Rainfall-Storage-Drain(RSD) model for Runoff Control

Rainwater Tank System Design in Building Rooftop

Youngjin Kim*, Mooyoung Han**

Rainwater Research Center in Seoul National University

* mukta73@snu.ac.kr, ** myhan@snu.ac.kr)

Abstract
Surcharge runoff over the capacity of drainage pipes is one of main flood breaking factors
in urban area. Change into larger pipes is costly and time consuming solution. Rainwater
tanks can be a sustainable solution for the control of heavy runoff
The Rainfall-Storage-Drain(RSD) model is developed for the design of rainwater tanks of
flood control. Design rainfall of Huffs distribution and runoff analysis on building rooftop
makes inflow curves into the tanks in analysis of the rainfall process. Through a water
balance equation, tank volume versus peak flow curves can determine the minimum tank
volume needed for controlled peak flow value of a design period in the storage and the
drain process analysis
The case study in Seoul City shows a tank of 29 /m2 can control the runoff of 30-years
with the drainage pipes of 10-years design period. The RSD model can give simple and
easy curves to understand for tank capacity determination in a local flood condition
Keywords: urban flood, rainwater management, rainwater tank, RSD model, decentralized system

1. Study background and goals

1.1 Flooding in Seoul

In cities all over the world, the characteristics of floods are changing markedly because of
increasing urbanization and climate change. Figure 1 shows the drainage districts of the
metropolitan area of Seoul in 2001. The Seoul storm sewerage consists of 16 districts, 13
of which were inundated by the flood of 2001. Among these, three had problems of
overflow in the final receiving water channel or as a result of their topographical condition,
and the remaining nine districts were flooded by runoff in excess of the maximum capacity
of their storm sewerage (Seoul, 2002). Sewer flooding is the main reason of inundation in
Seoul

no flood, 4

flooded for
surcharge runoff,
9

flooded for other

reasons, 3

Figure 1. Seoul city flood in 2001

Because only one or two storm sewers overflowed in the nine districts with inadequate
drainage capacity, increasing the size of the sewers using conventional engineering
techniques was the only solution proposed. This option is very costly and will take a long
time to complete, because Seoul is a high-density city with heavy traffic and complex
industrial activities. Furthermore, a larger sewer of fixed capacity may not be enough for
heavier rainfall in the future (Kim, 2006).

1.2 Rainwater management system in Seoul National University

Seoul National University (SNU) has three systems in her campus which are two rainwater
storage and supply systems with building rooftop catchment area and one infiltration and
retention system with a lawn and a small dam reservoir catchment.
A dormitory complex has 200 ton rainwater storage tank for toilet and gardening water,
which is constructed in November 2003 (Han et al, 2006). This system is almost the 1st
practical rainwater system for domestic water consumption of toilet in Korea. Rainwater
on three buildings roof of 2,098
has been harvested and used successfully until now
without serious failure or complain from residents. Figure 2 is the schematic diagram of
the system structure.

Figure 2. Rainwater harvesting system in SNU dormitory complex

In October 2005, other rainwater system was constructed in education research building in
SNU (Mun et al, 2006, 2007). This dual tanks system of 250 ton stores rainwater from
catchment area consisted of building roof of 2,828
and terrace of 824 . The water
harvested in this system is supplied into toilets of this building. Figure 3 shows the system
scheme.

Figure 3. Rainwater harvesting system of the education research building in SNU

Rainwater harvesting and infiltration system in Beo-deul valley is constructed in April

2004 as the eco campus project. SNU campus is located in Kwanak mountain of which
land slope is high. This system is designed to control runoff from Beo-deul valley and use
the water for lawn gardening (Han et al, 2004). Figure 4 is the diagram of the system.

Figure 4. Rainwater harvesting and infiltration system in Beo-deul valley

1.3 Review on urban flood control studies

(1) Review on decentralized management studies
Decentralized and localized management of water is an innovative strategy developed in
recent years. Localized and integrated flood management was studied and modeled at the
watershed scale by R. H. McCuen (McCuen, 1974). He concluded that a regional approach
to stormwater could reduce peak runoff while individual-site approaches could result in
flood increases. Low-impact development strategies (LID) were introduced in a manual by
3

the county of Prince George County in 1999 (Prince Georges County, 1999). In LID,
urban stormwater can be controlled through a variety of methods of detention and retention
instead of the conventional end-of-pipe or in-the-pipe strategies.
In Korea, Choi, Shin and Kim and coworkers pointed out the need for development of a
new drainage control system for surcharge runoff control that takes account of the unique
condition of small confined areas (Choi et al, 2001; Shin et al, 2005; Kim, 2005). Recently,
Miyazaki pointed out that an on-site, decentralized rainwater detention system using a lot
of tanks was effective in control of the flow in urban drain pipes. Engineering planning
methods are necessary to decide the location of each tank and its capacity (Miyazaki,
2006).
During recent years, decentralized and integrated water management has developed into a
remarkable strategy for sustainable and safe urban infrastructure. Sophisticated scientific
and engineering methodologies are required to implement the system.
(2) Review on rainwater tank design
The design of rainwater tanks follows the same principles as conventional detention tank
design in storm sewerage. McCarthy proposed that the water balance equation could be
applicable in the design of water detention tanks (McCarthy, 1938). Recently, Guo
developed a modified design method for flood control detention tank volume (Guo, 1999,
2001).
However, the design of rainwater tanks for capture and storage of water is not easy in the
sophisticated routing method for detention tanks. The classical and simple water balance
equation was developed for rainwater tank design. Hermann and Schmida calculated the
flood control effect of a rainwater tank as its volume varies (Hermann and Schmida, 2000).
Vaes and Berlamont analyzed the effect of rainwater tanks on storm sewerage design in
Belgium (Vaes et al, 2001). They concluded that detention in existing rainwater tanks can
control a peak runoff flow of 5-year frequency.
Mooyoung Han published a study on rainwater tank volume determination under Korean
rainfall conditions (Han et al, 2004). Mun showed that decentralized small rainwater tanks
could be more effective in control of floods than one large tank, using SWMM modeling
analysis in an apartment house complex (Mun et al. 2005).
A number of studies conclude that numerous small tanks can reduce urban flooding and
reduce inundation of the inner city. Nevertheless, rainwater tanks continue to be designed
only as a private water supply facility.
This study analyzed the urban flood control effect of a rainwater tank system. The RSD
model was developed to analyze tank capacity and the peak runoff flow. Furthermore, the
RSD model can calculate the minimum tank volume to control a hypothetical design storm
of any duration.

2. Introduction of the RSD system

The RSD system is a roof runoff control system using rainwater tanks (Figure 5). The RSD
system consists of rainfall, storage and drainage processes. During rainfall, the rooftop of a
building catches the water and collects it into a rainwater tank, thereby controlling the
flow. Flow over the tank storage capacity goes through the storage process directly into the
urban drain system.

Figure 5. The RSD rainwater management system

The RSD model is a new numerical model for the analysis and design of each process in
the RSD system. This study develops the RSD model and shows a case study of RSD for
Seoul City.

3. Methodology of RSD model development

The design of a facility for flow control must follow a detailed and reasonable design that
considers the local rainfall conditions. In this study, the RSD model is a tool for calculating
the variation of outflow from storage with different tank sizes and rainfall conditions.

3.1 Precipitation and runoff analysis in the rainfall process

The rainfall process analyzes the rainfall conditions and runoff from building rooftop
catchments. The rainwater tank controls the runoff by accumulating flow from the roof,
similar to the operational conditions of conventional drain control basins in present urban
drainage systems. Thus, the design rainfall curves for a conventional basin can be used in
the RSD model. Huffs distribution and the Yen & Chow method are reasonable for design
rainfall analysis for Korean metrological conditions (MOCT, 2000). This study draws
design rainfall curves on the 2nd-quartile, 50% rainfall cumulative curve of Korean Huff
curves. This pattern and probability is proposed for the design of most work for
hydrological facilities by Huff (Huff, 1967). Return periods of 10-year, 20-year and 30year rainfall curves of Seoul are analyzed for the following case study of rainfall tank
design. Figure 6 shows the design rainfall curves for a 10-year-frequency storm in the
Seoul City area.


 
     
   
"
%!
G

@ <F
DE

BC
@= A
= <?>
;<


$
#
"
!


&!(')*$+, " $- ./0# +1+2!3+'0!0!+4! " ! - /.//,/#

5 6 7%89 :

Figure 6. Design rainfall curves of 10-year-frequency in the Seoul city area

Rainfall on a catchment surface causes runoff to flow from the area into the rainwater
tanks. The inflow into the tank is one of the governmental conditions on the tank design.
The quantity of water collected during the rainfall process is controlled by the surface and
the structural design condition of the buildings roof. The conditions of rooftop surface
design were studied with respect to runoff from the rainfall events of the design curves.
Figure 7 shows the detailed factors of the rooftop.

Figure 7. Scheme of runoff during rainfall

This study developed a simple equation for runoff curves from rooftops. Conventional
hydrological runoff models can be applied to larger areas, over 1.3 ha (Ponce, 1989). A
rooftop area is generally about only 0.1~0.3 ha. The rooftop is an artificial surface
designed to drain rapidly and completely. The assumptions in Table 1 should be reasonable
for the conditions of rooftops. They are the same as the conditions for ideal runoff surface
in Mulvaneys equation, meaning that the runoff flow is the same as the rainfall quantity
(Huggins et al, 1982). The runoff curves have the same pattern as the original rainfall event.
The calculated runoff curves were synchronized with monitored curves in their pattern and
peak breaking points in a field study of the rainwater system of SNU (Kim, 2006). Table 1
shows the structural conditions of roof drain and the assumptions of the RSD model.
6

Table 1. Factors and assumptions for runoff analysis

Factors
Catchment
area
Material
Drain
condition
Flow time
Depression
storage
Runoff
formula

Field conditions
0.1~0.3 ha
Impervious material, i.e., concrete or
waterproof sealing
Architectural design for rapid and
complete drainage
Design for short time flow

Notes and assumptions

Minimum limit for conventional
models = 130 ha

inflow from rainfall = runoff flow

Time for concentration = 0
No depression storage

0.5~0.1 inch(ASCE, 1992)

Q(t): runoff flow = I(t) : intensity per unit time A: catchment area

3.2 Tank capacity design in the storage process

The storage process needs an analytical method for determining the peak values that vary
with tank capacities. Figure 8 is a scheme of a simple rainwater tank structure and its water
flow. In this system, a water balance equation can calculate the flow variation against tank
capacities.

Figure 8. Structure and water flow of a simple rainwater tank

Vt i =

ti
t0

(Qin Q out Q sup ) dt

(1)

Equation 1 is the water balance equation for the system of Figure 5. Qin is the inflow
quantity to the rainwater tank, the same as the runoff from the roof. Qout is the overflow
from the tank into the urban drainage. Qsup, the water flow of supply water to the building,
can be assumed to be zero because the water demand during a rainfall event would be
random and small, having no relation to the rainfall event.

Figure 9. Tank capacity vs peak outflow curves for 10 years in Seoul City

Figure 9 shows the tank storage capacity versus the maximum outflow value curves of 10year rainfall events in Seoul. The unit of tank capacity is expressed as catchment depth;
that is, tank volume (m3) per unit catchment area (m2). The deep black dotted line shows
the maximum peak outflow values following tank volumes, which is the general tank
storage capacity vs maximum outflow curve for a 10-year-frequency storm in Seoul city.
Although only three rainfall times are shown in Figure 6, events of many other durations
and intensities are possible.

4. Case study of RSD model in Seoul

Figure 10 shows the tank volume versus peak outflow curves for 10-, 20- and 30yearfrequency rainfall events according to the method of Figure 6. On the curves, the minimum
necessary tank volumes to peak outflow values can be found.

Figure 10. Tank volumes vs peak overflows as rainfall frequencies in Seoul

As an example of the tank design from Figure 10, Table 2 shows the analysis of the runoff
control effect of a 29 /m2 rainwater tank.
8

Table 2. Runoff control effect analysis of 29L/ H

tank in Seoul from Figure 7

W/O tank
W/ tank of 29 /m2
Rainfall
Peak runoff flow
frequency Peak runoff flow
Peak
Peak
SR
SR
(year)
frequency
frequency
( IKJ LTM3O PUJ
( IK2J LNM3O PQJ
10
0.22
10 year
0.16
< 10 year
20
0.26
20 year
0.20
< 10 year
30
0.28
30 year
0.22
10 year

Peak value
ratio
0.7
0.9
1.0

In the case of the RSD system with a tank of 29 /m2 in Seoul, the rooftop peak runoff for
a 10- or 20-year-frequency rainfall event can be controlled. The RSD system could control
the peak outflow of the 10 year rainfall event using only about 70% of the capacity.
Additionally, the 30-year peak flow can be reduced to almost the same value as a 10-yearfrequency rainfall event. This means that an RSD system with a storage tank of 29 /m2
capacity can control the runoff of a 30-year rainfall event from a unit catchment area with a
drainage pipe designed for a 10-year rainfall event.

5. Conclusion and discussion

In this study, the rainwater tank system for runoff management was designed using the
RSD model to manage heavier rainfall runoff without changing the existing urban drainage
system. The present urban flood problem can be solved by engineered rainwater tanks
without the lengthy and costly operation to increase the capacity of the storm sewerage.
The RSD model can be a simple and practical design method for rainwater tanks for field
engineers who are not accustomed to the RSD system.
This model depends on the 2nd quartile, 50% curve design rainfall, which is proposed as a
good for most of case in hydrological structure design by Huff (Huff, 1967). However, The
tank volume (29 /m2) from this study would be an underestimated value for real Korean
heavy rainfall. Further study on Huffs distribution for rainwater tanks will be developed.
Effective flood control in flood-prone districts needs a harmonious system of decentralized
rainwater management that can be realized by scientific planning and decision-making
tools. Rainwater tanks can be located in optimum positions for the best control. The RSD
model can be developed into an integrated engineering model for the planning and design
of decentralized rainwater management systems.

References
Seoul Metropolitan City Government, 2002, 2001 flood disaster white book, Korea Water Resources
Association
Younjin Kim, Mooyoung Han, 2006, Development of RSR system for urban drain capacity improvement,
The 6th International Workshop on Rainwater Harvesting & Management proceeding book, Seoul, Republic
of Korea.
Mooyoung Han, Sungwon Park, Haryong Kim, 2006, An analysis of rainwater quality in rainwater
harvesting system at dormitories in Seoul National University, Seoul, Korea, The 6th International Workshop
on Rainwater Harvesting & Management proceeding book, Seoul, Republic of Korea.
Jungsoo Mun, Hyoungkeun Yoo, Jaeyoun Jang, Mooyoung Han, 2006, Particle behavior and removal in a
rainwater storage rank and suggestion for operation, The 2nd Rainwater Harvesting and Management
Workshop, International Water Association 5th World Water Congress and Exhibition, pp213~221
Jungsoo Mun, Hyoungkeun Yoo, Mooyoung Han, 2007, Particle Removal in a Rainwater Storage Tank and
Suggestion for Operation & Design, Journal of the Korean Society of Water and Wastewater, Vol. 21, No. 1,

pp131~138
Mooyoung Han, Eunkyung Roh, Sangrae Kim, Inbong Choi, 2004, Rainwater Harvesting Demonstration
Projects in Seoul National University, 2004 KSWQ-KSWW Joint Spring Academic Conference proceeding
book, Korean Society of Water Quality, Korean Society of Water and Wastewater
Richard H. McCuen, 1974, A Regional Approach to Urban Storm Water Detention, Geophysical Research
Letters 1(7)
Gye Woon Choi, Jung Hun Park, Seugnkyu Choi, 2001, The Causes and Measures on the Flood in Incheon
Area, 2001 KSCE Regular Academic Conference proceeding book, Korean Society of Civil Engineers
Sang-Yong Shin, Chang-Geon Yeo, Chang-Hyun Baek, Yoon-Jong Kim, 2005, Mapping Inundation Areas
by Flash Flood and Developing Rainfall Standard for Evacuation in Urban Settings, the Journal of Korean
Association of Geographic Information Studies 8(4), The Korean Association of Geographic Information
Studies.
Geunyoung Kim, 2005, Interactive Strategies between Disaster Management and Urban Planning for the
Damage Mitigation of Urban Floods, the Journal of Korean Society of Civil Engineers. 53(9)
Makoto Miyazaki, 2006, The flood impact of reservoir, Journal of Hydrological system 61(1), Japan
G. T. McCarthy, 1939, The Unit hydrograph and Flood Routing, Army Engineer District, Providence, RI,
USA
Yiping J. C. Guo, 1999, Detention storage volume for small urban catchments, Journal of Water Resources
Planning and Management 125(6)
Yiping J. C. Guo, 2004, Hydrology-based approach to storm water detention basin using new routing
schemes, Journal of Hydrologic Engineering 9(4)
T Hermann, U. Schmida, 2000, Rainwater utilization in Germany : efficiency, dimensioning, hydraulic and
environmental aspects. Urban Water 1(4)
G Vaes, J. Berlamont, 2001, The effect of rainwater storage tanks on design storms, Urban Water 3(4)
Mooyoung Han, Myoungsil Han, Sangrae Kim, 2004, A Consideration in Determining the Tank Size of
Rainwater Harvesting System in Buildings, Journal of the Korean Society of Water and Wastewater, Vol. 18,
No. 2, pp99~109
Jungsoo Mun, Myungsil Han, Dongwon Ki, Mooyoung Han, 2005, Design of Rainwater Tank Capacity for
Preventing sewer pipe Flood, 2005 KSWQ-KSWW Joint Spring Academic Conference proceeding book,
Korean Society of Water Quality, Korean Society of Water and Wastewater
Ministry of Construction and Transportation, 2000, 1999 report of research and survey on water resources
management and development method, Republic of Korea
F. A. Huff, 1967, Time distribution of rainfall in heavy storms, Water Resources Research 3(44)
American Society of Civil Engineers, 1992, Design and Construction of Urban Stormwater Management
Systems, New York, ASCE
V. M. Ponce, 1989, Engineering Hydrology principles and practices, Prentice-Hall
L. F. Huggins, J. R. Burney, 1982, Surface runoff, storage and routing, Hydrologic modeling of small
watersheds, C. T. Haan, ASAE

10