The gold standard of watersheds requires keeping two questions separate: how much water passes an outlet, and how quickly it arrives. Two catchments can deliver the same runoff volume while producing very different peak flows. A total describes an amount. A hydrograph describes the changing rate through time. Understanding their relationship makes rainfall stories, river graphs and storage proposals much easier to question intelligently.
This guide develops runoff hydrographs through original worked examples. You will calculate volume from interval flow records, compare equal-volume events, test a simple storage account and see why synchronised tributaries can change a downstream peak. All numbers and named catchments are fictional teaching data. They are not forecasts, drainage designs or observations of a real flood.
The Watersheds subject guide explains outlets, contributing land and nested catchments. This article takes its volume-and-timing distinction further. It supports the educational aim of Learning That Transfers into Life: defining a quantity, explaining a relationship and recognising what changes when the conditions change.
eduKateSG · Watersheds
Find your next learning step
Choose a task-based route, then follow the expanded teaching chapters below.
Chapters 1–3 · Read the quantities
Connect the outlet, units and graph.
Chapters 4–6 · Calculate volume and timing
Compare totals, peaks and interval definitions.
Chapters 7–10 · Follow water through the model
Check background flow, catchment depth and storage.
Chapters 11–14 · Question and transfer
Examine limits, missing data and reference flows.
Chapters 15–16 · Questions and onward routes
Chapter contents
Read the quantities · Chapters 1–3
Calculate volume and timing · Chapters 4–6
Follow water through the model · Chapters 7–10
Question and transfer · Chapters 11–14
Questions and onward routes · Chapters 15–16
CHAPTER 1 OF 16 · Read the quantities
1. Start with the outlet and the time question
A river observation belongs to a location. Before reading a graph, identify the outlet or monitoring point and the land contributing to it. A flow record at an upstream station and one below a tributary junction describe different boundaries. Their numbers should not be compared as though the same water necessarily passed both points during the same interval.
Now identify the question. Are you interested in the total water passing during an event, the maximum recorded rate, the time of the maximum, or the duration above a stated level of flow? Each requires a different reading of the record. A high total is not automatically the highest peak, and a high peak does not automatically imply the largest total.
For the fictional Brook A, the question is how much event-related flow passes its outlet during four hours. For Brook B, the question is whether the highest interval-average flow occurs early or late. Both records concern water, but the requested answers differ. Name that difference before calculating.
USGS describes a hydrograph as a graph of water-related conditions through time, commonly stage or streamflow. Its Streamflow and the Water Cycle explanation provides the physical context. In this guide, every worked hydrograph uses discharge: volume passing a point per unit time.
Put a short description beside the graph: station, contributing area if known, time reference, units, interval and data status. These are part of its meaning. A graph without them may look precise while leaving the reader unable to decide which calculation or comparison is justified.
CHAPTER 2 OF 16 · Read the quantities
2. Distinguish stage, discharge, volume and rainfall
Stage is the height of the water surface relative to a stated reference. Discharge is a rate of flow. Volume is an amount accumulated over an interval. Rainfall depth is an input measured over a surface. These quantities can be related, but none should silently replace another.
A stage record of two metres does not mean two cubic metres passed the station. A discharge of two cubic metres per second does not mean that the river contains only two cubic metres. It describes how quickly volume passes the chosen point at that time or over the stated averaging interval.
USGS’s How Streamflow Is Measured explains the relationship between stage observations, discharge measurements and site-specific interpretation. A learner should therefore resist applying one universal height-to-flow conversion to every channel. The site and measurement method matter.
Rainfall adds another boundary. Ten millimetres over one square kilometre represents 10,000 cubic metres of rainfall input under a uniform-depth assumption. That calculation does not say how much reaches the outlet during the next hour. Storage, infiltration, other pathways and the period of observation affect the relationship.
For a quick check, label four statements. The water surface rose by 0.3 metres: stage change. The outlet averaged 2 m³/s for an hour: discharge over a defined interval. A total of 7,200 m³ passed during that hour: volume. The area received 12 mm of rain: rainfall depth. The exercise is simple, but it prevents a large proportion of misleading comparisons before any difficult arithmetic begins.
CHAPTER 3 OF 16 · Read the quantities
3. Read the hydrograph as a changing relationship
On a discharge hydrograph, time appears on the horizontal axis and flow rate on the vertical axis. A rising part shows increasing discharge. A falling part shows decreasing discharge. The highest displayed value is the highest value resolved by that record, not necessarily the exact instantaneous maximum of the real event.
The area beneath a discharge curve represents volume because rate multiplied by time gives an amount. This is a physical and mathematical relationship, not merely a graph convention. If the vertical unit is m³/s and the horizontal interval is seconds, the product is cubic metres.
An interval table needs an explicit interpretation. A value may be an instantaneous observation at the interval boundary, an average over the interval, or an output from a model. Those meanings support different volume calculations. Do not draw rectangles under point observations and call the result exact without stating the approximation.
The first examples here use interval-average discharge. Each value represents the average throughout its stated interval. Multiplying that average by the interval duration gives the modelled interval volume. Later, a point-observation example shows why a different assumption is needed.
Ask learners to explain the axes before touching a calculator. Then ask what one rectangle beneath the graph represents. If they can name its height as a rate, its width as a time and its area as a volume, the calculation has a clear foundation. If they only remember to multiply two numbers, a unit change or unequal interval can easily break the method.
CHAPTER 4 OF 16 · Calculate volume and timing
4. Calculate volume from equal time intervals
Brook A has four successive one-hour intervals with average discharges of 1, 3, 3 and 1 m³/s. Each hour contains 3,600 seconds. The first interval therefore contributes 1 × 3,600 = 3,600 m³. The second contributes 10,800 m³, as does the third. The final interval contributes 3,600 m³.
Adding the volumes gives 28,800 m³. Because all intervals have the same duration, you can also add the four rates and multiply their sum by 3,600 seconds: (1 + 3 + 3 + 1) × 3,600. This shortcut depends on equal interval lengths and interval-average values.
| Successive interval | Average discharge | Duration | Interval volume |
|---|---|---|---|
| Hour 1 | 1 m³/s | 3,600 s | 3,600 m³ |
| Hour 2 | 3 m³/s | 3,600 s | 10,800 m³ |
| Hour 3 | 3 m³/s | 3,600 s | 10,800 m³ |
| Hour 4 | 1 m³/s | 3,600 s | 3,600 m³ |
| Total | Not a rate sum to report as flow | 14,400 s | 28,800 m³ |
The mean discharge over the complete four-hour period is 28,800 ÷ 14,400 = 2 m³/s. The highest interval average is 3 m³/s. The record does not establish whether a brief instantaneous peak exceeded 3 during either central hour.
Check the answer in both directions. A steady 2 m³/s over 14,400 seconds would also total 28,800 m³. That confirms the average calculation while illustrating its limit: the average has replaced the pattern with one equivalent constant rate.
The result is a volume through the stated outlet and period. Do not call it the entire storm’s runoff if the supplied record omits an earlier rise or later recession. The beginning and end of the account are assumptions that belong beside the answer.
CHAPTER 5 OF 16 · Calculate volume and timing
5. Compare equal volumes with different peaks
Brook B has four one-hour interval averages of 0, 1, 6 and 1 m³/s. The rate sum is again eight, so the modelled volume is again 28,800 m³. Its four-hour mean is also 2 m³/s. Yet its highest interval average is 6 m³/s, twice Brook A’s 3.
The two events have equal totals under the supplied definitions while concentrating that volume differently through time. Brook B places three quarters of its total in one hour: 21,600 ÷ 28,800 = 75%. Brook A places 37.5% in either of its two central hours. This gives a more precise description than saying that B has more water.
| Teaching event | Hourly average sequence, m³/s | Four-hour volume | Highest interval average |
|---|---|---|---|
| Brook A | 1, 3, 3, 1 | 28,800 m³ | 3 m³/s |
| Brook B | 0, 1, 6, 1 | 28,800 m³ | 6 m³/s |
Would Brook B necessarily produce greater damage? The table does not answer that. Consequences depend on water levels, channel and structure behaviour, exposed people and assets, warnings, duration and other conditions. A flow comparison is one input into a wider flood-risk question.
Would a larger total necessarily produce a larger peak? Construct a counterexample. A steady 2 m³/s for six hours totals 43,200 m³, more than Brook B’s four-hour event, while its maximum under this model remains 2. The example separates amount, concentration and duration without pretending that every real catchment behaves like these simplified sequences.
The learner’s task is to preserve the valid comparison and stop before the evidence runs out. Equal volume does not mean equal hydrograph. Different hydrographs do not, by themselves, quantify different damage.
CHAPTER 6 OF 16 · Calculate volume and timing
6. Handle unequal intervals and point observations
Suppose a third record contains 30 minutes at an average 2 m³/s, followed by 90 minutes at an average 4 m³/s. The first interval lasts 1,800 seconds and contributes 3,600 m³. The second lasts 5,400 seconds and contributes 21,600 m³. Total volume is 25,200 m³ over 7,200 seconds.
The period mean is 25,200 ÷ 7,200 = 3.5 m³/s. Simply averaging 2 and 4 gives 3, which gives equal weight to intervals that lasted different lengths of time. The longer high-flow interval needs a larger weight.
Now change the meaning of the data. A station reports instantaneous values of 2 m³/s at noon and 4 m³/s at 13:00. If you assume a straight-line change between them, the trapezoidal estimate is ((2 + 4) ÷ 2) × 3,600 = 10,800 m³. That is an estimate under an interpolation assumption. The two endpoint observations do not prove that the flow followed a straight line.
A brief unobserved surge could increase the true volume and peak. A dip could reduce the volume relative to the straight-line estimate. The arithmetic can be exact for the assumed shape while the assumed shape remains uncertain.
For a changed-context check, give learners the same numerical values once as interval averages and once as point observations. Ask why their explanation must change even if a particular calculation happens to produce the same number. The method belongs to the data definition, not merely to the appearance of a table with a time column.
CHAPTER 7 OF 16 · Follow water through the model
7. Separate event flow from a background contribution
A river can carry flow before rain begins and after the event response subsides. If a question concerns additional event-related volume, the background contribution needs a stated treatment. Removing a constant value from a classroom dataset is a model assumption, not a universal method for separating real river processes.
Imagine the observed four-hour sequence is 2, 4, 4 and 2 m³/s. Its total volume is 43,200 m³. The exercise explicitly supplies a constant background contribution of 1 m³/s. Over four hours, that background contributes 14,400 m³. Subtracting it gives an additional event component of 28,800 m³.
The same calculation can be made interval by interval, giving 1, 3, 3 and 1 m³/s for the supplied event component. The agreement checks the arithmetic. It does not independently validate the constant-background assumption.
Real background contributions can vary, and different separation methods may answer different questions. The responsible hydrological analysis needs suitable observations and methods. A school exercise should make its simplifying assumption visible instead of encouraging the learner to subtract whatever baseline makes the graph look tidy.
Ask what happens if the supplied background is 0.8 rather than 1 m³/s. The background volume becomes 11,520 m³ and the calculated additional component becomes 31,680 m³. A modest change in the assumption changes the interpretation even though the observed record remains identical.
This is a useful lesson about models: a derived result inherits assumptions from its inputs. When a report gives an event volume, ask how the event boundary and any background contribution were defined before comparing it with another report.
CHAPTER 8 OF 16 · Follow water through the model
8. Convert runoff volume into an equivalent depth
An equivalent runoff depth spreads a volume uniformly across the contributing area for comparison. It does not claim that the water actually stood at that depth everywhere. The calculation is volume divided by area, with compatible units.
Take the fictional event component of 28,800 m³ from a 2 km² catchment. Two square kilometres equal 2,000,000 m². The equivalent depth is 28,800 ÷ 2,000,000 = 0.0144 metres, or 14.4 millimetres.
Suppose the exercise also supplies a uniform rainfall depth of 24 mm over the same area and relevant event period. Rainfall input is 0.024 × 2,000,000 = 48,000 m³. The ratio of the specified event-flow component to rainfall input is 28,800 ÷ 48,000 = 0.6, or 60%.
The ratio is a description of this modelled event account. It is not a permanent property of every storm on the catchment, and it is not automatically a design coefficient for another formula. The period, background treatment, rainfall representation and water pathways all affect its meaning.
If the observation stops while the river still carries additional event-related flow, the calculated event component is incomplete. If rainfall varies across the land, a single gauge may not represent the area mean. If water enters from another source or is transferred across the boundary, the simple input-output comparison needs revision.
A strong student answer includes the arithmetic and one relevant condition. For example: the supplied model gives an equivalent event depth of 14.4 mm, assuming the 2 km² boundary and complete event record are appropriate. That sentence keeps the useful number connected to the land and time account that produced it.
CHAPTER 9 OF 16 · Follow water through the model
9. Follow tributaries without adding their separate peaks
Two tributaries can reach their maxima at different times. Adding their individual maximum values may therefore invent a downstream condition that never occurred. Start with compatible time intervals and a stated combination model.
In a fictional four-hour exercise, Tributary East has interval averages of 0, 4, 0 and 0 m³/s. Tributary West has 0, 0, 4 and 0. Assume their contributions combine in the same intervals at the junction without additional travel delay, storage or local inflow. The combined sequence is 0, 4, 4 and 0, whose highest interval average is 4.
Adding the separate maxima gives 8, but those maxima happened in different intervals. Now shift West’s contribution into the second interval. The combined sequence becomes 0, 8, 0 and 0. Total volume is unchanged, while the highest interval average doubles.
This demonstrates synchronisation. It does not establish that changing one real tributary’s timing would produce this exact response. Real routing, channel storage, local contributions and changing hydraulic conditions need appropriate analysis. The simplified exercise isolates the time relationship so a learner can see why it matters.
When comparing actual records, align the clock, time zone, interval definition and station locations. A time offset in data processing can create or conceal an apparent coincidence. A peak at an upstream station may arrive later at the junction.
The practical reading question is therefore: are these values simultaneous contributions at the boundary being discussed, or are they separate maxima extracted from different places and times? A correct answer to that question should come before the addition.
CHAPTER 10 OF 16 · Follow water through the model
10. Work a simple storage account without promising a design
Storage can temporarily hold water when inflow exceeds outflow. A simple account begins with an opening amount, adds inflow volume and subtracts outflow volume. The closing amount becomes the next interval’s opening amount. This explains a relationship; it does not size a real detention system.
Consider a fictional basin starting empty. Four one-hour intervals have inflows of 2, 6, 2 and 0 m³/s. The teaching model allows an outflow up to 2 m³/s whenever enough water is present. Assume that this rule and the interval representation are exactly valid for the exercise.
| Interval | Inflow volume | Outflow volume | Closing storage |
|---|---|---|---|
| Hour 1 | 7,200 m³ | 7,200 m³ | 0 m³ |
| Hour 2 | 21,600 m³ | 7,200 m³ | 14,400 m³ |
| Hour 3 | 7,200 m³ | 7,200 m³ | 14,400 m³ |
| Hour 4 | 0 m³ | 7,200 m³ | 7,200 m³ |
After four hours, 7,200 m³ remains. A fifth hour at 2 m³/s would empty it under the model. Reporting only the first four outflow intervals would omit part of the eventual release. The storage has shifted timing, not made water disappear.
The largest interval-boundary storage in this simplified account is 14,400 m³. Do not call that an adequate real basin capacity. Within-interval variation, outlet behaviour, initial conditions, later rain, overflow paths and applicable design requirements are outside this model.
The most useful check is conservation: total inflow equals total outflow plus the change in storage. At four hours, 36,000 m³ entered, 28,800 m³ left and 7,200 m³ remains. If the numbers do not balance, inspect units and boundaries before explaining the result.
CHAPTER 11 OF 16 · Question and transfer
11. Recognise what a lower peak can leave unresolved
A lower peak may be a useful outcome for a defined purpose, but it does not describe the entire event. The duration of elevated flow, total volume, downstream timing, water quality and ecological effects may still matter. A proposal can improve one measure while requiring further examination of another.
In the storage exercise, outflow continues after inflow has ended. Someone reading only the maximum might describe a successful reduction. Someone considering the full receiving period also asks how long the outflow persists and whether another contribution arrives during that extension. Both questions are legitimate, and neither can be settled by the maximum alone.
Similarly, reducing a peak at one location does not prove that every downstream location improves by the same amount. Additional tributaries, local runoff and routing can alter the combined response. A catchment explanation needs the boundary of the claim.
Keep flood hazard and flood consequence distinct. Flow and water level help describe physical conditions. Risk also involves what is exposed and vulnerable, together with other relevant arrangements. The Flood Risk Engineering guide develops that wider route.
For a learner, ask for a three-part conclusion: what changed in the supplied model, what stayed the same and what remains unknown. For example, the outflow peak is lower under the stated release rule, the total eventual released volume is unchanged, and the real downstream consequence is not established. This format encourages a clear result without an inflated promise.
CHAPTER 12 OF 16 · Question and transfer
12. Diagnose missing records and misleading summaries
A missing interval is unknown, not automatically zero. If a station records no value during a storm peak, inserting zero reduces the calculated volume and may erase the highest flow. Leaving the gap visible is more honest than drawing a smooth curve that implies observation.
Suppose four hourly interval averages are 1, missing, 3 and 1 m³/s. The known intervals contribute 18,000 m³. That is the volume represented by the available intervals, not the complete four-hour total. A gap-filling method would need its own assumptions and uncertainty.
A daily mean can also conceal short-lived variation. If the question concerns an hourly peak, the daily series may be too coarse even when it contains no missing days. Data completeness and data resolution are different questions.
Check revisions as well. Public river data can be provisional or subsequently corrected. Preserve the source, extraction date and relevant status when using a record. A later revised value should not be compared with an earlier version as though a physical change had occurred between the two downloads.
Look for mismatched units, timestamps at interval starts versus ends, and cumulative totals misread as interval amounts. Each can create a plausible-looking graph with the wrong meaning. A cumulative volume should generally increase as more volume is counted; a discharge series can rise and fall. The label matters.
For a practical evidence check, give students a graph and its table. Ask them to reproduce one interval volume and locate one uncertainty. The ability to challenge a neat presentation through its underlying definitions is a stronger skill than merely describing the curve as rising and then falling.
CHAPTER 13 OF 16 · Question and transfer
13. Build a learning sequence that survives a new context
Begin with a steady-flow rectangle. Ask what 2 m³/s over ten seconds means, and check that the answer is 20 m³. Then change seconds to minutes. The learner should notice the unit conversion rather than multiply the visible numbers unchanged.
Next introduce two unequal intervals. Ask why a simple average of the rates can mislead. Only after that relationship is secure should the learner compare the equal-volume Brook A and Brook B events. This progression keeps a new difficulty attached to an already understood quantity.
For stabilisation, mix tasks. One asks for total volume, another for highest interval average, another for equivalent depth and another for the additional information needed to identify an instantaneous peak. Do not label every question with the method it requires. The learner needs to choose what the question actually asks.
For extension, supply two tributaries with a time offset and one missing interval. Ask for a bounded conclusion rather than a complete numerical answer at any cost. A strong response can identify the part that is calculable and explain why the remainder is unresolved.
Finally change the context to electricity demand or vehicles arriving at a depot. A rate through time can still produce a total, and the same total can still hide a peak. The physical meaning and relevant constraints differ. Transfer means preserving the mathematical relationship while rebuilding the domain interpretation.
This is the connection with the Gold Standard of Education. Understanding appears when a student can explain the same structure in a new setting and identify which assumptions no longer travel with it.
CHAPTER 14 OF 16 · Question and transfer
14. Compare duration above a stated reference flow
A peak and a total can still leave an important time question unanswered: how long does the supplied flow remain above a specified reference? The reference must have a stated purpose. In this teaching exercise it is an arbitrary comparison value of 4 m³/s, not a channel capacity, warning threshold or design requirement.
Return to Brook A’s hourly averages of 1, 3, 3 and 1 m³/s, and Brook B’s 0, 1, 6 and 1. If the model explicitly treats each hourly rate as constant throughout that interval, Brook A spends no time above 4, while Brook B spends one hour above 4. Both still deliver the same 28,800 m³ total.
The modelled volume above that reference during Brook B’s high interval is (6 − 4) × 3,600 = 7,200 m³. This is the area between the modelled discharge line and the arbitrary reference where the discharge exceeds it. It is not automatically an overflow volume. An actual overflow depends on the physical system and its behaviour, which the reference line does not describe.
Now consider a third fictional sequence, 0, 4.5, 4.5 and 0 m³/s, also represented as constant within each hour. It spends two hours above 4. Its maximum is lower than Brook B’s, but its above-reference duration is longer. Its above-reference volume is (0.5 + 0.5) × 3,600 = 3,600 m³, smaller than Brook B’s.
Three comparisons now differ: peak, duration above reference and volume above reference. Brook B has the higher peak and larger above-reference volume; the third sequence has the longer above-reference duration. Asking which is worse without identifying the actual consequence would hide the question rather than answer it.
Change one assumption. Suppose the table values are hourly averages, but the flow is allowed to vary within each hour. An hour averaging 3 m³/s can include a brief period above 4. An hour averaging 4.5 can include some time below 4. The exact duration above the reference can no longer be read from the averages alone.
That distinction is subtle and useful. Interval averages are sufficient to calculate interval volume, because the average is defined through the interval’s total. They are insufficient to reconstruct every feature of the within-interval shape. The earlier constant-rate version added a modelling assumption to answer the duration question.
For a learner exercise, ask for two possible within-hour patterns with the same average of 3 m³/s. One can remain at 3 throughout. Another can use 6 for half the hour and 0 for the other half. Both total 10,800 m³ during the hour. Their peak and time above 4 differ.
The check is not asking learners to predict a real river. It asks them to understand which information an average preserves and which it discards. That reasoning travels to sound exposure, electricity demand and service queues, where equal totals can likewise conceal different concentrations through time.
When reporting a threshold-based measure, state the reference, its source or teaching purpose, the time resolution, interpolation or shape assumptions, and the physical claim actually supported. A line on a graph becomes useful when its meaning is explicit. It becomes misleading when a convenient numerical reference quietly acquires the authority of a real engineering limit.
CHAPTER 15 OF 16 · Questions and onward routes
15. Questions about peaks, totals and rainfall
Can equal runoff volumes produce different peaks?
Yes. The Brook A and Brook B examples distribute equal volume differently through time. The highest interval averages differ even though the four-hour means and totals agree.
Is the highest table value the true peak?
It is the highest value resolved by that table. If values are hourly averages, a brief instantaneous maximum may be higher. Describe the record’s interval and method before calling the result a peak.
Can I convert rainfall directly into a river peak?
Rainfall input alone does not determine the outlet hydrograph. The contributing area, pathways, storage, prior conditions, timing and appropriate model or observations matter. This guide supplies arithmetic examples rather than a forecasting method.
Does detention remove runoff volume?
The simple teaching storage account shifts release through time and conserves eventual volume. Real systems may involve additional pathways, but those need their own account. A lower peak should not be described as water disappearing.
Why not add the maximum flow of each tributary?
Their maxima may occur at different times and locations. Combine compatible simultaneous contributions under a stated routing model. Separate maxima do not necessarily describe a condition the downstream system experienced.
What is the first thing to check in a river graph?
Identify the station, quantity, units, period, interval meaning and data status. Then ask which question the record can answer. That sequence prevents a polished graph from carrying a claim beyond its evidence.
CHAPTER 16 OF 16 · Questions and onward routes
16. Continue through water, engineering and source routes
For land and processes, follow Watersheds, Hydrology, Rivers and Rainfall. The first defines the contributing land; the others deepen the pathways, channel processes and rainfall observations used in a hydrograph explanation.
For measurement interpretation, read River Rating Curves and Uncertainty and Streamflow, Rating Curves and River Discharge. For the urban application, Flood Drainage, Rainfall Intensity and Storage and The Detention Tank develop different physical and practical questions from this guide’s worked hydrograph comparisons.
For storage and decisions, use Reservoirs, Water Resources Engineering and Flood Risk Engineering. For receiving environments, explore Wetlands and Ecology.
The How X Works library offers wider connections. Return to Learning That Transfers into Life when planning the teaching sequence.
The USGS resources linked in chapters one and two support the limited definitions of hydrographs, stage, discharge and streamflow measurement. USGS: Watersheds and Drainage Basins supports the contributing-land context. Source pages were checked on 21 October 2026. Every Brook A, Brook B, tributary and storage calculation is original fictional teaching material. No real waterway, including any illustrative photograph, should be inferred to have these modelled conditions.