Your dataset
Every reading from the fieldwork day, and your sub-hypotheses. Fill any gaps from the shared set and note which readings are not your own.
- One clean spreadsheet, raw readings and calculated values kept apart
- Discharge and hydraulic radius calculated for every site
- A decision about which sites and variables you are using, and why
- Any anomalous readings identified, and kept
This is where a shared dataset becomes your dataset.
Everyone has the same readings. Which sites and variables you keep, what you calculate and what you do about the odd number are your decisions, and Criterion C (Quality and treatment of information collected) marks the result.
| Quantity | How | Units | What it tells you |
|---|---|---|---|
| Cross-sectional area | Sum of segment areas: each depth × the interval between readings. Or mean depth × occupied width | m² | How much channel there is to carry water |
| Discharge | Cross-sectional area × mean velocity | m³ s⁻¹ (cumecs) | The volume passing a point each second, the headline downstream variable |
| Hydraulic radius | Cross-sectional area ÷ wetted perimeter | m | Channel efficiency. Higher means proportionally less water in contact with the bed and banks, so less friction |
Summing segment areas follows the real shape of the bed. Mean depth × width treats the channel as a rectangle, and the error is worst where the channel is most asymmetric, on a meander bend.
Everything below is how we suggest you actually do it.
Use less data, better
You collected more than you need. Pick only the variables your sub-hypotheses need: two or three analysed properly across nine sites beat eight described in turn. The top band of Criterion C (Quality and treatment of information collected) asks for data that is all directly relevant, and the top band of Criterion D (Written analysis) for analysis with no or only minor gaps in its evidence. Both get harder the more you carry.
The same goes for sites. You may leave out a site measured in unrepresentative conditions (a stagnant backwater, a concrete culvert, a reach being worked on), but say in the report that you did, and why. Never drop a site silently.
A tape along the bed lifts off between stones, so check your wetted perimeter against the flowmeter manual’s estimate (manual linked in step 3): wetted perimeter ≈ width + (2 × mean depth). Report the taped figure. If the two are far apart at one site only, re-read that site.
Work one site by hand, then put the same depths through the calculator. If they agree, your spreadsheet formula is right for the other eight. Never let it do the nine sites for you or paste its diagrams into your report: Criterion C (Quality and treatment of information collected) marks your own processing.
Eight variables, each with a scattergraph and a paragraph, none tied to a hypothesis: about a hundred words each of the analysis’s 850.
Enough to describe a graph and nothing else.
Two or three variables, each tied to a sub-hypothesis, with the right graph and one statistical test if it fits: about 650 words between them, 100 on anomalies, 100 drawing it together.
The same 850 words, with room to explain.
One spreadsheet, raw and processed kept apart
Build one sheet with a row per site, including latitude and longitude in decimal degrees (step 5 maps them). Keep raw readings and calculated values visibly apart, in separate blocks or under a clear label.
Use formulas, not typed values. A formula can be checked, and if a depth was copied wrongly, everything calculated from it corrects itself.
Pick one unit for each variable and keep it in the sheet, every graph and the write-up. Bedload in millimetres or centimetres is fine, as long as it is the same everywhere. The class sheet records in metres and the callipers read in centimetres, so convert once, in the sheet, with a formula, and put the unit in the column header.
If any velocities came from a float, correct them here: multiply by the surface-velocity coefficient you chose in step 3, commonly 0.85, and record which value you used. Never mix corrected and uncorrected velocities in one column.
This cleaned table, with any calculations, goes in your appendix. There are no data tables in the body of the report, where data appears as graphs and maps. An appendix of only raw data and calculations does not count towards the word limit.
Wherever you took repeat readings (three flowmeter trials, three float runs, thirty stones), the spread is data. Put the highest and lowest beside the mean.
Velocity trials that disagree wildly tell you about turbulent flow at that site. A mean with its range also gives your evaluation a limitation you can measure, instead of vague "human error".
The reading that looks wrong is your most valuable one
Some site will refuse to behave: a velocity that drops, a channel that narrows downstream, angular pebbles far from the source. Do not assume you measured it wrong and lose it.
The top band of Criterion D (Written analysis) asks for anomalies, if present, to be explained and linked to the question, the hypotheses, the theory, the fieldwork location and the methods used. The band below only lists them.
On the Asse you have a good chance of explaining one: a site in an engineered reach has a channel shape somebody chose, a second, human, control on your variables.
Check your sites against three kinds of place before calling anything an error.
Engineered or restored reaches. The flood-protection and renaturation scheme covers about 4.5 km inside Nyon and is built in stages, so check what exists at your site. A reworked channel can be wider, shallower or slower than distance from the source predicts.
Tributary junctions. The Calèves was reconnected to the Asse in 2019. Discharge steps up below a confluence rather than rising smoothly.
Structures and local conditions. Bridges, weirs and culverts fix the channel’s width and bed. A pool sampled as if it were a riffle gives a slack velocity unrelated to its place on the river.
Check every odd reading against a map and your own photographs. A slipped tape belongs in the evaluation; a rebuilt channel belongs in the analysis, and is worth far more.
Decide about a statistical test now
A statistical test is optional. The criteria only say you “may” use one, “if appropriate”. Judge whether one fits your question and data, then get it right.
If you run one, Spearman’s rank correlation coefficient usually fits: it measures the strength and direction of a relationship between two ranked variables, such as a variable and distance from the source. For counts of stones in roundness classes, chi-squared may suit better (step 6).
Decide before you make your graphs: the choice sets your null hypothesis, and a test chosen to fit a result you have already seen is obvious to a reader.
Report the coefficient (between −1 and +1) and whether it is significant. Without the second, the reader cannot judge the first.
With n = 9, Spearman’s needs a coefficient of about 0.6 (one-tailed; about 0.70 two-tailed) to be significant at the 95% level, so a relationship you can see on the graph can still fail. That is a limitation of the design, not your mistake.
Say so where you report the result: "rs = [your value], which does not reach significance at the 95% level with n = 9" is stronger than a coefficient alone, and Criterion F (Evaluation) rewards it.
Check the critical value in a table and say which you used. It depends on the number of pairs and on one-tailed or two-tailed testing; your directional hypotheses from step 1 should justify that choice.
The top band of Criterion D (Written analysis) asks for techniques “appropriate to the data”. Two facts about this dataset make that a real question.
The sites are not independent. Sites 5 and 6 are 40 m apart and sites 8 and 9 are 200 m apart, and a site’s discharge is partly the site above it. Correlation assumes independent observations.
Neither rules a test out: name both when you justify your test, and report a non-significant result honestly.
Your large samples are inside each site. Thirty stones per site means a bedload comparison between two sites rests on sixty observations, not nine: a test there has far more power (step 6 shows how).
The 3–4 band of Criterion D (Written analysis) reads: “Any statistical techniques used either are not relevant to the question formulated or contain errors.” A wrong test pulls you into that band, below work with no test at all.
Run one only if you can answer yes four times: Can you say why this test suits this data? Run it correctly? Interpret the result? Link it to a hypothesis you wrote? Any no, and spend the words on explanation instead.
Three tests do not show three times the skill. One test, chosen for a reason and interpreted properly, is what the descriptors ask for.
It can explain what a statistic does, what a spreadsheet formula means, or why an equation is built the way it is.
It cannot do the arithmetic. Numbers worked out from a pasted table look right and cannot be checked, and an error here spreads into every graph and sentence of your analysis.
Ready for step 5?
0 of 13This is the last step where a mistake is cheap: a spreadsheet error becomes an error in every graph, statistic and sentence of the analysis.
Step 5 is the other half of Criterion C (Quality and treatment of information collected): presentation techniques that suit each variable, made well enough that the pattern shows before a word of analysis.
Anything marked “our river” is specific to our fieldwork on the River Asse at Nyon. Photographs are the author’s own.
