west bengal : geoinformatics for sustainable ...

WEST BENGAL : GEOINFORMATICS FOR SUSTAINABLE ENVIRONMENT MANAGEMENT [ Volume - I ]

Editor Biraj Kanti Mondal

NETAJI SUBHAS OPEN UNIVERSITY DD-26, Sector-I, Salt Lake City, Kolkata-700 064

West Bengal : Geoinformatics for Sustainable Environment Management (Vol-I) Editor : Biraj Kanti Mondal

© Netaji Subhas Open University No part of the book be reproduced either electronically or by any means without written permission from the publishier/editor.

First Published : December, 2017 Published by: The Registrar, Netaji Subhas Open University DD 26, Sector I, Salt Lake City, Kolkata - 700064 Website: www.wbnsou.ac.in

ISBN: 978-93-82112-63-1

Price: ` 750.00 $ 15

Printed by: Cyber Graphics 36/1, Nainan Para Lane, Baranagar Kolkata - 700 036

Disclaimer : The views expressed by the authors/contributors are personal and do not necessarily represent the views of the University. It is the sole responsibility of the authors/contributors for any legal ramifications for the contents in the research papers.

Contents Message from the Vice-Chancellor – Professor Subha Sankar Sarkar Vice Chancellor, NSOU Foreword – Professor Sunando Bandyopadhyay Professor of Geography, University of Calcutta From the Desk of Director – Professor Kajal De Director, School of Sciences, NSOU Editor’s Note – Biraj Kanti Mondal Assistant Professor of Geography, School of Sciences, NSOU PART - I : Geoinformatics and its Electric Applications Application of Machine Learning in Environmental Management – An example of Landslide Susceptibility Zonation – A. R. Ghosh, Kajori Parial and Debananda Biswas

17-26

Monitoring of Flood Situation by Applying Remote Sensing Techniques – Asit Kumar Sarkar

27-40

PART - II : Studying the Himalayan Bengal : From Snowline to the Terai Morphometric Diversity and its Implication to Landslide Susceptibility in the Balason River Basin, Darjeeling Himalaya – Subrata Mondal and Sujit Mandal

41-68

PART - III : The Red Soil : Introspecting Western Bengal Acreage analysis of Betel vine crop through Boroj detection from high resolution imagery and internal architectural layout in Moyna Block of Tamluk Subdivision, Purba Medinipur – Manas Hudait and Priyank Pravin Patel

69-82

Impact and Vulnerability of Drought on the Agriculture of Bankura District in West Bengal – Tanmoy Basu

83-108

Impact of Stone Mining on Environment and Local Livelihoods: A Case Study of Mohammad Bazar Block, Birbhum – Firdousi Rahaman Siddika and Krishnendu Gupta

109-124

Influence Zone Analysis of Some Selected Colleges in Paschim Barddhaman District, West Bengal – Sarbendu Bikash Dhar

125-136

The Emerging Issue of Forest Degradation in Purulia District – Mau Manna and Biraj Kanti Mondal

137-156

PART - IV : The Riverine Plains and the Erosional Garnishing : Picturesque of West Bengal Implication of Geoinformatics and GIS to identify the Change Detection of East Kolkata Wetlands – Biraj Kanti Mondal

157-174

Identification of Optimum Band Combination for Land Surface Water Mapping Using Landsat TM Digital Data: A Statistical Approach – Sakti Mandal

175-194

A Remote Sensing and GIS based approach in Identifying Channel Changes and Fluvio-Geomorphic Features in Nabadwip and its Surroundings, West Bengal – Aparupa Sinha and Anupam Das

195-212

Sustainable Wetland Resource Management: A Case Study of Bariti Beel, North Twenty Four Parganas, West Bengal – Madhumita Basu

213-228

An Appraisal of Temporal Changes in Agricultural Land use Pattern of Krishnanagar Subdivision, Nadia District, West Bengal – Susmita Mandal and Anupriya Chatterjee

229-244

PART - V : Magnifying Delhi through the lences of Geoinformatics : A Special Focus Impact of Land use and Land Cover Change on Land Surface Temperature Using Remote Sensing Techniques – Dipanwita Dutta, Atiqur Rahman and Arnab Kundu

245-258

IMPACT OF LAND USE AND LAND COVER CHANGE ON LAND SURFACE TEMPERATURE USING REMOTE SENSING TECHNIQUES Dipanwita Dutta Atiqur Rahman Arnab Kundu

Abstract Considering the rapid changes in urban areas worldwide along with micro-climatic changes urban land use dynamics and its impact has been a major topic of concern among all countries. The present study aims to identify the spatio-temporal pattern of land use/land cover and its impact on land surface temperature using satellite based remote sensing technique. Landsat data of two different periods, 2003 and 2014 were used for extracting the prevailing land use and land cover. In order to classify the images a relatively new and robust support vector machine algorithm was applied. The spatio-temporal pattern of urban built-up surface was compared with the land surface temperature maps to identify their interrelationships. The kappa coefficient estimated for accuracy assessment of the SVM classified was found significantly high (>0.85) proving the efficiency of the classifier to extract land use and land cover information. The study reveals a close association between built-up surface and land surface temperature of the area.

Key Words: Land use/Land cover, Land Surface Temperature, Landsat data, Support Vector Machine. Introduction: Urbanization mainly considered as an important human activity which directly affects the environment at local, regional and global scales (Turner et al. 1990). Delhi megacity is not the exception of that and it is being changed everyday due to havoc anthropogenic activities. Urbanization and urban growth has 245

West Bengal : Geoinformatics for Sustainable Environment Management

been perceived as a complex process of spatio-temporal change in both socio-economic and physical components at different scales (Han et al. 2009). The socio-economic components of urban growth include spatio-temporal changes in population and economy whereas the physical component indicates spatial expansion of urban area as a result of changes in land use and land cover within urban area. Urban population in India is mainly confined in class 1 cities, especially in the megacities where urban population has concentrated and increased steadily over the last few decades (Kundu, 2006). Although the process of urbanization in Delhi has been continuing from the Mughal era, it was accelerated after being designated as the capital of India in 1911 (www.gisdevelopment.net). Remote sensing and geographical information system has been considered as an essential tool for mapping land use and land cover of an area since the last couple of decades. It has many advantages over conventional cartographic techniques such as efficient database management system, spatial analysis and high accuracy. Unlike the field based methods of data collection and mapping it provides detailed information of earth surface without taking much time and cost (Da Costa, 1999). Since remote sensing technique allows acquisition of spatial data in multi-resolution, multi-spectral and multi-temporal form, it has been accepted as an essential tool for mapping and monitoring land use land cover dynamics (Kushwaha et al. 1996). This technique provides vivid scope to explore land features, vegetation dynamics, agricultural productivity, water quality, spatio-temporal changes in land use and land cover and their impact on environment with good accuracy. While remote sensing is all about acquiring digital information based on spectral reflectance of earth surface, GIS is a computer-based information system with having capability to integrating data from various sources for providing the information necessary for effective decision-making (Han and Kim, 1989). 246


Urban areas are characterized by distinct climatic condition which is absolutely different from its surrounding areas. The micro climatic difference in rural and urban areas is mainly attributed to the varying temperature of the areas. Increasing temperature in cities is mainly caused by high outgoing radiation from impervious surface cover (Roetzer et al. 2000). This discrepancy between temperature pattern of urban and rural areas can be identified by remote sensing derived land surface temperature (LST). In order to estimate LST several sensors i.e., Moderate Resolution Imaging Spectroradiometer (MODIS), Advanced Very High Resolution Radiometer (AVHRR), Advanced Spaceborne Thermal Emission and Reflection (ASTER), Landsat Thematic Mapper/Enhanced Thematic Mapper Plus (TM/ETM+), are being popularly used by the researchers (Lo and Quattrochi, 2003; Han et al. 2007). In a study, Chaudhuri and Mishra (2016) compared the spatio-temporal pattern of LST and LULC in the bordering cities of India and Bangladesh and found that rate of change in LST was much higher during 2005-10. Based on the interrelationship between LST and NDVI, Mallick et al. (2008) proposed a prediction model for estimating land surface temperature. It is also proved that land surface phenology of rural-urban transitional areas is strongly correlated with surface temperature (Han and Xu, 2013). The main objective of this study is to identify the spatio-temporal pattern of land use and land cover and their impact on land surface temperature in Delhi NCT area. Materials and Methods: 1. Study area: The study area comprises Delhi national capital territory (NCT) which is the capital of India and located at the heart of the country (Fig. 1). Delhi NCT covers an area of 1483 sq. km. 247


extending between 28° 23’ 17’’ to 28° 53’ 00’’ North latitude and 76° 50’ 24’’ to 77° 20’ 37’’ East longitude. It falls under semi arid zone with extreme climatic condition. Generally this area is characterized by very hot summer (April-July) with maximum temperature greater than 40 degree Celsius and cold winter (December-January) with minimum temperature less than 6 degree Celsius. On an average, the city receives about seventy centimeter rainfall annually. As per the census record (2011) Delhi NCT is having a population of 46,05,555. It has total nine districts and among these, the north east district is characterized by highest population density. Data: 1. Satellite Data: The Landsat data of 2003 and 2014 were chosen for the present study (Table 1). Although the Landsat archive in GLOVIS provides a large number of temporal datasets for the study area, Landsat data of these years were selected considering availability of the data in similar season. For the present study, the data of late winter season (February-March) was targeted due to availability of cloud free and good quality images during the time. The Landsat 8 provides eleven spectral bands among which first nine bands including one PAN band are of onboard Operational Land Imager (OLI) sensor and remaining are of Thermal Infrared Sensor (TIRS). The Landsat 8 operates in the visible, near-infrared, short wave infrared and thermal infrared region of electromagnetic spectrum and having spatial resolution of 30 and 15 meters for multispectral and PAN images respectively.

248


Table 1. Characteristics of the Landsat 7 TM and Landsat 8 sensor Sensors

Spectral bands

Spatial Spectral region Time of resolution acquisition

Band 1: 0.45-0.52μm Band2: 0.53-0.60μm Landsat Band 3: 0.63-0.69μm 30 m TM Band 4: 0.75-0.90μm Band 5: 1.55-1.75μm Band 6: 1.04-12.5μm 120 m Band 7: 2.09-2.35μm 30 m Band 1:0.45-0.52 μm Band 2:0.52-0.60 μm Landsat Band 3:0.63-0.69 μm 30m OLI/ Band 4:0.77-0.90 μm TIRS Band 5:1.55-1.75 μm Band 6:10.40-12.50 μm Band 7:2.09-2.35 μm 60(30)m

VNIR

SWIR TIR SWIR Visible Visible Visible Near-Infrared Near-Infrared Thermal Mid-Infrared

2003, February

2014, March

Field Survey: Field verification was carried out during February, 2014 and fifty sample points were collected using a Trimble GPS. Among these fifty sample points, forty percent of the points were kept separately for accuracy assessment and rest of the sample points used as training sets. 2. Methodology: Since the raw images consist of radiometric and geometric errors, before classification they were pre-processed using the FLAASH atmospheric correction followed by image to image rectification. Image pre-processing is essential prior to change detection as noises present in satellite image may provide wrong information. The Principal Component Analysis was applied on both images for transforming the original 249


multispectral bands into principal components. Finally, the images were classified using support vector machine classifier for analyzing the land use and land cover of the study area. In order to assess the long term spatio-temporal changes, a simple change detection analysis was performed. The Land Surface Temperature of the year 2003 and 2014 were calculated using the following equation:

Where, =at satellite temperature. =wavelength of the emitted radiance. =h*c/ σ (1.438 * 10-2 m K). h =Planks constant (6.626 * 1034 J s). σ = Boltzmann constant (1.38 * 10-23 J/K). c =velocity of light (2.998 * 108 m/s) = Emissivity The brightness temperature values were derived from the spectral radiance map assuming the earth’s surface as blackbody (Chander et al., 2009)

250


Where, Tb=sensor brightness temperature Lλ = sensor spectral radiance K1 and K2 values of the Landsat ETM+ and OLI sensor applied as per the user guide of Landsat data. Result and discussions: Land use and land cover of Delhi NCT area in 2003 and 2014: The Landsat TM image of 2003 was classified using the support vector machine supervised classifier (Fig. 2). It can be observed that agricultural areas are located in the northern, western part and along the Yamuna River whereas builtup areas are mainly confined in central, eastern and southern part of NCT. Although, high density built-up areas are mainly located in the core and central part of the city, small patches are also visible towards the outskirts of NCT region. Distinct patches of low density built-up areas can be detected in the extreme east and southern part of peri-urban areas. It is seen that agricultural land and sparse vegetation occupies 465.07 sq. km. (31.42%) and 248.83 sq. km. (16.81%) in Delhi NCT (Table 2). Dense vegetation in Delhi NCT was about 131.94 sq. km (8.91%). It was observed that area under high density built-up was about 130.03 sq. km. (8.79%) in Delhi NCT whereas low density built-up area was estimated 377.76 sq. km. (1.24%). The study shows that more than one third of the total Delhi NCT was under built-up class.

251


Table 2. LULC in Delhi NCT area (2003) LULC classes High Density Built-up Low Density Built-up Dense Vegetation Agricultural Land Sparse Vegetation Waterbody Wasteland Total

Delhi NCT (2003) Area in sq. km. Percentage 130.03 8.79 377.76 25.52 131.94 8.91 465.07 31.42 248.83 16.81 22.86 1.54 103.58 7.00 1480.07 100.00

It is clearly evident from the classified map (Fig. 3) that most of the area within Delhi NCT was under built-up class during the year 2014. It was observed that these areas have experienced a huge real estate boom in recent years followed by mushrooming of high rises and cluster of new apartments. Table 3. LULC in Delhi NCT area (2014) LULC classes High Density Built-up Low Density Built-up Dense Vegetation Agricultural Land Sparse Vegetation Waterbody Wasteland Total

Delhi NCT (2014) Area in sq. km. Percentage 420.59 28.42 281.38 19.01 157.16 10.62 399.01 26.96 124.51 8.41 21.79 1.47 75.62 5.11 1480.07 100.00

The areas under different land use and land cover classes were estimated and shown in Table 3. The high density built-up area was found as dominant class in Delhi NCT region. It was 252


estimated that about 420.59 sq. km. (28.42 %) of the total NCT region was under high density built-up land which, together with the low density built-up area (19.01%) represents almost half of the total area. The estimated area under agriculture was about 399.01 (26.96%). It is noteworthy that Delhi NCT region having many distinct patches of dense vegetation regardless of its high rate of urbanization. Accuracy Assessment of LULC maps: In order to examine the accuracy of the classification, accuracy assessment was carried out for individual LULC maps. It can be seen that accuracy of classified LULC map of 2014 (K^=0.91) was significantly high due to better radiometric and spectral resolution of Landsat 8 data (Table 4). The SVM classifier with principal components has proved a close agreement between classified and referenced data. Table 4. Classification accuracy and Kappa statistics Land use and land cover Class

High Density Built-up Low Density Built-up Dense Vegetation Agricultural Land Sparse Vegetation Waterbody Wasteland Overall Classification Accuracy Overall Kappa Statistics

2003 2014 Produc- Users Condition- Produc- Users Conditioners Ac- Accura- al Kappa ers Ac- Accura- al Kappa curacy cy (%) for each curacy cy (%) for each (%) Category (%) Category 90.20

92.00

0.91

97.92

94.00

0.93

95.65

88.00

0.86

93.88

92.00

0.91

88.24

90.00

0.88

84.91

90.00

0.88

90.38

94.00

0.93

97.87

92.00

0.91

85.19

92.00

0.91

88.46

92.00

0.91

95.92 89.36

94.00 84.00

0.93 0.82

97.96 86.54

96.00 90.00

0.95 0.88

0.91

0.92

0.89

0.91

253


Land surface temperature (LST): The effect of urban heat island can be observed and distinguished spatially through the satellite derived land surface temperature (LST). High LST values can be found in the impervious areas of Delhi NCT and its surrounding periurban centres (Figs. 4a, 4b). Agricultural lands of this area especially towards the northern zone shows less LST. High LST values can be found in the sparse vegetation and wasteland class which can be explained by high albedo of bare soil and rocky surface present in those areas (Fig. 4). The LST maps derived from Landsat ETM+ (2003) and OLI (2014) reflects good agreement with the spatial pattern of the land use land covers. Conclusion: Estimation of urban growth has been a major topic of research considering its demand in urban planning as well as policies. In this study, spatio-temporal pattern of urban growth has been analyzed by classifying multi-temporal satellite images and quantifying the amount of area under various land use and land cover classes. It has been observed from LULC maps of different years that few areas have experienced notable urban growth during the last decades which is closely associated with the pattern of change in LST. This increasing growth of impervious land was occurred at the cost of pervious lands of bounding zones that caused increase in land surface temperature. The good agreement between the pattern of LULC and LST certainly proves the robustness of satellite based data in studying urban issues.

254


References: 1.

Census of India (2011), accessed on 26th March, 2016.

2.

Chander, G., Markham, B.L., Helder, D.L. (2009) Summary of current radiometric calibration coefficients for Landsat MSS, TM, ETM+, and EO-1 ALI sensors. Remote Sensing of Environment. 113:893-903.

3.

Chaudhuri, G., Mishra, N.B. (2016) Spatio-temporal dynamics of land cover and land surface temperature in Ganges-Brahmaputra delta: A comparative analysis between India and Bangladesh, Applied Geography, 68:68-83.

4.

Da Costa, S. M. F., Cintra J. P. (1999) Environmental Analysis of Metropolitan Areas in Brazil, ISPRS Journal of Photogrammetry and Remote Sensing, 54:41-49.

5.

Han, G., Xu, J. (2013) Land Surface Phenology and Land Surface Temperature Changes Along an Urban–Rural Gradient in Yangtze River Delta, China. Environmental Management, 52:234–249.

6. Han, G.F., Xu, J.H., Yuan, X.Z., Wang, Z.H. (2007) Spatiotemporal change of vegetation distribution in central area of Chongqing City in 1988–2001. Chinese Journal of Ecology, 26:1412–1417. 7.

Han, H., Lai, S., Dang, A., Tan, Z., Wu, C. (2009) Effectiveness of urban construction boundaries in Beijing: An assessment DOI: dx.doi. org. Journal of Zhejiang University, Science A, 10:1285–1295.

8.

Han, S. Y., Kim, T. J. (1989) Can expert systems help with planning? Journal of the American Planning Association, 55:296–308.

9.

Kundu, A. (2006) Globalization and the Emerging Urban Structure: Regional Inequality and Population Mobility, India: Social Development Report, Oxford, New Delhi.

10. Kushwaha, S.P.S., Subramanian, S.K., Chennaiah, G. C., Murthy, J. R., Rao, S. V. C. K., Perumal, A., Behera, G. (1996) Interfacing remote sensing and GIS methods for sustainable development. International Journal of Remote Sensing, 17: 3055-3069. 11. Lo, C., Quattrochi, D. (2003) Land-use and land-cover change, urban heat island phenomenon, and health implications: a remote sensing approach. Photogrammetric Engeering and Remote Sensing, 69:1053– 1063. 12. Mallick, J., Kant, Y., Bharath, B.D. (2008) Estimation of land surface temperature over Delhi using Landsat-7 ETM+, Journal of Indian Geophysical Union, 12:131-140. 255

West Bengal : Geoinformatics for Sustainable Environment Management 13. Roetzer, T., Wittenzeller, M., Haeckel, H., Nekovar, J. (2000) Phenology in central Europe-differences and trends of spring phenophases in urban and rural areas. International Journal of Biometeorology, 44:60–66. 14. Turner, B.L., Clark, W.C., Kates, R.W., Richards, J.F., Mathews, J.T., Meyer, W.B. (Eds.). (1990) The Earth as Transformed by Human Action: Global and Regional Changes in the Biosphere Over the Past 300 Years. Cambridge Univ. Press, Cambridge. 15. www.gisdevelopment.net, accessed on 26th March, 2016.

Fig. 1 Study area 256


Fig. 2 LULC classes of Delhi NCT (2003)

Fig. 3 LULC classes of Delhi NCT (2014) 257


Fig. 4 (a)

Fig. 4 (b) Fig. 4 LST pattern of Delhi NCT in 2003 (a) and 2014 (b) 258