Sampling Techniques: Quadrats and Transects

An ecosystem is the interaction of a community of living organisms with the non-living (abiotic) parts of their environment. To study one, ecologists describe it at increasing scales of organisation.

Level	Meaning	Example
Individual	A single organism	One oak tree
Population	All the organisms of one species in a habitat	All oak trees in a wood
Community	All the populations of different species living together	Oaks, ferns, beetles, fungi
Ecosystem	The community plus the abiotic environment	The whole woodland

Within a community, species depend on one another for food, shelter, pollination and seed dispersal. This is called interdependence: removing one species can affect many others.

Organisms also compete for resources that are in limited supply. Plants compete for light, water, space and mineral ions; animals compete for food, mates and territory. Sampling lets us measure how many organisms there are and where they live — the first step in understanding these relationships.

A quadrat is a square frame of known area (often $1{\text{ m}}^2$ or $0.25{\text{ m}}^2$) used to sample the organisms inside it. Counting every organism in a large habitat is impractical, so we sample a fraction of it and scale up.

For the sample to represent the whole habitat fairly, quadrats must be placed randomly. Choosing "good-looking" spots introduces bias — for example, deliberately placing quadrats where daisies are dense overestimates the population.

Random sampling means every part of the habitat has an equal chance of being sampled. This removes bias and makes the estimate representative.

To place quadrats randomly:

1. Lay two tape measures at right angles along two edges of the area to form a grid of coordinates.
2. Generate pairs of random numbers (from a calculator, random number table, or app).
3. Use each pair as $(x,y)$ coordinates and place a quadrat at that point.
4. Count the organisms of the target species inside each quadrat.

Quadrats are best for organisms that do not move (plants) or move very little (e.g. limpets, barnacles).

Once you have counts from several random quadrats, find the mean number per quadrat, then scale up to the total area.

$\text{Estimated total}=\frac{\text{total count in all quadrats}}{\text{number of quadrats}}\times \frac{\text{total habitat area}}{\text{area of one quadrat}}$

Worked example 1 — A field is $200{\text{ m}}^2$. A student places ten $1{\text{ m}}^2$ quadrats at random and counts daisies:

Quadrat	1	2	3	4	5	6	7	8	9	10
Daisies	4	7	5	9	6	3	8	5	6	7

Step 1 — total count: $4+7+5+9+6+3+8+5+6+7=60$.

Step 2 — mean per quadrat: $\frac{60}{10}=6$ daisies per $1{\text{ m}}^2$.

Step 3 — scale up. Each quadrat covers $1{\text{ m}}^2$, so $200{\text{ m}}^2$ contains $\frac{200}{1}=200$ quadrat-areas:

$6\times 200=1200\text{ daisies (estimated)}$

The more quadrats you sample, the closer the mean gets to the true value and the more reliable the estimate. A single quadrat is never enough.

The spec requires you to understand and calculate the mean, and to know what median and mode mean. They are all averages but answer different questions.

Average	What it is	How to find it
Mean	The arithmetic average	Add all values, divide by how many there are
Median	The middle value when data is in order	Sort the values; take the centre one
Mode	The most frequent value	The value that appears most often

Worked example 2 — Counts of buttercups in seven quadrats: $3,5,5,6,8,5,10$.

Mean: $\frac{3+5+5+6+8+5+10}{7}=\frac{42}{7}=6$ buttercups per quadrat.

Median: put the values in order: $3,5,5,5,6,8,10$. With seven values the median is the 4th (middle) value $=5$.

Mode: the value $5$ appears three times — more than any other — so the mode $=5$.

When data has a few very high or very low outliers, the mean can be misleading. The median is less affected by extreme values.

A transect is a line laid across a habitat along which quadrats are placed at regular intervals. Transects answer a different question from random quadrats: not "how many in total?" but "how does distribution change across an environmental gradient?"

A gradient is a gradual change in an abiotic factor — for example light decreasing from open ground into woodland, or moisture and salt changing from the sea up a beach.

Use random quadrats to estimate abundance or population size. Use a transect (a line of quadrats at set intervals) to study how distribution changes along a gradient. The transect line itself is not placed randomly — it is laid out along the factor you are investigating.

Method (belt transect):

1. Lay a tape measure in a straight line from one end of the gradient to the other.
2. Place a quadrat every fixed interval (e.g. every $2\text{ m}$).
3. In each quadrat, count the target species (or estimate percentage cover for plants that are hard to count individually).
4. Record the abiotic factor at each point (e.g. light level with a light meter).

Required practical 9 asks you to measure a population's size and use sampling to investigate the effect of a factor on distribution. A typical write-up records species abundance against distance, then looks for a pattern linked to the abiotic factor.

Worked example 3 — A belt transect runs from the bottom of a sand dune (wet, sheltered) up to the top (dry, exposed). Percentage cover of marram grass is recorded:

Distance up dune (m)	0	2	4	6	8	10
Marram grass cover %	5	15	35	60	80	90

Reading the pattern: marram cover rises steadily as distance up the dune increases. Marram is adapted to dry, exposed conditions, so it is most abundant near the top and scarce at the damp, sheltered base.

The transect shows a clear correlation between the abiotic gradient (dryness and exposure) and the distribution of the species. To draw conclusions about the whole dune you would repeat the transect in several parallel lines and average the results, reducing the effect of anomalies.

A transect shows where organisms live along a gradient. It is the wrong tool for estimating a total population — for that you need randomly placed quadrats and the mean-and-scale-up method.

In a stable community, the numbers of predators and their prey rise and fall in linked cycles.

• When prey are plentiful, predators have lots of food, so the predator population grows.
• More predators eat more prey, so the prey population falls.
• With less food, predators starve and their numbers fall.
• Fewer predators means more prey survive, so the prey population recovers — and the cycle repeats.

Two features you must be able to read from a predator-prey graph:

1. The two curves are cyclical — they repeat up-and-down over time.
2. The peaks are out of phase: the predator peak comes shortly after the prey peak, because predators only increase once there is already extra food.

(Extra context — the classic data set is the Canadian lynx and snowshoe hare from Hudson's Bay Company fur records. Naming it is not required by AQA 8461.)

1. Confusing quadrats with transects

Random quadrats estimate abundance or total population; transects measure distribution along a gradient. Pick the method that matches what the question is investigating.

2. Placing quadrats non-randomly for an abundance estimate

For a population estimate, quadrats must be placed using random coordinates. Choosing where to put them introduces bias and invalidates the scale-up. (The transect line is the one case where placement is deliberate, not random.)

3. Forgetting to scale up — or scaling up wrongly

The mean is per quadrat, not the answer. Multiply the mean by how many quadrat-areas fit into the total area: $\text{mean}\times \frac{\text{total area}}{\text{quadrat area}}$. In worked example 1 that was $6\times 200=1200$, not $6\times 10=60$.

4. Mixing up mean, median and mode

Mean = add and divide. Median = middle of the ordered list. Mode = most common value. Read which one the question asks for.

5. Getting the predator-prey peaks the wrong way round

The predator peak follows the prey peak, not the other way round. Predators can only increase after there is extra prey to eat, so their curve lags behind.

6. Using too few quadrats

A reliable estimate needs many random quadrats so the mean is representative. One or two readings give an unreliable result that is easily skewed by an anomaly.