Simulations Show Greening Could Cool Japan’s Fukuyama Station Square by Nearly 0.6°C

A sweltering station square in western Japan could become measurably cooler if planners cover its bus-dominated tarmac with grass and trees, according to a new open-access study that combines five days of field measurements with high-resolution numerical simulations. The research, published in the journal City and Built Environment, examined the square south of Fukuyama Station, a site where the city is actively planning a spatial reorganization, and translated its simulation results into concrete redesign proposals structured around practical redevelopment processes.

Fukuyama, the second-largest city in Hiroshima Prefecture with roughly 460,000 residents, sits about seven kilometres from the open sea, with its southern boundary facing the Seto Inland Sea. Its station handles approximately 34,000 passengers per day, and the adjacent square functions as a major public space. At present, much of that square is occupied by transportation infrastructure, including a bus rotary. In March 2023 the city published a basic policy for redeveloping the square into a more pedestrian-oriented space, and although that policy does not explicitly address urban heat, its conceptual illustrations feature grassy open spaces and water basins, elements directly relevant to heat mitigation.

The research team, Makoto Yokoyama and Kazunari Watanabe, first characterized the existing summer conditions through short-term field measurements. Four standalone weather stations recorded air temperature, wind speed and wind direction at 10-second intervals at roughly two metres above ground, the height experienced by pedestrians, from 11:00 on 9 September to 18:00 on 13 September 2024, all under clear skies. During this period the mean daily maximum air temperature reached approximately 36 °C and the mean daily minimum about 27 °C, indicating persistently hot conditions day and night. Although the four measurement points lay within roughly 130 metres of one another, the maximum inter-site temperature difference approached 2 °C, with the hottest readings consistently at Point C and the coolest at Point A or B.

The wind data explained much of that spatial variation. Wind speeds were lower at night and higher during the day, and Point B recorded the highest speeds while Point D recorded the lowest, a ranking that mirrored the temperature patterns. Wind roses showed a predominantly southerly prevailing direction, which the authors attribute to sea-breeze circulation from the Seto Inland Sea. Points more directly exposed to those southerly winds tended to be cooler than points closer to the station building, demonstrating that even within a small square, localized wind environments strongly shape near-surface temperatures.

For the simulation component, the team used the Multi-Scale Simulator for the Geoenvironment, an atmosphere-ocean coupled model developed by JAMSTEC that runs on the Earth Simulator. Four nested domains were established with one-way nesting; Domains 1 through 3 functioned as mesoscale meteorological models, while Domain 4 operated at two-metre horizontal resolution as a time-dependent Large Eddy Simulation that explicitly resolves building geometry. This atmospheric component was coupled with a three-dimensional radiative transfer model in which tree crowns are represented explicitly, so shading and transpiration cooling are incorporated, with tree height, crown base height and leaf area density prescribed at 20 metres, 10 metres and 1.0 m² m⁻³ respectively. The target date was 31 July 2021, a clear-sky summer day, with analysis focused on the 13:30 to 14:30 window representing peak thermal stress.

Validation against the field measurements was deliberately framed as preliminary. For wind speed the mean absolute error was 0.41 and the correlation coefficient a moderate 0.58, suggesting the spatial variation among the four points was reasonably reproduced. For air temperature the correlation was relatively low at 0.25, partly because the simulated mean at Point B exceeded the observed value by about 1.5 °C; excluding that point raised the correlation to 0.69. The authors note that the observation days and simulation date differed, and that further verification of absolute accuracy under temporally aligned conditions is required. Nevertheless, because the model partially reproduced the observed spatial patterns, they judged it suitable for the subsequent scenario analyses.

The baseline simulation revealed stark contrasts within the study area. Wind speeds reached about 4 m/s near mid- and high-rise buildings, where downdrafts associated with upper-level winds descended to the surface, an effect apparently reinforced by the Tenmaya Store, which intercepts upper-level winds and aligns with the adjacent roadway. Air temperatures peaked near 37 °C in sunlit areas, with a maximum spatial difference of roughly 2 °C, while wet-bulb globe temperature, the heat-stress index widely used in Japan and standardized under ISO 7243, reached around 30 °C in sun-exposed spots, corresponding to the official

“Severe Warning” category, whereas shaded areas registered near 25 °C, at the “Warning” level. The index combines natural wet-bulb temperature, globe temperature and air temperature in a 0.7, 0.2 and 0.1 weighting, and in Japan it serves not only as a research metric but as a practical guideline for outdoor activities and sports, with risk levels ranging from “Almost Safe” below 21 °C to “Danger” at 31 °C and above.

Averaged over the analysis window, the two sub-areas of the study site diverged noticeably. Mean wind speed was higher in the area in front of the Ines FUKUYAMA building than in the station square itself, primarily because of downdrafts from upper-level winds. Mean air temperature, by contrast, was about 0.6 °C higher in the station square, where elevated temperatures prevailed across most of the surface except for the shade cast by the NEW CASPA building to the northwest. The contrast in heat stress was sharper still: the mean WBGT difference between the two areas reached approximately 1.6 °C, with the station square averaging 28.44 °C, within the “Severe Warning” category, while the area in front of Ines FUKUYAMA fell within the “Warning” category. The authors attribute this marked difference to the strong reduction of WBGT along building-shaded zones, suggesting that shading and downdrafts together can yield greater reductions in thermal stress than either process alone.

To test the cooling potential of greening, the researchers built two scenario models by modifying land-use classifications and tree placements in the current-condition model, with reference to Fukuyama City’s future vision illustrations. Total green area increased substantially from 756 m² under current conditions to 9,440 m² in the square-greening scenario and 11,568 m² in the combined square-and-road scenario. Tree canopy area rose to 372 m² and 548 m² respectively, and a 100 m² water feature was introduced in both cases, although its extent remained small compared with the increase in vegetated area.

The simulation results showed a clear cooling effect on air temperature. Average air temperatures across the two areas fell to 35.43 °C and 35.40 °C under the two greening scenarios, reductions of roughly 0.6 °C relative to the 35.97 °C baseline. The reduction was more pronounced within the station square than in front of Ines FUKUYAMA, where pre-existing shading and downdraft effects had already moderated conditions. Slight cooling was also observed in non-greened locations, such as along the railway tracks, suggesting a downwind propagation of the cooling influence. In the temperature-band analysis, the share of locations exceeding 35.5 °C declined while the proportion below that threshold increased, with the 36.0 to 37.0 °C range shrinking substantially.

The effects on heat stress were more modest. WBGT decreased by more than 1 °C only where new trees were introduced, an effect the authors attribute primarily to additional shading and note was spatially localized. Areas that became sunlit after the removal of bus stops showed locally elevated WBGT values, and in unshaded portions of the square the decrease was less than 0.5 °C. Although the proportion of locations exceeding 29.5 °C declined, overall values remained within the “Severe Warning” category, indicating that surface greening alone is insufficient to shift the overall risk level. Wind speeds, meanwhile, changed little, though the removal of wind-obstructing structures such as bus stops allowed the southerly inflow to penetrate slightly further into the square.

From these findings the authors derived a two-stage redesign strategy organized around how redevelopment decisions are actually made, with facility placement addressed before detailed spatial design. Thermal mitigation measures should be prioritized in the sun-exposed northern station square, where a grass square concentrated in exposed areas is expected to produce meaningful air-temperature reductions, while strategic shading aligned with pedestrian circulation is necessary to address heat stress. Because the cooling influence of grass extended downwind, placing greenery in upwind positions near Ines FUKUYAMA may further enhance performance, though the authors treat this as a secondary priority. In the spatial design stage, structures and trees must be placed and configured so as not to obstruct airflow, since improper tree placement can reduce ventilation efficiency, a caution illustrated by the existing bus stops that partially impeded wind dispersion.

The authors acknowledge several limitations. The analysis covered a limited number of observation days and time periods, the field measurements and simulations were conducted on different dates, and the validation focused on reproducing spatial patterns rather than absolute values, so further verification under temporally aligned conditions is needed. Because the scenarios were built from conceptual illustrations, the cooling potential of water features was not fully examined, and the authors suggest that future work explore a wider range of redevelopment configurations and develop visualization-oriented planning tools to communicate results to other urban districts.

Subject of Research: Simulation-based assessment of summer thermal conditions and greening scenarios for the redevelopment of Fukuyama Station Square in Japan

Article Title: Assessment of summer thermal environment and greening effects in Fukuyama Station Square: proposal for square reorganization considering thermal environment improvement

Article References: Yokoyama, M., & Watanabe, K. (2026). Assessment of summer thermal environment and greening effects in Fukuyama Station Square: proposal for square reorganization considering thermal environment improvement. City and Built Environment, 4(1), Article 7. https://doi.org/10.1007/s44213-026-00073-y

Image Credits: AI Generated

DOI: 10.1007/s44213-026-00073-y

Keywords: urban heat island, Fukuyama Station Square, thermal comfort, WBGT, green infrastructure, large eddy simulation, sea breeze, shading, urban redevelopment, Japan, air temperature, city planning