Using quadrats
A quadrat is a square frame (commonly 0.5 m x 0.5 m or 1 m x 1 m) placed on the ground to define a sample area. The number of organisms of each species within the quadrat is counted. Multiple quadrats are placed in the habitat and the results are averaged to estimate population density.
Quadrats must be placed randomly to avoid bias. This can be done by using a random number generator to produce coordinates on a grid laid over the habitat. Non-random placement (such as choosing "interesting" areas) gives biased results.
Estimating population size
To estimate the total population in a habitat:
1. Count the number of organisms in several randomly placed quadrats.
2. Calculate the mean number per quadrat.
3. Calculate the area of one quadrat.
4. Divide the total area of the habitat by the area of one quadrat to find how many quadrats would fit.
5. Multiply the mean number per quadrat by the number of quadrats that would fit.
Population estimate = mean number per quadrat x (total area / area of one quadrat)
More quadrats give a more reliable estimate. The estimate assumes uniform distribution, which is rarely perfect, so it is an approximation.
Using transects
A transect is a line placed across a habitat, often where there is an environmental gradient (such as from a pond edge to dry land, or from a footpath into a field). Quadrats are placed at regular intervals along the transect, and species are recorded in each one.
A line transect records only the organisms touching the line. A belt transect uses quadrats placed beside the line at regular intervals, giving more data. Transects show how species distribution changes across a habitat and can be linked to changes in an abiotic factor such as light intensity, soil moisture or soil pH.
Sources of error and reliability
To improve reliability: use a large number of quadrats, ensure random placement, use the same identification criteria for each species, and repeat the investigation at different times. Percentage cover can be estimated for species that are difficult to count individually (such as grass or moss).
Sources of error include: non-random placement (bias), too few quadrats (unreliable mean), difficulty identifying species, counting organisms that are on the edge of the quadrat inconsistently, and environmental changes between sampling sessions.