Integrated Watershed Modeling using Numerical Method, Geographical Information Systems

1
Integrated Watershed Modeling using Numerical
Method, Geographical Information Systems Remote
Sensing Scope for Urban Flood Assessment
Modeling

• Dr. T.I. Eldho
• Associate Professor
• Department of Civil Engineering
• Indian Institute of Technology Bombay
• Email eldho_at_iitb.ac.in
• Phone (91-22) 25767339 Fax [phone]
• http//www.civil.iitb.ac.in

2
Content

• Introduction
• Watershed modeling
• Remote sensing, GIS and numerical methods
• Development of Integrated Watershed Model Using
  FEM, GIS Remotely Sensed Data
• Applications
• Scope for Urban Flood Assessment Modeling

3
INTRODUCTION

• Need for proper planning and
• management of available water resources /
• Flood hazards management/ urban floods.
• Watershed is the basic scientific unit.
• Number of models Black box models/ lumped/
  distributed
• For better water management/ flood hazard
  assessment Distributed models models based on
  
• physical laws.
• Digital revolution
• An integrated watershed model Recent advances
  in modeling - use of numerical methods/ remote
  sensing and GIS.

4
WATERSHED MODELING

• Watershed Characteristics.
• Hydrology of watershed.
• Modeling.
• Watershed modeling steps
• 1. Formulation
• 2. Calibration/verific
  ation
• 3. Application
• Watershed model constitutes
• 1. Input function
• 2. Output function
• 3. Transform function

5
Fig 1 Flowchart of simple watershed model
(McCuen, 1989)
6
Fig 2. Flow in a watershed Typical flow
pattern
Fig 3 General concept of flow modeling
7
REMOTE SENSING, GIS AND NUMERICAL METHODS

• Remote sensing The remote sensing data are
  capable of solving the problem of scarcity of
  data -capability of observing several
  hydrological variables - over large areas on
  repetitive basis
• GIS Database development Management enhances
  the ability to incorporate spatial details
• Numerical methods Rainfall-runoff simulation
  FEM, FDM etc.

8
Integrated Watershed Model Using FEM-GIS RS

• Event-based rainfall - runoff model
• FEM based kinematic and diffusion wave models
• Overland flow and channel flow
• Overland flow model verification
• Hypothetical watershed
• Infiltration models GAML and Philip
• Interception and interflow models
• Composite model
• GIS - Database preparation
• RS - LU/LC preparation
• Model application
• Six watersheds of different physiographic
  regions with
• different model combinations
• Sensitivity analysis

9
Frame work for the model development
10
Model Formulation

• Interception
• Infiltration
• Overland flow - Two/one dimensional
• Channel flow - One dimensional
• Interflow
• Component models coupled to get the runoff

Flow in a watershed Typical flow pattern
11
Start
Input Excess rainfall from infiltration model
and duration, Number and size of elements,
Roughness coefficients, Slopes, Bed width, Time
step, Duration of simulation etc.
Initialization of variables (Initialize
nonzero depth at time t0)
Calculation of element matrix for channel and
overland flow
Generation of global matrix by assembling element
matrices and applying boundary conditions for
channel and overland flow
Solving the system for overland flow
If

No
Yes
Solving the system for channel flow
if
No
Yes

Yes
If
No
Stop
Flow chart for coupled overland and channel flow
model
12
Model Applications

• Application of different combination of runoff
  and infiltration models
• Kinematic-GAML
• Diffusion-GAML
• Kinematic- Philip
• Diffusion-Philip
• With/without Interception model
  Depending upon availability of data
• With/without Interflow model
  Depending upon watershed type
• Watershed Model combination
• Catsop Kinematic-GAML
  with interception
• Peacheater Creek Diffusion-Philip with
  interception and interflow
• Banha, Amba Kinematic- Philip
• Harsul, Khadakhol Diffusion-GAML

13
Study Area Banaha Watershed

• Location- Chatra district in Jharkhand State,
  India
• East Longitudes of
  85o12'15? and 85o 16'15?
• North Latitudes of 24o
  13' 45? and 24o 17'
• Area- 16.72 km2
• Major Soil class Sandy loam.
• Hydrological Data- Mr. Guy Honore, Project
  coordinator, Indo
• German
  Bilateral Project
• Remotely Sensed Data- IRS 1D LISS III imagery of
  January,
• 
  1998
• Thematic Maps- Drainage, Slope and LU/LC

14
WATERSHED CHARACTERIZATION

• Map generation and analysis- ERDAS IMAGINE and
  ArcGIS
• Slope map- ArcGIS
• LU/LC map - ERDAS IMAGINE
• Mannings roughness map- Based on LU/LC map
• Finite Element Grid map- ArcGIS
• Grid map has been overlaid on slope and Mannings
  
• roughness maps
• Mean value of slope and Mannings roughness- Each
  
• element of the grid
• Nodal values- Average of adjacent element values

15
Drainage map of watershed
16
Digital Elevation Model map
17
Slope map of Banha watershed
18
False Colour Composite of Banha Watershed
19
Land Use/ Land cover map of Banha watershed
20
Finite element grid map of Banha watershed
21
Results and Discussion

• Diffusion wave- Philip model
• Calibration - 4 Rainfall events
• Validation - 3 Rainfall events

Calibrated parameters for rainfall events (Banha
Watershed)
22
Calibration event, July 24, 1996
Validation event, August 17, 1996
Observed and simulated hydrographs of rainfall
events (Banha)
23
Model results for rainfall events (Banha)
24

• Calibration events
• Variation in Volume of flow
  - 34
• 
  Peak runoff - 1
  Time to peak runoff
  - 9
  With
  few exceptions.
• Validation events -Mixed performance
• Variation in Volume
  of flow - 30
• 
  Peak runoff - 39
  Time to peak runoff
  - 25
• Large variations in observed and simulated values
  for validation
• - Characteristics of rainfall events
• - Parameter uncertainty

25
Sensitivity Analysis

• Altering the calibrated parameters of ,
  , and of the watershed by 10
• Peak runoff and time to peak are most sensitive
  to followed by .
• Time to peak runoff is least affected by all
  these parameter changes

26
Effect of change in model parameters on computed
values of volume of runoff, peak runoff and time
to peak runoff for the event of August 17, 1996
27
Study Area Harsul Watershed

• Location- Nashik district, Maharashtra, India
• East Longitudes of 73o
  25' and 73o 29'
• North Latitudes of 20o
  04' and 20o 08'.
• Area- 10.929 km2
• Major Soil class Gravelly loam
• Hydrological Data- Mr. Guy Honore, Project
  coordinator, Indo
• German
  Bilateral Project
• Remotely Sensed Data- IRS 1D LISS III imagery of
  January,
• 
  1998
• Thematic Maps- Drainage, Slope and LU/LC

28
Drainage map of Harsul watershed
29
(No Transcript)
30
Slope map
Digital Elevation Model map
Harsul watershed
31
(No Transcript)
32
False Colour Composite
Land Use/ Land Cover map
Harsul watershed
33
(No Transcript)
34

• Overland flow elements - 144
• Overland flow nodes -188
• Channel flow elements - 22
• Channel flow Element length - 0.25 km
• Average bed width - 18 m
• Slope
• Overland flow
• Channel flow
• Mannings roughness
• Overland flow
• Channel flow

Finite element grid map
35
Results and Discussion

• Diffusion wave- GAML model
• Calibration - 3 Rainfall events
• Validation - 2 Rainfall events

Calibrated parameters for rainfall events (Harsul)
36
August 22, 1997
September 23, 1997
September 26, 1997
Observed and simulated hydrographs of calibration
rainfall events (Harsul)
37
August 21, 1997
August 23, 1997
Observed and simulated hydrographs of validation
rainfall events (Harsul)
38
Model results for rainfall events (Harsul)
39
Scope for Urban Flood Assessment Modeling

• Rainfall-runoff modeling of urban watersheds
  considering the overland flow, channel flow,
  retention basins and tidal influence
• Integrate the simulation model with the database
  including remotely sensed data within GIS
  environment.
• Develop a database framework for a possible cyber
  infrastructure of urban watersheds with specific
  case studies

40
Concluding Remarks

• A watershed model which simulates event based
  runoff using FEM, GIS and remote sensing
  techniques using kinematic/diffusion wave
  equation is presented .
• Philips/ GAML model is used for estimation of
  infiltration.
• Model has been calibrated and validated on
  Banaha/ Harsul watershed, India.
• Developed model has fairly simulated the
  hydrographs at the outlet of watershed.
• Model is useful for simulation of hydrographs in
  small ungauged watersheds.

41

• THANK YOU

Dr. T. I. Eldho Associate Professor, Department
of Civil Engineering, Indian Institute of
Technology Bombay, Mumbai, India, 400 076.
Email eldho_at_iitb.ac.in Phone (022) 25767339
Fax [phone] http//www.civil.iitb.ac.in
42
Interception Model

• To calculate the effective rainfall after the
  interception loss
• LISEM model

Cumulative interception during a rainfall event
is given by
43
Infiltration Model

• Two infiltration models
• -GAML model
• -Philip model
• Philip Infiltration Model To calculate
  infiltration rate and subsequent
  excess rainfall

The rate of infiltration is given by
Infiltration sorptivity
44
Interflow Model

• Interflow model -Through flow equations given by
  Jayawardena and White (1977)

Continuity equation
can be expressed by Darcys law
Interflow
Finite element formulation
45

• Governing equations for overland flow
• 1.Continuity equation
• 2. Momentum equation
• Kinematic wave form -
• Diffusion wave form -
• Finite element formulation
• - Galerkins criterion is used

For diffusion wave modeling
46

• Governing equations for channel flow
• The equation of continuity
• Momentum equation
• Kinematic wave form
• Diffusion form
• Mannings equation
• Finite element formulation

For diffusion wave modeling
47
Green-Ampt Mein Larson Infiltration model

• The Green-Ampt equation for infiltration rate
• Mein and Larson (1973) modification
• Upto time of ponding
  infiltration rate
• Cumulative infiltration at
  ponding time
• Equation given by Chu (1978)

48
Solution methodology

• Duration of rainfall is divided into many short
  periods
• Each sub period of rainfall
• (1) No ponding at the start of the
  period
  (a) Non ponding continues up to the end of sub
  period
• 
  (b) Ponding occurs during the sub period
• (2) Ponding at the start of the period
• (a) Ponding continues up
  to the end of sub period
• (b) Ponding ceases
  during the sub period

49
Conclusions

• -Based on model application to four watersheds
• Reasonably simulated the hydrographs at the
  outlet of the watersheds.
• Simulation results (Calibration)
• Variation in volume of flow 60
• peak flow
  48
• time to peak
  27
• Variations in observed and simulated values for
  validation
• - Characteristics of rainfall events and
  parameter uncertainty.
• Sensitivity analysis
• For most of the watersheds is the most
  sensitive parameter. However, for some
  watersheds, is the most sensitive parameter.
• Decrease in grid size, the peak runoff
  increased and the time to peak runoff decreased.
  
• Decrease in time step, the volume of runoff
  and peak runoff decreased where as the time to
  peak runoff is increased.