Main

Rooftop photovoltaic solar panels (RPVSPs) have been promoted both locally and globally to address energy demand1,2 as RPVSPs material advancements3 hold the promise of higher efficiency and reduced costs, making them accessible worldwide4. However, the effects of city-scale deployment of RPVSPs on the urban microclimate remain uncertain. RPVSP efficiency diminishes in high ambient temperatures and humidity, raising concerns about their environmental impact and influence on the urban thermal environment4. Recent studies suggest that city-scale RPVSP deployment might either decrease or increase urban temperatures5,6,7. For example, it has been observed that the deployment of RPVSP on a building not only curtailed greenhouse gas emissions but also reduced annual cooling loads8. Similar results have been observed in simulation-based studies and in the case of medium-to-large-scale RPVSP deployment9,10,11,12. For instance, in the Paris metropolitan area during summer, RPVSPs reduced the urban heat island effect by 0.2 °C during the day and up to 0.3 °C at night, resulting in energy savings for air conditioning of up to 12% (ref. 9). In Sydney, RPVSPs reduced summer maximum temperatures by up to 1 °C because the need to import energy is offset by local generation10. Similarly, the cooling effect with RPVSPs in Los Angeles was found to be up to 0.2 °C (ref. 11). Further, in Phoenix, Tucson and Arizona, RPVSPs lowered the 2 m air temperature compared with conventional roofs and resulted in 8–11% reduction in cooling energy demand12.

However, studies suggesting an increase in urban microclimatic temperature have observed that RPVSPs on urban rooftops might elevate the temperature at building scale, causing discomfort in outdoor urban environments13,14. Similar findings appear from the simulation studies5,7,14,15. For instance, large-scale RPVSP deployment can increase outdoor temperatures up to 0.5 °C measured by the universal thermal comfort index during heatwaves7. RPVSPs can induce a temperature rise of up to 1.5 °C during the daytime due to less efficient heat storage and increased sensible heat flux (Qsensible) from both RPVSPs and roof surfaces15. Additionally, RPVSP systems can significantly elevate daytime city temperatures and marginally lower them at night, depending on the climate zone and building types14. In Sydney, city-scale RPVSP deployment at the local district scale increased daytime air temperature by 1.5 °C while reducing it by 2.7 °C at night5.

While existing literature reports on the impact of RPVSP on the urban environment, most are based on in situ field experiments or building-scale simulations, lacking a comprehensive multicity-scale analysis. These studies also neglect convective heat transfer between the roof surface and the back of solar panels. Even most of the modeling studies with RPVSP were purely based on simulations and were not evaluated in any shape or form. Our study addresses these gaps by incorporating new parameterizations for RPVSPs, including convective heat transfer, resulting in more aligned results with other studies incorporating similar considerations. The presence of RPVSPs on urban buildings, particularly rooftops, can have complex effects on the local thermal environment. When RPVSPs are installed on roofs, they absorb a significant amount of solar energy, converting some of it into electricity but also generating heat in the process. This heat is released into the surrounding air, leading to an increase in air temperature around the panels. Moreover, the elevated installation of RPVSP creates two hot surfaces: the top surface of the panels and the underside surface. As air flows over these RPVSPs, it picks up heat more efficiently than it would from typical building or ground surfaces. Observational studies in the literature have shown that areas with RPVSP arrays can experience higher daytime air temperatures compared with reference sites without RPVSP. However, the effect on nighttime temperatures is generally minimal.

Given these conflicting studies, it is imperative to comprehensively analyze the impact of city-scale RPVSP deployment in diverse cities globally. Our study focuses on the advanced capabilities of the latest weather research and forecasting (WRF) model integrated with multilayer urban schemes, such as a building energy model (BEM) coupled with building effect parameterization (BEP), that is, WRF/BEP + BEM, especially in tropical cities. This model accurately captures the energy balance of RPVSPs, an aspect overlooked in previous studies. To address this concern, we have validated the WRF/BEP + BEM-simulated RPVSP surface temperature against an experimentally validated model. Evaluating RPVSP at a city scale is crucial for sustainable energy solutions as it assesses broader impacts on urban microclimates and energy consumption patterns. Through city-wide implementation scenarios and simulation studies, we are suggesting improvements to RPVSP efficiency, which can lead to a decrease in convective heat generation. This is crucial because the efficiency of RPVSPs decreases by approximately 0.5–0.8% for each 1 °C rise above standard test conditions (25 °C)16. Our comprehensive assessment includes parameters such as surface energy budgets, near-surface meteorological fields, boundary layer dynamics and sea breeze circulations. The findings can inform policy discussions and lead to redesigning policies for RPVSP implementation in cities. By addressing overheating issues and advocating for advanced RPVSP technologies and efficient cooling strategies, we aim to promote sustainable urban development practices and enhance the understanding of RPVSP deployment’s broader impacts.

Results

Regional impacts on urban surface energy budgets

City-scale deployment of RPVSPs led to changes in the Qsensible from urban surfaces. At 14:00 LT, Qsensible peaked at 415.9 W m−2 for RPVSPs 25%, 469.2 W m−2 for RPVSPs 50%, 533.9 W m−2 for RPVSPs 75% and 595.2 W m−2 for RPVSPs 100%. Similarly, at 18:00 LT, the average Qsensible values were 245.6 W m−2, 271 W m−2, 285.4 W m−2 and 306.6 W m−2 for RPVSPs 25%, RPVSPs 50%, RPVSPs 75% and RPVSPs 100%, respectively. This significant rise was primarily attributed to RPVSPs’ low albedo value, leading to amplified solar energy absorption, predominantly emitted as sensible heat rather than converted into usable electricity. During the early morning and nighttime, minor fluctuations in Qsensible were observed, revealing the intricate interplay of RPVSPs with the urban heat environment. Furthermore, the latent heat flux (Qlatent) from the urban surface exhibited a corresponding increase owing to the energy mechanisms of the RPVSPs array. At 14:00 LT, the peak Qlatent reached 100 W m−2 for RPVSPs 25%, 107.8 W m−2 for RPVSPs 50%, 113 W m−2 for RPVSPs 75% and 121.5 W m−2 for RPVSPs 100%. Likewise, at 18:00 LT, the average Qlatent ranged from 66.3 W m−2 to 80.3 W m−2 for RPVSPs 25% to RPVSPs 100%, highlighting the increased Qlatent resulting from RPVSP installation. These alterations were associated with a marginal temperature rise surrounding the RPVSPs, potentially influencing evaporation and significantly contributing to the Qlatent of urban areas. Moreover, the installation of RPVSPs over building roof surfaces led to augmented ground heat storage (Qstorage) on urban surfaces. In the early morning, the average Qstorage ranged from 53.5 W m−2 to 56.6 W m−2 for RPVSPs 25% to RPVSPs 100%, respectively. The maximum Qstorage during peak hours ranged from −270.2 W m−2 to −310.8 W m−2 for RPVSPs 25% to RPVSPs 100%, with the maximum increase during this time ranging from 18.9 W m−2 to 59.5 W m−2. The convention taken is that the surface energy fluxes are counted positive when they are upward (relative to the surface). Diurnal changes in Qstorage showed an average increase during the daytime, ranging from 9.9 W m−2 to 32.8 W m−2 for RPVSPs 25% to RPVSPs 100%, and during nighttime, the average increase ranged from 4.9 W m−2 to 10.4 W m−2 for all RPVSP scenarios. This is due to the additional radiation and heat transfer between the RPVSP and roof surface beneath the panel. Additionally, the widespread deployment of RPVSPs at a city scale altered the surface energy balance on the urban surface by transferring heat from RPVSPs to the urban atmosphere, as evidenced by changes in net radiation (Qnet) (Fig. 1). This describes the total radiation emitted to the atmosphere, including the radiation (longwave) from RPVSPs to the urban atmosphere. Early morning average Qnet ranged from −29.2 W m−2 to −8.1 W m−2 for RPVSPs 25% to RPVSPs 100%, while the peak Qnet varied from 786.1 W m−2 to 1,027.5 W m−2. This enhanced Qnet during peak hours ranged from 94 W m−2 to 335.4 W m−2 over the urban surface. Furthermore, the daytime maximum increase of Qnet ranged from 51.9 W m−2 to 193.5 W m−2, with a slight nighttime increase ranging from 10 W m−2 to 31.9 W m−2 for all RPVSP scenarios. These findings underscored the multifaceted impact of RPVSPs on the surface energy balance of urban areas.

Fig. 1: Changes in simulated Qnet.
figure 1

a–d, The changes (building rooftops with RPVSPs scenarios minus building rooftops without RPVSPs scenarios) in simulated Qnet during daytime (at 14:00 LT, top) and nighttime (at 02:00 LT, bottom) for RPVSPs 25% (a), RPVSPs 50% (b), RPVSPs 75% (c) and RPVSPs 100% (d) over an urban grid cell. The urban area of Kolkata is marked by the black line inside the box. CTRL refers to control case.

Full size image

Regional impacts on near surface standard meteorological fields

An examination of various atmospheric parameters, including ambient air temperature at 2 m height, urban surface skin temperature (Tsurface), relative humidity (Hrelative) at 2 m height and wind speed (Wspeed) at 10 m, was undertaken to assess the regional impact of widespread deployment of RPVSPs. Analysis of spatial average for urban ___domain suggests for the control case that the peak 2 m ambient air temperature (Tambient) occurred around 42.8 °C at 14:00 LT, while the minimum temperature was approximately 30.8 °C at 01:00 LT. Similarly, the maximum Tsurface of the urban area was approximately 47 °C at 14:00 LT, with a minimum temperature of around 34.8 °C at 01:00 LT. These temperature fluctuations reflected typical conditions during a strong heatwave in the tropical coastal urban environment. The adoption of RPVSPs induced changes in the energy balance, significantly affecting the urban air temperature. The maximum Tambient ranged from 43.5 °C to 44.3 °C at 14:00 LT, with an average increase of 0.5–0.8 °C at 18:00 LT (Fig. 2). Concurrently, the Tsurface exhibited a maximum increase ranging from 1.7 °C to 3.2 °C at 14:00 LT, with an average increase ranging from 1.3 °C to 2.1 °C at 18:00 LT (Fig. 3). Additionally, Hrelative at 2 m height experienced reductions of 1.2–5.6% during the day and 0.7–2.1% at night, while Wspeed increased by 0.5–1.2 m s−1 during the day and decreased by 0.2–0.6 m s−1 at night in urban areas (Supplementary Figs. 1 and 2). These alterations in atmospheric conditions were closely linked to the thermal properties of the RPVSPs, which significantly influenced the temperature, humidity and Wspeed in the urban environment, particularly during the diurnal cycle. The changes (experiment minus control) of Tambient and Tsurface have also been calculated (Supplementary Table 1). The increase in the 2 m Tambient and Tsurface during the daytime can be attributed to the heat absorption capacity of the RPVSPs and the subsequent release of stored heat during the nighttime, leading to a decrease in surface temperatures. The changes in Hrelative were influenced by alterations in atmospheric stability, induced by the modified energy balance due to RPVSP deployment. Furthermore, the shifts in Wspeed were influenced by the changes in the surface temperature and subsequent alterations in boundary layer mixing induced by RPVSP deployment. These findings stress the importance of evaluating RPVSPs’ impact on urban microclimates, especially in hot and humid regions. It highlights the need for effective mitigation strategies and emphasizes the significance of RPVSPs’ thermal properties in future urban planning and climate-sensitive designs.

Fig. 2: Changes in simulated Tambient.
figure 2

a–d, The changes (building rooftops with RPVSPs scenarios minus building rooftops without RPVSPs scenarios) in simulated 2 m Tambient during daytime (at 14:00 LT, top) and nighttime (at 02:00 LT, bottom) for RPVSPs 25% (a), RPVSPs 50% (b), RPVSPs 75% (c) and RPVSPs 100% (d) over an urban grid cell. The urban area of Kolkata is marked by the black line inside the box.

Full size image
Fig. 3: Changes in simulated Tsurface.
figure 3

a–d, The changes (building rooftops with RPVSPs scenarios minus building rooftops without RPVSPs scenarios) in simulated Tsurface during daytime (at 14:00 LT, top) and nighttime (at 02:00 LT, bottom) for RPVSPs 25% (a), RPVSPs 50% (b), RPVSPs 75% (c) and RPVSPs 100% (d) over an urban grid cell. The urban area of Kolkata is marked by the black line inside the box.

Full size image

Regional impacts on urban boundary layer dynamics

The combined effects of rapid urbanization and regional climate change-induced warming are expected to intensify heatwave events, impacting the planetary boundary layer (PBlayer) height in cities. The simulation demonstrates that widespread RPVSP deployment alters the lower atmospheric PBlayer structure over urbanized areas, such as Kolkata, especially during daytime (Supplementary Fig. 3). Generally, surface heating significantly influences PBlayer dynamics, playing a key role in shaping the PBlayer state in cities17,18. The changes in PBlayer depth by the adoption of RPVSPs are seen in the Kolkata urban region during the heatwave period. It is evident that the city-scale deployment of RPVSPs increases the PBlayer height over the urban region and the maximum increase (~415 m to ~615.6 m) was simulated by the most aggressive scenario (that is, RPVSPs 100%). During the heatwave period, the resulting increases in the convective PBlayer height indicate that the lower atmosphere has a great potential to mix air pollutants vertically, with possible consequences for air quality. The diurnal variability of PBlayer depth was simulated with the effects of the implementation of RPVSP technology at the city scale. The height of PBlayer depth increases significantly by city-scale deployment of RPVSP technology. The maximum increase of PBlayer height at 14:00 LT is about 165.3 m for RPVSPs 25%, 354.4 m for RPVSPs 50%, 456.4 m for RPVSPs 75% and 615.6 m for RPVSPs 100%. In the late afternoon (18:00 LT), the average increase of PBlayer depth over urban regions is about 126.4 m, 235.5 m, 328.6 m and 415.6 m for RPVSPs 25%, RPVSPs 50%, RPVSPs 75% and RPVSPs 100%, respectively. In the early morning, the PBlayer depth increased significantly over the urban region and the values are 105.4 m for RPVSPs 25%, 220.3 m for RPVSPs 50%, 310.1 m for RPVSPs 75% and 356.4 m for RPVSPs 100%. City-scale deployment of RPVSPs absorb solar energy, increasing Qsensible and turbulence, which raises the urban surface temperature and deepens the PBlayer in urban areas. The impact of RPVSPs on boundary layer structure is demonstrated through the simulated vertical profile of temperature, Hrelative and horizontal Wspeed was investigated during peak heatwave events (Fig. 4). RPVSPs intensify urban warming and convective mixing, potentially aiding air pollution dispersion. During peak heatwave events, implementing 100% RPVSPs over building roofs impacts the atmosphere up to 1.2 km. Hrelative decreases within the urban boundary layer, attributed to higher urban surface temperatures and reduced evaporation. RPVSPs enhance vertical Wspeed, reaching up to 1.2 m s−1 during daytime, extending up to 0.8 km in the urban boundary layer. The technology exacerbates urban temperature and lowers Hrelative, driven by changes in vertical mixing. Stronger vertical winds indicate enhanced energy transport from the surface to higher atmospheric layers. RPVSPs increase Qsensible, intensifying vertical mixing and deepening the PBlayer over the urban ___domain. In addition, the weaker vertical mixing over urban areas results in the advection of moist air coming from surrounding rural areas to urban areas and, consequently, the atmosphere beyond a certain height over the urban ___domain is barely affected by surface condition19. The weaker vertical mixing further developed stronger advection. However, this study does not indicate the stronger advection of moist air from rural areas because of the stronger vertical mixing of RPVSP scenarios. Surface heating of urban areas by RPVSP scenarios is the main important factor for controlling the vertical Wspeed in urban areas that enhances further development of a large convective roll over urban areas20.

Fig. 4: Changes in simulated vertical meteorological variables.
figure 4

a–c, The changes show the vertical distribution profile in simulated standard surface meteorological fields air temperature (a), Hrelative (b) and wind fields (c) between building rooftops with RPVSP scenarios (100% coverage) and building rooftops without RPVSP scenarios along point A to point B (marked in Supplementary Fig. 8) during peak hour (14:00 LT) over Kolkata.

Full size image

Regional impacts on sea breeze circulations

The height of the PBlayer is closely associated with the advection of the sea breeze in tropical coastal cities. RPVSPs have the potential to modify sea breeze circulation by altering surface conditions and the boundary layer structure, potentially triggering localized circulation patterns in urbanized regions. The urban surface affects the Qlatent, Qsensible and momentum flux, while boundary layer height and surface convergence help to modify the sea breeze circulation impact on the urban atmosphere. Recently, the intensification of sea breeze circulation is strongly regulated by the city-scale synoptic background climate, which plays a critical role in modulating the prevailing wind near the surface21. The local circulation is generated by the thermal incongruity between land and sea22,23 and, on the synoptic scale, wind. This local circulation can extend more than 100 km horizontally inland24 and vertically may reach up to 3–4 km during the mid-afternoon. Due to the effect of urban signature, the advancement of sea breeze circulation can be accelerated for many hours25. RPVSPs alter these factors, increasing PBlayer height and Wspeed in urban areas. The adoption of RPVSPs accelerates sea breeze circulation, increasing vertical Wspeed by 3–5 m s−1 over urban regions. Due to drag forces, high-rise buildings and high surface roughness delay the arrival of sea breeze in urban cores (Supplementary Fig. 4). RPVSPs induce sensible heating and enhanced wind fields, affecting the urban climate during heatwaves. City-scale RPVSP adoption alters the thermal pressure gradient, raising Tambient by up to 1.5 °C and horizontal Wspeed by 1.2 m s−1. Therefore, increasing the coverage of the roof surface by RPVSPs induces sensible heating and increased wind fields, resulting in feedback within the urban local climate during peak hours in heatwave events. RPVSPs enhance the advective flow between the city and its environs, accelerating local warming in urban environments. They create a regional low, which can increase the horizontal and vertical Wspeed over the urbanized region.

Discussion

We conducted a comprehensive assessment of the Kolkata RPVSPs study by comparing its findings with similar data from 30 existing case studies across various scales, including urban, continental and global contexts. The comparison is to assess the response of our model when using RPVSPs at the city scale, with results obtained from other cities with different background climate conditions. The case studies were categorized into two groups: (1) simulation-based studies and (2) observational studies. Within each group, the studies were further classified on the basis of their thermal effects into four categories: daytime warming, nighttime cooling, daytime cooling and nighttime warming (Fig. 5 and Supplementary Table 2). These case studies utilized both observational and modeling approaches, providing critical insights for the RPVSP community. The focus of these studies predominantly revolved around the assessment of first- and second-generation RPVSP systems. Among the 30 case studies, 17 reported daytime warming ranging from 0.1 °C to 1.5 °C. Only two studies found that RPVSPs might warm the atmosphere at night by 3.1–4 °C. On the other hand, 13 studies indicated that RPVSPs could reduce Tambient throughout the day by 0.02–1 °C, while 11 studies suggested that RPVSPs may decrease nighttime temperatures by 0.12–2.7 °C. Notably, five studies found that RPVSPs could contribute to both daytime warming by 0.34–1.5 °C and nighttime cooling by 0.24–2.7 °C. Our findings in the tropical region of Kolkata suggest that the widespread adoption of RPVSPs can significantly raise urban temperatures during the day, while typically cooling the urban environment at night. During summertime heatwaves, near-surface air temperatures in urban areas can increase by up to 1.5 °C during the day. However, it has been observed that RPVSPs can reduce nighttime maximum near-surface air temperatures by up to 0.6 °C when rooftops are fully covered with RPVSPs. Moreover, the extensive implementation of RPVSPs can lead to an increase in TSurface of urban grid cell of up to 3.2 °C during peak hours, with an average cooling effect of up to 1.4 °C during summer heatwaves at night. This increased temperature refers specifically due to the increased fraction of the RPVSPs on urban rooftops at the whole city scale. The RPVSPs absorb approximately 90% of solar energy, converting up to ∼20% of it into electricity, while the remainder contributes to their warming. As a result, these RPVSPs often become very hot surfaces due to RPVSPs surface temperature reaching ∼70 °C during peak hours in the urban building rooftops (Fig. 6). Typically installed elevated above roof surfaces, RPVSPs have both a hot top surface and a hot underside. As air flows over these RPVSPs, it absorbs heat from both surfaces, amplifying the heating effect nearly twice as much as if the same temperature were present on a standard building surface. Additionally, this heating is because the RPVSPs (monocrystalline types) we considered are thin (6.55 mm), resulting in lower thermal inertia compared with traditional roof surfaces. As a result, air temperatures around RPVSPs tend to remain higher during the day, particularly in urban building environments. Moreover, higher air temperature was also observed in the gap between the RPVSPs and roof (~0.3 m) compared with the ambient air temperature8,26. This increase of 3.2 °C is indicative of the overall impact of the RPVSPs on urban temperature during peak hours at the city scale. Conversely, during summer nights, the cooling effect of the RPVSPs, along with other environmental factors, leads to an average decrease in surface temperature of 1.4 °C. This is because the combined effects of radiative cooling and enhanced convective cooling lead to a net temperature reduction of the shaded rooftop surface. Although the RPVSPs reduce nighttime radiative cooling, the increased convective cooling compensates for this, resulting in the observed urban temperature being 1.4 °C cooler at night27,28. These results align closely with a previous review study14, which yielded valuable insights into the influence of RPVSPs on their surrounding environment. The review indicates that RPVSPs tend to warm the environment during the day and cool it at night. Nonetheless, compared with earlier similar studies, the uniqueness of our current analysis lies in its meticulous approach to accurately represent the energy balance of RPVSPs, the influence of roof surfaces and the local climate of tropical Kolkata. These existing studies have revealed two diverging viewpoints: one group of studies asserts that solar panels can significantly cool the urban environment on a diurnal scale, while another group demonstrates that solar panels elevate local urban temperatures during the day and reduce them at night. The inconsistencies in these findings are largely attributed to the erroneous reporting of the RPVSPs energy balance, the local climate and the nature and characteristics of the mounting surface. Notably, the heat convection between the back surface of the panels and the underlying surfaces (roofs or grounds) was not considered in these studies. It is worth mentioning that the temperatures of RPVSPs can reach up to 70 °C during peak solar hours14. According to publicly available information on first- and second-generation RPVSP systems, they can convert energy at a rate of 15–20%, while the majority of the balance, around 80–85% of panel-absorbed solar energy, can be stored as heat on the panel surface and then released as heat in the urban environment through thermal convection. This excess heat, which accounts for approximately 40% of the energy, cannot be converted into electric energy by first- and second-generation RPVSPs systems, with the remaining losses occurring through direct transmission and reflection. The management of this excess heat poses a significant challenge for current RPVSPs technology and should be evaluated in an ideal environment with an accurate RPVSPs energy balance to assess the practical implications. While considering the shadowing effects of RPVSPs over the surface (roofs or ground), it is crucial to account for variables, such as tilt angle, orientation, installation height and local climate and environment.

Fig. 5: The reported results of 30 evaluated case studies for RPVSPs.
figure 5

The effect of RPVSPs on Tambient from evaluated case studies, including 30 large-scale implementations (Supplementary Table 2). These studies have been grouped into four categories: daytime warming, nighttime warming, daytime cooling and nighttime cooling. The boxes illustrate the central tendency and variability of daytime warming, nighttime warming, daytime cooling and nighttime cooling, while the whiskers represent the overall range of reported warming and cooling effects, indicating changes in Tambient (°C).

Full size image
Fig. 6: The design of RPVSPs within the framework of WRF/BEP + BEM.
figure 6

A schematic representation of energy exchanges among building rooftops, the back surface of RPVSP and the ambient environment15. In this study, the parameterization used to incorporate the effects of RPVSPs within WRF/BEP + BEM assumes that the panels are parallel to and detached from the roofs, forming a single layer. The schematic also illustrates the time derivative of the RPVSP temperature.

Full size image

In addition, we conducted city-specific RPVSP sensitivity studies for four other cities—Sydney, Austin, Athens and Brussels—to compare their findings with those from Kolkata. This comparison illustrated in Supplementary Figs. 5 and 6 highlights the diurnal phenomenon characterized by RPVSPs surface temperature and associated RPVSP-induced nighttime cooling and daytime warming, respectively. It also shows the changes in other urban state variables after the implementation of RPVSPs at a city scale versus without RPVSPs for five cities—Sydney, Austin, Kolkata, Athens and Brussels. The data show that RPVSPs lead to a significant decrease in temperatures during the night and an increase during the day in cities such as Sydney (−0.8 °C to 1.9 °C), Austin (−0.7 °C to 1.8 °C), Kolkata (−0.6 °C to 1.5 °C), Athens (−0.4 °C to 1.2 °C) and Brussels (−0.3 °C to 1.1 °C), underscoring their impact on the urban thermal environment. The robust correspondence observed among the data from these cities supports the hypothesis that covering rooftops with RPVSPs significantly and consistently affects the urban microclimate. Thus, it would not be inaccurate to suggest that if an entire city is enveloped by RPVSPs, the prevailing microclimatic conditions would be largely influenced by their presence. Therefore, the detailed results regarding the impact of large-scale deployment of RPVSPs on urban thermal environment conditions in a tropical city, such as Kolkata, can be generalized to other global cities undergoing or planning for large-scale RPVSP implementation, considering that prevailing microclimatic conditions would be significantly influenced by RPVSPs’ presence.

The impact of city-scale deployment of RPVSPs in cities, specifically in Kolkata, under heatwave conditions has been investigated. City-scale RPVSPs integration to cities presents a serious thermal challenge. RPVSPs can locally heat the urban environment, influencing heat absorption and contributing to localize urban warming. During a hot day with a heatwave, RPVSPs alter the urban energy budget, leading to a maximum increase of 241.6 W m−2 in Qsensible and a significant rise in Qlatent (35.3 W m−2). Maximum increase of heat flux and changes of heat flux have also been calculated (Supplementary Tables 3 and 4). RPVSPs increase Tambient by 1.5 °C during the day but cause a slight decrease (0.6 °C) at night. RPVSPs lower thermal capacity and high thermal conductivity resulted in higher urban TSurface during the day, which decreases slightly at night due to rapid heat release. RPVSPs affect Hrelative, decreasing it by 5.6%, increasing thermal stress and enhancing Wspeed in the urban atmosphere. This is because a very slight increase in latent heat occurs due to higher evaporation caused by the temperature rise. However, this insignificant phenomenon does not significantly affect the temperature in the canyon, which is more impacted by the substantial increase in sensible heat at the roof level. The increase in canyon temperature leads to a slight decrease in Hrelative. RPVSPs trap longwave radiation, warming the urban surface and raising the urban Tambient. Modification of RPVSP technology intensifies urban warming, increasing Qsensible, turbulence and vertical Wspeed, elevating the PBlayer height by 615.6 m. This alteration influences convective rolls and accelerates Wspeed, impacting the urban atmosphere during heatwaves.

This study is solely based on simulations due to the absence of available observational data for urban RPVSPs at the whole city scale. To address this concern, we have validated the WRF/BEP + BEM-simulated RPVSP surface temperature using an experimentally validated model. A key limitation of our study is the assumption that most roofs in Kolkata are flat, despite the presence of some historical buildings with sloped roofs. Furthermore, it emphasizes the necessity of updated urban representations within the WRF/BEP + BEM model, particularly concerning construction materials and operational performance indicators. By identifying these areas for improvement, our study fosters a proactive approach toward enhancing modeling accuracy, ultimately contributing to more refined analyses and valuable insights. Despite uncertainties in accurately representing urban properties in simulation modeling, it underscores the importance of informed decision-making for sustainable city-scale solutions by balancing RPVSPs’ environmental impact with their benefits. Future research should prioritize enhancing roof reflectance and developing advanced systems to mitigate local urban heating during daytime. Additionally, the integration of cutting-edge materials science with RPVSP breakthroughs holds significant potential for creating cool photovoltaic systems that operate at lower temperatures, thereby reducing heat emissions into urban environments. Increasing roof reflectance through the use of cool roofs or super cool roofs in urban installations of RPVSPs could significantly boost the energy production of solar panels. Cool photovoltaic technology promises a thermally optimized, modular and compact solar solution. This solution leverages innovative materials, such as phase change materials, photothermal plates and nano-fluids, to maximize RPVSP efficiency and address the challenges of urban heat.

Methods

WRF/BEP + BEM modeling system and configuration

The WRF model29,30 was employed, utilizing one parent ___domain and two nested domains with resolutions of 18 km, 6 km and 2 km, focusing on the 2 km resolution within Kolkata metropolitan city and other four cities—Sydney, Austin, Athens and Brussels (Supplementary Fig. 8). The WRF/BEP + BEM31,32 was used to assess the impact of RPVSPs on urban temperature at the city scale33. The coupling calculated heat exchange, momentum, humidity and turbulent kinetic energy flux between buildings and the outdoor environment under stable atmospheric conditions. The model’s configuration included the WRF single-moment six-class for microphysics34, Noah-multi-physics model for land surface35, BEP + BEM system for the urban canopy layer36, Janjić scheme for turbulent kinetic energy37, Dudhia scheme for shortwave radiation38, rapid radiative transfer model for longwave radiation39, Bougeault–Lacarrere scheme for PBlayer40 and Kain–Fritsch scheme for cumulus parameterization41 (Supplementary Table 5). Initial and lateral boundary conditions were defined using high-resolution National Centers for Environmental Prediction Global Forecasting System data42. An updated urban land use/land cover map derived from ESA Sentinel-2 imagery with 10 m resolution was used for the innermost ___domain (D03) of each model simulation43. Critical parameters of BEP + BEM were updated on the basis of each city using multisourced datasets (Supplementary Table 6). In our study, we reported only the results for Kolkata, while the other four cities—Sydney, Austin, Athens and Brussels—were used for comparison and verification of Kolkata’s results with RPVSP at the city scale due to lack of urban-scale RPVSP data in Kolkata. This comprehensive approach ensures that the findings are not limited to a specific climatic zone, but are applicable to a wide range of urban climates, enhancing the overall understanding of RPVSP impacts on urban thermal dynamics. Simulations were conducted during the high temperature period, analyzing RPVSPs’ impact on meteorological and thermal urban environments for each city. These specific periods were chosen to ensure that the model captures the seasonal variations and urban climate dynamics effectively. By selecting different times of the year for each city, the study accounts for the diversity in climatic conditions—ranging from the hot and humid summer in Kolkata to the dry summer in Austin, the mild summer in Sydney, the hot and dry summer in Athens and the moderate summer in Brussels. The first 24 h of simulation were considered spin-up time and excluded from postprocessing and analysis for each city.

RPVSPs parameterization in WRF/BEP + BEM

The parameterization of RPVSPs in the multilayer BEP + BEM scheme in the WRF model is assumed to be horizontally adopted on rooftops to avoid unnecessary complexities in model simulation9. Recently, a new parameterization scheme of RPVSPs in the latest version of the WRF model was developed and tested through idealized simulation15. The new parameterization of RPVSPs is assumed to be parallel and detached from the building roof surface, so that not only the profile of buildings can remain simple, but also the shading effect of RPVSPs impacted on roof surface temperature. The heat flux from urban surface and building roof surface and its association with RPVSPs temperature parameters are determined by the time derivative (Fig. 6)5,15. The parameters in the time derivative are expressed in watts per square meter and are defined in Fig. 6.

$$\begin{array}{rcl}{C}_{\mathrm{module}}\frac{\partial {T}_{\mathrm{RPVSP}}}{\partial t} & = & (1-{a}_{\mathrm{RPVSP}}){\mathrm{S}}{\mathrm{W}}_{\mathrm{sky}}^{\downarrow }+{{{\epsilon }}}_{\mathrm{RPVSP}}^{U}{\mathrm{L}}{\mathrm{W}}_{\mathrm{sky}}^{\downarrow }-{\mathrm{L}}{\mathrm{W}}_{\mathrm{RPVSP}}^{\uparrow }\\&& +{\mathrm{L}}{\mathrm{W}}_{\mathrm{roof}-{\mathrm{RPVSP}}}^{\uparrow\downarrow } -{E}_{\mathrm{RPVSP}}-{H}^{\uparrow }-{H}^{\downarrow }\\ && +(1-{\mathrm{VF}})[(1-{a}_{\mathrm{RPVSP}}){\mathrm{S}}{\mathrm{W}}_{\mathrm{DIFF}}+{\mathrm{L}}{\mathrm{W}}_{\mathrm{sky}}^{\downarrow }]\end{array}$$
(1)

The configuration of new parameterization of RPVSPs is implemented in WRF/BEP + BEM to arrange the coverage fraction over the building. Compared with the parameterization in previous studies9,11, the new parameterization scheme of RPVSPs, the temperature on RPVSPs roof (TRPVSP) is solved numerically from the derivative equation (1) instead of through its dependence on shortwave solar radiation, contributed by all associated components7,44. Other heat flux parameters were updated and proceed to other sections in the multilayer canopy scheme after the TRPVSP values are computed numerically. The material properties of RPVSPs used in this study are summarized in Supplementary Table 7.

Numerical design and experiments

Five experiments were conducted to assess the regional impact of extensive RPVSPs deployment during the current heatwave month in Kolkata. The control simulation used a roof albedo of 0.15 and no RPVSPs. The other experiments varied the proportion of rooftops covered by RPVSPs, ranging from 25% to 100% in uniform distribution across urban grid cells. The study utilized a mesoscale weather model to simulate hourly meteorological fields resulting from RPVSPs deployment. The experiments explored RPVSPs scenarios with coverage fractions of 0.25, 0.50, 0.75 and 1.0 (representing 25%, 50%, 75% and 100% coverage, respectively) on city rooftops (Supplementary Table 8). Standard RPVSP parameters, including albedo, conversion efficiency and emissivity, were set to 0.11, 0.19 and 0.95, respectively9,12,15,44. According to a previous study11, the value of the albedo of the RPVSPs ranges from 0.06 to 0.1. We used the measured value9 of the albedo, which is 0.11.

Model evaluation and validation

The coupled WRF/BEP + BEM model’s performance in surface meteorological fields was evaluated using 2 m air temperature and 2 m Hrelative data from the base case simulation, compared with local observations in Kolkata’s urban grid cells. Hourly meteorological data were obtained from ten local observation stations maintained by Kolkata Municipal Corporation. The evaluation considered spatial variability between the 2 km × 2 km grid cells and various land use categories covered by the stations. Daytime and nighttime were defined as 6:00–18:00 LT and 19:00–5:00 LT, respectively. Statistical metrics including mean bias error (MBE), mean absolute error, root mean square error (RMSE), correlation coefficient (r) and the index of agreement (IOA) were calculated to assess the model’s accuracy. These comparisons aimed to validate the WRF/BEP + BEM model’s ability to accurately represent surface meteorological conditions within the urban boundary, considering the specific temporal and spatial factors of the study area.

To evaluate the impact of RPVSPs, conducting a field study at a building scale and incorporating the findings into a mesoscale model to understand the city-wide impacts would be desirable in our future study. In light of these limitations, we conducted sensitivity analyses for four other cities—Sydney, Austin, Athens and Brussels—to compare their RPVSP surface temperatures and urban state variables on an hourly scale with those of Kolkata’s urban rooftop, as shown in Supplementary Figs. 5 and 6, respectively, alongside observed climate data including Tambient and Hrelative. This approach allows us to demonstrate the reliability of our Kolkata findings across other four cities. Additionally, we have included a validation of WRF/BEP + BEM-induced RPVSP surface temperature using experimental model45, considering RPVSP surface temperature (°C), Wspeed (m s−1), Hrelative (%) and solar irradiation (W m−2) during solar hours for Kolkata (Supplementary Fig. 7).

T ambient

The near-surface temperature above the ground at 2 m height Tambient, T2 is calculated by using the Monin–Obukhov similarity theory46 in equation (2)

$${T}_{2}={T}_{\mathrm{surface}}+({T}_{\mathrm{ambient}}-{T}_{\mathrm{surface}})\frac{U}{{U}^{\ast }},$$
(2)

where Tsurface is the surface temperature, Tambient is defined as the first layer model level, U is the velocity of first model level and U* defines the friction velocity at 2 m height.

The control simulation accurately represented 2 m air temperature, validated against data from ten local meteorological stations throughout the simulation period (Supplementary Table 9). The coupled WRF/BEP + BEM model closely matched observed temperatures at different stations (mean r = 0.96; mean bias of 0.92 °C). The control simulation demonstrated statistically significant agreement with local observations (P < 0.05), with MBE and RMSE ranging from 0.49 °C to 1.4 °C and 0.95 °C to 1.83 °C, respectively. The IOA ranged from 0.93 to 0.98, averaging 0.96 across all ten observation stations within the urban ___domain. The simulated 2 m air temperature exhibited a warm bias, particularly during nighttime compared with daytime, consistent with previous studies in tropical cities. The model’s slight overestimation of nighttime temperatures could be attributed to anthropogenic heat release from the coupled WRF/BEP + BEM system and indoor air conditioning, causing urban areas to be warmer at night than during the day47,48,49.

H relative

The coupled WRF/BEP + BEM model accurately simulated near-surface Hrelative above an urban grid cell, with an MBE of −6.73% in urbanized regions compared with observation stations. The MBE and RMSE of Hrelative ranged from −5.7% to −7.43% and 6% to 8.1%, respectively. The IOA ranged from 0.89 to 0.97, averaging at 0.92 across all observation stations. Despite this accuracy, the simulated 2 m Hrelative exhibited a dry bias in urbanized regions due to the WRF model assuming urban grid cells to be grassland, neglecting trees and vegetation and underestimating surface evaporation and latent heat source from anthropogenic heating. Similar dry biases50 were observed in other tropical cities51,52,53. Despite these biases, the coupled model realistically represented the urban thermal environment and can be utilized to investigate the regional impact of RPVSP technology on the urban environment during heatwave conditions.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.