Introduction to

hydrological modelling
with HEC-HMS
Part 2: Introduction to hydrological modelling

Dr. Claudia Bertini, [email protected], with slide courtesy of Dr. Bhattacharya

 Hydroinformatics: managing flow of information
Data  Models  Knowledge  Decisions

Data modelling, Access to

Earth observation, Numerical Weather Decision
integration with modelling
monitoring Prediction Models support
hydrologic and results
hydraulic models

Map of flood probability

Modellng

• Represent one (or more) physical processes through mathematical equations

∂Q ∂  Q 2  ∂h
+   + gA − gAS o + gAS f = 0
∂t ∂x  A  ∂x
Rainfall-Runoff
process and
modelling
Rainfall-Runoff process (1/2)
Rainfall-Runoff process (2/2)
• Complex
• Depends upon many factors
• Geology
• Terrain
• Land use
• Soil type
• Constitutes of the following processes:
• Infiltration
• Overland flow
• Unsaturated and saturated zone flow
• Evapotranspiration
Rainfall-Runoff modelling (1/18)
• Primary model for flood forecasting and flood management
• 5 categories of R/R models:
• Distributed physically based
• Quasi-physical conceptual/lumped
• Systems based models
• Rational method
• Unit hydrograph
• Data-driven (Artificial Neural Networks, etc.)
• Hybrid
Rainfall-Runoff modelling (2/18)

Distributed physically based

• Main physical processes considered are:
hydrology and hydraulics of ground surface,
infiltration in the unsaturated zone, and
groundwater in the saturated zone
• Mathematical representation
• Equations and parameters distributed in
space
• Example: MIKE-SHE
Rainfall-Runoff modelling (3/18)
Lumped conceptual
• Only temporal variation, no spatial variation
is considered
• Catchments are often divided into sub-
catchments to make it semi-distributed
• Popularly used
• Well-known tools: HEC-HMS, NAM (or
MIKE-NAM), HBV, SWAT
Rainfall-Runoff modelling (4/18)
Data-driven actual Y
output
value
• A complex data-driven modelling
method may be used to approximate model
the function 𝑌𝑌 = 𝑓𝑓 (𝑋𝑋) predicts
• Example of a complex method: new output
artificial neural networks value
• The internal parameters are found by
solving an optimization problem X
(training)
• The model then predicts output for new
the new input without actual input
knowledge of what drives 𝑌𝑌 value
 Rainfall-Runoff modelling (5/18)
Event-based vs Continuous models
P Event-based P Continuous
t t
Q Q t
t
Event-based models:
• Developed to represent a specific event, e.g. increasing of discharge due to a
high rainfall event

Continuous models:
• Developed to represent all-year processes
Rainfall-Runoff modelling (6/18)
Catchment modelling

• It refers to modelling spatially distributed hydrological processes over

the entire catchment

• Any catchment model can be used for rainfall-runoff modelling

• To compute runoff for forecasting: use rainfall-runoff models (simpler,

lumped-conceptual, empirical or data-driven models)

• To carry out scenario analysis: use catchment models

Rainfall-Runoff modelling (7/18)
Rainfall loss and excess rainfall Excess rainfall
(red)

Rainfall (mm/hr)
Rainfall (mm/hr)

Initial abstraction Rainfall loss

(blue)

t t

• Rainfall fallen on the catchment surface is stored in: vegetation, surface

depressions (e.g. lakes, ponds), infiltration
• Above three together: rainfall loss
• Excess rainfall (also effective rainfall) = Total rainfall – rainfall loss
• If there is excess rainfall that ends up as direct runoff
• Initial abstraction: rainfall amount before excess rainfall starts
Rainfall-Runoff modelling (8/18)
Excess rainfall and direct runoff

Discharge
Excess rainfall
Rainfall (mm/hr)

Direct runoff
(red)
Base flow

t Rainfall loss
(blue) t
Initial
abstraction
 Rainfall-Runoff modelling (9/18)
Steps in lumped conceptual rainfall-runoff modelling

Discharge
Rainfall (mm/hr)

Excess rainfall (red)

Direct runoff

Base flow

t Rainfall loss (blue)

t
Initial abstraction

• Estimate loss (blue part of rainfall)

• Compute excess rainfall = rainfall- rainfall loss (red part of rainfall)
• Estimate direct runoff (coming from excess rainfall) (red part of the hydrograph)
• Estimate baseflow (grey part of the hydrograph)
• Total flow = baseflow + direct runoff
 Rainfall-Runoff modelling (10/18) SCS is only good for event based modelling and not
continuous modelling

Rainfall losses estimation: SCS Curve Number approach

• A number of approaches are there to

Direct runoff
M3/s
estimate rainfall loss
• One of them is due to Soil Conservation
Service (SCS) method from USA
• SCS carried out a large number of
experiments on catchments
• SCS prepared rainfall vs direct runoff curves 950

• From the chart you can find direct runoff for 600
any rainfall 300
100
Direct runoff (mm) = Excess rainfall (mm) = Rainfall – Rainfall loss (mm) 10 20 30 40 Rainfall (mm)
 Rainfall-Runoff modelling (10/18)
Rainfall losses estimation: SCS Curve Number approach
• A number of approaches are there to

Direct runoff
M3/s
estimate rainfall loss
• One of them is due to Soil Conservation
Service (SCS) method from USA
• SCS carried out a large number of
experiments on catchments
• SCS prepared rainfall vs direct runoff curves 950

•Assignment:
From the chart you can find direct runoff for 600
1. any
Rainfall
rainfall= 40 mm, find out the direct runoff (based on SCS method)
300
2. Rainfall = 20 mm, what is the rainfall loss? 100
Direct runoff (mm) = Excess rainfall (mm) = Rainfall – Rainfall loss (mm) 10 20 30 40 Rainfall (mm)
 Rainfall-Runoff modelling (11/18)
Rainfall losses estimation: SCS Curve Number approach

The relation depends upon:

Direct runoff

- Land use
M3/s

- Soil type

• This means that we need many

curves!
Rainfall
• Each curve corresponds to a particular
(mm)
combination of land use and soil type
Rainfall-Runoff modelling (12/18)
Rainfall losses estimation: SCS Curve Number approach

Assignment:
1. CN= 45, rainfall = 9 inches,
find out the direct runoff
2. CN = 95, rainfall = 6 inches,
find out the direct runoff
 Rainfall-Runoff modelling (13/18)
Rainfall losses estimation: SCS Curve Number approach
• The Curve Number (CN) model of the Soil Conservation Service is a popular
method in hydrology in determining excess precipitation in a catchment
• The CN of a catchment is estimated mainly as a function of land use and soil
type
• The maximum retention capacity (S) of a catchment is empirically connected
with the Curve Number
• This empirical approach is widely used by hydrologists, particularly in the USA

Further reading:
Chow, V. T., Maidment, D.R., and Mays, L.W. (1988). Applied hydrology (pp 147-151). McGraw-Hill, New York.
Mishra, S.K. and Singh, V.P. (2010). Soil Conservation Service Curve Number (SCS-CN) Methodology. Kluwer Academic Pub,
Dordrecht.
 Rainfall-Runoff modelling (14/18)
Rainfall losses estimation: SCS Curve Number approach
• 𝑃𝑃 = total rainfall Ia = loss before runoff starts e.g infiltration, evaporation, depression storage
• 𝐼𝐼𝑎𝑎 = Initial abstraction

Rainfall rate (mm/hr)

Excess
• 𝐹𝐹𝑎𝑎 = Additional depth of water retained in the rainfall (red)
(Pe)
catchment after the runoff begins
• Rainfall loss 𝐼𝐼𝑎𝑎 + 𝐹𝐹𝐹𝐹
• 𝑃𝑃𝑒𝑒= effective rainfall
• 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 = 𝑃𝑃 = 𝐼𝐼𝑎𝑎 + 𝐹𝐹𝐹𝐹 + 𝑃𝑃𝑃𝑃 t

• 𝑆𝑆 = Max potential retention Rainfall loss

beyond initial
• Hypothesis: Initial abstraction (Ia) abstraction
𝐹𝐹𝑎𝑎 𝑃𝑃𝑒𝑒 (purple) (Fa)
=
𝑆𝑆 𝑃𝑃 − 𝐼𝐼𝑎𝑎
 Rainfall-Runoff modelling (15/18)
Rainfall losses estimation: SCS Curve Number approach
𝐹𝐹𝑎𝑎 𝑃𝑃𝑒𝑒 • The precipitation excess and obviously the runoff is zero
=
𝑆𝑆 𝑃𝑃 − 𝐼𝐼𝑎𝑎 until the accumulated rainfall > 𝐼𝐼𝑎𝑎
• Using data of many small experimental catchments, an
empirical relationship of 𝐼𝐼𝑎𝑎 and 𝑆𝑆 has been developed by
≡ 𝑃𝑃 − 𝐼𝐼𝑎𝑎− 𝑃𝑃𝑒𝑒 𝑃𝑃𝑒𝑒 SCS:
=
𝑆𝑆 𝑃𝑃 − 𝐼𝐼𝑎𝑎 𝐼𝐼𝑎𝑎 = 0.2 𝑆𝑆
• S of a catchment is empirically related through the curve
2 number (CN) (valid for SI)
≡ 𝑆𝑆𝑆𝑆𝑒𝑒 = 𝑃𝑃 − 𝐼𝐼𝑎𝑎 − 𝑃𝑃𝑒𝑒 (𝑃𝑃 − 𝐼𝐼𝑎𝑎 )
25400 − 254 CN
(P − I a )2 S=
≡ Pe = CN
P − Ia + S
Assignment: Total rainfall = 50 mm, CN= 50, find out excess rainfall
Rainfall-Runoff modelling (16/18)
Direct runoff

Rainfall event Direct runoff

Discharge per unit

excess rainfall

t
 Rainfall-Runoff modelling (17/18)
Direct runoff with Unit Hydrograph
Approximated direct runoff using unit hydrograph
Rainfall event Direct runoff
t
r
t
p Qp
Discharge per unit

Discharge per unit

t
excess rainfall

excess rainfall
t t
Rainfall-Runoff modelling (18/18)
Total river discharge
Total discharge Approximated total discharge
Rainfall event

tr

t
p
Q
p

Total discharge
Total discharge

Baseflow Baseflow

t t
Thank you!