The Environmental Effects of Urban Development in Hanoi

Preparing to load PDF file. please wait...

0 of 0
100%
The Environmental Effects of Urban Development in Hanoi

Transcript Of The Environmental Effects of Urban Development in Hanoi

sustainability
Article
The Environmental Effects of Urban Development in Hanoi, Vietnam from Satellite and Meteorological Observations from 1999–2016
Thi Mai Nguyen, Tang-Huang Lin * and Hai-Po Chan
Center for Space and Remote Sensing Research, National Central University, Taoyuan 32001, Taiwan; [email protected] (T.M.N.); [email protected] (H.-P.C.) * Correspondence: [email protected]; Tel.: +886-3-422-7151 (ext. 57633)
Received: 4 March 2019; Accepted: 22 March 2019; Published: 23 March 2019
Abstract: Since 1990 the Hanoi capital region (or Hanoi metropolitan area) in Vietnam has undergone rapid development, which has gone together with increasing socio-economic growth and prosperity. However, the environmental degradation that has accompanied urban development has raised considerable concern from the public in recent years. This research investigates the effects of urban development on urban sprawl, urban heat island (UHI), and metropolitan weather phenomena that are related to the quality of urban life in the period from 1999–2016. To achieve these objectives, remote sensing technologies were applied to satellite images at three time points (i.e., 1999, 2009, and 2016) that were associated with the meteorological dataset from ground-based stations. The spatial distribution evolution was examined for the land use/land cover changes while using the normalized difference vegetation index (NDVI) and normalized difference built-up index (NDBI). The increasing impact of urban sprawl on UHI intensity is determined based on the land surface temperature (LST) in multi-temporal forms. Increasing urbanization with the development of gradual outward and northward expansion from the city centre intensified the correlation analysis shows that the UHI. The potential formation of new UHI areas in the near future is also indicated. Furthermore, more than 30% of the metropolitan area is decaying in ecological quality according to an assessment of the urban thermal field variance index (UTFVI). With respect to metropolitan weather, the urbanization in Hanoi affected the observation of meteorological parameters revealed that the relative humidity, total rainfall, temperature, and wind speed over both urban and rural areas. The overall results imply that urban development and its environmental effects and impacts have imposed pressing issues and new challenges to sustainable development in the Hanoi metropolitan area.
Keywords: Hanoi Vietnam; urbanization; remote sensing; urban heat island effect; ecological assessment; metropolis weather

1. Introduction
The world has recently experienced the most rapid pace of urbanization since the industrial revolution in the second half of the twentieth century, especially in developing countries [1]. Rapid urbanization has strongly promoted economic and social development, but many environmental issues have emerged as a consequence of this development [2,3]. The urban heat island (UHI) phenomenon is one of the major negative impacts of urbanization, which indicates a metropolitan core that has higher air temperature and land surface temperature (LST) than the surrounding rural area. The surface UHI (SUHI) and atmospheric UHI are two types of UHI [4]. Atmospheric UHI is observed based on air temperature and SUHI is observed based on LST [5]. Together with the urbanization process, the rapid spread of urban areas and changes in land use have led to significant and remarkable UHI

Sustainability 2019, 11, 1768; doi:10.3390/su11061768

www.mdpi.com/journal/sustainability

Sustainability 2019, 11, 1768

2 of 24

phenomena [2]. Urbanization via the adjustment of natural surfaces and atmospheric conditions has an effect on local weather and climate systems, and UHI affects the quality of urban life through impacts on human health and ecosystem functions [6,7]. Thus, a top priority has been the identification and assessment of urbanization impacts, and recent research has been performed with a view towards a better understanding of the impacts and problems that are related to urbanization [2].
Hanoi, the capital of Vietnam, occupies an important geographical and political position. Since 1990, industrialization and modernization began to accelerate in Hanoi, creating a firm foundation for development in the 21st century. The city has achieved excellent results during this period, and the results in the field of economic development and modernization of urban infrastructure are most noteworthy. However, Hanoi also faces serious environmental problems. Issues that are related to urbanization in Hanoi are multi-faceted and many problems related to urban transition still require additional research [8]. The city has rapidly urbanized since the country adopted economic reforms that drove built-up land to exponentially expand, especially in the buffer zone between 10 and 35 km from the city centre [9]. The change in land use in Hanoi has increased the number of hotspots and also increased the average night air temperature in new urban areas by up to 1◦C [10].
Hanoi has significantly increased in population (4643.4 thousand in 17 years, a 46.37% increase) and experienced urbanization, as indicated by the increase of residential and industrial classes as land uses and rapid increases in socio-economic indicators from 1960 to 2013, such as the newly built-up areas of residential housing (4987 thous.m2, a 99.48% increase), industrial establishments (97092 industrial establishments, a 97.75% increase), and electricity per capita (1531 kWh/pers, a 93.75% increase) [11]. The environmental degradation that accompanied urban development has increased great concern from the public. Therefore, the investigation and assessment of urban development and its environmental effects in the Hanoi metropolitan area are essential and urgent issues in environmental sustainability.
With the advantages of high spatial and temporal resolution, availability, and free and easy access, numerous studies of UHI have been based on LST that were observed from satellite imagery [12]. For UHI studies at the local scale, thermal data (including Landsat 5 TM and Landsat 8 OLI with spatial resolutions of 120 and 100 m, respectively) were widely used [13]. LST is governed by a complex pattern of landscape structure that is directly linked to land use/land cover (LULC) variables [14]. To examine the spatial and temporal variations of LST that are caused by land cover changes, the normalized difference vegetation index (NDVI) and the normalized difference built-up index (NDBI) are among the most commonly used landscape indices [15]. In addition, based on remote sensing data, a number of available thermal comfort indices have been used to evaluate the effects of UHI on urban quality of life, namely, the temperature humidity index (THI), the physiological equivalent temperature [16], the wet-bulb globe temperature (WBGT), and the urban thermal field variance index (UTFVI) [17–20]. The UTFVI was used in the ecological evaluation of UHI zones because of the application that was previously tested for Landsat data [7]. Thus, Landsat imagery was used to monitor the development of the urban sprawl and UHI in this research. Retrievals of satellite observations and in situ meteorology data were combined to explore the urbanization effects and the UHI phenomenon in Hanoi.
In this study, the impacts of urban development on the environment in the Hanoi metropolitan area during the period from 1999 to 2016 were examined regarding four aspects. First, we analysed the impact of urbanization on the development of SUHI: (a) analysis of the change in urbanization indicators; (b) mapping of the LULC and the detection of the related changes over time; (c) retrieval of LST, calculation of NDVI and NDBI from Landsat data, and the production of LST, NDVI, and NDBI spatial distribution maps; (d) mapping of the UHI zone and non-UHI zone based on the LST; and, (e) Establishment of the correlation of LST with NDVI and NDBI, LST for each land type, and UHI with urbanization indicators. Second, we explored the development trend of atmospheric UHI via changes in air temperature and UHI index over time. Third, we evaluated the impact of UHI on the quality of urban life in Hanoi based on the UTFVI. Finally, we assessed the impacts of urbanization

Sustainability 2019, 11, 1768

3 of 24

on the changes in regional weather, including relative humidity, total rainfall, temperature, and wind speed in both urban and rural areas.
2. Materials and Methods
Hanoi, as the political centre of Vietnam, is located in the northern region and it covers an area of 3358.9 km2. The city is the country’s second largest city by population, with 7328.4 thousand people in 2016. Hanoi has a subtropical climate with abundant precipitation. With the typical climate of northern Vietnam, the city experiences four distinct seasons. Hanoi’s administrative boundaries changed in 2008, during which Ha Dong was transformed into an urban area, Ha Tay was merged into Hanoi, and Son Tay was downgraded to a district-level town (Figure 1b). Hence, the city is divided into 12 urban districts, one district-level town, and 17 rural districts (Figure 1a). At the metropolitan level, Hanoi is developing based on a centre core and satellite urban areas [21]. In addition to the eight existing industrial parks, five new large-scale industrial parks and 16 small- and medium-sized industrial clusters are under construction by the city. Figure 1 shows the study area.

Figure 1. (a) Geographic location and administrative districts of the study area in Hanoi city, (b) comparison of Hanoi before and after expansion of the administrative boundaries in 2008.
The datasets that were used in this study include (1) Landsat images of the study region at three time points (1999, 2009, and 2016), as collected from United States Geological Survey (USGS), and Table 1 shows the descriptions of the Landsat images; (2) Socio-economic statistical data, including the newly built area of residential housing in the year, the number of industrial establishments, electricity per capita during 1961–2013, and the population and population density during 1999–2016, as sourced from the Statistical Yearbook, Statistical Office in Hanoi city; and, (3) Climatic data during 1961–2016, including temperature, rainfall, wind speed, and relative humidity, as supplied by the Vietnam Meteorological Bureau.

Sustainability 2019, 11, 1768

Table 1. The Landsat imagery datasets used in this study.

Sensor Landsat 5 TM Landsat 5 TM Landsat 8 OLI/TIRS

Scene ID
LT51270451999266BKT00 LT51270461999266BKT00
LT51270452009309BJC00 LT51270462009309BJC00
LC81270452016281LGN01 LC81270462016281LGN01

Acquisition date 23/09/1999 05/11/2009 07/10/2016

Season Autumn Autumn Autumn

3. Methodology The overall methodology of this study has been presented in a flowchart (Figure 2).

4 of 24

Figure 2. Flowchart of this study.
3.1. Change of Urbanization Indicators
In this study, population and population density data from 1999 to 2016 were used to analyse the population change and population distribution in Hanoi. Furthermore, the 2016 population density map was produced.
Socio-economic statistical data, including the newly built area of residential housing in the year, the number of industrial establishments, and electricity per capita in Hanoi during the period 1960–2016, were analysed to indicate development over time.
3.2. Image Processing
3.2.1. Pre-processing
First, the single-band images of Landsat 5/Landsat 8 were combined into a multi-band image (excluding the thermal band) while using the layer stacking tool.
Second, two Landsat images were merged each year to cover the entire study area. However, the study area is only a small portion of that region. Therefore, while using the clip tool in ArcGIS software [22], we created a copy of only the data that fell within the study area and discarded the information that lies outside of the area.

Sustainability 2019, 11, 1768

5 of 24

Third, the quick atmospheric correction module in the ENVI software was used to perform atmospheric correction. Finally, all of the image data were re-projected to a universal transverse mercator (UTM) coordinate system, datum WGS84, zone 48. The Landsat images were interpreted into product datasets at a scale of 1:450.000.
3.2.2. Mapping Land Use Land Cover (LULC)
In the early 1970s, visual analysis was the first land cover classification method that was applied to Landsat imagery and, subsequently, pixel-based classification methods, including unsupervised and supervised classification, were used [23]. Furthermore, over the most recent four decades, the supervised classification method using maximum likelihood classifier (MLC) became one of the most commonly used classification methods for Landsat imagery [23,24]. Therefore, in this study, MLC was conducted on the optical bands of Landsat images to construct the land use/land cover maps with eight categories of land use classes (Table 2). In addition, the changes in land use types over time were detected and analysed from the resulting maps.

Table 2. Eight categories of land use/land cover classes that were used in this study.

Classes

Include

1 Residential

Rural residential land and urban residential land

2 Industrial, Commercial Land for industrial parks, industrial clusters, export processing zones; land for trading and service

3 Active agriculture

Land for cultivation of perennial trees, land for cultivation of annual crops, paddy land

Inactive agriculture

Wetland—Paddy land: pump water in the field to prepare for

4 (Wetland and Dry land) farming works Dry land—Land for cultivation of perennial

trees, land for cultivation of annual crops

5 Water

Land with rivers, streams, canals, springs (located in rivers, natural lakes)

6 Forestry

Land for special-use forests Land for protection forests Land for production forests

7 Sandbars

Riparian land

8 Bare soil

Unused land and Unused Mountain Land

After classification was performed, the LULC map in 2016 was compared with the land use map in 2016, as supplied by the Hanoi Environment and Natural Resource Department as reference data to assess the accuracy of the classification. The result of the classification map was assessed based on the simple random sampling method, as the traditional and simplest accuracy assessment [25]. In general, 50 samples for each class are suggested [26]; therefore, the validation of classification was calculated from 400 comparison samples of the classified image. The evaluation quantitative classification performance analysis result is based on the total accuracy, the error matrix (including the producer accuracy and user accuracy), and the Kappa index [27].
3.2.3. Calculation of NDVI and NDBI
NDVI is one of the most widely used vegetation indices and it has a long record of merit within the remote sensing community [28,29]. This index is a measure of the difference in reflectance between the red band (RED) and near-infrared band (NIR) of the images. The NDVI ranges between −1.0 to +1.0, where positive values indicate vegetated areas and negative values denote non-vegetated areas [30]. NDVI is expressed in the following Equation (1):
NDVI = NIR − RED (1) NIR + RED

Sustainability 2019, 11, 1768

6 of 24

where NIR corresponds to Landsat TM band 4 (0.76–0.90 µm) and Landsat OLI band 5 (0.85–0.88 µm) and RED corresponds to Landsat TM band 3 (0.63–0.69 µm) and Landsat OLI band 4 (0.64–0.67 µm).
NDBI is an effective technique for mapping urban built-up areas, with a total accuracy of 92% [31]. NDBI is calculated based on the difference between NIR and MIR (middle-infrared band) of the images. This index is obtained from Equation (2) [32]:
NDBI = MIR − NIR (2) MIR + NIR
where MIR corresponds to Landsat TM band 5 (1.55–1.75 µm) and Landsat OLI band 6 (1.57–1.65 µm.

3.2.4. Retrieval of LST
To derive LST from Landsat data, a variety of algorithms has been developed [13]. The technique that is presented in this study is used in retrieving the LST of a given Landsat image based on the input of the RED, NIR, and TIR (thermal infrared band). Firstly, the proportional vegetation P(v) is calculated from the NDVI values, as Equation (3) [33].
P(v) = NDVI − NDVImin 2 (3) NDVImax − NDVImin
where the NDVImin (minimum NDVI) and NDVImax (maximum NDVI) stand for the dynamic range of NDVI value.
For the LST retrieval, LSE (land surface emissivity) is required and is performed while using the following equation [34]:

ε = mP(v) + n (4) m = (ε − ε) − (1 − εσ)Fεv and n = εσ + (1 − εs)Fεv

where εs and εv are the soil emissivity and vegetation emissivity, respectively. The values of m = 0.004 and n = 0.986 were obtained from the final result of a previous study [34].
After converting the DN values to at-sensor spectral radiance (Lλ), LST can be retrieved with LSE (ε) from Equation (4) while using the following equation:
LST = K2 − 273.15 (5) Ln εL∗Kλ 1 + 1
where LST is the land surface temperature in Celsius (o C); K1 and K2 are the thermal constants of the TIR, which can be identified in the metadata file that is associated with the satellite image [35]; Lλ is the spectral radiation value; and, ε is the emissivity calculated from Equation (4). To supply the results in Celsius, it is necessary to revise the data by adding absolute zero, which is approximately equal to −273.15. The relationships between land cover types and LST in Hanoi were analyzed accordingly.

3.2.5. Mapping of the UHI zone
Not all cities have separate UHIs, and their distribution and development depends on many factors. Therefore, it is necessary to indicate the spatial distribution and development trends of the UHI areas. The intensity of the UHI is identified as the difference between the LST of the UHI areas, and that of the rural areas while using the following equations [36]:

LST > µ + 0.5∗δ

(6)

referred to UHI area

0 < LST ≤ µ + 0.5∗δ

(7)

Sustainability 2019, 11, 1768

7 of 24

denoted as non-UHI, where µ and δ are the mean and standard deviation of the LST in the study area, respectively.
However, the UHI zone distribution that is based on this method might be influenced by many factors. Therefore, to increase accuracy, a combined method using the LULC map and LST map is used to determine the heat island zone. In this study, the heat island region is defined as containing both of the following elements:
LST > µ + 0.5 * δ and the LULC type is Residential and Industrial referred to UHI area.
3.2.6. Urban Thermal Field Variance Index (UTFVI)
In urbanized areas, the UHI effect is an important factor in environmental and public health studies [37]. A number of thermal comfort indices are available for the evaluation of UHI impacts on the quality of urban life. In this study, UTFVI was chosen and it is estimated by Equation (8) [20].
UTFVI = Ts − Tm (8) Ts
where Ts is the LST and Tm is the mean of the LST of the study area. The UTFVI values were divided into six categories, each with the corresponding interpreted
ecological valuations. [20,38]. Table 3 shows the thresholds in the six UTFVI categories.

Table 3. Thresholds of the six urban thermal field variance index (UTFVI) levels used in this study.

UTFVI
<0.000 0.000–0.005 0.005–0.010 0.010–0.015 0.015–0.020
>0.020

Urban Heat Island (UHI) Phenomenon
None Weak Middle Strong Stronger Strongest

Ecological Evaluation Index
Excellent Good Normal Bad Worse Worst

3.3. Variation of Meteorology
Atmospheric UHI was observed based on the air temperature that was collected from four stations in Hanoi from 1999–2016. Two stations represent urban areas, namely, Lang and Ha Dong, and two stations represent rural areas, namely, Son Tay and Ba Vi.
To analyse the climatic effects of urbanization in Hanoi, two stations, i.e., Lang and Son Tay, were selected as the urban and rural representative stations, respectively. We aimed to analyse the difference of the temporal variations (between urban and rural areas) in the mean annual temperature, total rainfall, relative humidity, and wind speed from 1961–2016. The variation in annual temperature was defined as the UHI index.
3.4. Statistical Analysis
Scatter diagrams were plotted with regression analyses to examine the correlations of LST with NDVI and NDBI. The points for analysis were extracted from the pixel values of the LST, NDVI, and NDBI maps over study area. The UHI index and the relationships between the UHI index and urbanization indicators, including the newly built area of residential housing in the year, the number of industrial establishments, and electricity per capita, were also analysed and discussed.

Sustainability 2019, 11, 1768

8 of 24

4. Results and Discussion
4.1. Evolvement of Urbanization in Hanoi
From 1998 to 2016, Hanoi experienced a considerable change in the total population and population density (Figure. 3). The total population significantly increased from 2685 thousand people in 1999 to 7328.4 thousand people in 2016 (Figure 3a), exhibiting an increase of 4643.4 thousand people after 17 years, with a rate of increase of 273.14 thousand people per year. A double jump in population occurred in 2008 with an increase of 3153.3 thousand people as compared with 2007, causing administrative expansion at this time. In addition, this event is related to rapid urbanization and economic development, i.e., people from other provinces left their hometowns and then moved to Hanoi with the desire to find jobs with higher incomes.
Population density also continuously increased in all districts of Hanoi during the study period (Figure 3b). According to the population density map in 2016 (Figure 3c), the five districts of Ba Dinh: 26249 people/km2, Cau Giay: 20931 people/km2, Dong Da: 40331 people/km2, Hai Ba Trung: 31308 people/km2, Hoan Kiem: 29471 people/km2, and Thanh Xuan: 29295 people/km2 have the highest population densities, which tend to increase over time. These locations are the central areas of Hanoi, where the business and economic conditions are favourable, and this situation has driven the density population in these areas much higher than in the surrounding regions.

Figure 3. Cont.

Sustainability 2019, 11, 1768

9 of 24

Figure 3. (a) Total population (thousands), (b) population density (people/km2), and (c) population density map in 2016 (people/km2) in Hanoi from 1999 to 2016.
Other aspects, including the newly built area of residential housing in the year, the number of industrial establishments, and electricity per capita, were also used to describe the process and extent of urbanization in Hanoi. From 1960 to 2013, these indicators significantly increased in Hanoi (Figure 4), especially after 1990. The newly built area of residential housing (Figure 4a) increased by 4987 thousand m2 in the 53 years from 1960 to 2013, with a rate of increase of 94.09 thousand m2 per year. Since the beginning of rapid development in 1990, the construction and rehabilitation of the city’s infrastructure have led to the rapid increase in new buildings in Hanoi and, after 23 years, this value has increased by 4870.5 thousand m2, equivalent to an increase rate of 211.76 thousand m2 per year. At the same time, the number of industrial establishments (Figure 4b) has increased from 1128 in 1960 to 98220 in 2013, an increase of 97092 industrial establishments. As a result, the number of industrial establishments after 53 years has increased by 1831.93 industrial establishments per year. In the period of 1990–2013, the value has rapidly grown, with 90,163 new industrial establishments after 23 years, which is equivalent to an increase rate of 3920.13 industrial establishments per year. Finally, from 1960 to 2013, electricity per capita (Kwh/pers) in Hanoi (Figure 4c) has increased by 1531 Kwh/pers in 53 years, with an increase rate of 28.89 Kwh/pers per year. During the period from 1960–2013, electricity per capita increased by 1281 Kwh/pers after 23 years, with an increase rate of 55.7 Kwh/pers per year. The growth tendency of these socio-economic indicators is closely related to the economic development of the city, especially after the boom of 1990.
It is clear that the capital has experienced rapid urbanization since 1990, with continuous increases in population and socio-economic indicators over time.
Figure 4. Cont.

Sustainability 2019, 11, 1768

10 of 24

Figure 4. Variances in socio-economic indicators of Hanoi from 1960 to 2013, (a) newly built area of residential housing in the year (thousand m2), (b) industrial establishment (number of establishment), and (c) electricity per capita (Kwh/pers).
4.2. Changes of LULC
Table 4 shows the result of the accuracy assessment of the land use/land cover classification map in 2016. The overall accuracy (90.75%) and Kappa index (0.88) are sufficiently high to believe in the classification result. Therefore, we classify the land use/land cover for 1999 and 2009, based on the 2016 classification map.

Table 4. Results of validations of land use/land cover (LULC) classification in 2016.

Types of LULC
Residential Industrial Active agriculture land Inactive agriculture land
Water Forestry Sand-bar Bare soil Overall Accuracy (%) Kappa Index

User’s (%) Producer’s (%)

94

96

93

92

90

94

81

78

88

90

94

94

96

98

88

84

90.75

0.88

In general, the classification results reveal that, agricultural land, including active and inactive agriculture land, has the largest area of land use/land cover type and it is distributed throughout the city (over 50% in all three time points). Forests occupy a small area, which is mainly located in the west and southwest of Hanoi. Residential areas are found in the city centre and the northwest in 2009 and 2016.
Figure 5 displays the LULC map of Hanoi in 1999, 2009, and 2016, with eight land use/land cover classes of residential, industrial, commercial, active agriculture, inactive agriculture, water, forestry, sandbars, and bare soil. Furthermore, the temporal variations of land use/land cover in fractions at
HanoiLstDevelopmentUrbanizationUhi