispersal ability of the mammals, and so control how quickly they can re-colonize a patch after a disturbance event. You will be examining how changing these parameters affects the size and persistence of the mammal metapopulations over time.
Start EcoBeaker and open the situation file called "Metapopulation". A number of windows will be opened once the situation file is loaded. The main window (upper left) will show the location and arrangement of the habitat patches, and will provide visual feedback about the location, growth and movement of the mammal species ("rats"), which are represented by blue dots which move over the landscape. (Donít worry if they window is blank now, you will see the landscape and the rats once you start the simulation) The rats depend on a plant as their food source, which is represented by green dots. The plants only grow in the patches of preferable habitat. Rats can live in both the patches or in the spaces between them, but each rat can only store a finite amount of energy (fat) and must eat on a regular basis to survive, move, and reproduce. Because their food does not grow between patches, rats can not survive in these regions indefinitely. They must move to a patch with food before their energy stores run out or they will die.
1. The main window should be labeled "single patch". If it is not, use your mouse to open the pull-down menu called "Markers" and select the marker called "single patch". The small window on the upper right is the control panel where you can start, stop, and reset each simulation. Use your mouse to press the "go" button in this window.
2. You should see a single orange square in the main window. This represents a single large patch. The population of small mammals living in this habitat is not a metapopulation and this simulation is meant only as a reference, since this type of population structure is probably very rare.2
3. After starting the simulation, you should see green dots (the plants) filling in the orange square and at a certain time step, rats (blue dots) are released into the patch. (Sometimes the rats fail to get established. If this happens, press "reset" and then "go" to start the simulation over again.) The rats will move around the patch, eating the plants and reproducing once they have enough energy. Under the main window are two graphs which show you the population size of the rats over time.
4. Soon, you will see red dots moving across the patch. These dots represent a catastrophic disturbance (a "plague"). They will kill all the rats in the patch, but do not affect the food plant.
Now you will examine two heterogenous landscapes, one with a few large patches of preferable habitat and one with many, smaller patches. This type of population structure is very common in small mammals2 and is also called a "source-sink" population, because the "overflow" of individuals from the preferable "source" patches disperses into the less-preferable "sink" between the patches. A well-studied example of this sort of population structure is that of voles (Microtus spp.) in North American grassland, where patch quality is determined by the food-quality of the plants growing there.
1. Open the "Markers" pull-down menu and select "Patches".
2. Start the simulation by pressing the "go" button in the control window.
3. You should see four square patches, growing the food plant. When the rats are released, they will begin to eat and reproduce and you will see some of them disperse between patches. Periodically, the plague will wipe out all the rats in one of the habitats. (You will also see yellow dots in the spaces between the patches. These dots behave the same way as the plague does within the patches.) After the last rat in a patch has died the plague in that patch dissipates (because it has no more rats to live in) and the patch can be recolonized.
4. Now, you will change the frequency of disturbances. You do this using the controls in the window called "Disturbance and Dispersal", which should be located below the main control window. Click in the box next to the words "time between disturbances" and change the value from "50" to "10". Then click the "change" button at the bottom of the window. How does this change the size of the rat population? How does it look over time? (You can stop the simulation using the main control window and then click on the "graph" window beneath the main window. There is a scroll bar beneath the graph which allows you to examine the rat population vs. time plot over the whole course of the simulation. You can also scroll backwards within the "rat population" window. You can also reset the simulation at any point, using the "reset" button in the main control window). Experiment with different values of "time between disturbances". How small can you make this number without all the rats going extinct within 100 time steps? How quickly does the population go extinct if there is only one timestep between disturbances? Change the number back to 50 before you go on to the next part of the exercise.
5. Now, you will change the dispersal ability of the rats. You do this by clicking on the "habitat numbers" button next to the words "dispersal ability" in the "Disturbance and Dispersal" window. Another window will pop up. In that window, change the number in the box next to the words "open field" and then click the "OK" button. You are changing the speed of the rats in the "sink" habitat, so a higher number means they have a better ability to disperse. Experiment with different dispersal abilities. How does increasing or decreasing the dispersal ability change the shape of the population vs. time graph? How does it affect the extinction probability of the rats? Can you balance out the effects of a higher frequency of disturbance by making the rats better dispersers? Can you set it up so that the rats can survive a disturbance interval of only one time step?
6. Now, you will examine how changing the same parameters (disturbance frequency and dispersal ability) affects a population of rats in a very patchy landscape. To do this, open the "Markers" pull-down menu and select "Many patches". Press the "go" button to start the simulation.
7. This habitat is divided into 49 patches. Now two patches will be hit by the plague at every time step. In this simulation, You can control the number of patches disturbed as well as the frequency of disturbance and the the dispersal ability of the rats.
8. Experiment with the different parameters to see how each one affects the population size and the shape of the population vs. time graph. Only change one parameter at a time, so that you can separate its effects from the other two.
9. Because there is more habitat in this simulation, the population size of the rats will be larger than in the 4-patch habitat, but you can compare the shapes of the population vs. time graphs between the two simulations. In the many-patches simulation is it easier or harder to change the parameters so that the rat population is steady over time? What values cause the population to go extinct?
Now you will examine a landscape in which one patch is invulnerable to extinction. This type of population structure is called the "mainland-island" model. It not only applies to actual mainland and island populations, but is also seen in habitats which contain one very large patch and many smaller "satellite" patches. An example (from reference 2) is from a study of a population of pikas (Ochotona princeps; small mammals related to rabbits) living on abandoned mine tailings (basically piles of rocks) in the Sierra Nevada mountains. Two very large tailings were always occupied, but smaller tailings around the larger ones were only occupied during some of the population surveys. Tailings which were over 300 m. from the large piles were almost never occupied.
1. Open the "Markers" pull-down menu and select "Mainland Islands". This simulation contains a larger, central "Mainland" patch (red) surrounded by smaller orange habitat islands. The island patches are connected to the mainland by dispersal "corridors". In this simulation, the rats can only disperse to the islands along the corridors.
2. Press go and begin the simulation. Notice which islands are usually occupied and which are almost never occupied. You can change the parameters within this simulation using the same controls as in the previous 2. When you change the dispersal ability of the rats in this simulation, change the speed within the "corridors".
Subpopulations connected by dispersal corridors
The final simulation you will examine represents several supopulations connected by dispersal corridors. This population structure is similar to the island-mainland model, except that there is no "mainland" population which is invulnerable to extinction. An example of this type of population structure is chipmunks (Tamias striatus) in patches of woods separated by fields. The chipmunks do not disperse through the fields, but only along fences which connect some of the patches of woods.2
1. Open the "Markers" pull-down menu and select "subpop.wcorridors". In this simulation, there are 9 habitat patches connected by corridors. Again, you can change the disturbance and dispersal parameters and examine their effects on the rat population over time. Notice that some patches are connected to more than one other patch, while others are only connected to one. How does this affect how often the patches are inhabited?
2. What happens when you change the dispersal ability in the corridors to 0? This sort of situation occurs when habitats that used to be at least marginally supportive to an animal ("sinks") become uninhabitable. This can happen, for example, when high-elevation areas become surrounded by water after a dam backs up a river or when human development destroys all but certain parts of a populationís range. Studying metapopulation dynamics may help conservationists design wildlife preserves more effectively by including corridors for dispersal, or at least by providing more than one habitat patch as insurance against catastrophic disturbance.
1 Hanski, I. and Gilpin, M. (1991) Metapopulation dynamics: concepts, models and observations, Biological Journal of the Linnean Society 42, 89-103
2 Krohne, D.T. (1997) Dynamics of metapopulations of small mammals, Journal of Mammalogy 78, 1014-1026