Storm Hydrographs and Flood Analysis

Spec mapping: AQA 7037, Paper 1 (Physical), §3.1.1 — runoff variation and the storm hydrograph, the impact of physical and human factors (notably urbanisation) on the drainage-basin cycle, and the consequences of change. This depth lesson treats the hydrograph quantitatively, develops flood-frequency / recurrence-interval analysis, and anchors the theory in a real UK flood. AOs exercised: AO1 (hydrograph anatomy, Manning's equation, the Weibull formula), AO2 (explaining why urbanisation and antecedent conditions reshape the hydrograph), AO3 (calculating recurrence intervals and exceedance probabilities, interpreting hydrograph data). Synoptic links run strongly to Hazards (the flood hazard, risk and management) and to Changing places (vulnerability of urbanised floodplains).

A storm hydrograph plots river discharge at a gauging station against time during and after a rainfall event. It is the integrated output of everything in the previous lesson — infiltration, soil moisture, antecedent conditions, baseflow — and reading it fluently, in numbers rather than adjectives, is one of the most heavily rewarded quantitative skills in the specification.

---

Storm Hydrograph Components

flowchart LR
  RAIN[Rainfall hyetograph] -->|lag time| PEAK[Peak discharge Qp]
  BASE[Baseflow] --> RISE[Rising limb]
  RISE --> PEAK
  PEAK --> FALL[Falling / recession limb]
  FALL --> BASE2[Return to baseflow]
  RISE -.bounded above baseflow.-> SF[Stormflow / quickflow volume]
  FALL -.bounded above baseflow.-> SF

• Baseflow: the sustained groundwater-fed discharge, shown as the gently rising line beneath the peak.
• Rising limb: discharge climbing in response to rainfall; a steep limb signals rapid quickflow delivery.
• Peak discharge ($Q_{p}$Qp​): the maximum discharge, in cumecs (m³/s).
• Lag time: the interval between peak rainfall and peak discharge — the single most diagnostic feature, because it measures how long the catchment takes to route water to the gauge.
• Falling (recession) limb: discharge declining as stores drain; its gradient reflects storage and permeability.
• Stormflow (quickflow): the volume between the hydrograph and the projected baseflow line — the storm's runoff contribution.
• Time to peak: from the start of the rise to $Q_{p}$Qp​.

Flashy versus subdued hydrographs

Feature	Flashy hydrograph	Subdued hydrograph
Rising limb	Steep	Gentle
Peak discharge	High	Low
Lag time	Short (< ~6 h)	Long (> ~12 h)
Recession limb	Steep	Gentle, extended
Stormflow:baseflow	High	Low
Typical catchment	Small, steep, impermeable, urban	Large, gentle, permeable, vegetated

---

Factors Affecting Storm Hydrographs

Basin (drainage-basin) characteristics

Factor	Effect on the hydrograph
Basin size	Larger basins → longer lag, lower peak per unit area (water travels further)
Basin shape	Circular basins → flashier (tributaries deliver simultaneously); elongated basins spread runoff over time
Drainage density	High density → more channels → faster delivery → flashier
Slope	Steeper → faster overland flow → shorter lag
Soil	Clay (impermeable) → flashy; sand/chalk (permeable) → subdued
Geology	Permeable bedrock (chalk, limestone) absorbs water → longer lag, lower peak, higher baseflow
Vegetation	Interception + infiltration → longer lag, lower peak
Antecedent moisture	Saturated soil → reduced infiltration → faster, higher peak

Storm characteristics

Factor	Effect
Intensity	High intensity exceeds infiltration capacity → infiltration-excess (Hortonian) overland flow
Duration	Prolonged rain saturates soils → saturation-excess overland flow, contributing area expands
Type	Convective (intense, localised) vs frontal (moderate, widespread) produce different shapes and durations
Snow vs rain	Snow delays delivery; rapid melt (especially rain-on-snow) yields delayed but large peaks

The interaction between storm input and catchment state is the heart of A-Level flood analysis: a high-intensity storm on a saturated, urbanised, steep, circular catchment is the worst-case combination, and it is exactly the combination behind the case study below.

---

Bankfull Discharge and the Dominant Discharge

Bankfull discharge is the maximum a channel conveys without overtopping. It is a critical threshold:

• Discharge above bankfull spills onto the floodplain — i.e. flooding.
• In most natural rivers bankfull recurs about every 1–2 years (Leopold, Wolman & Miller, 1964).
• It is the dominant (channel-forming) discharge — large enough to do significant geomorphic work yet frequent enough that its cumulative effect shapes the channel over time. The product of magnitude and frequency, not the rare catastrophic flood alone, sculpts channel form (the magnitude–frequency concept, a synoptic link to Hazards).

Exam application: Because bankfull recurs every 1–2 years, floodplains are inundated frequently by definition — which is precisely why building on them is inherently risky and why floodplain land is naturally fertile (silt deposition).

---

Flood Recurrence Intervals

The recurrence interval (return period, $T$T) is the average interval between floods equalling or exceeding a given magnitude.

The Weibull formula

$T = \frac{n+1}{m}$T=mn+1​

where $n$n is the number of years of record and $m$m is the rank of the event (1 = largest).

Worked example. With 50 years of annual maximum data, the largest flood ($m=1$m=1) has

$T = \frac{50+1}{1} = 51\ \text{years}.$T=150+1​=51 years.

The second largest ($m=2$m=2): $T = 51/2 = 25.5$T=51/2=25.5 years; the tenth largest: $T = 51/10 = 5.1$T=51/10=5.1 years. Plotting discharge (log scale, $y$y) against $T$T (log scale, $x$x) gives a near-straight flood-frequency curve, which can be extrapolated to estimate the discharge of events not yet observed (the design "1-in-100-year" flow, say).

Exceedance probability — the crucial AO3 point

The annual exceedance probability is the reciprocal of the return period:

$P_{(\text{exceed in any year})} = \frac{1}{T}.$P(exceed in any year)​=T1​.

A 100-year flood therefore has a 1% chance in any single year — not a guaranteed 100-year gap. The probability of at least one such flood over a period of $L$L years is:

$P_{(\geq 1\ \text{in}\ L\ \text{years})} = 1 - \left(1 - \frac{1}{T}\right)^{L}.$P(≥1 in L years)​=1−(1−T1​)L.

Over the lifetime of a 75-year mortgage-era house exposed to a 100-year flood:

$1 - \left(1 - \frac{1}{100}\right)^{75} = 1 - 0.99^{75} = 1 - 0.470 = 0.530,$1−(1−1001​)75=1−0.9975=1−0.470=0.530,

so there is a 53% chance of at least one "1-in-100-year" flood within 75 years. This counter-intuitive result — that a "rare" flood is more likely than not over a lifetime — is exactly the kind of manipulation that secures top AO3 marks.

Exam tip: "1-in-100-year flood" means 1% annual exceedance probability, not a fixed 100-year interval; two such floods can fall in consecutive years (as parts of the UK saw in the 2010s). Stating this precisely is a reliable discriminator.

---

Manning's Equation

Average channel velocity can be estimated with Manning's equation:

$V = \frac{1}{n}\,R^{2/3}\,S^{1/2}$V=n1​R2/3S1/2

where $V$V is mean velocity (m/s), $n$n is Manning's roughness coefficient (dimensionless), $R = A/P$R=A/P is the hydraulic radius (m), and $S$S is the energy slope (m/m).

Channel type	Manning's $n$n
Smooth concrete	0.012
Earth channel, straight	0.020–0.025
Natural stream, clean	0.030–0.040
Natural stream, vegetated	0.050–0.100
Floodplain with trees	0.100–0.200

Worked illustration. For a channel of $R = 1.5$R=1.5 m, $S = 0.001$S=0.001, and $n = 0.035$n=0.035:

$V = \frac{1}{0.035}\,(1.5)^{2/3}\,(0.001)^{1/2} = 28.6 \times 1.31 \times 0.0316 \approx 1.18\ \text{m/s}.$V=0.0351​(1.5)2/3(0.001)1/2=28.6×1.31×0.0316≈1.18 m/s.

If the channel is straightened and lined with concrete ($n$n falls to 0.012), velocity rises to ~3.45 m/s — roughly tripling — illustrating exactly how channelisation accelerates flow and transfers flood risk downstream. Manning's equation thus quantifies a key management trade-off rather than merely describing it.

---

Urbanisation Effects on Flood Hydrology

Urbanisation reshapes the hydrograph through several reinforcing mechanisms:

• Impermeable surfaces (roads, roofs, car parks) prevent infiltration; in a heavily urbanised catchment 80–95% of rainfall becomes runoff, versus 10–30% under forest.
• Efficient artificial drainage (gutters, storm drains, sewers) delivers water to channels far faster than natural pathways, slashing lag time.
• Channel modification (culverting, straightening, concrete lining) raises velocity (Manning's $n$n falls).
• Reduced interception and evapotranspiration as vegetation is removed.
• Soil compaction during construction lowers infiltration capacity.

Quantified effects: peak discharge can rise 2–5×; lag time can shorten by 50–75%; total runoff volume increases (less ET and recharge); and baseflow often falls because reduced infiltration starves the groundwater store — so urban rivers can be simultaneously more flood-prone and lower in drought.

Supporting example: the Upper Mersey, NW England

The Upper Mersey basin urbanised heavily through the twentieth century (Greater Manchester conurbation). Documented changes include more frequent minor flooding, shortened lag times, lower summer baseflows from reduced recharge, and elevated sediment loads from construction phases — a textbook demonstration of the urban hydrograph transformation. Sustainable urban drainage systems (SuDS) — permeable paving, swales, retention ponds, green roofs — are now used to partially reverse these effects by restoring infiltration and storage.

---

Case Study (real data): Storm Desmond, December 2015, Cumbria

Storm Desmond struck north-west England on 4–6 December 2015, drawing a deep "atmospheric river" of subtropical moisture onto the saturated, steep Cumbrian fells — a near-perfect worst-case combination of intense input and primed catchment.

The numbers (Met Office / Environment Agency):

Variable	Value
24-hour rainfall at Honister Pass	341.4 mm — a UK record for 24 h
48-hour rainfall at Thirlmere	~405 mm
River Eden peak at Sheepmount, Carlisle	among the highest on record
Properties flooded (NW England)	~5,200 in Cumbria; ~16,000+ across the storm region in total over winter 2015–16
Estimated UK economic cost (winter 2015–16 storms)	~£1.6 billion

Why the hydrograph was so flashy. Antecedent conditions were critical: November 2015 had been exceptionally wet, so soils were already at or above field capacity and the soil-moisture deficit was effectively zero. With infiltration capacity exhausted, almost all of the record rainfall became saturation-excess overland flow, routed rapidly down steep, high-drainage-density valleys into the Eden, Kent and Derwent. Lag times were short and peaks enormous; defences in Carlisle that had been upgraded after the 2005 flood were overtopped because the event exceeded their design standard.

Management response and evaluation. The flood exposed the limits of hard engineering designed to a fixed standard: Carlisle's defences had been built to roughly a 1-in-100-year standard yet were overtopped. The aftermath accelerated investment in catchment-scale Natural Flood Management (upland tree planting, leaky dams, peat restoration in the Eden catchment) and a shift towards "making space for water" rather than relying solely on walls. It also fuelled debate about non-stationarity: a flood of this magnitude, following the 2005 and 2009 Cumbrian floods, suggested the design return periods were already out of date.

Synoptic link (Hazards): Storm Desmond illustrates the risk equation — hazard (extreme rainfall) × exposure (floodplain settlement) × vulnerability (defences exceeded) — and the tension between hard and soft engineering that recurs in the management lessons.

---

AO3 skills exemplar: ranking annual maxima

Given annual maximum discharges (m³/s) for a station with $n = 9$n=9 years: 210, 185, 340, 160, 295, 175, 410, 150, 230.

Manipulate. Rank them ($m=1$m=1 largest): 410, 340, 295, 230, 210, 185, 175, 160, 150. The recurrence interval of the largest:

$T = \frac{n+1}{m} = \frac{10}{1} = 10\ \text{years},$T=mn+1​=110​=10 years,

and its annual exceedance probability is $1/10 = 10\%$1/10=10%. The 295 m³/s event ($m=3$m=3) has $T = 10/3 = 3.3$T=10/3=3.3 years.

Explain and evaluate. With only nine years of record, the largest observed flood is assigned a return period of just 10 years — yet the true 1-in-100-year flow is far higher and entirely unsampled. Short records systematically underestimate extreme floods, which is why the Weibull plot must be extrapolated and why design standards carry large uncertainty. Combined with non-stationarity (below), this means historic data alone is an unreliable guide to future risk — a key evaluative point.

---

Flood Frequency Analysis: Limitations and Non-Stationarity

• Short records cannot constrain rare events (the example above).
• Non-stationarity: the assumption that hydrological statistics are fixed in time is undermined by climate change and land-use change. Milly et al. (2008) argued, in a landmark paper, that "stationarity is dead" — historic flood frequencies can no longer be assumed to represent future risk.
• Land-use change (urbanisation, deforestation, drainage) alters the rainfall–runoff relationship over the record period itself.
• Measurement error rises exactly when it matters most: gauges can be damaged, bypassed or submerged during extreme events, truncating the very peaks that matter.

---

A contrasting case study (real data): the Somerset Levels, winter 2013–14

Where Storm Desmond was a flashy, high-energy upland flood, the Somerset Levels flood of winter 2013–14 was the opposite end of the hydrograph spectrum — a prolonged, low-lying, slow-onset inundation — and the contrast is itself an examinable analytical point.

The setting. The Somerset Levels and Moors are a very low-gradient floodplain and former wetland (much of it at or below high-tide level), drained for centuries by an artificial network of rivers (notably the Parrett and Tone), rhynes and pumping stations. With almost no gradient, the rivers have very low velocity (Manning's equation: tiny $S$S) and limited capacity, so water leaves the system only slowly.

The event. The winter of 2013–14 was the wettest on record for parts of southern England, with a relentless succession of Atlantic depressions. Some areas received around 350 mm of rain over December–January, roughly double the average. The flat, saturated catchment could not drain: river levels stayed above bankfull for weeks.