Skills in the study of ESS practice questions.

A student sampling percentage cover of heather on a moor has already fixed the quadrat positions using random coordinates. The soundest way of deciding how many quadrats to record is

• adding quadrats until the running mean stops changing
• recording five quadrats, one at each corner and one in the centre
• using as few quadrats as the available time allows, so that each one is read carefully
• matching the number of quadrats to the number of species expected on the moor

Answer: adding quadrats until the running mean stops changing

The mean of a patchy variable jumps about while the sample is small and settles once enough ground has been covered, so a plot of the running mean against quadrat number is the direct evidence that the sample is large enough. A number fixed in advance is arbitrary: it may be far too few on patchy ground and wasteful on uniform ground.

Seaweed species on a rocky shore form clear bands between the low-water mark and the splash zone. The sampling design that will record how abundance changes with height on the shore is

• random quadrats scattered across the whole shore without regard to height
• a belt transect of quadrats running up the shore from the water
• one large quadrat placed in the middle of the shore
• kick sampling at three points along the water's edge

Answer: a belt transect of quadrats running up the shore from the water

A transect keeps the samples ordered along the environmental gradient, so the position of each band can be recorded. Random quadrats give a good overall mean for the shore but average away the very pattern being investigated.

A Secchi disc is lowered into a lake until its pattern just disappears, and that depth is recorded. The depth is used as a field estimate of

• the dissolved oxygen concentration of the surface water
• the transparency of the water, from which euphotic depth is estimated
• the biochemical oxygen demand of the water column
• the concentration of nitrate reaching the lake from surrounding farmland

Answer: the transparency of the water, from which euphotic depth is estimated

Secchi depth is a measure of turbidity: the disc vanishes at roughly the 10 to 20 per cent light level, whereas the euphotic zone reaches down to the 1 per cent level, about two to three times deeper. The reading therefore gives transparency directly and euphotic depth only as an estimate, and it says nothing about dissolved gases or nutrients.

In the determination of biochemical oxygen demand the sealed sample bottles are stored in the dark. The reason for excluding light is that

• darkness slows the decomposers down so that the test takes exactly five days
• light would raise the temperature of the sample above 20 degrees Celsius
• algae in the sample would otherwise photosynthesise and release oxygen
• darkness prevents nitrate from being converted into nitrite

Answer: algae in the sample would otherwise photosynthesise and release oxygen

BOD is the oxygen consumed by decomposers, so any oxygen added by photosynthesis would mask that consumption and give an artificially low value. Keeping the bottle dark is what suppresses photosynthesis; the separate conditions of 20 degrees Celsius and a fixed incubation period standardise the rate and the duration so that samples can be compared.

Invertebrates in two stretches of a stream are compared using kick sampling. For the two sets of results to be comparable, the student must keep constant

• the water temperature at the two sampling points
• the number of invertebrates collected from each sample
• the species of fish present in each stretch
• the kicking time, net mesh size and area disturbed

Answer: the kicking time, net mesh size and area disturbed

Kick sampling gives a relative, not an absolute, measure of abundance, so equal sampling effort is what makes two sites comparable. Temperature is a natural variable that cannot be fixed, and the invertebrate counts are the data being collected, not a control.

A geographic information system produces an erosion risk map by overlaying a soil map published at a scale of 1:250 000 with the boundaries of individual 2 hectare fields. The conclusion that the output can legitimately support is

• a broad risk pattern for the catchment, not a rating per field
• the exact area of each field that falls into the highest risk class
• which individual fields should be taken out of cultivation next season
• the soil type at any chosen point, to within a few metres on the ground

Answer: a broad risk pattern for the catchment, not a rating per field

An overlay can never be more precise than its coarsest input: a soil polygon drawn at 1:250 000 is generalised over hundreds of metres, so a 2 hectare field sits well inside a single mapped unit. The GIS still draws crisp boundaries on the output, and that false precision is the standard trap when layers of different scale are combined.