Rain Gardens and Barrels Slash Urban Flooding Nearly 30% in Dense Tokyo Suburbs

Researchers from Shibaura Institute of Technology (led by figures including Jiajing Lin and Hitoshi Nakamura) have studied rain gardens (RGs), often combined with rain barrels (RBs), as an effective, balanced green-infrastructure approach for managing urban stormwater and reducing pluvial flooding in densely built- up residential areas near Tokyo, such as Miyahara-Cho and other sites in Saitama City.

Saitama City lies roughly 20- 30 km north of central Tokyo and includes low-lying, densely developed residential zones vulnerable to flooding from intense or prolonged rainfall, a risk heightened by climate change and extensive impervious surfaces.

Traditional gray infrastructure (sewers, pipes) can be overwhelmed; small-scale, decentralized low- impact development (LID) measures that restore more natural water processes are therefore of interest.

Key findings from the studies

Using the U.S. EPA Storm Water Management Model (SWMM), the team simulated short-duration and long-duration rainfall events:

• Rain gardens alone meaningfully reduced flood volume (e.g., ~13% in one analysis) and peak runoff, with greater effect than rain barrels alone (~1% flood- volume reduction in the same comparison).
• Combining RBs (for temporary storage) and RGs (for infiltration, filtration, and delayed release) performed best. e.g., ~29% reduction in flood volume rate in one case study, and better attenuated total runoff and early peaks during longer events. The combination helped keep excess water from overwhelming sewers and downstream rivers.⁠
• Benefits were stronger for the first peak of multi- peak storms; effectiveness declined once storage and infiltration capacity was exceeded.

A later comprehensive assessment applied the Analytic Hierarchy Process (AHP) to weigh multiple criteria across scenarios in three flood- prone Saitama residential areas:

• Flood- control benefits (peak timing, flood conditions, total runoff).
• Economic factors (lifespan, installation, management, and maintenance costs).
• Social and environmental benefits (aesthetics, education, outreach value, biodiversity support, carbon reduction), evaluated partly with tools such as B£ST and SWMM.⁠

This multi- criteria framing supports the idea of rain gardens (and RB and RG combinations) as a balanced solution rather than a purely technical one: they deliver measurable runoff and flood reduction while adding co- benefits that pure gray infrastructure typically lacks, at costs and scales suitable for dense residential settings.

Follow- on work from the same group has explored “smart” variants (sensor- and weather- forecast- controlled rain gardens or barrels) that further improve infiltration and reduce runoff relative to conventional designs by actively directing water to optimal soil zones or pre- emptying storage.

Practical context

Rain gardens are shallow, planted depressions that capture runoff from roofs, driveways, and streets, hold it temporarily, and allow it to infiltrate or evaporate while plants and soil filter pollutants.

In the Japanese suburban context studied, they can be installed at household or neighborhood scale (e.g., ~20 m² plots in simulations) and complement existing drainage. Related real- world efforts in Tokyo (e.g., Setagaya Ward initiatives) also promote household- scale rain gardens for flood- risk reduction, heat- island mitigation, and community engagement.

Overall, the Shibaura research indicates that rain gardens, especially when combined with storage measures and evaluated across flood, cost, and social dimensions, offer a practical, multi- benefit option for denser urban and suburban areas facing increased stormwater challenges near Tokyo.

Implementation still depends on local soil permeability, available space, maintenance, and integration with larger drainage systems.

In the Shibaura Institute of Technology study (Lin et al., 2023) on Miyahara-Cho, Saitama City, EPA SWMM version 5.1 was used to model sewer pipes, subcatchments, and LID controls (rain barrels and rain gardens).

Key setup and parameters include the following.

Subcatchment and catchment parameters

• Area: Measured from Google Earth Pro for each catchment.
• Width: Calculated as $0.8\times \sqrt{\text{Area}}$0.8×Area​.
• % Impervious: Set to 60% (based on the Saitama City infiltration capacity map and SCS Curve Number method).⁠

Sewer elements (manholes, pipes, outfalls) used invert elevations, max depths (pipe diameters), and lengths from local sewer maps. Model calibration and validation compared simulated outlet water levels to Saitama City observed data, with good agreement at peaks.

Precipitation scenarios used AMeDAS data: a short- duration event (July 15, 2021; peak 24 mm/h) and a long- duration event (July 2, 2021).

Rain barrel (RB) parameters (Table 1 in the paper)

RBs were modeled as storage LID units, one per detached house:

• Barrel height: 1000 mm
• Unit area (bottom): 0.2 m²
• Capacity: 200 L
• Total installations: 9600 (total area 1920 m²)
• Offset: 6 mm
• Flow coefficient: 0
• Flow exponent: 0.5
• Drain delay: 6 h⁠

(These draw from typical residential sizing and prior modeling references such as Maya et al. 2010.)

Rain garden (RG) parameters (Table 2 in the paper)

RGs were modeled primarily with surface and soil layers (bioretention- style):

• Area per RG: 20 m²
• Total installations: 1080 (total area 21,600 m²)
• Berm height: 200 mm
• Soil thickness: 700 mm
• Conductivity (saturated hydraulic conductivity of the media): 87 mm/h
  (This implies roughly 12 hours for full infiltration after the garden reaches capacity.)⁠

Local soil saturated hydraulic conductivity near the site was noted as $3.0\times {10}^{-5}$m/s.

Combination scenario

Rainwater from roofs is routed first into the RBs. When barrel capacity is exceeded, overflow is directed via pipes into the RGs. Separate sub catchments were defined for buildings (RBs), gardens and open spaces (RGs), and parks (no LID, direct to sewer).

These parameters produced the reported flood- volume- rate reductions (including the ~29% figure for the combined short- duration case).

Exact layer- by- layer SWMM LID inputs (e.g., porosity, field capacity, vegetation volume fraction, surface roughness) beyond the values above are referenced to standard literature defaults or the cited sources but are not exhaustively tabulated in the publicly extractable text of the paper.

Comprehensive evaluation of the role of rain barrels and rain gardens in densely built- up residential areas near Tokyo

The ~29% (specifically 29.26%) figure refers to the reduction in flood volume rate achieved by combining rain barrels (RBs) and rain gardens (RGs) in a SWMM simulation of a short- duration rainfall event in the Miyahara- Cho case study (Saitama City, near Tokyo).

How the reduction works

Researchers modeled the local sewer system and surface runoff under real or representative rainfall (including a short-duration event around July 2021). They compared a baseline “current condition” (no added RBs and RGs) against scenarios with:

• RBs only (typically capturing rooftop runoff into storage tanks, with overflow directed elsewhere),
• RGs only (shallow planted depressions that detain, infiltrate, and filter surface runoff from roofs, drives, and paved areas),
• and the combination (RBs and RGs).

Key quantitative results for the short- duration event (flood volume rate reductions relative to baseline):

• RBs alone: ~0.93%
• RGs alone: ~13.32%
• Combination: 29.26%

The combination also delayed peak runoff (e.g., by about 10 minutes in some RG scenarios) and reduced total sewer- system runoff volume more effectively than either measure alone.

Mechanisms behind the numbers

1. Storage and controlled release (RBs): Barrels temporarily hold a portion of rooftop rainwater. When full, overflow can be directed toward the rain gardens rather than straight into the sewer. This reduces the immediate volume entering the drainage network during the intense part of a storm.
2. Detention, infiltration, and filtration (RGs): Rain gardens create temporary surface storage in a planted basin with engineered or amended soil. Water slows down, soaks into the ground (reducing peak flows), and is partly taken up by plants or evaporated. This cuts the volume and speed of runoff that would otherwise overwhelm sewers or cause surface flooding (pluvial flooding).
3. Synergy of the combination: RBs capture concentrated roof runoff early; excess is then routed to RGs for further infiltration and detention. Together they intercept more total volume before it reaches the sewer system, lowering the “flood volume rate” (the excess volume that causes surcharge, overflow, or ponding). The effect is strongest early in a storm while storage and infiltration capacity remain available. In longer multi- peak events the first peak is still attenuated, but later peaks see diminished returns once capacity is saturated.

The model used the U.S. EPA Storm Water Management Model (SWMM) with local parameters for soil infiltration rates, barrel sizes, garden dimensions (e.g., berm height, soil depth), contributing drainage areas, and sewer capacity.

Results are therefore specific to the dense residential layout, topography, and rainfall pattern of the study area; absolute percentages would differ under other conditions, rainfall intensities, or implementation scales.

In short, the 29% reduction is a modeled outcome showing that pairing temporary storage (barrels) with distributed infiltration and detention (gardens) removes a substantial share of the runoff volume that would otherwise contribute to urban pluvial flooding in that setting.

Journal: ScienceDirect, City and Environment Interactions (2026)

DOI: 10.1016/j.cacint.2026.100480

Provided: Shibaura Institute of Technology

Authors: Jiajing Lin (Regional Environment Systems, Shibaura Institute of Technology, Japan),

Hiroyuki Taguchi (Department of Planning, Architecture and Environmental Systems, Shibaura Institute of Technology, Japan) and

Hitoshi Nakamura (Department of Planning, Architecture and Environmental Systems, Shibaura Institute of Technology, Japan)

Abstract

Low Impact Development (LID) has become an important approach for mitigating urban pluvial flooding by reducing stormwater runoff while providing additional environmental and social benefits. Although numerous studies have evaluated LID performance using hydrological simulation, economic assessment, or ecosystem-service valuation, relatively few have integrated these approaches into a unified decision-making framework for densely built-up residential areas.

This study develops an integrated evaluation framework that combines the Storm Water Management Model (SWMM), the Benefit Estimation Tool (B£ST), economic life-cycle assessment, and the Analytic Hierarchy Process (AHP) to compare three residential-scale LID scenarios: Rain Barrels (RB), Rain Gardens (RG), and their combined implementation (RB + RG). Three flood-prone residential catchments in Saitama City, located near Tokyo, were selected as case studies. Flood-control performance was evaluated using peak runoff, total runoff volume, and peak runoff timing obtained from SWMM simulations. Economic performance was assessed using life-cycle implementation and maintenance costs, while social benefits, including aesthetic value, educational value, biodiversity enhancement, and carbon reduction, were quantified using B£ST. All indicators were normalized and incorporated into the AHP framework to generate comparable comprehensive scores.