The fieldwork
Your sub-hypotheses, so you know which variables you actually need. If you have not settled them, do that first: the method section only describes the methods behind the data you use.
- A definition of each variable you are using
- A description of how each was measured, and with what
- The main source of error in each technique
- The sampling strategy, and an honest account of what we actually did
- Your own copy of every reading
You were told what to measure. You still have to say why it was worth measuring.
The descriptor does not ask you to invent a method. It asks you to describe the methods, explain how the combination of data collected is relevant to your question, and show that the techniques were used correctly. Every one of those is available to somebody who followed instructions carefully and paid attention while doing it.
| Variable | How it is measured | Direct or calculated | Main source of error |
|---|---|---|---|
| Occupied channel width | Tape across the water surface, bank to bank | Direct | Where exactly the water's edge is, on a shelving bank |
| Bankfull width | Tape between the tops of the banks | Direct | Judging the bankfull line, especially where banks are vegetated or engineered |
| Wetted perimeter | Tape or chain laid along the bed and up both banks | Direct | The tape lifting off the bed, or cutting corners between boulders |
| Channel depth | Metre rule at ten fixed intervals across the channel, starting from the right bank as you look downstream | Direct | Rule pushed into soft sediment; the operator's own feet disturbing the bed |
| Gradient | Clinometer between two ranging poles a set distance apart | Direct | Poles not vertical, or sighted to the wrong height on the far pole |
| Cross-sectional area | Mean depth × occupied width, or the sum of segment areas | Calculated | Inherits every depth error; segments too wide to catch the channel's shape |
| Velocity | Impeller flowmeter at right bank, centre and left bank, three trials each; plus a float timed over a set distance, three runs | Direct (then converted) | Surface floats overestimate mean velocity; one trial is not enough |
| Bedload size and shape | Long axis with callipers; roundness against the Powers index | Direct | Picking the pebbles your eye is drawn to |
Everything below is how we suggest you actually do it.
Explain the combination, not just the list
The commonest shape for this section is eight paragraphs, one per variable, each saying what was measured and with what. That is a description of methods, and it sits in the middle band.
The top band asks for something slightly different: why this set of measurements answers this question. Group them by what they let you calculate and you are most of the way there.
Together they give cross-sectional area and hydraulic radius, the shape of the channel, and how efficiently it carries water.
With cross-sectional area, velocity gives discharge. With gradient, it tests the assumption that steeper means faster.
Evidence of the processes doing the work: attrition and abrasion rounding and shrinking the load as it travels.
Only describe the methods behind the data you actually use. If you are not testing a bedload hypothesis, the callipers do not need a paragraph. This is the single easiest way to get 300 words to fit.
An annotation of ten words or fewer does not count towards the word limit. Annotate a photograph of the flowmeter in the channel with four short notes and you replace forty words of prose describing where it was held.
The distinction is worth getting right, because only one of them earns anything. A label names something in a word or two, bankfull, riffle. An annotation says something about it: held at 0.6 of depth, mid-channel. Label a photograph and a reader knows what they are looking at; annotate it and they know what you did and why.
Take those photographs on the day. You cannot reconstruct them afterwards, and a diagram of equipment you have downloaded from somewhere else says nothing about how you used it.
If you are running a statistical test, say so here
This is the one piece of examiner advice that costs a sentence and is almost always missed. If you intend to run a statistical test on your data later, describe it in the method section, not only at the point in the analysis where the result appears.
It is a small move that does real work. Naming the test here forces you to say what you are collecting the data for, which is exactly the link between method and question that separates the top band of Criterion B (Method(s) of investigation) from the middle one. Turning up with a test in step 6 that the method never mentioned reads as an afterthought, because usually it was one.
One or two sentences is enough. Which test, on which pair of variables, and what its null hypothesis will be. Step 4 is where you decide whether to run one at all.
Almost everyone describes their methods. What separates a 2 from a 3 is whether the description is linked: to the fieldwork question, to the hypotheses, and to what each measurement is going to be used for.
Most reports lose the mark not by leaving something out but by staying general. “We measured velocity with a flowmeter” is a description. “We measured velocity at three points across the channel so that the mean could be multiplied by cross-sectional area to give discharge, the variable the first hypothesis is about” is a description that is linked, and it is barely longer.
Sampling: what we should do, and what we did
Three sampling strategies, and each one belongs somewhere different in this investigation.
Now the honest part, and the reason this beat exists. Our sites are not systematically spaced. They are where the river is safe to enter and legally accessible, a bridge, a footpath, a public bank. That is opportunistic sampling, and pretending otherwise is worse than admitting it.
Saying so is not a confession, it is material for Criterion B (Method(s) of investigation) and Criterion F (Evaluation). What matters is that you can state the ideal, state what was actually possible, and say what the difference does to your data: sites cluster where access is easy, gaps open where it is not, and any reach with no public access is simply missing from your dataset.
Nine sites across 8.9 km, from 711 m down to 413 m. The spacing is nothing like even: 2.3 km between the first two, 405 m between sites 2 and 3, gaps of well over a kilometre through the middle, and just 40 m between sites 5 and 6, either side of the Paléo festival site.
Two pairs are worth knowing about before you plot anything: sites 5 and 6 are 40 m apart, and sites 8 and 9 are 200 m apart. Each pair sits almost on top of itself on every scattergraph, and where two such readings differ, the difference cannot be caused by distance downstream; it has to be something about those two places.
Look at your site list before the fieldwork day, not after. Knowing where the gaps and the clusters are tells you which pairs of sites can be compared fairly, and gives you a specific, evidenced sampling weakness for the evaluation instead of a generic one.
The two velocity methods, and why the float lies
A float measures the velocity of the water surface, in the middle of the channel, where flow is fastest. Friction against the bed and banks slows the water everywhere else, so a surface float systematically overestimates the mean velocity of the whole cross-section.
The fix is a surface-velocity coefficient: multiply the float figure by a number less than one to estimate the mean velocity of the whole cross-section. The USGS reference work on streamflow puts that coefficient in the range 0.84 to 0.90, depending on the shape of the vertical-velocity profile and how close you are to the banks, and 0.85 is the value most school fieldwork uses. Rougher beds sit at the low end of the range, smoother sand or mud beds at the high end. An uncorrected float velocity fed into a discharge calculation makes your discharge too big at every site, so if you use float data, say which coefficient you applied and why.
The impeller flowmeter avoids that by being held at a stated depth in the flow, and by being read at several points across the channel. Its count converts to a velocity with the calibration equation for the instrument:
velocity (m s⁻¹) = 0.000854 c + 0.05, where c is the count in one minute. That is the manufacturer’s own calibration line, fitted in a flume tank against two reference flowmeters, and its published chart runs from 0 to 2,000 counts a minute, which is 0.05 up to about 1.76 m s⁻¹.
Note the intercept. A count of zero does not give a velocity of zero; the equation is a calibration line fitted to the instrument, not a law of nature, and it stops being meaningful in very slow water.
A site gives these readings: width 3.2 m, mean depth 0.4 m, wetted perimeter 4.1 m, flowmeter count 410 revolutions in one minute. Work out, in this order: the cross-sectional area; the velocity, using the equation above; the discharge; and the hydraulic radius.
Then state the units of each answer, and say which of the four you would expect to increase with distance from the source. If any of those four numbers is a struggle now, it will be a much bigger struggle across nine sites in a spreadsheet.
Whichever method you use, repeat it. Three trials and a mean is the minimum defensible practice, and having the three separate values lets you say something about how variable the flow was, which is far more useful in the evaluation than a single number.
The coefficient depends on bed roughness, and bed roughness is one of the things changing along your river: coarse and angular at the top, finer and rounder near the mouth. So a single 0.85 applied at every site is itself a small, systematic, downstream-varying error.
You are not expected to fix that. You are expected to notice it, and it makes a far better evaluation point in step 7 than anything about the weather.
Velocity is not the same all the way down a river. Friction against the bed slows the water at the bottom, and the surface runs fastest, so where in the water column you hold the impeller decides what your number means.
The manual gives the standard answer: in water shallower than 0.6 m, a single reading taken at 0.6 of the depth below the surface is a reliable estimate of the average for that whole vertical, because the fast surface flow and the slow bed flow average out at roughly that point. Its trick for finding it without arithmetic in mid-river: rest the stick on the bed, pinch the water line, lift it out, and put a rubber band 40% of the way down from your pinch to the base. Submerge to the band.
Whatever you did, write it down. “Velocity was measured at 0.6 of the depth” and “the impeller was held just under the surface” are two different variables, and only the first is the mean velocity your discharge calculation assumes.
Two numbers off that page do real work in your evaluation. The first is the ±5% accuracy: a velocity of 0.60 m s⁻¹ could be anywhere from 0.57 to 0.63, and because discharge is velocity multiplied by area, that uncertainty passes straight into every discharge figure you calculate. If two sites differ by less than that, you cannot claim one is faster than the other.
The second is the 0.5 to 1.5 m s⁻¹ working range. The Asse mostly flows below it: on our fieldwork day only site 8, at 0.512 m s⁻¹, sat inside the band. A reading outside that band is the manufacturer telling you it is less reliable, which is a specific, evidenced limitation with a source behind it.
That is the difference between a top-band evaluation and a weak one. “Human error” is worth nothing; “the flowmeter is quoted at ±5% and eight of our nine sites sat below its stated working range” is worth a great deal.
The sheet asks for three stones at each of ten positions across the channel, thirty per site, and left to itself almost everybody picks the large, distinctive, easy-to-grab ones. Small pebbles are harder to see, awkward to measure and boring, so they are quietly under-sampled at every site.
The result is a bedload dataset that is biased in the same direction everywhere, which is the worst kind of bias: it looks consistent.
The fix is a rule you follow without looking: throw a quadrat blind, or take the pebble nearest each metre mark, and measure whatever you get.
The six classes come from a 1953 paper by M. C. Powers in the Journal of Sedimentary Petrology, and they are a visual comparison chart: you hold your stone against a row of silhouettes and pick the closest. Very angular, angular, sub-angular, sub-rounded, rounded, well rounded.
That makes roundness an ordinal measurement, not a continuous one. The classes are in a genuine order, but the gaps between them are not equal, and there is no unit. It matters more than it sounds: a “mean roundness” of 3.7 is not really a number, because you invented the 1-to-6 spacing when you typed it into the spreadsheet. Counts of stones in each class are honest; an average of class numbers is a convenience that a careful reader can challenge.
It is also judged by eye, so two people classify the same stone differently. The cheap fix is a rule you set before you start: one person classifies every stone, or the group agrees a reference stone for each class and keeps it in the tray. Either way, say in your method which you did, because an ordinal scale applied by several people is a real limitation with a specific name.
Powers gave the chart a second axis almost nobody uses: sphericity, high or low, running across all six roundness classes. Roundness is how worn the edges are; sphericity is the overall shape. A flat, well-worn disc is rounded but has low sphericity, and two stones in the same roundness class can travel very differently.
Increasing roundness downstream is evidence of attrition and abrasion during transport, which is why the variable is worth collecting at all.
Recording sphericity alongside it gives you something to say when roundness alone refuses to trend.
On the day: write everything down yourself
Every member of the group records every measurement. Not one person with the clipboard. This is the rule we care about most on the fieldwork day, and it is not about mistrust; it is about what you will need three months later.
The IB is explicit that fieldwork information may be collected and exchanged as a group, and that the emphasis then moves to individual work: the writing, the justification of methods, the analysis and the conclusion must be entirely your own. Having your own record of the numbers is what makes the individual part possible.
It also matters because you will be evaluating these methods in step 7, and the material for that evaluation is what you noticed while doing it. Which reading you had to take twice, the site where the tape kept lifting off the bed, the pool that was obviously slack water. None of that survives in a shared spreadsheet.
All nine sites are measured on the same day, which is a real strength: you are comparing sites under the same weather and the same flow conditions rather than comparing a wet week with a dry one.
It is also a real limitation: you have a single snapshot, and the same river in a different season would give different numbers. Both halves of that belong in your evaluation, and noticing it now is easier than remembering it later.
It can explain a technique or a piece of equipment properly, how an impeller flowmeter works, what wetted perimeter is measuring, why roundness indicates transport distance. The techniques here are standard and well documented, so this is safe ground.
It cannot evaluate a method it did not watch you use. "The tape may have sagged" is a plausible sentence about any tape in the world. What Criterion B (Method(s) of investigation) rewards is the error you saw happen at your site, and only you were standing in the river.
Ready for step 4?
0 of 10Boxes m1 to m7 can all be ticked by someone who followed instructions and paid attention. m8 to m10 can only be ticked on the day; if you are reading this before the fieldwork, they are the ones to plan for.
Step 4 turns a field notebook into a dataset: the calculations, which sites and variables to use, and what to do about a reading that looks wrong.
Anything marked “our river” is specific to the fieldwork we do together on the River Asse at Nyon: the sites, the equipment and the way we collect the data. Site coordinates and altitudes are from our own fieldwork records and the exact list can change from year to year, so check them against the sheet you are given on the day. Photographs are the author’s own, taken at the river.