Ecological Fieldwork and Population Estimates

To estimate the population of a plant species, you need a quadrat, two tape measures, a random number generator and a way to record results, such as paper on a clipboard.

1. Measure the study area. Lay two tape measures at right angles along its edges to provide a coordinate system.
2. Generate two random numbers within the coordinate ranges of the area. These give the location of the sample.
3. Place the quadrat at that location using the same positioning rule each time.
4. Identify, count and record the individuals of the chosen species inside the quadrat.
5. Generate new coordinates and repeat. For example, take ten samples spread across the study area.

Choosing patches with lots of plants would overestimate the population; choosing mostly bare patches would underestimate it. Random coordinates reduce this selection bias. A larger number of samples usually gives a more representative estimate, especially when plants grow in patches.

Use the same quadrat size and counting method throughout. If comparing sites, also use the same number of samples and the same coordinate-selection procedure.

Scaling sample counts up to a population estimate

First calculate the mean number of organisms per quadrat:

$$\text{Mean per quadrat}=\frac{\text{Total number counted}}{\text{Number of quadrats sampled}}$$

Then work out how many quadrat-sized areas fit into the whole habitat:

$$\text{Estimated population}=\text{Mean per quadrat}\times\frac{\text{Total habitat area}}{\text{Area of one quadrat}}$$

The habitat area and quadrat area must use the same units. For example, a quadrat with sides of 25 cm has sides of 0.25 m, so its area is $0.25\times0.25=0.0625\text{ m}^2$.

Consider a teaching example in which five randomly placed 1 m² quadrats in a 500 m² field contain 5, 2, 6, 3 and 4 daisy plants. The total counted is 20, giving a mean of $20\div5=4$ plants per quadrat. There are $500\div1=500$ quadrat-sized areas in the field, so the estimated population is $4\times500=2000$ daisy plants.

This does not mean that exactly 2000 daisies have been counted. It assumes that the sampled mean is representative of the field as a whole.

Another equivalent approach is to calculate population density: divide the total count by the total sampled area to find plants per square metre, then multiply by the habitat area. This also works with raw counts from equal-sized quadrats along a belt transect. However, a transect deliberately crossing an environmental gradient may not represent the whole habitat, so its results should not automatically be scaled up to the entire area.

Investigating an environmental gradient with a belt transect

A belt transect samples a strip of habitat using quadrats along a line. Quadrats may be placed next to one another for continuous sampling, or at regular intervals for an interrupted belt transect. Regular spacing makes it possible to compare counts at known distances.

For example, investigate plant distribution from beneath a tree into open ground. Light intensity is the abiotic factor being studied. Begin with a testable hypothesis, such as: ‘The number of plants of this species increases as light intensity increases.’ This is a prediction to test, not a result to assume.

1. Lay a tape measure in a straight line from the tree into the open area.
2. Place the quadrat at 0 m, with a corner touching the mark.
3. Count the chosen plant species and measure light intensity using a light meter.
4. Repeat at regular distances, such as every 5 m, keeping the quadrat size and placement rule unchanged.
5. Repeat the investigation along comparable transects and calculate the mean count at each distance.

Record results in a table with columns for distance along the transect / m, number of plants and light intensity / lux. Measuring light matters: distance from a tree is not itself a measurement of light intensity.

Turning results into a conclusion

Plot distance along the transect on the horizontal axis and mean number of plants per quadrat on the vertical axis. This shows how abundance changes across the habitat. Comparing plant counts with measured light intensity helps assess the hypothesis.

For a concrete teaching example, quadrats placed every 10 m towards woodland contain 12, 10, 11, 8 and 6 poppy plants. The overall count falls from 12 to 6, although it rises slightly between 10 m and 20 m. Describe this as an overall decrease, not a decrease at every point. Increasing shade near the woodland could help explain fewer plants because less light is available for photosynthesis and growth.

A relationship between light and plant counts supports an explanation, but does not establish light as the only cause. Other conditions may also change along the transect. Repeating transects improves confidence that the pattern is not just a feature of one narrow strip.

Making fieldwork dependable and safe

Small or difficult-to-identify plants may be missed or counted as the wrong species. Use a consistent identification and counting method. Too few samples may fail to represent patchy vegetation, so increase the number of samples rather than relying on one unusually crowded quadrat.

Light readings can vary with time of day, and plant abundance can vary with season. Consider these when comparing results. Some plants can cause cuts or allergic reactions; wear gloves when handling them.

$$\text{Estimated population}=\text{Mean per quadrat}\times\frac{\text{Habitat area}}{\text{Quadrat area}}$$

Alternatively: total count ÷ total sampled area = population density; density × habitat area = estimated population.

Use matching area units. Scaling up assumes the samples represent the habitat.

Quality and interpretation

• Keep quadrat size, counting method and sampling effort consistent.
• More samples improve representativeness; repeated transects strengthen evidence for a pattern.
• Plot mean count against distance and compare the trend with the hypothesis and measured environmental factor.
• Consider missed plants, identification errors and other changing environmental factors.
• A transect across a gradient is not automatically representative of the whole habitat.