Sampling design diagrams
(scroll down to view all nine)
The following diagrams will give you some examples of what sampling designs look like. For use in a full study, these designs would be repeated at
one or more additional sites.
Symbols are as follows: irregular lines represent the boundaries of habitat types (such as a meadow, a particular type of forest, a particular type of wetland or pond, etc.)—they also represent the boundaries of a "field site"; squares or rectangles represent sampling plots; solid lines represent transects; dashed lines represent trails. Note that the scale varies considerably among the different diagrams. For clarity, the diagrams have been simplified—sampling designs for real field studies can get more complex.
Feel free to use one of these designs as a model for your study, but keep in mind that each organism, abiotic factor, habitat type, field site, etc. is unique. And most importantly, make sure the method you choose addresses your question or hypothesis! If you do use one of the following designs, you'll likely need to modify it to suit your needs.
Design for a small field site (roughly 15 m x 15 m). The four transects are 10 m long, and the sampling plot dimensions are 0.5 m x 0.5 m. This is an example of
with the transects spaced 4 m apart, and the sampling plots placed at 2 m intervals along the transects. This design would be most appropriate for very small organisms (such as soil or litter invertebrates or perhaps small forest herbs) or possibly sampling abiotic factors in the soil, leaf litter, or humus layer.
Design for a larger field site (roughly 30 m x 30 m). The three transects are 20 m long, and the sampling plot dimensions are 1 m x 1 m. This is another example of
with the transects spaced 10 m apart, and the sampling plots placed every other meter along the transects. It has the advantage of being easy to implement, but using more transects with fewer plots (such as five transects with six plots each) might do a better job of more evenly samping the field site. This design would be most appropriate for small organisms such as salamanders (depending on the habitat and type of salamander) or herbaceous plants. It also may be appropriate for insects, soil invertebrates, or abiotic factors in the soil, leaf litter, or humus layer.
Similar design to Figure 9b; again for a field site roughly 30 m x 30 m in size. The big difference between this design and the one shown in Figure 9b is the way plots and transects are located. In Figure 9b,
is used. The design shown here is based on
(a random sampling
design would look similar to this). Like in Figure 9b, there are a total of thirty 1 m x 1 m sampling plots. However, here the plots are placed along six 15 m transects located at different distances from the edge of the field site and with different spacing between the transects. The five sampling plots per transect are located at randomly selected intervals along the transect. Like in Figure 9b, this design would be most appropriate for small organisms such as salamanders (depending on the habitat and type of salamander) or herbaceous plants. It also may be appropriate for insects, soil invertebrates, or abiotic factors in the soil, leaf litter, or humus layer. The advantage of this design is that it should give a better
especially if there is some type of variability at the site not captured by uniform design (such as if there were particularly wet spots between the transects in Figure 9b). The disadvantage is that this design would take more time to implement in the field.
This design is based on the one used by Professors Koning and Singleton for their study of Glossy Buckthorn, an invasive shrub on the Franklin Pierce University Campus. The large plots are 20 m x 20 m and are designed for sampling forest trees; the medium sized plots are 5 m x 5 m and designed for sampling shrubs; the smallest plots are 1 m x 1 m and designed for sampling forest understory plants. Placement of the large plots is roughly haphazard and constrained by the edge with a different forest type on the other side of the stone wall (it's best not to sample too close to an edge unless you're focusing on edge habitat or edge effects) and changes in the forest type beyond the boundary of the field site. This diagram (and the next) illustrates that conditions at a particular field site may restrict your ability to implement a truly uniform or random sampling design. However, this design should do a good job of capturing an
unbiased representative sample
of the plant communities within the non-edge habitat of the post-agricultural forest. This type of nested plot design would be appropriate for studies aiming to sample all life forms of plants within a forest community.
This diagram is based on a study in Monteverde, Costa Rica that is designed to sample fruiting avocado trees and tropical bird communities. The diagram shows just three of forty five sampling plots distributed at fifteen different field sites. For consistency in the study, all sampling plots are 20 m x 50 m. At this particular site it was very difficult to fit three plots within the small remaining patch of rainforest. In addition, it was not possible to entirely exclude the trail from the sampling plots, but the amount of trail within plots was minimized. More generally, the use of 20 m x 50 m plots could be appropriate for sampling trees, some types of birds, and perhaps medium sized mammals depending on their density, mobility, and the sign they leave.
The field site shown in this diagram is relatively large, roughly eight hectares in size (1 ha = 100 m x 100 m). In this sampling design, a four hundred meter long section of trail running through the site is being used as a transect. This approach might be appropriate for doing bird or animal surveys, though it's important to consider whether sampling along a trail will result in a biased sample. Of course, the answer to this depends on whether the animals being studied are affected by the trail.
This diagram shows the same field site as that in Figure 9f. Here a four hundred meter long transect has been positioned to minimize sampling near the trail. In addition, the sample area extends 10 m out from both sides of the transect. This type of transect is often referred to as a belt transect. It is essentially the same as a 400 m x 20 m sampling plot. Typically the transect and ten meter boundaries along either side would be marked by temporary flagging tape. Clearly it would be impractical to try to use a giant pvc sampling plot! This type of design could be appropriate for sampling sign left by animals such as moose or deer, the presence of certain types of plants, insects or fungi, or for sampling other less common features of the habitat of interest.
This diagram shows a design intended to sample along an environmental gradient—in this case, distance from the edge of a vernal pool. In this design, distance from the edge of the pool may represent an independent variable (such as soil moisture) that could affect the organisms (or other variables) being sampled. In this design, plots are 1 m x 1 m and transects are 20 m long with ten plots per transect located every two meters. The size of plots and length and number of transects could be modified depending on the organisms or other factors being studied and the size and shape of the field site.
This diagram shows two transects extending at least ninety meters through a field site (the dashed lines represent the fact that the transects could be extended as far as the field site permits). Another unique feature of this design is that it does not rely on sampling plots, but rather uses
In some cases, transects with meter tapes can be used to sample features along the ground such as cover for salamanders or animal sign; wherever the meter tape crosses the feature of interest, that feature would be recorded (an example of the
In this diagram, a sample point is located every 15 m along the transect. These points represent specific spots where variables are measured (such as soil moisture, soil pH, light coming through the canopy, etc.). In addition, this is a great approach for using the
for sampling forest trees. (This method can also be modified to sample other features, such as small mammal burrows or animal sign.) In some cases, you may be able to avoid laying out a meter tape by walking a straight line through a habitat along a compass bearing. Under this scenario, distance is marked by walking paces. If you use this approach, be sure you practice walking a consistent pace distance, and measure how long your pace is that is so you can report the distance sampled!
Copyright 2008-2019, ecologyandevolution.org