Assessing Future Changes in Mean Radiant Temperature: Considering Climate Change and Urban Development Impacts in Fredericton, New Brunswick, Canada, by 2050

Abstract

Urban development and climate change are two main impacting factors in the thermal environment of cities. This study aims to analyze future changes in Mean Radiant Temperature (MRT), one of the main contributors to human thermal comfort and the concept of Urban Heat Island (UHI), considering climate change and urban development scenarios in the study area, Fredericton, New Brunswick, by 2050. The analysis utilizes the SOLWEIG (Solar and Longwave Environmental Irradiance Geometry) model from the Urban Multi-scale Environmental Predictor (UMEP) platform to calculate MRT values. By integrating these two impacting factors, this research provides insights into the potential future changes in MRT levels and the resulting thermal conditions and geohazards in the study area. The analysis enables the identification of areas susceptible to increased radiant heat exchange due to the proposed changes in land cover, urban morphology, and air temperature. Furthermore, this study contributes to a better understanding of the complex interactions between climate change, urbanization, and urban microclimates. By incorporating MRT assessments and prioritizing thermal comfort, cities can develop strategies to mitigate the negative effects of UHI and create sustainable and livable urban environments for future generations.

1. Introduction

The Canada’s Changing Climate Report reveals that Canada is experiencing a rate of warming that is twice the global average [1,2]. In addition, the upward trends in urbanization may result in a change in land cover, thereby exacerbating a number of environmental issues, including UHI [3]. UHI is a phenomenon representing the tendency of the urban areas to become warmer than their rural surroundings. This occurs due to higher heat absorption rates of constructed urban surfaces in urban environment and results in urban residents experiencing higher levels of heat stress [4].

UHI and the potential impacts of climate change have necessitated the investigation of thermal comfort in urban environments. The physical characteristics of the built structures in urban areas impact the thermal comfort levels as a result of the alterations in sunlight exposure and surface materials [2,5]. Sky view factor and wall aspect are two main geometry factors involved in spatial patterns of heat stress in the study area [6,7,8]. In the literature, Mean Radiant Temperature (MRT) is identified as the main indicator of the harmful heat stress condition on the human body [9,10]. MRT measures the combined effect of all radiant heat energy sourced by the objects and surrounding surfaces of a given body. It takes into account the temperature of the surrounding objects (warmer surfaces emit higher radiation according to the Stefan–Boltzmann Law), their emissivity (how efficiently they radiate energy compared to a perfect blackbody), and their geometric arrangement (where objects that are closer and have a more direct orientation have a bigger impact) [9]. By considering the geometry of the urban environments, including building heights, orientations, and surface characteristics, MRT is estimated as the radiant exchange between different surfaces and the human body. Unlike the small changes in the geospatial variation of the air temperature, MRT has substantial variations in geospatial distribution [8,11,12]. By comprehending the patterns and trends in MRT, it becomes possible to estimate the thermal comfort levels that individuals may experience.

There have been valuable efforts in studying UHI in different aspects. Sharma et al. provide projections for short- and long-term climate changes in Delhi [13]. Their study also shows that the UHI effect has a significant impact on heat stress indices, with higher temperatures and sultriness in urban areas leading to higher heat stress. Keppas et al. employ climate simulations using the Weather Research and Forecasting (WRF) model to obtain daily maximum temperature, air temperature, and humidity data to assess the increase in heat stress levels resulting from the UHI effect within two Mediterranean cities of Thessaloniki and Rome [14]. Their study reveals that both cities are projected to undergo a substantial rise in temperature and heat stress under future scenarios by the years 2050 and 2100, with urban areas experiencing greater effects compared to the surrounding rural areas. Shen et al. studied predicting the future urban heat island intensity (UHII) and distribution based on land use and land cover, land surface temperature, and landscape composition and configuration [15]. The authors used a combination of linear regression and non-linear machine learning models to detect the relationship between Surface UHII and landscape patterns. Lindberg et al. used a method that combines regional climate model outputs with meteorological observations to create future scenarios used to model outdoor thermal comfort and evaluate the effectiveness of different mitigation strategies [16].

While numerous studies, such as the aforementioned ones, have explored the factors affecting thermal comfort and UHI, and predicted the future levels in UHI, few have simultaneously considered the combined influence of climate change and urbanization scenarios on MRT, which is the objective of this study. This research, conducted as a 3D GIS approach to studying UHI, addresses this gap by employing the SOLWEIG model to assess the spatiotemporal distribution of MRT under two scenarios. The first scenario represents the current conditions, providing a baseline for comparison, while the second scenario incorporates projected changes due to climate change and city development by the year 2050. This year was chosen due to aligning with near-term climate change evaluations as well as the benchmark year for climate policies such as net-zero emission goals [17,18]. Hence, by examining both present and future values of MRT, we are aiming to gain insights into the potential thermal impacts on the urban environment by the end of the study period. The findings of this study will contribute to a better understanding of the potential impacts of climate change and urbanization on human thermal comfort, and inform future urban planning policymakers for creating sustainable and livable cities.

2. Materials and Methods

The methodology adopted in this study involves a series of pre-processing steps to provide necessary input data for the model calculations for the two scenarios being considered, including: (1) current situation of UHI levels; and (2) predicting the impacts of climate change and urban development on the UHI levels for the year 2050. As Figure 1 illustrates, the first step in this process is to provide surface and meteorological data for both scenarios, followed by calculations of sky view factor and wall and surface albedo. These steps will then enable the model to calculate MRT levels for both scenarios under study, providing necessary information for a comparative analysis.

2.1. Pre-Processing

The main sets of data needed to perform calculations in SOLWEIG and generate MRT values include (1) topographic models such as Digital Terrain Models (DTM), ground and building DSMs, and Canopy DSMs; (2) meteorological data such as air temperature and incoming shortwave radiation; (3) environmental parameters such as albedo and emissivity of ground and walls; (4) sky view factor; and (5) wall aspect and wall height. The data utilized in the preparation phase of the topographic information were obtained through the collection and processing of pre-classified lidar point clouds from GeoNB data catalogue (2023). A total of 104 tiles of 1 Km by 1 Km lidar point clouds, accompanied by 350 tiles of orthophotos, building footprints, and water body polygons were obtained from GeoNB [19]. The lidar data included 32 classes, including ground, buildings, vegetation, and other relevant features, providing detailed information for subsequent analysis. To prepare the data for further analysis, DTM, DSM, and CDSMs were generated from the classes ‘ground’, ‘buildings’, and ‘vegetation’ of the pre-classified lidar point clouds with a 5 m resolution using the CloudCompare 2.12.4, ArcGIS Pro 3.1, and QGIS 3.32 software packages. Additionally, sky view factor, wall aspect, and wall height raster layers were generated using the prepared DSM and CDSM maps, providing the orientation and height of building walls for the calculations of direct and reflected heat flux [20].

A land cover map, Figure 2, was created through Support Vector Machine classification on the orthophotos with 0.5 m resolution, categorizing the land cover classes required as an input for the SOLWEIG model. To ensure the map correctly represents the built environment and the local hydrography, shapefile polygons of building footprints and water bodies were inserted into the land cover map. Additionally, 5 m of buffer zone was created along the shapefile for the streets. For the accuracy assessment of the created land cover map, 100 points were randomly created throughout the area to compare the classified and the ground truth pixels from the orthophotos. The result of this assessment proved a good level of agreement, with the Kappa coefficient being equal to 0.75.

2.2. Study Area

This study has focused on a specific region, Figure 3, within the city of Fredericton, New Brunswick, chosen due to its significance in terms of high urban development rates and potential climate change impacts. Reports suggest that the number of residents of Fredericton is increasing to as many as 93,600 by the year 2041 [21]. It has also been mentioned that Fredericton has a growth rate higher than the national average in Canada [22]. The city has recently seen CAD 187 million worth of building permits, which proves the momentum of the city towards the projected growth rates [23]. Analysis of historical building construction data revealed that the majority of building construction activities have been concentrated in the western and eastern parts of the city. This information guided the selection of the study area, ensuring it captured the areas experiencing significant urban expansion.

2.3. City Development Simulation

To estimate the number of new buildings to be added in the study area for the future urban development scenario, a trend analysis of building construction over time was conducted for the period between 1990 and 2023. Based on the historical data for the buildings’ construction year, a quadratic curve fitting was performed to predict the number of buildings for the year 2050.

The prediction for the buildings being constructed by 2050 yielded a total number of 23,616 buildings. A detailed breakdown of the number of buildings constructed over time is provided in

Table 1. The trend in building construction and the future anticipated number of buildings by 2050.

Year	Number of Buildings
1990	16,272
1995	17,361
2000	18,221
2005	19,305
2010	20,599
2015	21,316
2020	21,581
2050	23,616

. Additionally, a diagram representing the trend in building construction is shown in Figure 4, reflecting the anticipated future urban development scenario.

Considering the trend in the spatial distribution of buildings being constructed since 1990, it was determined that the western and eastern parts of the city have experienced the highest rate of development, with approximately 2000 buildings. Therefore, 2000 artificial residential buildings were created in those areas due to their significance in terms of urban development, representing the future urbanization scenario. These buildings, Figure 5, were created after copying residential areas from the existing building footprints into the western and eastern parts of the city. The height for these buildings was chosen to be an average 10 m, with the shapes of the buildings represented at Level of Detail 1 (LOD1). Additionally, an average buffer zone of 10 m was created around the buildings, which were defined as paved surfaces in the land cover map, to make the scenario as close as possible to the real world.

After incorporating the artificial buildings, the DSM and land cover maps were updated to reflect the revised urban morphology. The new DSM was created by inserting the 10 m-high artificial buildings into the previous DSM raster. Similarly, the new land cover map was created by inserting the artificial buildings and the buffer zones into the previous land cover raster file, assigned with the classes ‘Building’ and ‘Paved’. Sky view factor, wall height, and wall aspect were regenerated accordingly to capture the updated characteristics of the urban environment. This step ensured that the simulation accurately represented the potential changes in the urban landscape and its influence on MRT levels.

2.4. Climate Change

Climate change projections were incorporated using the Climate Change World Weather File Generator (CCWorldWeatherGen) tool which employs the Third Assessment Report of the Intergovernmental Panel on Climate Chane (IPCC TAR) model summary data of the Hadley Centre Coupled Model version 3 (HadCM3 A2) experiment ensemble [24]. This tool utilizes morphing methodology for climate change transformation of weather data developed by Belcher, Hacker, and Powell [25]. For this study, the Representative Concentration Pathway 8.5, representing a high greenhouse gas emissions pathway scenario where concentrations of greenhouse gases continue to rise rapidly throughout the 21st century, was chosen to capture the highest impact of climate change on MRT values. Considering the projected climate change effects, the warmest month of the year 2050 is July, with an average temperature of 22.6 °C and maximum temperature of 35.5 °C, compared to the year 2022, when the average temperature was 20.2 °C and the maximum temperature was 32.3 °C. Therefore, the analysis was implemented on the month of July of 2022 and 2050 during daylight time, which is approximately from 7 a.m. to 8 p.m. (ADT) during July. Finally, the prepared data were imported into the SOLWEIG model to calculate MRT values.

2.5. Mean Radiant Temperature (MRT)

To calculate MRT values in this study, the SOLWEIG model was employed. SOLWEIG is a widely used and validated model for estimating outdoor MRT and other related parameters [2,26,27,28]. The model considers various factors, including solar radiation, sky view factor, wall aspect, wall height, and air temperature, to estimate the MRT at different locations within the study area. Wall height and wall aspect—i.e., the orientation of walls—are two important factors in calculating MRT based on solar access and radiative exchanges in the vicinity of building walls [21]. To calculate MRT initially, it is essential to take into account the mean radiant flux density (Sstr). This refers to the total of all long- and shortwave radiation across three dimensions, Figure 6, factoring in the angular and absorption characteristics of a human body. A summary of the calculation of MRT in SOLWEIG involves the following formulas [26]:

• Calculation of shortwave radiation:

I_glob = I_dir + I_diff

(1)

R = ρ × I_glob

(2)

where R is the reflection of shortwave radiation, ρ is surface albedo (default value 0.15), and I_glob is global radiation.

K_i = (1 − ρ_i) × I_glob + R_i

(3)

where K_i is the shortwave radiation flux for surface i, ρ_i is albedo of surface i, and R_i is the reflection of shortwave radiation from surface i.

• Calculation of longwave radiation:

L_out = εσT_s^4

(4)

where L_out is the outgoing longwave radiation, ε is the emissivity of the ground surface, σ is the Stefan–Boltzmann constant, and T_s is surface temperature.

L_in = εσT_a^4

(5)

where L_in is the downward longwave radiation and T_a is air temperature.

• Calculation of net longwave radiation:

L_net = L_in − L_out

(6)

• Calculation of net radiation:

R_net = SW − LW

(7)

where SW is shortwave radiation fluxes and LW is longwave radiation fluxes. Equations (1)–(6) are all for each specific surface i considering there are 6 surfaces surrounding the human body, including west, east, front, back, up, and down (i = 1–6).

S_str = ζk × ∑Ki + εp × ∑Li

(8)

where S_str is mean radiant flux density, Ki and Li are the short- and longwave radiation fluxes of surface I, respectively (i = 1–6), ζk is the absorption coefficient for shortwave radiation (standard value 0.7), and εp is the emissivity of the human body (standard value 0.97).

Once S_str is calculated, MRT can be determined using the following formula.

• Calculation of MRT in Kelvin:

MRT = (S_str/σ)^1/4

(9)

SOLWEIG utilizes these formulas to calculate the incoming shortwave and outgoing longwave radiation fluxes from the sky, ground, and surrounding surfaces.

Using the pre-processed data, including DSM, land cover maps, sky view factor, wall aspect, and wall height, MRT values for both the current and future scenarios were calculated after running the model. Meteorological data such as air temperature and incoming shortwave radiation, etc., for the present were obtained from the weather station at Fredericton airport. The emissivity and albedo of the ground are calculated by the model based on the provided land cover map, while the emissivity and albedo of the walls are default values of 0.90 and 0.20, respectively. The model’s calculations, incorporating the aforementioned formulas, provide valuable insights into the spatiotemporal distribution of MRT and allow for the evaluation of thermal conditions and potential thermal discomfort experienced by individuals within the study area.

It is important to note that while the SOLWEIG model has been widely used and validated in various urban contexts, it is still subject to certain limitations and uncertainties, including the accuracy of input data and assumptions made in the model’s algorithms. These limitations should be taken into account when interpreting the results and considering the implications for urban planning and design.

3. Results and Discussion

MRT values were calculated for two scenarios, showing a noticeable increase across the entire study area in the second scenario, indicating the potential impacts of urban development and climate change on thermal conditions.

To visually depict the spatial distribution of the changes, raster files were generated for the current and the future hourly daytime levels of MRT, as seen in Figure 7. This figure provides a visual representation of the baseline and projected hourly daytime MRT values averaged throughout the warmest month of the two study periods, namely, July 2022 and July 2050, allowing for a comparative analysis of the spatial patterns and magnitude of changes. It is evident from this figure that MRT values will experience an average rise of approximately 3 °C in the study area over the one-month study period compared to the baseline scenario. The presence of the artificially created buildings is seen to have an impact on the heat load in the area, with high values adjacent to the walls and within the buffer zones created around the building polygons in particular. These areas demonstrated a notable elevation in MRT levels, likely due to the effect of longwave radiation emitted from the building walls and pavement material surfaces. This suggests an overall increase in the heat load on the urban environment, which could have implications for human well-being. It is also evident from the figure, where vegetation and building areas exist, that land cover materials have an important impact on the levels of MRT. This indicates that planting trees is an effective strategy for lowering the heat stress in urban areas, as the areas with more vegetation cover are associated with less heat stress [9]. Additionally, the results indicated that in some areas throughout July 2050, the average MRT values surpass the 47.6 °C line, where an increase in heat-related mortality is seen for the age group 80+ [9]. This highlights the importance of strategic city planning to prevent the dangerous impacts of this phenomenon.

Looking closer at the daily distribution of MRT levels rather than monthly, it is evident that there are several days with projected MRT values passing 55.5 °C and 59.4 °C thresholds, which have been associated with 5% and 10% increase in heat-related mortality rates among the elderly [9]. Figure 8 shows a particular day, 1 July 2050, where the MRT values are rising to an alarming rate for a couple of hours during the daytime.

The identification of these localized hotspots is critical for understanding the thermal conditions and potential thermal discomfort experienced by individuals in these areas, especially considering the rise in temperature in the future climate change scenario. Such findings underscore the importance of considering building design, materials, and urban planning strategies to mitigate the impacts of longwave radiation and minimize the risk of thermal discomfort.

It is important to note that the observed rise in MRT provides a general understanding of the thermal changes across the study area. However, further analysis and interpretation are required to examine the spatial patterns in detail, identify potential hotspots, and evaluate their implications for human thermal comfort. Additionally, the climate models used in this research, while based on well-established methodologies and datasets, are subject to inherent uncertainties in their predictions and assessments. These uncertainties arise from a few factors including model assumptions, emission scenarios, spatial and temporal resolution, and the complex interactions within the climate system, which can influence the precision of future projections [29].

In addition to the previous results, a diagram was created along a proposed route, Figure 9, to inspect the effects of UHI, starting from green (sub-urban) areas in the western part of the study area, moving towards the central areas of the city, and ending at the green areas of the southeastern part of the study area. The values along this route were illustrated in Figure 10, showing the effect of UHI during 1 July 2050 as a phenomenon indicating the difference in heat comfort levels between the central parts of the city and the rural (sub-urban) areas.

4. Conclusions

In In conclusion, this study provides valuable insights into the spatiotemporal distribution of MRT under urban development and climate change scenarios in our study area, Fredericton, NB, Canada. The findings indicate a significant increase in mean radiant temperature, with localized hotspots observed near building walls and within buffer zones around the buildings. These results emphasize the importance of considering mean radiant temperature in urban planning, design, and climate change adaptation strategies.

The contributions of this research lie in the integration of future urban development scenarios, climate change impacts, examination of local landscape and land cover patterns, and consideration of MRT as a crucial factor in assessing thermal comfort. By highlighting the areas experiencing elevated MRT levels, this study informs targeted mitigation strategies to enhance urban thermal comfort and contribute to the creation of sustainable and liveable cities.

Future research endeavours should focus on refining the modelling approach, incorporating additional factors influencing thermal comfort, and addressing the uncertainties and limitations identified in this study. By further advancing our understanding of urban microclimate dynamics, we can develop more effective strategies to mitigate the impacts of rising temperatures and ensure the well-being of urban populations.