Flood Modeling using Gis and LiDAR of Padada River in

Southeastern Philippines

Joseph E. Acosta

, Ryan Keath L. De Leon

, Judy Rose D. Hollite

Richard M. Logronio

and Genelin Ruth James

LiDAR 1.B. 13, University of the Philippines Mindanao, Mintal, Davao City, Philippines

Department of Mathematics, Physics, and Computer Science, University of the Philippines Mindanao,

Mintal, Davao City, Philippines

Keywords: LiDAR, Light Detection and Ranging, HEC-RAS, Modeling, Flood Hazard, Flood, Hazard, Floodplain, River

Analysis System, Digital Terrain Model.

Abstract: Located along the typhoon belt in the Pacific, Philippines is visited annually by an average of 20 typhoons.

This brought frequent flooding to communities and caused damage to flood prone residents. This study is

aimed to generate GIS flood model to simulate flood inundation using Synthetic Aperture Radar (SAR) data

and Digital Terrain Model (DTM) derived from Light Detection and Ranging (LiDAR) datasets. Open-

sourced HEC-HMS and HEC-RAS Models were used. HEC-GeoRAS in ArcMap was used to preprocess

geospatial data of Padada River basin in the Southern Philippines. The result of the study has identified

inundated areas and communities for a 24-hour 5, 25, and 100- year return periods as well as flooded critical

facilities in the flood plains. The output of this study can be used in planning and predicting flood risks for

mitigation.

1 INTRODUCTION

Flood is one of the most disastrous phenomena that

can occur in an area since it has damaging impacts to

human lives, properties, and infrastructures such as

death, spread of diseases, and destruction of

livelihood (Siddayao et al, 2015).

Padada watershed is situated in the southern part of

the Philippines with a total land area of 1,186.80 sq.

km which drains to Padada river and occasionally

flooding the downstream municipalities of Hagonoy,

Padada, and Sulop affecting hundreds of households,

damaging crops, livestocks and properties. Flooding

aggravated hunger and resulted to poor health

conditions among the villagers in these communities

( General (Philippines, 2009):

To reduce flood effects and risks, prediction of the

flood event is a vital information to locate the areas at

risk and to quantify the damages it may result to. The

extraction of this data can be acquired through

simulating the event in a computer using flood model.

Flood modeling requires accurate assessment of

water movement within flood-prone areas as a

reliable basis for mitigation and resource

management (Kim et al, 2016). Several techniques

considering different parameters were used in

practice. One of which is the Hydrologic Engineering

Center-River Analysis System (HEC-RAS).

HEC-RAS is a one-dimensional model which is

able to perform an unsteady flow simulation of a river

(Hammerling et al, 2016). It has been applied to a

range of applications such as bridge and culvert

design, channel modification and floodplain

management (Yuan and Qaiser, 2011). In

implementing this tool, several studies have utilized

high-resolution datasets like digital elevation models

as inputs to capture the detailed profile of river

channels and to achieve more precise results (Turner

et al, 2013).

Light Detection and Ranging (LiDAR) is a digital

elevation model with a topographic horizontal

resolution of 1m. Due to its higher resolution, LiDAR

gives more details and near to accurate topographic

information compared to other DEMS such as

Synthetic Aperture Radar (SAR) and Interferometric

Synthetic Aperture Radar (ifSAR) (Suarez et al,

2014).

In the Philippines, some areas has experienced

Acosta, J., Leon, R., Hollite, J., Logronio, R. and James, G.

Flood Modeling using Gis and LiDAR of Padada River in Southeastern Philippines.

DOI: 10.5220/0006378103010306

In Proceedings of the 3rd International Conference on Geographical Information Systems Theory, Applications and Management (GISTAM 2017), pages 301-306

ISBN: 978-989-758-252-3

301

distinct changes in the weather patterns with

disastrous results such as flooding (Yumul et al,

2013). With the disaster risk management program in

place, a study was conducted that generates a flood

model using Hydrologic Engineering Center-River

Analysis System (HEC-RAS) as a tool and Light

Detection and Ranging (LiDAR) data sets as digital

elevation model. A case study is done in Padada River

floodplain in Southern Philippines as in Figure 1.

Figure 1: Location of Padada floodplain in Philippines.

2 METHODOLOGY

Presented in Figure 2 is the method of preparing the

HEC-RAS Model of Padada floodplain. The process

involves four major activities namely (i) Generating

the River Digital Plot, (ii) Flood Model

Preprocessing, (iii) Flood Modeling, and

(iv)Identifying Flood Prone Communities.

Figure 2: Process flow in generating HEC-RAS Model.

2.1 Generating the River Digital Plot

One of the prerequisites in flood model preprocessing

is the river digital plot (RDP) which are polyline

vector files that follow the centerline, left flow path,

right flow path, left bank, and right bank of Padada

River. In this study, LiDAR DTM with a 20 cm

vertical accuracy and various satellite images were

mainly used to verify and generate the RDP. The

digitization of the RDP started from the project point

St art

Generate the River Digital Plot

Digital River Plot

Flood Model

Prep rocessing

Discharge Data of

Padada River

RAS Data

Padada LiDAR Data

Flood Modeling

(HEC-RAS)

Flood Inundation M ap of

Padada Floodplain

Feature Extracted

Shapefiles of Padada

Flood

lain

Data Overlay for

Identification of Flood

Prone Communitie

Flood Prone Communities

End

Hypothetical

Scenarios

GISTAM 2017 - 3rd International Conference on Geographical Information Systems Theory, Applications and Management

302

where the discharge data was gathered going down to

the outlet of the river. Once done, the generated river

digital plot in Keyhole Markup Language (KML) file

format was converted to a vector file.

2.2 Preprocessing

In generating the flood model, the LiDAR DEM and

the river digital plot were prepared for processing the

physical features of the river. The LiDAR DEM was

initialized as the elevation data while the river digital

plot was used to define the extent of the river. The

digitized river lines were grouped into river data,

containing the centerline, River banks data,

containing the left bank and right bank, and flowpaths

data, containing the right flow path, left flow path,

and centerline of the river. The river centerline was

assigned with a river code and a reach code to provide

a name for the river. To define the flow direction of

the river, each paths from the flowpaths data were

assigned whether a right flow path, a left flow path,

or a channel path. Using the LiDAR DEM as terrain

and the river data as the centerline of the stream, the

stream profile was generated to extract the changes in

elevation along the river channel from upstream to

downstream locations.

After generating the geometric attributes along the

rivercenterline, the construction of the cross-section

cutlines followed. Each cross-section cutline

represents a station in the river where the data for the

one-dimensional flood simulation was based on.The

initial set of cross-section cutlines were uniformly

distributed along the river centerline. Furthermore,

these were adjusted so that there were no lines

intersecting another line and that they have crossed a

river bank only once. This was to avoid the

overlapping of flood inundation among stations.

Having the river data as the centerline of the stream,

LiDAR DEM as terrain, and cross-section cutlines,

the cross-section cutline profile was generated to

extract the elevation data across the river in each

station. In preparation for the export of the GIS data

for HEC-RAS modeling, the generated datasets were

initialized including the terrain, centerline of the

stream, cross-section cutlines, cross-section cutline

profiles, bank lines, flow path, and stream profiles.

2.3 Flood Modeling

The previously generated GIS Data in flood model

preprocessing was converted into geometric data with

a metric system of Système international d'unités or

SI. The Manning’s n values were supplied and Cross

Section Point Filter was set in the process to initialize

the roughness value and reduce the elevation data

points that can be processed by HEC-RAS model for

each river station, respectively. For the simulation,

unsteady flow were used in utilising the discharge

data and simulated 5, 25, and 100 Year Rain Return

hypothetical scenarios of the target river. On the other

hand, a friction slope value of 0.001 was set. Once all

the previous parameters were set properly,

computation of unsteady flow simulation followed.

Afterwards, the result of the flood model simulation

was produced. Lastly, generation of a flood

inundation was done.

2.4 Identifying Flood Prone

Communities

The 5, 25, and 100 Year flood inundation maps that

were generated were overlaid to the feature extracted

vector files of Padada floodplain to identify the flood

prone communities. These were done in ArcMap

10.2.2.

3 RESULTS AND DISCUSSION

The digitization of the river digital plot started from

Diversion Bridge (06°40’35.4”N, 125°19’47.9”E).

The LiDAR elevation datasets used in preprocessing

the RAS Model has a 1-meter resolution with 20cm

vertical accuracy and is covering the Padada

floodplain as in Figure 3. For the development of the

model, the discharge data used was gathered in the

same bridge with a base flow discharge of 3.35 m

and was specifically set as input as the normal flow

of Padada River.

Figure 3: LiDAR Coverage of Padada River floodplain.

Riverbed cross-sections or the XS-Cutlines of the

target watershed are crucial in the setup of HEC-RAS

Flood Modeling using Gis and LiDAR of Padada River in Southeastern Philippines

303

Model. These cross-section data (Figure 4) were

defined using the Arc GeoRAS tool and the

LiDARDEM and were post-processed in ArcMap

10.2.2.

Figure 4: Padada riverbed cross-section data.

The Manning’s n is dependent by the nature of the

channel and surface. It is important to determine the

correct roughness coefficient of the channel to have a

good water flow simulation. The values used were

based on the floodplain’s land cover type and

standardized Manning’s n look-up table (Table 1).

Table 1: Look-up Table for Manning’s n Values.

Land-cover Class Manning’s n

Barren Land 0.030

Built-up Area 0.015

Annual Crop 0.040

Perennial Crop 0.035

Fishpond 0.018

Inland Water 0.030

Grassland 0.030

Mangrove Forest 0.120

Shrub Land 0.100

The HEC-RAS Flood Model produced water level for

every cross-section based on the discharge data that

served as input. Initially, the actual discharge data

were used in running the model to assure good

simulation and afterwards, hypothetical scenarios of

5, 25, and 100 year rain return periods were then used

as the input discharge. Right after running three

simulations, 5, 25, and 100 year rain return flood

inundation maps of Padada floodplain were generated

as shown in Figures 5-7.

Figure 5: Padada Floodplain 5 Year Flood Inundation Map.

Figure 6: Padada Floodplain 25 Year Flood Inundation

Map.

GISTAM 2017 - 3rd International Conference on Geographical Information Systems Theory, Applications and Management

304

Figure 7: Padada Floodplain 100 Year Flood Inundation

Map.

The generated flood inundation maps, together with

the feature extracted vector files of the floodplain,

were overlaid to identify the number of structures that

will be affected in a given return period. For each

rainfall return period, flood levels were categorized

into low, moderate, and high. Structures affected with

low flood level will experience below 0.5m flood

depth while structures in the moderate level will have

flooding between 0.5m and 1.5m deep. Structures

with high flood level will encounter more than 1.5m

flood depth. The result is shown in Table 2.

Table 2: Number of structures in Padada floodplain that will

be affected during 5, 25, and 100 Year RIDF.

Flood

Level

Number of Structures Affected in:

5-Year 25-Year 100-Year

Low

3,704

Structures

4,977

Structures

5,406

Structures

Moderate

951

Structure

1,800

Structures

2,999

Structures

High

Structures

143

Structures

4 CONCLUSIONS

Flood model geometric data preparation and model

parameterization were made through HEC-RAS and

ArcMap extension. The 24-hour 5, 25, and 100

Rainfall Intensity Duration Frequency (RIDF) curves

of Padada river were used as input for the setup of

HEC-RAS model. The HEC-RAS flood model

produced a water level at every cross-section for

every time step and for every flood simulation

created. The resulting model was used in determining

the flooded areas. The model demonstrated change in

the behavior for the three different return periods.

There are two communities that will be affected

resulting from the hypothetical scenarios. A total of

4,678 structures in the floodplain will be flooded in a

5-year rain return period, 6,842 in a 25-year rain

return period, and 8,548 for a100-year rain return

period. The simulated results provided the maximum

flood extent and inundation levels. As a result, the

identification of those flood-prone areas at risk will

help in the information and planning of a more

effective emergency responses. The output of this

study can help the local planning units to forecast and

assess future flood risk for sustainable development.

REFERENCES

General (Philippines): Flood aggravates hunger and poor

health of the villagers in Davao. (2009, February 5)

Retrieved from Asian Human Rights Commission:

http://www.humanrights.asia/news/hunger-alerts/AHR

C-HAG-002-2009/

Hammerling, M., Walczak, N., Walczak, Z., and Zawadzki,

P. “The Possibilities Of Using HEC-RAS Software For

Modelling Hydraulic Conditions Of Water Flow In The

Fish Passexampled By The Pomiłowo Barrage On The

Wieprza River,” Journal of Ecological Engineering,

Vol. 17, No. 2, Apr. 2016, pp. 81–89.

Kim, S., Yoon, B., Kim, D., and Kim, S. “Accuracy

Analysis of HEC-RAS for Unsteady Flow Simulation

considering the Flow Pattern Variations over the Side-

weir of Side-Weir Detention Basin,” in Journal of

Korea Water Resources Association, 2016, vol. 49 no.

1, pp. 29-39.

Siddayao, G., Valdez, S., and Fernandez, P. “Modeling

Flood Risk for an Urban CBD Using AHP and GIS” in

International Journal of Information and Education

Technology, Vol. 5, No. 10, October 2015, pp. 748-

753.

Suarez, J. K. B., Santiago, J. T., Muldong, T. M. M.,

Lagmay, A. M. A., Caro, C. V., and Ramos, M.

“Comparison of High Resolution Topographic Data

Sources (SAR, IfSAR, and LiDAR) for Storm Surge

Flood Modeling using Gis and LiDAR of Padada River in Southeastern Philippines

305

Hazard Maps,” in American Geophysical Union, Fall

Meeting 2014, abstract #ED11B-3405.

Turner, A., Colby, J., Csontos, R., and Batten, M. “Flood

Modeling Using a Synthesis of Multi-Platform LiDAR

Data,” in Water 2013, Vol. 5, No. 4, pp.1533-1560.

Yumul, G., Dimalanta, C., Servando, N., and Cruz, N.

“Abnormal weather events in 2009, increased

precipitation and disastrous impacts in the Philippines,”

Climatic Change June 2013, Vol. 118, No. 3, pp. 715–

727.

Yuan, Y. and Qaiser, K. “Floodplain Modeling in the

Kansas River Basin Using Hydrologic Engineering

Center (HEC) Models” in U.S. Environmental

Protection Agency, National Exposure Research

Laboratory, Environmental Sciences Division,

Landscape Ecology Branch, Las Vegas, NV 89119.

GISTAM 2017 - 3rd International Conference on Geographical Information Systems Theory, Applications and Management

306