Overview of Hydrological Modeling

C.M.Bhatt
[email protected]
Scientist
Disaster Management Studies Department
Indian Institute of Remote Sensing
ISRO | Dept. of Space | Govt. of India
Hydrological Cycle
In meteorology, Precipitation is any product of the condensation
of atmospheric water vapour that falls under gravity.

Precipitation occurs when a portion of the atmosphere becomes

saturated with water vapor, so that the water condenses and
"precipitates“ (fog and mist are not precipitation but suspensions,
because the water vapor does not condense sufficiently to
precipitate).
 Interception
Interception refers to precipitation
that does not reach the soil, but is
instead intercepted by the leaves,
branches of plants and the forest
floor.
• Canopy interception
• Forest floor interception
P - Precipitation
C - Canopy interception
L - Litter interception
T - Through fall
Rain moves through the canopy in two-ways
S – Stem flow
R - Net rain entering into surface

R=P-I
R=P-(C+L)
Components of interception
 Interception
Factors influencing interception: Interception losses equation
Li = S + KEt
(i) Type of Vegetation:
Li is the total volume of water intercepted.
(ii) Wind Velocity: S is the interception storage (0.25-1.25mm).
(iii) Duration of Storm: K is the ratio of the surface area of the leaves to the
(iv) Intensity of Storm: area of the entire canopy.
E is the rate of evaporation (mm/hr) during the
(v) Season of the Year: precipitation event, and
(vi) Climate of the Area: t is time (hr).

Interception loss is large for small rainfall and

levels of to constant value for larger storms.

Coniferous trees interception>deciduous ones

Interception can amount up to 15-50% of

precipitation, which is a significant part
of the water balance. Interception Rate versus Time. (Source:
Singh, 1994)
 Depression Storage
Refers to small low points in undulating terrain that can store
precipitation that otherwise would become runoff. The
precipitation stored in these depressions is then either removed
through infiltration into the ground or by evaporation.

Factors Affecting Depression Storage

(1) Nature of terrain (Linsley 1982)
(2) Slope
(3) Type of soil surface
(4) Land use
(5) Antecedent rainfall
(6) Time

Typical values for depression storage range from 1 to 8 mm (0.04 to

0.3 in) with some values as high as 15 mm (0.6 in) per event.
 Infiltration & Percolation
The process by which water on the ground surface enters
the soil Infiltration.

The infiltrated water first meets the soil moisture

deficiency if any & excess water moves vertically
downwards to reach the groundwater table. This vertical
movement is called Percolation.
 Infiltration
Factors affecting:
Soil properties (texture, porosity, hydraulic conductivity)
Condition of soil surface; Vegetative cover
Soil temperature; Antecedent soil moisture

Infiltration Rate (f): Velocity or speed at which water enters into the soil.
Usually measured by depth (mm) of water layer that can enter soil in 1 hour.

Infiltration Capacity (fc) : Max rate at which ground can absorb water
where
f = fc when i ≥ fc fc = infiltration capacity (cm/hr)
f = i when i < fc i = intensity of rainfall (cm/hr)
f = rate of infiltration (cm/hr)
Field Capacity: Volume of water that soil can hold a few days after having been
wetted and after free drainage has stopped
Infiltration
Infiltration is the movement of water into the soil

If rainfall intensity < infiltration capacity, If rainfall intensity > infiltration capacity

infiltration capacity = rainfall intensity infiltration capacity = Max. field capacity

 Surface runoff
Surface runoff or overland flow is that water which travels over the
ground surface to a drainage channel.

Most surface runoff flows to first-order channels because they

collectively drain the greatest area of the drainage basin.

SR also includes that precipitation that falls directly on water flowing

in the channel.

Surface runoff is principal contributor to the peak discharge from a

storm event. First to reach the channel and, hence, forms the rising
limb and peak of the hydrograph.
 Interflow
Interflow also called subsurface storm flow, is that surface water
that infiltrates the surface layer and moves laterally beneath the
surface to a channel.
Direct Runoff: It is that part of runoff which enters the stream
immediately after the precipitation. It includes surface runoff, prompt
interflow and precipitation the channel surface. In the case of snow-
melt, the resulting flow entering the stream is also a direct runoff.
Sometimes terms such as direct runoff and storm runoff are used to
designate direct runoff

Base flow: The delayed flow that reaches a stream essentially as

groundwater flow is called base flow. Base flow is a portion of stream
flow that is not directly generated from the excess rainfall during a storm
event. In other words, this is the flow that would exist in the stream without
the contribution of direct runoff from the rainfall. The delayed flow that
reaches a stream essentially as groundwater flow is called base flow.

Estimation of base flow and direct runoff is useful to understand the

hydrology of a watershed, including interaction of surface and sub-surface
water, role of urbanization on runoff generation and the health of aquatic
habitat within a stream.
 Evapotranspiration (ET)
EVAPOTRANSPIRATION EVAPORATION + TRNSPIRATION

WATER MOVEMENT INTO WATER BODIES, SOIL, THROUGH THE

ATMOSPHERE FROM VEGETATIVE SURFACES PLANTS

Over 61% of rainfall is lost to evapotranspiration, so mapping its variance is

an important part of understanding the global water cycle

• Evapotranspiration (ET) is a crucial variable in the hydrologic cycle and one of

the major hydrological fluxes.

• ET provides a measure of water loss from land to the atmosphere over space and
time and varies significantly due to local controlling factors.

• Satellite based ET estimation method depends on remote sensing based

information of land cover, land surface temperature, albedo, vegetation
indices/Leaf Area Index and emissivity.
 Hydrograph
•A hydrograph is a graph showing the rate of flow versus time past
a specific point in a river, channel, or conduit carrying flow.

•The rate of flow is typically expressed in cubic meters or cubic

feet per second.
 Discharge

Volume of water flowing through a river channel; measured at any

given point in cubic metres per second.
Factors affecting river discharge: the storm hydrograph
 Hydrological Modeling
• Rainfall-runoff modeling is an
important aspect of hydrologic
analysis and design.

• Exact scale models of the hydro-

climatological systems to
perform experiments to
understand the nature of its
operations on rainfall and its
eventual transformation into
runoff not feasible.
• Mathematical modeling approach is an alternative path.

• A model is a simplified representation of a complex system and

helps in making decisions.
 Hydrological Modeling
• Hydrological modelling basically tries to simulate the natural
hydrological processes existing in a catchment.
• Hydrological models are essentially mathematical models where
the physical processes of hydrologic cycle are described by a set
of mathematical equations, logical statements, boundary
conditions and initial conditions, expressing relationships
between inputs, variables and parameters.

• A model’s structure determines how runoff is calculated. Some

are easily used with few variables, while others require a vast
number of interconnecting variables. Model structure varies from
simple to complex, based on the governing equations
• Not a replacement for field observations and accuracy of model
depends of input parameters correctly chosen and adjusted, to
extract maximum amount of information from the available data.
 Types of Hydrological Models

• Based on physical process:

(1) empirical (data driven models),
(2) conceptual
(3) physically based.

• Based on spatial representation:

(1) Lumped
(2) Semi-distributed
(3) distributed.

• Based on aspect of randomness:

(1) deterministic
(2) stochastic
 Empirical (black-box)
• Empirical models, sometimes called data-driven models, use
non-linear statistical relationships between inputs and outputs.

• They are observation-oriented and depend heavily on input

accuracy. For simple rainfall-runoff regression models, inputs are
rainfall and historical runoff, with outputs of runoff at a specific
location. General governing equation for empirical models

• Black box models, meaning very little is known about the

internal processes that control how runoff results are
determined.

• The function used to transform rainfall to runoff is either an

unknown procedure (as in machine learning) or without any
reference to the physical processes (as in CN method.)
 Empirical (black-box)
• Empirical runoff models are best used when other outputs are
not needed; for example: the distribution of runoff values
between upstream and downstream areas cannot be calculated
with this model type.

• Ungauged watersheds are also best modeled by an empirical

method due to lack of specific information about the watershed.

• Simplicity of implementation, faster computational times, and

cost effectiveness are reasons for empirical models.

• Since they are data-driven, input data are a main source of error
because input data distortion produces serious ramifications in
the modeled output.

• SCS-Curve Number used in SWAT, regression equations, and

machine learning used by Artificial and Deep Neural Networks.
 Conceptual Model (grey box)
• Conceptual models interpret runoff processes by connecting
simplified components in the overall hydrological process.

• They are based on reservoir storages and simplified equations of

the physical hydrological process, which provide a conceptual
idea of the behaviors in a catchment.

• Conceptual models represent the water balance equation with

the conversion of rainfall to runoff, evapotranspiration, and
groundwater. The general governing equations for conceptual
models are versions of the water balance equation which control
surface water and storage fluctuations.
 Physically based (white-box)
• Physical models, also called process-based or mechanistic
models, are based on the understanding of the physics related to
the hydrological processes.

• Physically based equations govern the model to represent

multiple parts of real hydrologic responses in the catchment.

• The general physics laws and principles used include water

balance equations, conservation of mass and energy,
momentum, and kinematics. St. Venant, Boussinesq’s, Darcy, and
Richard’s are some of the equations adopted by physical models.
 Physically based (white-box)
• Spatial and temporal variations within the catchment are
incorporated into physical models.

• A physical model has a logical structure similar to the real-world

system. The greatest strength of a physical model is the
connection between model parameters and physical catchment
characteristics which make it more realistic.

• They are best used when precise data are available, physical
properties of the hydrological processes are accurately
understood, and applied on fine scales due to computational
time.

• A large number of physical and process parameters are needed

to calibrate the model. VIC , MIKE SHE , PIHM
 Physically based (white-box)
• Models build on sound mathematical knowledge; attempt
to represent the natural phenomenon by physical laws
governing the conservation of mass and
momentum/energy.
• These are also called mechanistic models that include the
principles of physical processes. It uses state variables
which are measurable and are functions of both time and
space.
• The hydrological processes of water movement are
represented by finite difference equations. It does not
require extensive hydrological and meteorological data for
their calibration but the evaluation of large number of
parameters describing the physical characteristics of the
catchment are required.
 Physically based (white-box)
• In this method huge amount of data such as soil
moisture content, initial water depth, topography,
topology, dimensions of river network etc. are required.

• Physical model can overcome many defects of the

other two models because of the use of parameters
having physical interpretation.

• It can provide large amount of information even

outside the boundary and can applied for a wide range
of situations.

• SHE/ MIKE SHE model is an example.

 Hydrological Models: Based on physical process
Empirical Conceptual Physically-based
Metric/Data based/Black box Parametric model/Grey box Mechanistic model/White box
Include physical basis semi-
Observation oriented models;
empirical equations and based on Evaluation of parameters
Involve derive the value from
modelling of reservoirs describing physical
existing time series and
characteristics, based on the
mathematical equations
spatial distribution
Require data about the initial
Low consideration of features Parameters are derived from
state of the model and
and processes of the system calibration and field data
morphology of the catchment
Complex model. Require human
Low explanatory High and Really simple, can be easily
expertise and computation
depth predictive power implemented in computer code
capability
It cannot be generated in other Require large hydrological and Suffer from scale-related
catchments meteorological data problems
Calibration involves curve fitting
Valid within the boundary of make difficult physical
the given domain interpretation Valid for several situations
Small number of parameters is Simple model structure, Easy to Incorporate temporal and spatial
required, fast run time, high calibrate variability
accuracy
ANN, unit hydrograph, SCS-
HBV , TOP, Tank Models
Curve Number SHE or MIKE SHE model, SWAT
Hydrological Models: Based on spatial representation

Lumped Semi-distributed Distributed

Parameters partially allowed to
Spatial variability is disregarded; the
Procedure whole catchment is modelled as a unit vary in space dividing basin to Spatial variability is considered
small sub basins
Both averaged and specific data
Inputs All averaged data by catchment
by sub-catchment
All specific data by cell
Fast calculation time, Perform well at Represents important features in Physically related to hydrological
Strengths simulating average conditions the catchment processes
A lot of assumptions perform weakly
Averages data into sub- Data intense, long computational
Weaknesses for large areas, loss of spatial catchment, spatial resolution loss time
resolution
Conceptual and some physical
Empirical and conceptual models, ML Physically distributed models,
Examples (Machine Learning); SCS-CN based
models, TOP MODEL, SWAT,
MIKE SHE
SWMM
Sub-areas categorized by land use or cover type, soil type, and surface slope, a combination called hydrological reaction units (HRU)
Hydrological Models: Based on aspect of randomness
• Deterministic model: A deterministic hydrological model
is one in which the processes are modelled based on
definite physical laws and no uncertainties in prediction
are admitted.
• Deterministic models permit only one outcome from a
simulation with one set of inputs and parameter values.
This model does not consider randomness.
• In general, if the probability of the effect of variables on
the occurrence of a process is not considered and the
model does not follow the laws of probability, then this
model is deterministic.
Hydrological Models: Based on aspect of randomness
• Stochastic model: Stochastic models allow for some
randomness or uncertainty in the possible outcomes due
to uncertainty in input variables, boundary conditions or
model parameters.
• In stochastic hydrological models, different values of
output can be produced for a single set of input values. If
the probability of occurrence of the desired variables
and probability concepts are involved in the construction
of the model then it is stochastic.
 Calibration & Validation of Model
• Model calibration refers to the adjustment of
parameters of a particular model using historical input-
output records to reproduce the observed response of
the catchment within the range of accuracy specified in
the performance criteria.
• Model performance is typically evaluated from
comparison of simulated & observed discharge data in
terms of mean, std. deviation, maximum daily discharge
& total discharge using commonly used indices.
• Statistical indices used to evaluate model performance:
Coefficient of determination (R2 ), Nash and Sutcliffe
efficiency (E), index of agreement (d), relative error of
the stream flow volume (RE).
 Calibration & Validation of Model
• Model validation refers to the judgment on the
calibrated model performance to simulate the response
with sufficient accuracy for a period other than the
calibration period. Various statistical indices and graphs
based on the observed and computed output are used.
 Model Selection
Scale of Application:
Small-Lumped; Medium-Semi Distributed; Large-Distributed

Terrain/Relief:
Flat/Plain-Lumped; Moderate/Hilly-Semi Distributed; Mountainous-
Distributed

 Temporal Resolution:
Event based (single storm); Time-step(minutes to hours);Continuous
time step (daily)

Processes Integrated:
Rainfall-runoff, Glaciers, Snow, Groundwater, Evapotranspiration,
Reservoir management, Dynamic vegetation growth, Routing etc.
 Model Selection
Model Components
HEC-HMS(Hydrologic Engineering Semi-distributed model for event
Center-Hydrologic Modeling simulation
System)
SWMM Storm Water Management Semi-distributed model for
Model) continuous simulation.
SWAT (Soil and Water Distributed model for continuous
Management Tool) simulation.
WEPP (Water Erosion Prediction Distributed model for simulating
Project) USDA surface flow, water balance, plant
growth, erosion etc
TOP MODEL Distributed model for continuous
simulation
MIKE-SHE (Mike –Systems Distributed model for continuous
Hydrologique Europee) simulation of surface and
groundwater flow
 SWAT (Soil and Water Assessment Tool)
Input Data Required for SWAT modelling: General Methodology:

1. DEM
2. LULC
3. Soil Map
4. Daily Temperature
5. Daily Rainfall
6. Daily Relative Humidity
7. Daily Solar Radiation
8. Daily Wind Velocity
9. Daily Discharge
SCS-CN Method
 SCS-CN Method
The Soil Conservation Service Curve Number approach is frequently used empirical
methods to estimate direct runoff from a given rainfall event from small agricultural,
forest and urban watersheds.
For a given rainfall the ratio of actual infiltration (F) to the maximum possible
infiltration (S) is equal to the ratio of actual runoff (Q) to the maximum possible
runoff (R).
i.e _F_ = _Q_ or, _P-Ia-Q = _Q__
S R S P - Ia
if Ia = 0.2 S, Then, Q = (P - 0.2 S)2 -------- (1)
P +0.8 S

S = _25400_ _ 254 --------(2)

CN
Q is the surface runoff [mm]
P is the precipitation [mm]
Ia is initial abstraction
S is the potential water retention [mm]
CN (dimensionless number ranging from 0 to 100) is determined from a table,
f(Soil, Land, Vegetal Cover, Antecedent Moisture Condition).
 SCS-CN Method
Antecedent Moisture Conditions
The soil moisture condition in the drainage basin
before runoff occurs is another important factor
influencing the final CN value.
a) AMC I: Soils are dry but not to wilting point;
satisfactory cultivation taken place.
b) AMC II: Average conditions.
c) AMC III: Heavy rainfall, or light rainfall and low
temperatures have occurred within last 5 days;
saturated soil.

Hydrological Soil Group

This parameter, indicates a soil’s runoff
potential, is the qualitative basis of the
classification of all soils into four groups.
Land Use or Cover Land Treatment Practice in relation to
Land use represents the surface conditions in a Hydrological Condition
drainage basin and is related to the degree of Land treatment applies mainly to agricultural
cover. land uses. It includes mechanical practices
such as contouring or terracing, and
management practices such as rotation of
crops, grazing control, or burning.
 References
• https://link.springer.com/chapter/10.1007/978-981-19-1898-8_8#Tab2
• https://www.grdjournals.com/uploads/article/GRDJE/V03/I10/0020/GRDJEV03I
100020.pdf
• An Overview of Rainfall-Runoff Model Types: EPA/600/R-14/152| September
2017| www.epa.gov/research
Books-
• Engineering Hydrology: K Subramanya
• Applied Hydrology:Ven Te Chow, David R. Maidment, Larry W. Mays
 Acknowledgements

Definition, Figures & Statistics utilized in this Lecture from various

sources available in open domain are thankfully acknowledged.

Lecture slides made are only for Capacity Building

Disaster Management Studies Department

Indian Institute of Remote Sensing (IIRS)
Indian Space Research Organization (ISRO)
Department of Space, Dehradun