Application of a web-based GIS model for

urban flood simulation

INDIAN CONFERENCE ON

GEO-SPATIAL TECHNOLOGIES AND APPLICATIONS

Department of Civil Engineering Indian Institute of Technology Bombay

Mumbai, April 12-13, 2012

Jyotirmoy M., JRF

Prof. T.I. Eldho, Supervisor

Prof. E.P. Rao, Co-supervisor

Prof. B.K. Mohan, Co-supervisor

Kulkarni A.T., Research Scholar (Presenter)

Outline

Introduction

Overview of Web Gram Server

Hydraulic model development

Model frame work

Application to Study Area

Summary

July 2005 floods in Mumbai

Floods in Surat-2007

INTRODUCTION

Overview

Major cities in India are subjected to urban flooding viz: Ahmedabad (2001), Delhi (2002, 2003), Chennai (2004), Mumbai (2005), Surat (2006), Kolkata (2007), Jamshedpur (2008) (NDMA, 2010)

“...The frequency of heavy rainfall events are decreasing in major parts of Central and North India while they are increasing in peninsular, east and north east India...” (Guhathakurta et al 2011)

Coastal Urban areas are vulnerable to flooding under combined influence of heavy rainfall and high tides

Need for flood plain zoning, early warning system and flood simulation

Urban Flooding Problems

Loss of Life, damage to property

Disruption of infrastructure services & transportation

systems.

Can cause erosion and instability of soils

on steep slopes threatening building foundations

Extreme events result in inundation for a

prolonged duration

Heavy rainfall, tidal influences and lack of

adequate drainage system is a serious problem

affecting on many coastal cities.

Due to the complexity of the problem, modelling & simulation necessary

Objectives

Demonstrate application of web-GIS based

flood model

Present model frame work for integration of web-GIS

and developed hydraulic model

Details of hydraulic model

Application with a case study

WEB GIS

Web GIS

It is an application that utilizes web and networking systems to facilitate access, processing and dissemination of spatial information

Fig. Typical working of Web-GIS model (Alesheikh etal 2002)

E.g. Commercial web-GIS packages: ArcGIS server, ERDAS Apollo, MapExtreme etc

Open source package: Mapserver, Geoserver, MapGuide etc.

Web Gram Server

Web Gram Server (WGS) is a server-based geographic information system (GIS). A product by BHUGOL GIS Pvt. Ltd.

Has web enabled services for spatial data management, geo data services allow administrators to

publish geographic data

Visualization: map viewing with features like zoom in/out capability, zoom window, info tool etc.

Spatial analysis: Performs queries based on attribute data and generate theme based maps

Web Gram Server offers access to extensive GIS capabilities that enable organizations to publish and share geographic data, maps, and analyses over web.

WGS is OGC compatible GML 2.1.2 file standard as well industry standard ESRI Shape file format

HYDRAULIC MODEL

Modeling Philosophy

Overland flow

computed considering

overland flow grid

Channel flow

computed using FEM

nodes

Tidal Boundary

condition for Channel

flow

Overland flow added

to channel nodes

Fig: Schematic Diagram for

Modelling Flow Processes

Components of Hydraulic Model

Overland Flow- based on Mass Balance

Inflow-Outflow = Change in Storage per unit time

This can be written as

Where, re is the excess rainfall

q is the overland flow from the catchment into the stream

∆t is the time step in seconds

L is the length of the stream element in meters

Ac is the area of sub region in km2

t

VOLQI

tVOLLqAr ce /..

Overland flow- Numerical Formulation

The above non-linear equation is solved

iteratively to obtain water depth(Shahapure et al. 2010)

23

5

1 100 KddK tttt

co An

tLSK

2

2/1

1

3

5

1272

100 tttt

t dKrr

tdK

Channel Flow

Continuity Equation:

Momentum Equation:

Diffusion Wave Approximation:

0qx

Q

t

A

0

2

fS

x

HgA

x

A

Q

t

Q

cf

hS S

x

Initial condition, t=0

U/S B C, D/S B C

,0Q ,0A 0q

,0Q ,0A 0h

x

Channel Flow- Numerical Formulation

Diffusion wave form equation: (Hromadka et al. 1986;

Shahapure et al 2010)

Where

W is top width; H is water level in channel; A is

cross-sectional area

Element Level Approximation:

0

q

x

HK

xt

HW

2

1

3

2

mx

Hn

RAK

ttt

ttt

tt HHL

KH

t

LWqL

HHL

KH

t

LWqL

H

H

L

K

t

LW

L

KL

K

L

K

t

LW

122

211

2

1

222

222

222

222

Unknowns solved for system

of equations

Tidal Boundary Condition

Boundary Condition

Semi-diurnal tidal condition is considered based on

tidal equation

Where dm is mean tidal stage; h is half the tidal oscillation

range; tp is the period of one complete cycle

pm tthdd 2sin

Raster Based Flood Model (RBFP)

The basic component of Flood Plain (FP) model

is a DEM

When water level in channel exceeds bank level,

water ceases to be in main channel and water

spills into adjacent plain

Distributed routing over the flood plain is to

treat each cell as storage volume and solve

continuity eq.

Raster Based Flood Model … contd

rightleftdownup

t

ji

tt

jiQQQQ

t

VV

,,

xx

hh

n

hQ

jiji

fp

flow

right

2/1

,,1

3/5

Flow exchange between Channel and FP

ZwHjZwHi 3

2

2/3

1, 2 ZwHjgbQ ji

Free flow condition

ZwHjZwHi 3

2

Submerged condition

2/1

2, 2 HiHjZwHigbQ ji

Progress of RBFP model in space

APPLICATION TO CASE STUDY

Study Area: Navi Mumbai, Maharashtra, India

Fig: Few delineated catchments of Navi Mumbai

Study Area: New Panvel, Maharashtra, India

Panvel Catchment Area 8.27 Km2

North Latitudes 190 00’ 00” and 190 01’ 19”

East Longitudes 730 04’ 55” and 730 08’ 37”

Ground Elevation varies from 0 m to 231m above MSL

Dominant flow direction is west

Fig: Location map of New

Panvel, Maharashtra, India

HILLY AREA

Study Area superposed on GE

Hilly

Slopes Creek End

Fig: Study Area New Panvel, Maharashtra, India

DATABASE PREPARATION

Database Preparation

Data Acquired DEM of (~14 m pixel) obtained from digitizing topographic

map and spot levels

Multispectral images from IRS – LISS III dated 14th May 2005 (23.5 m)

Storm water plan for the study area was obtained from CIDCO

Data Processing Watershed delineated in ArcGIS

Image processing in ERDAS IMAGINE using supervised classification

DEM & Slope of the Catchment

Fig: DEM for New Panvel Fig: Slope Map for New Panvel

Grid & Land Use Map for Catchment

Fig: Grid Map of catchment Fig: Land Use Map of catchment

Discretization for Model

Overland Flow Grid Catchment divided into 38 overland flow elements

Min and Max overland flow element area are 103.48 ha and 1.26 ha

Channel discretized into 314 elements

Average channel slope is 1 in 1011

Fig: Channel

discretization for

New Panvel

RESULTS & DISCUSSION

Model frame work

Hydraulic Model.exe

Display of flooded channel

nodes as point file with

NodeID; Start time and

Duration (in mins.)

Processed in ArcMap

environment

Database for overland flow

elements

(in *.xls format)

Rainfall data

Channel geometry details

Channel-overland node connections

Tidal Boundary Condition

(all in *.xls format)

Model simulation time

and model time step thru

User Interface

(in mins)

WGS environment

GUI showing loaded GML layers

GUI prompt to entre model parameters

Simulation Results- Discharge & Stage Hydrographs

July 26, 2005

Peak discharge 112 m3/sec Time

to peak discharge 10.05 hrs

•Event of July 26, 2005 was

extraordinary

•Maximum rainfall intensity of

76 m/hr occurring over 1 hrs

•Total volume of rainfall 745

mm in 24 hrs

Fig. Simulated discharge and stage hydrograph for 26th

July 2005 event

Simulation Results- Water surface profile along the

channel for July 26, 2005

Flooded nodes as water surface

level is above the bank level

GUI with flooded nodes.xml

Flood Inundation Simulation ... 0 hr

Flood Inundation Simulation ... 10min

Flood Inundation Simulation ... 1 hr

Flood Inundation Simulation ... 4 hr

Flood Inundation Simulation ... 6 hr

Flood Inundation Simulation ... 8 hr

Flood Inundation Simulation ... 10 hr

Flood Inundation Simulation ... 12 hr

Flood Inundation Simulation ... 15 hr

Flood Inundation Simulation ... 17 hr

Flood Inundation Simulation ... 19 hr

Flood Inundation Simulation ... 24 hr

Issues in implementation of web-GIS model

Model run time

A 24 hr long event simulation takes ~1 hr for 8km2

catchment

Optimisation for multiple catchment simulation-->

parallel processing

Server Response time ~30 mins

Matching model run time and server response time

Display of model results: Flood inundation extent

As a Raster, vector or video file?

System Requirements

Supported Operating System Version

Windows 2003 Server Standard, Enterprise & Datacenter (32-bit and 64-bit

(EM64T))

SP2

Windows 7 Ultimate, Enterprise, Professional, Home Premium (32-bit and 64-

bit (EM64T))

SP1

Windows Vista Ultimate, Enterprise, Business, Home Premium (32-bit and 64-

bit (EM64T))

SP2

Windows XP Professional Edition, Home Edition (32-bit) SP2,SP3

Windows XP Professional Edition, Home Edition (64-bit (EM64T)) SP2

CPU Speed 2.2 GHz minimum or higher; Multi-core recommended

Processor Intel Pentium 4, Intel Core Duo, Intel Core i5

Memory/RAM 2 GB or higher

Disk Space 1 MB for War file of WEB SERVER; 460 MB for Application Server

Networking Hardware Simple TCP/IP, Network Card

Software requirement:

XAMPP SERVER:

WEBServer is built on purely PHP domain. So it require xampp server to launch the

application

Internet Explorer Requirement:

WEBServer requires a minimum installation of Microsoft Internet Explorer Version

7.0 or 8.0. If you do not have an installation of Microsoft Internet Explorer Version

7.0/8.0, you must obtain and install it prior to installing WEBServer.

MYSQL v5.0:

Database MYSQL is required to store data at the backend. It supports GIS data

transaction.

Summary

Demonstration of desktop web-GIS based flood

simulation tool is presented

Details on webgram server is presented

Details of the hydraulic model is presented

Acknowledgements

The authors acknowledge their sincere gratitude to Department of Science and Technology (DST), Govt. of India, New Delhi for sponsoring a project on Integrated Flood Assessment Modelling for Urban Watersheds using FEM, GIS and Remote Sensing. Authors are thankful to Head, NRDMS and Mr. P.S. Acharya, Scientist-F for all their support in execution of the project.

The authors are also thankful to Mr. S.C. Deshpande (Ex- CE), Mr. P.P. Joshi, Mr. D.R. Hartalkar, Mr. P.U. Natesh and Mr. V.U. Lathkar, engineers from the City and Industrial Development Corporation (CIDCO), Navi Mumbai, Maharashtra for providing the topographical, hydrological and storm water network layout related data for the study.

THANK YOU for your attention! Anand T Kulkarni

Research Scholar

Department of Civil Engineering,

Indian Institute of Technology Bombay,

Mumbai, India, 400 076.

Email: anand.kulkarni@itb.ac.in

Flow chart WGS

Study Area: Stretch of Thames,

Buscot

U/S

D/S

Source: Google

Earth

DEM of

study

area

Simulation Information

•Inflow discharge at U/S as BC

•Model simulated for 7.5 hours

•Model time step 0.1 sec

•Uniform flow depth at D/S

RBFP Model Validation

Water Surface Profile Comparison

Client-Server Interaction