📄 earth_science

Long-term Intensification of Surface Urban Heat Island Under Urban Expansion: A Remote Sensing-based Analysis

This study utilizes remote sensing data from 2003 to 2021 to demonstrate that rapid urban expansion in Sylhet City, Bangladesh, has significantly intensified the surface urban heat island effect through a 61% increase in built-up areas, with daytime temperature anomalies peaking during the monsoon and nighttime peaks in pre-monsoon and winter, driven primarily by the loss of vegetation and expansion of impervious surfaces.

Original authors: Md. Aminul Haque Laskor, Abeda Tabassum, Abdulla Al Kafy, Torikul Islam Sanjid, Mostofa Kamal, Patrick Ludwig, Tanvir Ahmed

Published 2026-08-10

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Md. Aminul Haque Laskor, Abeda Tabassum, Abdulla Al Kafy, Torikul Islam Sanjid, Mostofa Kamal, Patrick Ludwig, Tanvir Ahmed

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The City That Gets Too Hot: A Story of Concrete, Tea, and Invisible Blankets

Imagine the Earth as a giant, cozy blanket. Usually, this blanket breathes; plants and soil soak up the sun's energy during the day and release it slowly at night, keeping things comfortable. But when we build cities, we swap that soft, breathing blanket for a thick, heavy coat made of concrete, asphalt, and brick. These materials are like thermal sponges: they gulp down sunlight and hold onto it tightly. This creates a phenomenon scientists call the "Urban Heat Island." Think of a city as a giant, glowing ember sitting in a cooler field. The city stays hot long after the sun goes down, while the surrounding countryside cools off. This isn't just about feeling a little warm; it makes cities dangerous during heatwaves, wastes energy on air conditioning, and changes how the weather behaves.

To understand this, scientists use two main tools. First, they look at "Land Surface Temperature" (LST), which is basically how hot the ground feels to the touch, rather than the air temperature we feel on our skin. Second, they measure the "Surface Urban Heat Island Intensity" (SUHII). This is a simple math game: they take the temperature of the city and subtract the temperature of the nearby countryside. If the city is 5°C hotter than the country, the SUHII is 5°C. The bigger the number, the more the city is acting like a heat trap. Now, imagine a city in a humid, rainy part of the world, surrounded by lush green tea gardens and rivers. Does the city still get hot? Does the rain cool it down? That is exactly the mystery a team of researchers decided to solve in the city of Sylhet, Bangladesh.

The Heat Trap in Sylhet

This paper is a detective story about Sylhet City Corporation (SCC), a bustling city in northeastern Bangladesh. For nearly two decades, from 2003 to 2021, the researchers watched this city grow. They didn't just stand on street corners with thermometers; they used "eyes in the sky"—satellites like Landsat and MODIS—to take the city's temperature from space. They wanted to see how the city's expansion was changing the heat, day and night, and through all four seasons: pre-monsoon, monsoon, post-monsoon, and winter.

The City's Big Transformation
First, the researchers looked at how the city changed its "skin." In 2003, Sylhet was a mix of buildings, bare dirt, water, and green plants. By 2021, the city had undergone a massive makeover. The area covered by buildings (built-up areas) exploded by 61%. Where there used to be bare land and water bodies, there were now roads, houses, and shops. Interestingly, while the city grew, the total amount of green vegetation in the wider area actually stayed roughly the same or even grew slightly, mostly because the city ate up "bare land" rather than just chopping down forests. But inside the city center, the greenery was definitely taking a hit.

The Temperature Story
As the city grew, so did the heat. The researchers found that the city was getting hotter every year, but the story was different depending on whether the sun was up or down.

Daytime: During the day, the city acts like a frying pan. The concrete and asphalt soak up the sun and get scorching. The study found that the "heat island" effect (how much hotter the city is than the country) got stronger every year. In 2021, during the monsoon season, the city was a staggering 7.00°C hotter than the surrounding rural areas during the day! That's a huge difference. The hottest spots were always in the center of the city, where the buildings are packed tightest.
Nighttime: At night, the story flips. Usually, cities stay hotter than the country at night because the concrete releases the heat it stored during the day. In Sylhet, the city did stay hotter at night, but the effect was more subtle. The peak nighttime heat difference was 2.77°C in the pre-monsoon and winter of 2021. However, there was a twist: in the northeastern part of the city, where the famous tea gardens are, the temperature actually dropped below the rural average at night. The tea gardens acted like a giant air conditioner, keeping that specific area cool.

The Seasons and the Trends
The researchers used some fancy math (called the Mann–Kendall test) to see if these changes were just random or a real trend. They found that the city is definitely getting hotter over time.

The hottest part of the day in the city is getting hotter at a rate of 0.051°C per year.
The average heat difference between the city and the country is also creeping up, day and night.
The "hotspots" (the most dangerous heat zones) are stuck in the middle of the city. In the daytime, these hotspots are biggest in the winter. At night, they seem to stretch southward during the rainy monsoon season.

Why is this happening?
The paper connects the dots between the city's "skin" and its temperature. They used two special "thermometers" made from satellite data:

NDBI (Built-up Index): This measures how much concrete and buildings are in an area. They found a strong link: more concrete means much hotter days.
NDVI (Vegetation Index): This measures how green an area is. They found that losing green plants makes the nights hotter.

The big takeaway is that impervious surfaces (concrete) are the main villain for daytime heat, while losing vegetation is the main reason nights are getting warmer. The concrete traps the sun's energy, and without plants to cool things down with evaporation, the heat gets stuck.

The Verdict
This study confirms that as Sylhet expands, it is becoming a more intense heat trap. The city is not just growing; it is heating up. The researchers suggest that to fix this, the city needs to be smarter about its design. They recommend adding more green spaces, using materials that reflect sunlight instead of absorbing it, and planning neighborhoods that can handle the heat. It's a warning that if we keep paving over the ground without thinking about the heat, our cities will become increasingly uncomfortable ovens, especially as the world continues to warm up.

Technical Summary: Long-term Intensification of Surface Urban Heat Island Under Urban Expansion in Sylhet City Corporation

Problem Statement
While the Surface Urban Heat Island (SUHI) effect is well-documented in major metropolitan areas like Dhaka, Bangladesh, long-term spatio-temporal variations of Land Surface Temperature (LST) and SUHI intensity (SUHII) in secondary cities remain underexplored. Specifically, Sylhet City Corporation (SCC) lacks comprehensive analysis regarding diurnal, seasonal, and annual thermal patterns over extended periods. Existing studies often treat Sylhet as a single urban entity without resolving intra-urban variability, fail to define optimal rural reference boundaries objectively, or do not apply pixel-level screening to exclude non-rural land covers from reference zones. This study addresses these gaps by investigating the intensification of SUHII in SCC from 2003 to 2021 amidst rapid urban expansion, aiming to quantify the biophysical controls of vegetation and built-up surfaces on thermal patterns.

Methodology
The study employs a remote sensing-based approach integrating MODIS and Landsat data to analyze LST and SUHII dynamics over a 19-year period (2003–2021).

Data Sources:
- Thermal Data: MODIS MOD11A2 LST product (v006), an 8-day composite at 1 km resolution, covering 2003–2021. Data were filtered for clear-sky pixels with errors < 2 K.
- Land Use/Land Cover (LULC): Landsat-7 ETM+ (2003) and Landsat-8 OLI/TIRS (2021) surface reflectance data (30 m resolution) were used for LULC classification. Annual Landsat imagery (2003–2021) was utilized for biophysical indices.
- Auxiliary Data: MODIS MCD12Q1 annual land-cover product (500 m) was used for rural pixel screening.
Analytical Techniques:
- LULC Classification: Support Vector Machine (SVM) supervised classification in ArcGIS 10.8 categorized the area into built-up, vegetation, water bodies, and bare land. Ground-truth points were collected via GPS and Google Earth Pro.
- Rural Reference Zone Optimization: A systematic sensitivity analysis was conducted across five candidate buffer distances (3, 5, 7, 10, and 15 km) from the old SCC boundary. The optimal buffer was selected based on pixel count stability, statistical significance of SUHII differences between adjacent buffers, and local representativeness. Built-up and water pixels were excluded from the rural reference zone annually.
- SUHII Calculation: SUHII was computed as the difference between the mean LST of the urban area ( $T_U$ ) and the cleaned rural reference zone ( $T_R$ ): $SUHII = T_U - T_R$ . Calculations were performed separately for daytime and nighttime across annual and seasonal scales.
- Trend Analysis: The non-parametric Mann–Kendall (MK) test and Sen's Slope estimator were used to detect monotonic trends and quantify magnitudes. Linear regression was applied to estimate rates of change.
- Spatial Analysis: Global Moran's I assessed spatial autocorrelation, and Getis-Ord Gi* identified hotspots and cold spots at the 99% confidence level.
- Biophysical Controls: Normalized Difference Vegetation Index (NDVI) and Normalized Difference Built-up Index (NDBI) were derived from Landsat data, resampled to 1 km, and correlated with LST and SUHII using Pearson correlation coefficients.

Key Contributions

Objective Rural Reference Definition: The study introduces a sensitivity analysis to determine the optimal rural buffer distance (7 km) for SUHII computation in Sylhet, ensuring the reference zone represents a thermally stable, non-urban background while excluding non-rural pixels.
Intra-Urban Resolution: Unlike previous multi-city studies that treated Sylhet as a monolith, this research resolves intra-urban spatial and temporal variability, identifying specific hotspots within the SCC.
Diurnal and Seasonal Granularity: The analysis provides a detailed breakdown of SUHII variations across four distinct seasons (pre-monsoon, monsoon, post-monsoon, winter) and diurnal cycles, highlighting differential heating dynamics.
Biophysical Quantification: The study quantifies the specific roles of impervious surfaces (NDBI) and vegetation (NDVI) in driving daytime and nighttime thermal intensification, respectively.

Results

Urban Expansion: Between 2003 and 2021, built-up areas in SCC increased by 61%, while bare land and water bodies declined by approximately 68% and 25%, respectively. Vegetation cover showed a slight increase due to the conversion of bare land in the southeastern region, though the overall trend indicates a loss of natural surfaces in the urban core.
LST Anomalies: Daytime LST anomalies peaked at 4.08°C during the 2021 monsoon, while nighttime anomalies reached 1.94°C in the pre-monsoon and 1.5°C in winter. The old SCC region consistently exhibited the highest positive anomalies.
SUHII Trends:
- Annual: Statistically significant upward trends were observed in annual maximum daytime SUHII (0.051°C yr⁻¹) and mean daytime SUHII (0.005°C yr⁻¹). Nighttime mean SUHII also showed a significant increasing trend (0.003°C yr⁻¹).
- Seasonal: Daytime SUHII showed significant increasing trends in pre-monsoon, post-monsoon, and winter seasons. The highest daytime SUHII (7.00°C) occurred during the 2021 monsoon. Nighttime trends were significant only during winter.
Spatial Distribution: Hotspots are persistently concentrated in the central SCC. Daytime hotspots are most spatially extensive during winter, while nighttime hotspots exhibit significant southward expansion during the monsoon and post-monsoon seasons. The northeastern region, dominated by tea gardens, consistently acts as a cold spot, particularly at night.
Biophysical Drivers:
- NDBI: Exhibited a significant positive association with daytime SUHII ( $r = 0.45$ ), confirming that impervious surface expansion drives daytime heating.
- NDVI: Showed a stronger negative association with nighttime SUHII ( $r = -0.52$ ) compared to daytime ( $r = -0.17$ ), indicating that vegetation loss disproportionately impairs nocturnal radiative cooling.

Significance and Claims
The paper claims that the study provides a critical baseline for understanding the thermal impacts of urbanization in secondary cities in South Asia, which are often overlooked in favor of megacities. By establishing a robust methodology for defining rural reference zones and analyzing intra-urban variability, the study offers a template for similar assessments in rapidly urbanizing regions.

The authors assert that the observed intensification of SUHII is directly linked to the expansion of impervious surfaces and the loss of vegetation. They conclude that while the cumulative temperature rise over 19 years is modest in some metrics, the trend in maximum daytime SUHII (~0.97°C increase) has tangible implications for urban thermal comfort and heat stress vulnerability. The findings suggest that climate-adaptive urban planning in SCC must prioritize green infrastructure, reflective materials, and heat-resilient zoning to mitigate these effects. The study acknowledges limitations regarding MODIS spatial resolution and cloud cover during monsoons, recommending future work with higher-resolution thermal data and urban growth projection models.

The City That Gets Too Hot: A Story of Concrete, Tea, and Invisible Blankets

The Heat Trap in Sylhet

More like this