Abstract
Heatwave is one of the most dangerous “natural hazards” with the greatest impact on people and other living beings that rarely receive adequate attention worldwide. Climate change & intense landcover conversion result in extreme heatwaves in urban areas. Chattogram, which is located in the southeast of Bangladesh, is a prominent coastal city and key economic hub. Chattogram City Corporation (CCC) consists of 41 wards with a large population, and most of the heavy, medium, and light industries of Bangladesh are situated in Chattogram. Due to this commercial and industrial development, and rapid urbanization, the landcover of the CCC area has changed expeditiously which is responsible for the urban heatwaves hazard. The study intends to comprehend the future heat wave scenario. For a better understanding of heatwaves, it has been analyzed the temporal changing and relationship between the Land Surface Temperature (LST) and the Land Use & Land Cover (LULC) from 2000 to 2020. Based on this data, the Artificial Neural Network (ANN), Markov chain model (MC), and Cellular Automata (CA) models have been applied to the prediction of future LST and LULC in different two years, 2030 and 2040. The temporal analysis result shows that the LST of the CCC area has risen with the decrease of vegetated areas and water bodies. The model simulation result suggests that the total buildup area of the CCC will be increased around 45% and 60% in 2030 and 2040 respectively, and 13 and 43% of the CCC area will suffer temperatures higher than 36° C in 2030 and 2040, respectively which are considered to be more vulnerable to the people. Policymakers will find this research useful in interpreting the effects of LULC change on LST and in recommending management frameworks. It will also be favorable for analyzing future vulnerability in the CCC area due to heat stress.
Access provided by Autonomous University of Puebla. Download conference paper PDF
Similar content being viewed by others
Keywords
- Land Surface Temperature (LST)
- Land Use and Land Cover (LULC)
- Heatwave
- Markov Chain Model (MC)
- Artificial Neural Network (ANN)
1 Introduction
World temperature is getting warmer with the rhythm of time. Incidents of extreme heat are becoming more often and more intense on a global scale, and this trend is predicted to continue into the twenty-first century [1, 2]. Currently, 50% (or 3.5 billion) people live in urbanized regions around the world, and by 2030, that number is anticipated to rise to 60% (4.9 billion). This large populations with their significant social inequality will make them more vulnerable to climate change on a social level [3]. The frequency of heat waves and their harmful impacts would increase due to the rise in global temperatures. The effects of heat waves is reliant on the severity and duration of the individual events along with environmental and socio-demographic factors [4]. As we all know heat waves can cause numerous heat-related illnesses and deaths which can be a major problem for developing countries like Bangladesh. Land surface temperature (LST) is a major contributing factor attributed to an increase in urban heat and microclimatic warming [5]. The conversion of natural land cover into artificial materials, causes an increase in sensible heat flow as well as air temperature rise [6]. Between 2000 and 2016, at least 136,835 people in Europe died of heat-related health issues, which represents more than 87% of all disaster-related deaths in that area [7].
Bangladesh is one of the foremost countries susceptible to the unpleasant effects of global warming. In 2003 around 62 people died due to heat waves across Bangladesh and the death rates rise by approximately 20% during heat waves [8]. Heat wave hazard mitigation is becoming a significant policy concern for Bangladesh's policymakers, since it is a hazard to the country's ecosystem and habitats. Chattogram is the country's biggest port and the primary location for the development of heavy, medium, and light industries. Additionally, Chattogram is the site of Bangladesh's only oil refinery and steel factory. According to research, Chattogram city has observed significant increases in surface urban heat island intensity at night by 1.9 °C during the last 20 years [9].
This study has tried to understand the future heat wave scenario in Chattogram City for mitigating future heat related disaster. The present study aims to analyze the temporal change of land use & land cover (LULC) and land surface temperature (LST) during the year 2000–2020 in Chattogram City Corporation area. This research also aims to predict the future heat wave scenario in 2030 and 2040 in Chattogram City Corporation Area, Bangladesh applying the Markov chain model, Cellular Automata model, and Artificial Neural Network (ANN). Also, this study will help to investigate future heat wave vulnerability of this area.
2 Materials and Methods
The methodology of the study includes outline of the study area, data collection and analysis procedure have been mentioned during this section.
2.1 Study Area Profile
Chattogram is the ‘Commercial Capital’ and the second-biggest metropolitan city declared by the government of Bangladesh. It is considering the industrial, commercial and institutional potentials of the country. Chattogram is also known as a port city of Bangladesh and the third busiest international seaport in South Asia [10]. Most of the heavy, medium, and light industries of Bangladesh are situated in Chattogram. As well as Chattogram is also the location of Bangladesh's only steel mill and oil refinery. Chattogram City Corporation (CCC) has a total area of 160.99 km2 and 41 wards with a large population. The amount of vegetation and water bodies has been continuously decreasing because of the rapid urbanization, commercialization, and industrial development. And this the main reason for increasing land surface temperature as the urban and industrial areas cannot absorb heat. So, increasing land surface temperature is a threatening issue for this city for now and in the nearest future (Fig. 1).
2.2 Data and Methodology
In this research, three Multi-spectral Landsat satellite data were obtained the United States Geological Survey (USGS) website. Landsat Thermal Infrared Sensor (TIRS), Thematic Mapper (TM), and Operational Land Imager (OLI) data were collected from 2000 to 2020 in 10-year intervals with 0% cloud coverage to analyze the land use and land cover (LULC) change and land surface temperature (LST) of the study area.
The acquired satellite images were categorized into five broad land cover types (built-up 01, built-up 02, water body, vegetation & bare soil) for the year of 2000, 2010, and 2020 on the basis of the Maximum Likelihood Supervised Classification (MLSC) technique. After the classification, an accuracy assessment has done for each LULC image for the validation purpose. For the ground truth verification using Google Earth, a total number of 400 sample pixels have been generated based on random sampling [11]. The land surface temperature (LST) has been obtained from the thermal bands of Landsat-5 TM (band 6) and Landsat-8 OLI (band 10). All the LST extraction was done in ArcGIS 10.7 with the help of model builder and raster calculator tools.
For predicting future LULC changes, a combination of the multilayer perceptron (MLP) model, which is an artificial neural network model [12] and the Markov Chain (MLP-MC) model have been used. The Land Change Modeler (LCM) tool has been used to run the model in IDRISI Selva V.17 software. There are two variables in the MLP model which are the land cover images are input as dependent variables and the variables controlling the land cover changes are input as independent variables [12]. For this study, 6 independent variables have been used that control the land cover change such as DEM, slope, aspect, distance to water bodies, major roads, buildup area. Distance to major roads, buildup areas, and water bodies have a great impact on future urban growth or development. An accuracy assessment has performed using the existing dataset for ensuring the model’s acceptance. For the acceptance of the model, at first, the LULC of 2020 has been simulated. A validation has been done with the help of predicted LULC 2020 and existing LULC 2020. After validation with an acceptable level, the LULC 2030 and LULC 2040 have been simulated.
In this study, the prediction of LST for the years 2030 & 2040, an Artificial Neural Network (ANN) model has been applied with the assistance of MOLUSCE plugin in QGIS software. As input parameters, the study has been used LULC images, NDVI, NDWI, NDBI, NDBaI, SAVI, BUAI, and MNDWI. For the model’s acceptance, a validation process had completed with the simulated LST of 2020 and the existing LST of 2020. After validation with an acceptable level, the LST 2030 and LST 2040 have been simulated as the same way of 2020 prediction (Fig. 2).
3 Results and Discussions
According to the methodology, the spatiotemporal distribution of LST and LULC pattern as well as simulations of the future distribution LST and LULC were calculated for the area of study.
3.1 Temporal LULC and LST Change Analysis of CCC Area
Maximum Likelihood Supervised Classification method was applied to evaluate the LULC of year 2000–2020. The overall accuracies of the classified images (2000, 2010, and 2020) are, respectively 86.48%, 90.69%, and 94.83%, with Kappa coefficients of 0.86, 0.91, and 0.95. Figure 3 (1a–1c) represent the land classification of CCC area for three different years of 2000, 2010, and 2020. From the Fig. 4 it had found that between 2000 and 2020, the total area covered by the vegetation has reduced gradually. In year 2000, the total percentage of vegetation cover was around 57.37% which decreased into 39.01% in 2020. The buildup 01 areas which are mostly residential, commercial, roads, rail areas have increased 28.07% to 37.48% between 2000 to 2020. Most of the vegetation and waterbody area turned into buildup 01, buildup 02, and bare soil in the year 2020. The expansion of the urban area is the consequence of unplanned population expansion, industrialization and the trend for rural residents to move to the cities.
Landsat Thermal bands were used to determine the Land Surface Temperature for the years 2000 to 2020 using different formulas. The spatial distribution of LST in the CCC area present in Figs. 3(2a–2c). For better visual interpretation the LST range of each is divided into six uniform classes respectively <22, 22–24, 24–26, 26–28, 28–30, and >30 °C. The maximum LST of the year 2000 is 30.84 °C but there are fewer areas with this temperature. The LST of the maximum area in the year 2010 is more than 24 °C. In the year 2020, the value of maximum LST increased to 35.95 °C and the maximum area’s LST is more than 26 °C. The maximum LST in 2020 is very close to the mild heat wave according to Bangladesh Meteorological Department (BMD). According to Bangladesh Meteorological Department (BMD) ‘heat wave’ means when maximum day temperature attains 36 °C or more. From 2010 to 2020 the maximum LST increased around 5.11 °C between 10 years. The rapid rate of urbanization can be a major reason behind this increased LST in 2020. We have also found in the LULC map of the year 2020 that 45% of the total area of CCC is buildup area. These buildup areas generate more heat than other land uses as the conversion of the natural land cover into man-made materials like concrete and asphalt, increasing the surface temperature.
3.2 Simulation of LULC and LST of the Year 2030 and 2040
For the simulation of LULC of the year 2030 and 2040, the Multi-layer Perceptron- Markov Chain (MLP-MC) model were used. The Land Change Modeler (LCM) tool had been used to run this model in IDRISI Selva V.17 software. For the validation of the model, the LULC of the year 2020 were simulated first. The MLP model has two variables: independent and dependent. The LULC images used as the dependent variable and 6 independent variables had been used that control the land cover change such as DEM, aspect, slope, distance to water bodies, major roads, buildup area. All the input helped to run the model. After the prediction of LULC 2020, a validation process had done for the acceptance of the model with the help of exiting the LULC 2020 map. The overall kappa coefficient was 0.82 which is an almost perfect agreement [13].
After the LULC prediction of 2020, it can be said that the model is suitable for the prediction of 2030 and 2040. So, the LULC of 2010 and LULC of 2020 have input into the model as dependent variable and DEM, aspect, slope, distance to water bodies, major roads, and buildup area as the independent variable. The predicted LULC of 2030 and 2040 has been represented in Figs. 5. From the visual interpretation of the map in 2040, most of the area converted into buildup 01.
In 2030 44% of the total area will be converted into buildup 01 which are mostly residential, commercial, road, rail, and other infrastructures. And in 2040 it will be increased by 49% of the total CCC area (Table 1). In 2020 the 37% of the total area was buildup 01. The bare soil, vegetation, water bodies have decreased day by day. This rapid rate of buildup growth is the prime reason for the LST increase.
For the LST simulation of the year 2030 and 2040, the Artificial Neural Network- Cellular Automata (ANN-CA) model applied in QGIS using the MOLUSCE plugin. Same as the LULC prediction, the LST of 2020 predicted at first for the validation of the model. For the LST simulation of the year 2020, the LST map of 2000 used as the initial data and the LST map of 2010 used as the final data. Some spatial variable which influences the prediction process such as LLC images, NDVI, NDWI, NDBI, NDBaI, SAVI, BUAI, and MNDWI of the year 2000 and 2010 were used as the spatial variable. The kappa value of the model was 0.764 and the correctness with the existing LST of 2020 was 70.91%.
For 2030 and 2040 simulation the LST of 2010 input as initial data and LST 2020 input as final data. LULC images, NDVI, NDWI, NDBI, NDBaI, SAVI, BUAI, and MNDWI of the year 2000 and 2010 input as the spatial variable. Then the ANN model was run with the highest iteration of 1000 and 0.001 momentum for the model’s better performance. The current validation kappa value of the model was 0.7037 which is a substantial agreement. After running the ANN model, the result input into the CA simulation for the final prediction output.
From the simulation of LST of year 2030 and 2040, the maximum temperature of the CCC area will be 39.61 °C in 2030 which is moderate heat wave according to Bangladesh Meteorological Department (BMD). And in 2040 the maximum LST will be 41.326 °C which is Severe heat wave. In 2030 most of the area’s LST will be more than 34 °C and in 2040 it will be increased in 37 °C. In 2030 and 2040, around 13 and 43% of the CCC area will have temperatures higher than 35 °C.
4 Conclusion
With the increasing heat wave the normal lifestyle of people will be devastated in the future. From the analysis it can be seen that in 2030 and 2040 the temperature will be higher than 36 °C in the maximum area which is not suitable for our living condition. As well as from the analysis it can be realized that in 2040 the maximum area of the CCC will be converted into buildup area. Such continuous growth of buildup area along with LST will create several medical, economic, environmental problems for the city. So, from this study the most vulnerable area can be identified and proper mitigation strategies can be taken. From this study the similar study of mitigation and resilience capacity can be done by the researchers. It can be helpful to reduce the adverse effect of heat wave which is useful for inhabitants of these places.
References
Luber G, McGeehin M (2008) Climate change and extreme heat events. Am J Prev Med 35(5):429–435. https://doi.org/10.1016/j.amepre.2008.08.021
Weber S, Sadoff N, Zell E, de Sherbinin A (2015) Policy-relevant indicators for mapping the vulnerability of urban populations to extreme heat events: a case study of Philadelphia. Appl Geogr 63:231–243. https://doi.org/10.1016/j.apgeog.2015.07.006
Aubrecht C, Özceylan D (2013) Identification of heat risk patterns in the U.S. National Capital Region by integrating heat stress and related vulnerability. Environ Int 56:65–77. https://doi.org/10.1016/j.envint.2013.03.005
Macnee RGD, Tokai A (2016) Heat wave vulnerability and exposure mapping for Osaka City, Japan. Environ Syst Decis 36(4):368–376. https://doi.org/10.1007/s10669-016-9607-4
Maduako I (2015) Simulation and prediction of land surface temperature (LST) dynamics within Ikom City in Nigeria using artificial neural network (ANN). J Remote Sens GIS 5(1). https://doi.org/10.4172/2469-4134.1000158
Fabrizi R, Bonafoni S, Biondi R (2010) Satellite and ground-based sensors for the Urban Heat Island analysis in the city of Rome. Remote Sens 2(5):1400–1415. https://doi.org/10.3390/rs2051400
Nissan H, Burkart K, de Perez EC, Van Aalst M, Mason S (2017) Defining and predicting heat waves in Bangladesh. J Appl Meteorol Climatol 56(10):2653–2670. https://doi.org/10.1175/JAMC-D-17-0035.1
Rajib MA, Mortuza MR, Selmi S, Ankur AK, Rahman MM (2011) Increase of heat index over Bangladesh: impact of climate change. Int J Civ Environ Eng 5(10):434–437
Dewan A, Kiselev G, Botje D, Mahmud GI, Bhuian MH, Hassan QK (2021) Surface urban heat island intensity in five major cities of Bangladesh: patterns, drivers and trends. Sustain Cities Soc 71(April):102926. https://doi.org/10.1016/j.scs.2021.102926
Sarwar IM, Billa M, Paul A (2016) Urban land use change analysis using RS and GIS in Sulakbahar ward in Chittagong city, Bangladesh. Int J Geomat Geosci 7(1):1–10
Madow WG (1968) Elementary sampling theory. Technometrics 10(3):621–622. https://doi.org/10.1080/00401706.1968.10490610
Pereira e Silva L, Xavier APC, da Silva RM, Santos CAG (2020) Modeling land cover change based on an artificial neural network for a semiarid river basin in northeastern Brazil. Glob Ecol Conserv 21. https://doi.org/10.1016/j.gecco.2019.e00811
Das N, Mondal P, Sutradhar S, Ghosh R (2021) Assessment of variation of land use/land cover and its impact on land surface temperature of Asansol subdivision. Egypt J Remote Sens Sp Sci 24(1):131–149. https://doi.org/10.1016/j.ejrs.2020.05.001
Acknowledgements (if any)
This research is the output of the author’s undergraduate thesis. The authors gratefully acknowledge the Department of Urban and Regional Planning, Chittagong University of Engineering and Technology for facilitating all the requirements of this research.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2024 The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Das, T., Islam, M.I., Raja, D.R. (2024). Predicting Future Land Use and Land Cover Changes and Their Effects on Land Surface Temperature in Chattogram City of Bangladesh. In: Arthur, S., Saitoh, M., Hoque, A. (eds) Proceedings of the 6th International Conference on Advances in Civil Engineering. ICACE 2022. Lecture Notes in Civil Engineering, vol 368. Springer, Singapore. https://doi.org/10.1007/978-981-99-3826-1_14
Download citation
DOI: https://doi.org/10.1007/978-981-99-3826-1_14
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-99-3825-4
Online ISBN: 978-981-99-3826-1
eBook Packages: EngineeringEngineering (R0)