A River Measured Twice

A river can look fuller without telling us how much water passes a particular place each second. The shining surface gives us a level; movement gives us another question. Confusing those questions is easy because both concern the same water. This short book follows them separately, then brings them together through one complete paper example.

The river in the example is invented. No author stood in it, lowered an instrument from a bridge or collected the numbers below. Its two records were designed to make a calculation understandable. The scientific account draws on the US Geological Survey; the numerical exercise is original and hypothetical. You need paper and ordinary arithmetic, not access to a river. This is an introduction to reading measurements, not instructions for entering water or making safety decisions.

We will ask three things. What is the height of the water surface relative to a reference? How much water crosses a section per second? What can we responsibly conclude when a second record differs from the first? The third question matters because an answer can be mathematically correct and still tell a larger story than its evidence supports.

Imagine two sheets on a table. One bears an early record; the other a later one. Neither has a dramatic photograph or a headline about a flood. Their quiet usefulness lies in what they specify: areas, representative velocities and units. By the end, their difference will have an exact value. Its cause will remain a separate question, with a clear account of the additional evidence it would need.

Discharge is a rate of volume passing a cross section. Think of an imaginary window standing across the channel, perpendicular to the main flow. Water passes through it. The window is not a bucket holding the river, and its area alone cannot tell us how rapidly water travels through it.

In a simple area–velocity model, discharge Q equals cross-sectional area A multiplied by the average velocity through that area. Square metres multiplied by metres per second give cubic metres per second: m² × m/s = m³/s. The unit is part of the explanation. Area describes a slice; discharge describes volume per time. Neither is the volume of all water in the river.

Real cross sections contain different depths and velocities. USGS explains a method that divides the section into subsections, computes each contribution, then adds them. Our four cells are a deliberately small teaching model, not a claim that four observations would adequately describe a real channel. Each listed velocity is stipulated to represent the cell-average component normal to the section: the motion through our imaginary window.

That stipulation does important work. A fast patch on the surface is not automatically the average for the entire section. Nor does the speed of a floating object establish all the velocities below it. We will not turn those observations into a field procedure. For our paper exercise, representative values are supplied so that we can concentrate on what the calculation means and where a careless shortcut changes it.

The first hypothetical record divides the section into cells A, B, C and D. Their areas are 0.6, 0.8, 0.5 and 0.3 m². Their respective representative velocities are 0.4, 0.5, 0.6 and 0.3 m/s. Keep the order: an area and its velocity belong together.

For A, 0.6 × 0.4 = 0.24 m³/s. For B, 0.8 × 0.5 = 0.40 m³/s. For C, 0.5 × 0.6 = 0.30 m³/s. For D, 0.3 × 0.3 = 0.09 m³/s. Adding those four contributions gives Q = 1.03 m³/s. Adding their areas gives 2.2 m². We have accounted for all four cells; no exercise answer has been left for a later volume.

A tempting shortcut is to average the four velocities equally. Their sum is 1.8 m/s, so their arithmetic mean is 0.45 m/s. Multiply 0.45 by 2.2 and the result is 0.99 m³/s. That is not 1.03. The shortcut gave a small cell as much influence on the average as a larger one.

The area-weighted average consistent with our record is 1.03 divided by 2.2, approximately 0.46818 m/s. Multiplying that value by 2.2 recovers the discharge, allowing for rounding. There is no quarrel between Q = A × average velocity and adding cell contributions. The quarrel is with calling the wrong average representative of the whole area.

Notice one more detail: cell C has the fastest listed velocity, but B contributes the most flow. B has more area. A single superlative—fastest, deepest, widest—cannot replace the combination the question requires. The sheet becomes more intelligible when every number keeps its partner and its unit.

The second invented record uses areas of 1.0, 0.9, 0.7 and 0.4 m² for A through D. To isolate the effect of area in this model, the four representative velocities remain 0.4, 0.5, 0.6 and 0.3 m/s. This is a chosen teaching condition, not a prediction that real velocities remain unchanged as water rises.

The new contributions are 0.40, 0.45, 0.42 and 0.12 m³/s. Their sum is 1.39 m³/s; the total area is 3.0 m². The second discharge exceeds the first by 1.39 − 1.03 = 0.36 m³/s. The area increased by 0.8 m². Those differences answer two distinct questions and must not be exchanged simply because both are increases.

Equal averaging still gives 0.45 m/s, but now multiplying by the total area produces 1.35 m³/s. The correct weighted average is 1.39/3.0, approximately 0.46333 m/s. It is slightly lower than the first weighted average even though every individual velocity stayed the same. Changing the areas changed their weights. This small surprise is the reason for keeping both sheets in view.

What caused the greater area? The invented record does not say. Rainfall, an upstream release or another story cannot be established by our arithmetic. We can honestly say that the second model carries more volume through its section per second under the supplied assumptions. Naming an environmental cause would require observations we have not supplied.

A useful comparison therefore ends with a precise sentence rather than an exciting one: the hypothetical discharge increased by 0.36 m³/s, with the listed representative velocities held constant and the cell areas changed. Every part of that sentence can be traced back to the sheets.

Stage, also called gage height, describes the water surface relative to an established reference elevation. It does not automatically give the depth from the current riverbed to the surface. A reference can stay fixed while the bed changes. USGS describes the importance of maintaining a consistent datum and checking the structures used for measurement.

Return to the imaginary window. A water-level reading can help describe its upper boundary, but the number on a gage is not itself the area of the window or the velocity through it. Our sheets supplied areas directly; they did not supply stage values. We cannot work backwards and invent corresponding heights from Q = A × velocity alone.

At a particular station, measurements of stage and discharge can establish a relationship called a rating curve. This lets a stage record support estimates of discharge. It is a local relationship supported by measurements, not a universal conversion from metres of height to cubic metres per second. The same reported height at another station does not guarantee the same discharge.

Nor should doubling a gage height be treated as doubling discharge. Channel and floodplain shape help determine how a rising level changes the space available to flow. The USGS water-science explanation specifically cautions against treating the relationship as a simple linear rule. Our two sheets were not designed to fit a rating curve, and two invented flow values cannot establish one.

The gage and the calculation now have separate jobs. One records a surface level against a reference; the other describes a rate through a section. A supported relationship can connect them. Keeping the jobs distinct makes that connection more useful, not less.

A rating curve belongs to a physical channel. USGS identifies changes such as aquatic vegetation, shifting sediment, floods and construction that can affect the relationship between stage and discharge. Some changes call for adjustments; others may require a new curve. The drawn line is not a promise that the channel will always behave as it did when earlier measurements were made.

This gives the second measurement a purpose beyond checking arithmetic. A later observation may test whether the relationship still represents the place. A difference could also raise questions about instruments or procedure. Those possibilities should be investigated rather than compressed into an instant verdict that either the river or the earlier observer was wrong.

Our invented sheets do not reproduce that professional investigation. Their inputs have been declared, not measured, and they contain no uncertainty estimates. Writing 1.03 with two decimal places does not prove a real instrument could resolve a change of 0.01 m³/s. Numerical detail and knowledge of measurement uncertainty are different things.

Consider an editor comparing two reports. Before writing that a river's behaviour changed, the editor would want the location, time, method, reference information, data status and any relevant qualifications. A second number without those details may be difficult to compare with the first. This is a reading habit, not a substitute for hydrological training.

The fair question is not whether a record ever needs revision. It is whether its changes can be explained, documented and evaluated. A defensible correction can improve what a record tells us. Preserving the earlier version and the reason for change also makes the history of that knowledge easier to follow.

Suppose our first hypothetical discharge, 1.03 m³/s, remained constant for exactly one hour. One hour contains 3,600 seconds. Multiplication gives 1.03 × 3,600 = 3,708 m³. This is the volume crossing the section during that hour under the explicit constant-rate assumption. It is not the total water stored in the channel.

A single reading of 1.03 m³/s does not establish that hourly volume. The calculation needs the assumption as well as the number. If the rate changes, the interval must be represented appropriately; one isolated value cannot silently stand in for all the others. A daily average and an instantaneous value also describe different time scopes.

We can answer another paper question completely. If the second model's 1.39 m³/s stayed constant for the same hour, its volume would be 5,004 m³. The difference would be 1,296 m³, the same as 0.36 × 3,600. These are consistent calculations of the invented steady conditions, not measured totals for a particular day or river.

USGS labels provisional data as subject to revision until review and final approval. Its statement identifies possible instrument problems and physical changes at a site, with inspections and measurements informing subsequent review. A current-looking number may therefore carry a qualification that matters to its interpretation.

When reading a real record, keep its station, timestamp, time zone, units and status with the value. Do not discard those details to make a tidier caption. This book supplies no real-time conditions, flood threshold or judgement that a place is safe. Its calculations teach a distinction that remains useful precisely because they do not pretend to answer those larger questions.

Both sheets can now be read without guessing. The first model has an area of 2.2 m² and a discharge of 1.03 m³/s. The second has 3.0 m² and 1.39 m³/s. Their representative cell velocities were held constant by design. Their area-weighted mean velocities differ because the shares of area differ. Their discharge difference is 0.36 m³/s.

Those results belong to a complete calculation. They do not identify a storm, establish a station rating curve, supply a measured stage or certify the precision of real equipment. The boundary is clear enough to be useful. We know what was supplied, what was calculated and what would require more evidence.

Three questions can travel with you into a future chart or report. What quantity is this number describing? What place, interval and reference make it interpretable? What supports the step from this observation to the conclusion being proposed? They are questions to ask of evidence, not a demand that every ordinary reader become a river scientist.

The imagined river has done its job. At first it was a shining surface with an uncertain promise of information. On the table it became a cross section, four paired values, a weighted average and a comparison. It has not become less vivid because we can name those things. Its movement has simply acquired a description we can check.

A river measured twice offers more than a larger or smaller answer. It offers an opportunity to notice what stayed constant, what changed, and what the numbers have not yet told us. That is where this account ends: with a completed calculation and an open, properly bounded question about the world beyond it.

Research checked 1 November 2026. This complete short digital science book contains eight chapters and these notes as nine fixed content units, not printed pagination. Original AI-assisted prose and full translations; no author-performed fieldwork, human-authorship or independent native-language certification is claimed. Covers are original conceptual artwork, not maps or measured cross sections.

USGS, How Streamflow is Measured, dated 17 September 2026: selected main sections on stage, reference elevation, discharge and summing subsection contributions. https://www.usgs.gov/water-science-school/science/how-streamflow-measured?qt-science_center_objects=0

USGS, How is a rating curve used to convert gage height into streamflow?: main explanation of local relationships and physical changes; older linked interface instructions were not tested or adopted. https://www.usgs.gov/faqs/how-a-rating-curve-used-convert-gage-height-streamflow

USGS, Water Q&A: Does stage tell you how much water is flowing?, 20 June 2019: selected main explanation of nonlinear stage–discharge relationships. https://www.usgs.gov/water-science-school/science/water-qa-does-stage-tell-you-how-much-water-flowing

USGS, Provisional Data Statement: review status, possible changes and qualifications. https://water.usgs.gov/data/provisional.html

The four-cell records, all values, paired calculations and constant-hour examples are original hypothetical teaching material. They are not extracts from USGS station observations. No source photographs, diagrams, field procedures or current station statistics have been reproduced. Direct access to some additional USGS pages failed; those pages are not represented as fully read sources here.