Next Article in Journal
Quaternary Crustal Shortening of the Houyanshan Structure in the Eastern Chinese Tian Shan: Constrained from Geological and Geomorphological Analyses
Previous Article in Journal
Cyanobacteria Index as a Tool for the Satellite Detection of Cyanobacteria Blooms in the Baltic Sea
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Remote Sensing
Volume 15
Issue 6
10.3390/rs15061604
remotesensing-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Study of Land Cover Change in the City with the Fastest Economic Growth in China (Hefei) from 2000 to 2020 Based on Google Earth Engine Platform

by
Jun Geng
1,†,
Lichen Xu
1,2,†,
Yuping Wang
1,† and
Lili Tu
3,*
1
School of Civil Engineering, Hefei University of Technology, Hefei 230009, China
2
Anhui Institute of Geological Surveying and Mapping, Hefei 230022, China
3
School of Resources and Environment, Anhui Agricultural University, Hefei 230036, China
*
Author to whom correspondence should be addressed.
†
These authors contributed equally to this work.
Remote Sens. 2023, 15(6), 1604; https://doi.org/10.3390/rs15061604
Submission received: 3 February 2023 / Revised: 6 March 2023 / Accepted: 9 March 2023 / Published: 15 March 2023
Browse Figures
Versions Notes

Abstract

:
Hefei’s gross domestic product (GDP) growth rate ranks first among all cities in China, and it was the fastest-growing city in China from 2000 to 2020. The high-speed economic development inevitably led to rapid changes in land cover types, e.g., an increment in artificial features (built-up) and decrement in natural features (vegetation). However, (1) many previous studies focus on the land cover change in metropolis or at the global scale, yet few focus on underdeveloped but fast-growing cities; (2) land cover studies mainly focus on global variations, yet seldom on local characteristics. Thus, it is of great significance to monitor the land cover change for the city with the fastest economic growth in China based on the long time-series satellite images from both global and local perspectives. In this study, with support from huge amounts of data (including 719 Landsat TM/ETM+/OLI satellite images, land surface temperature, nighttime satellite images, DEM, multiple land cover products, and various auxiliary data), processing and parallel computing abilities of the GEE platform, classification maps of land cover in Hefei from 2000 to 2020 are produced based on a random forest machine learning method, and the spatio-temporal variations and driving factors are analyzed from both global and local viewpoints. The results show that: (1) the classification accuracy is excellent; the average overall accuracy is 93% and the Kappa coefficient is 0.88; (2) the general spatio-temporal variations in land cover in Hefei from 2000 to 2020 are obvious; the built-up area expanded from 419.72 km2 to 1530.20 km2, with a total growth rate of 264.58%. With the expansion of the built-up area, the vegetation coverage decreased by 16.61% (1652.56 km2); (3) the land surface temperature shows an increment trend in the new town yet a decrement trend in the old town due to the change in vegetation coverage and the decentration of administration centers; further analysis shows that the population and the social economy are two driving factors for land cover changes. It is worth noting that both the area and coverage of vegetation in the old town and water body area in Hefei increased significantly, although the fast urbanization inevitably caused a decrement in vegetation and water area in the whole city, indicating both the high-speed economic development and improvement in green surfaces simultaneously experienced in Hefei from 2000 to 2020.

1. Introduction

Land Use/Land Cover (LULC) change is the most basic and prominent feature to describe the impact of human activities on land surface change and plays an increasingly important role in the study of regional and global change in the environment and society [1]. Long-term satellite images are essential to help us understand land cover dynamics and urban expansion [2]. Recently, with an increment in massive remote sensing data, a great challenge in the storage and computing capacity of the computer causes extensive concern [3]. The emergence of a free remote sensing image data processing cloud platform, Google Earth Engine (GEE), provides a new method for geospatial analysis [4,5]. Conventional and time-consuming satellite image processes can be avoided, which greatly improves the efficiency of LULC monitoring, especially for long time series [6,7]. The first significant work on GEE was published in 2013, using more than 650,000 Landsat 7 satellite images to draw a global forest change map with 30 m resolution, from 2000 to 2012, with the support of the GEE platform [8]. Gorelick et al. (2017) provided the first comprehensive introduction of GEE [3]. During the past ten years, the GEE platform has played an essential role in monitoring the spatio–temporal variation in land cover and urban expansion (forest [8] , crop [9], grass [10], and wetland [11]), hazard (earthquake [12], flood [13], and drought [14]), and global change (sea-ice thickness [15], land surface temperature [16], carbon cycling [17]) at regional or even global scales.
It is widely accepted that the massive land in suburbs could be efficiently occupied and used during the urban expansion process and the high-speed economic development, which inevitably lead to the rapid changes in its land cover types, e.g., an increment in artificial features (i.e., built-up) and decrement in natural objects (e.g., vegetation and water) [18,19]. It is a crucial topic, how to maintain a balance between environment protection (e.g., preservation of green space) and rapid economic development and urban expansion [20]. High-speed economic development and urbanization occurred in China in the past thirty years under the government’s policy of reform and opening up, and studies of LCC variation have been widely reported in the literature [21,22,23,24]. Yet, most studies focus on the metropolis or developed cities (e.g., Beijing [25], Shanghai [24,26], Guangzhou [18], and Nanjing [19,27] et al.) and rarely on underdeveloped but fast-growing cities. In addition, most previous studies mainly focus on global variations but seldom on local characteristics (especially for the core area) and how the preservation of green space changes during the process of high-speed rapid economic development.
Since entering the new century, the total gross domestic product (GDP) of Hefei has rapidly increased from CNY 32.5 to 1004.5 billion (increased by nearly 30-times), and its GDP growth rate ranks first in China [28]. In this regard, Hefei is the most active city in China and has been the fastest-growing city economy in China in the past twenty years. The high-speed growth of the economy inevitably leads to urban expansion and dramatic changes in land cover change (LCC), especially for vegetation, built-up area, and water bodies. Therefore, it is of great importance to study the changes in LCC in the fastest-growing city in China. In this study, our main objectives are to (1) analyze the spatio-temporal variations in LCC based on multiple satellite images in the city with the fastest economic growth in China from 2000 to 2020; (2) find how the balance between fast economic development and natural objects (e.g., vegetation and water) changed, and whether there has been a massive loss of green space under the high-speed development in the past twenty years. Specifically, with the support of massive data resources from the GEE platform and interactive big data computing services, multi-year sample datasets are constructed. Thus, 716 Landsat 5 TM, Landsat 7 ETM+, and Landsat 8 OLI images are collected and processed, long time-series land cover in Hefei from 2000 to 2020 is mapped, and the spatio-temporal changes in mainland features are studied and analyzed from global and local viewpoints, respectively. Finally, the influence of social and economic driving factors on the change in land cover types in the study area is discussed.

2. Materials and Methods

2.1. Study Area

Hefei is the provincial capital of Anhui Province, China (spatial location is 30°57′–32°32′N, 116°41′–117°58′E). The average altitude is 20–40 m, and the terrain is mainly composed of plains and hills. It has nine counties and administrative districts, with a total area of about 11,445.1 km2 (Figure 1).
Hefei GDP increased from CNY 32.5 billion to CNY 1004.5 billion and the population increased from 4.3818 million to 7.7044 million from 2000 to 2020. The GDP and growth rates of major cities in China in 2000 and 2020 are shown in Table 1. From Table 1, Hefei GDP growth rate ranked first among all cities in China from 2000 to 2020. During the past 20 years, the government of China has paid more and more attention to the development of Hefei. Hefei has been defined as one of China’s four major science and education bases since 2002, one of three sub-central cities of the Yangtze River Delta urban agglomeration since 2014, and one of four comprehensive national science centers of China since 2017. The above great progress and performance of Hefei were closely related to the government policies and its rapid economic development. In this regard, Hefei is a significant representation of fast-growing cities in China from 2000 to 2020.

2.2. Data Acquisition

Data Sources

The series of Landsat satellites have provided continuous and relatively high spatial resolution remote sensing images for the past few decades [29]. All available Landsat Collection 1 Tier 1 top-of-atmosphere (TOA) image products from 2000 to 2020, with a step of 5 years, are collocated, including Landsat 5 TM, Landsat 7 ETM+, and Landsat 8 OLI. The other datasets include Shuttle Radar Topography Mission version (SRTM), MODIS land surface temperature (LST) products, nighttime satellite images: Defense Meteorological Satellite Program-Operational Linescan System (DMSP-OLS) [30], nighttime satellite images: National Polar-orbiting Partnership Visible Infrared Imaging Radiometer Suite (NPP-VIIRS) [31]. The topographic feature was used from SRTM, and the textural features and brightness were used from the nighttime images. MODIS land cover products (MCD12Q1), global surface coverage data (GlobeLand30) [32], and Global Land Cover with Fine Classification System (GLC_FCS30) [33,34,35] serve as comparisons. In addition, Hefei administrative division vector data were collected on the National Geomatics Center of China website (http://ngcc.cn/ngcc/ (accessed on 1 January 2020)). Hefei’s statistical yearbook was analyzed to explore the main factors driving changes in land cover types on the Hefei Municipal Bureau of Statistics website (http://tjj.hefei.gov.cn/ (accessed on 1 January 2020)). Detailed datasets in this study are listed in Table 2.
The detailed image processing steps of mapping LCC are summarized in Figure 2. In this study, the images with high quality (i.e., cloudless and non-snow) were selected. Based on the availability of filtered images, five image collections were created with each image collection. The corresponding time nodes were 2000, 2006, 2011, 2016, and 2020, respectively. Each time node, together with the previous year and the following year (in 2020, only the previous and current year’s images were collected), constituted a three-year period; detailed information is listed in Table 3.

2.3. Methodology

2.3.1. Technical Process

The technical process consisted of four main steps:
  • With the support of the GEE platform, the cloud cover function was used to process the multi-year image datasets to obtain the composite image data of TOA products without cloud, snow, and shadow covering in the five periods from 2000 to 2020 (1999–2001, 2005–2007, 2010–2012, 2015–2017, and 2019–2020).
  • The sample points were collected on GEE, filtered, and corrected with Google Earth high-resolution images. The training and validation samples were carefully deployed according to the “complete consistency” and “temporal stability” principles. The land cover attributes were extracted for the samples.
  • Based on the Landsat series satellite images, a variety of data and normalized difference index were used as input parameters of the RF algorithm classifier. Then, the training samples were used to produce the land use classification result of the study area in the five periods.
  • After obtaining the land use classification result of each period, the changes in various land features and the driving factors of LCC were analyzed.
The overall technical processes are shown in Figure 2.

2.3.2. Training Sample Point Selection

Supervised classification usually requires a certain number of training samples and verification samples. Typically, traditional studies often use manual visual interpretation to obtain points. For a region with a large area and large geographical diversity, this method presents considerable difficulties. This study uses a method to obtain high-precision sample points to reduce the workload. The method includes two parts: (1) online extraction and (2) offline inspection. The steps are described as follows:
  • According to the research interval (2000–2020), we selected the land cover products (e.g., GlobeLand30 and MCD12Q1.006) for stratified sampling based on the GEE platform. The remap function on GEE platform was used to remap the two images to the required feature types, and the stratified sample function was used to sample the images hierarchically.
  • In the selected sample points, some error sample points appear in the sample data selected online affected by individual error pixels. Therefore, we need to further improve the sample data through the offline Google Earth platform test. In the process of classification, 70% of the collected sample points were used as training sample points, and 30% were used as verification sample points to verify the accuracy of classification. The numbers of selected images and sample points are shown in Figure 3.

2.3.3. Study Method

Over the past few decades, many approaches for land cover classification algorithms have been proposed, including pixel-based, sub-pixel-based, object-oriented algorithms, random forest (RF), and artificial neural networks (ANNs) [36,37,38,39]. Among these algorithms, RF is a relatively novel machine learning algorithm and is an excellent classifier that uses multiple trees to train and predict samples [40].
To improve the classification accuracy, a variety of feature variables, e.g., multispectral bands, SRTM, DMSP/OLS, NPP/VIIRS, and spectral characteristic index, are selected as the input variables of the RF algorithm. Spectral characteristic indexes include normalized difference vegetation index (NDVI) [41], normalized difference water body index (NDWI) [42], modified normalized difference water body index (MNDWI) [43], and normalized difference building index (NDBI) [44]. The above operations can improve the classification accuracy of vegetation, water body, and built-up area and highlight the feature information [45]. The calculation formulas of each index are as follows:
NDVI = ( ρ NIR − ρ red ) / ( ρ NIR + ρ red )
NDWI = ( ρ green − ρ NIR ) / ( ρ green + ρ NIR )
MNDWI = ( ρ green − ρ SWIR 1 ) / ( ρ green + ρ SWIR 1 )
NDBI = ( ρ SWIR 1 − ρ NIR ) / ( ρ SWIR 1 + ρ NIR )
where ρgreen, ρred, ρNIR, and ρSWIR1 are the second, third, fourth, and fifth band reflectance of Landsat 5 and Landsat 7, respectively, and the third, fourth, fifth, and sixth band reflectance of Landsat 8, respectively.
The fractional vegetation cover (FVC) of the study area is retrieved based on NDVI using the pixel dichotomy model [46]. NDVI is a kind of index that reflects the condition of surface vegetation based on spectral information. NDVI spectral information of a pixel can be expressed as vegetation part and non-vegetation part. Therefore, combined with the pixel dichotomy model, the vegetation coverage estimation model based on NDVI is established according to [47,48]:
FVC = ( NDVI − NDVI soil ) / ( NDVIveg − NDVI soil )
where NDVIveg is the NDVI value of pure vegetation pixel and NDVIsoil is the NDVI value of the pure non-vegetation pixel. In this study, the maximum and minimum values of the given information interval are used to replace the values of NDVIveg and NDVIsoil. According to the statistical histogram, the NDVI value corresponding to the cumulative frequency of 5% is selected as NDVIsoil, and the NDVI value corresponding to the cumulative frequency of 95% is selected as NDVIveg [48].
The change in vegetation coverage was analyzed and calculated in the Hefei area in the most recent 20 years. The expansion of the city inevitably led to the change in LST. In this study, the MOD11A1 V6.1 product (see Table 2) is used to analyze the variation in LST in Hefei from 2000 to 2020. MOD11A1 V6.1 product is the 3.3 version of MODIS LST, including two products: LST and emissivity. The LST algorithms, including generalized split window algorithm and day–night LST algorithm, are developed based on the global land surface heat flux databases. The algorithm has been validated and widely applied in studies of LST at regional and global scales [46]. The night light information is a reflection of human economic activities and can be detected by the night satellite sensor in the visible and near-infrared bands. Generally, the intensity of human economic activities in the city is larger than that in rural areas [49]. Therefore, the satellite night images provide excellent information to classify the city boundary and to analyze the variation in urban expansion [46].

3. Results

3.1. Accuracy Assessment

Accuracy assessment is the most important aspect for any satellite image classification. Here, the four classification accuracies are shown in Table 4: the overall accuracy (OA), Kappa coefficient, user accuracy (UA), and producer accuracy (PA). A confusion matrix is calculated in five periods: 2000, 2006, 2011, 2016, and 2020.
From Table 4, the OA of land cover classification is larger than 91%, and the kappa coefficient is larger than 0.85. In all the classification features, the classification accuracy of surface water is higher (its UA and PA are all larger than 95%). The two kinds of water-normalized difference index have a greater impact on the classification results. Compared with the surface water, the classification accuracy of the built-up area and the barren area is relatively low. It is difficult to distinguish the built-up area and barren area because they have similar spectral characteristics. Although the PA of the built-up area and the barren area is low, the OA of classification results in each period is high, indicating that the four land cover types have strong consistency with the corresponding validation dataset. For the classification accuracy of vegetation features, the UA and PA are larger than 90%. The accuracy of classified images in different periods is enough to evaluate the regional land cover pattern and to analyze the land use change. A comparison of different land cover products (GLC_ FCS30 and GlobeLand30 product) with results in this study in 2020 also show the reliability of the classification accuracy in this study (Figure 4).

3.2. Spatio–Temporal Variations in Land Cover

The land cover classification results in the five periods (1999–2001, 2005–2007, 2010–2012, 2015–2017, and 2019–2020) from 2000 to 2020 are produced using the RF algorithm on the GEE platform (Figure 5).
Generally, the urbanization of the Hefei area presents a multi-core, multi-center radiation structure model to the surrounding areas, and the built-up areas spread outward with different centers. The built-up area in the city gradually expanded from 419.72 km2 to 1530.20 km2 (increased by 1110.48 km2) with a growth rate of 90.46% from 2000 to 2006, and a rapid growth rate of 35.33% from 2011 to 2016. The total area of the built-up area increased by 1110.48 km2, with a growth rate of 264.58%. The water area was increasing with an increment of 415.14 km2 and a growth rate of 39.54% from 2000 to 2020. The area of bare land was also increasing, with a total area of 126.94 km2. With the rapid expansion of the built-up area, the total area of vegetation decreased by 16.61% (1652.56 km2). The vegetation area continued to decrease rapidly, with the fastest decrease being about 6.65% from 2011 to 2016.
The transfer matrix of LCC in Hefei from 2000 to 2020 was obtained to analyze the change and transformation of LULC using the superposition function in ArcGIS 10.2 software (ESRI, Beijing, China). The transformation matrix is shown in Table 5.

3.2.1. Spatio-Temporal Variations in Vegetation Coverage

Generally, according to the LCC matrix from 2000 to 2020, 81.3% of the vegetation area (8062.25 km2) remained unchanged, 4.3% of the vegetation area (426.97 km2) was converted to the water body, and 12.9% of the vegetation area (1279.92 km2) was converted to the built-up area, which was also the largest part of the vegetation area converted to other land features. It means that the construction land area continues to increase and the urbanization process of Hefei is significant. Further, 1.4% of the vegetation area (142.89 km2) was converted to bare land, which also led to the conversion of part of the vegetation area to most of the bare land. It warned us to pay attention to environmental protection. On the whole, due to the rapid development of the economy and urbanization in Hefei, the total vegetation area decreased rapidly from 2000 to 2020 (about 18.7% of the vegetation area).
For further analysis, spatio-temporal variation in vegetation coverage from the local viewpoint, the boundaries of Hefei in 2000 and 2020 need to be differentiated. The boundaries of the built-up area of the city in 2000 and 2020 were plotted with the method in [50], respectively (as shown in the left and middle of Figure 6). The expansion of the city is shown clearly in Figure 6. The boundary of Hefei in 2000 can be deemed as an old town of Hefei (area in the dotted polygon in Figure 6) in 2020, and the expansion of the city from 2000 to 2020 (area between the dotted polygon and the solid polygon) can be deemed as the suburb and surrounding rural areas of Hefei in 2000, according to [51]. After plotting the two polygons, the spatio-temporal variation in FVC shows a clear pattern: 67.9% of the FVC change value in the area between the dotted polygon and the solid polygon is less than 0, indicating that the FVCs in these areas show a obvious downward trend. It is due to the construction of three new districts: a Binhu New District, a high-tech industrial development District, and a national-level economic development district by the Hefei municipal government from 2000 to 2020 (Figure 1 and Figure 6). It is worth noting that 66.1% of the FVC change value in the dotted polygon in Figure 6 is greater than 0, indicating that the FVC in the old town of Hefei shows an obviously upward trend. It shows that the ecological environment in the old town is significantly recovering and improving (see the right plot in Figure 6). It means that both rapid development of the economy and improvements in environmental protection have been experienced in Hefei during the past twenty years.

3.2.2. Spatio-Temporal Variation in Water Body

To evaluate the change in water bodies in the Hefei area from 2000 to 2020, the water bodies obtained via ground feature classification in 2000 were compared with those in 2020. Figure 7 shows the change in surface water spatial distribution in the Hefei area from 2000 to 2020.
Chaohu Lake occupied most of the surface water system in the study area, and its area change was not obvious from 2000 to 2020. However, Figure 7 showed that there were two obviously increasing areas. A new water body in the north of the central urban area of Hefei, which was related to the completion of the Dafangying reservoir in 2002, was constructed to increase water supply volume to meet the expansion of the city and accretion of the population. The other was a water body called Huangpi Lake in the south of Hefei, and the area of the water body nearby increased significantly from 2000 to 2020. According to the LCC matrix from 2000 to 2020, 94.9% of the water body (988.98 km2) remained unchanged; 3.1% of the water body (32.24 km2) was converted to vegetation; and 1.8% of the water body (19.20 km2) was converted to the built-up area; the rest was converted to bare land. In general, compared with vegetation and built-up area, the water bodies increased, but not obviously, during the past 20 years.

3.2.3. Spatio-Temporal Variations in Built-Up Area

To evaluate the changes in the built-up area in Hefei from 2000 to 2020, the images of the built-up areas obtained by the classification of the land objects in 2000, 2006, 2011, 2016, and 2020 were analyzed, respectively. The results show the spatial distribution of the built-up area changes in Hefei from 2000 to 2020 (Figure 8).
The built-up area was expanding rapidly and substantially. In general, the main urban area of Hefei expanded obviously. The built-up area in other places, such as Changfeng County in the north, Chaohu City in the east, and Lujiang County in the south of Hefei, was also increasing. With the policy of building a Binhu new district by the Hefei municipal government, the built-up area of Hefei gradually extended to Chaohu Area in the southeast. According to the LCC, the original main urban area was developed based on the four central areas of Hefei, including Shushan District, Luyang District, Baohe District, and Yaohai District, and the surrounding towns were developing rapidly as the extension of the main urban area. However, the fastest growth period of the built-up area was from 2000 to 2006 and from 2011 to 2016 (379.69 km2 and 356.82 km2, respectively), mainly distributed in Baohe District and Shushan District. The main reason for the decrease in built-up area was the demolition of urban and rural houses according to urban and rural planning. The purpose was to allocate land and space resources rationally, including urban transformation and township transformation.

3.3. LCC of Driving Forces in Hefei

The annual average of LST and nighttime light intensity in 2000 and 2020 was processed from MOD11A1 V6 product and nighttime images in ArcGIS 10.2 Software (ESRI, Beijing, China), respectively. The thermal environment of a city can be detected by the LST. In addition, the nighttime satellite images can reflect, to some degree, the economic activities. From Figure 9, the area with high temperature increases from 2000 to 2020, mainly focusing on the new town because of the urban expansion. The city LST is obviously higher than that in other regions in a certain period, showing the heat island effect. There is a significant difference in the variation in LST between the old town and the new town from 2000 to 2020. The mean LST in the new town increased 0.66 (Table 6) because of the rapid urban expansion. Yet, for the old town, the mean LST decreased 1.01 C (Table 6). It is mainly due to two reasons: (1) the administration centers (both Hefei municipal government and the Anhui province government) were moved from the center of the old town in 2000 to the south of the city in 2020. The movements of administration centers lead to the transfer of the highest LST from 2000 to 2020: from the center of city to the southern of city. (2) The variation in vegetation coverage. The vegetation coverage shows an increment trend in the old town from 2000 to 2020, as mentioned above. The urban boundary is greatly similar to the city night light intensity area, except for the Chaohu lake (which is a permanent lake) located in the southeast of Hefei and two reservoirs located in the northwest of Hefei. The increment in night light (Figure 10) intensity is focused on the new town, especially in the south of Hefei, which is the same to LST.
The socio-economic factors affecting the change in the land cover are complicated; therefore, selecting the appropriate driving factors is the key to analyzing the change in land cover [52]. Human economic activities affect the change in land cover. The population is an important factor in LCC. The increment in population leads to the change in different land cover types. Hefei municipal government continues to increase the development of strategic emerging industries, promotes the adjustment of industrial structure, and accelerates the pace of transformation from 2000 to 2020. The increment in GDP leads to the aggravation of land cover type change. The total population and GDP at the end of the year in Hefei were selected to study the relationship between the change in built-up area and vegetation area. The change in four parameters is shown in Figure 11.
Figure 11 shows that the population of Hefei rapidly increased from 2006 to 2011. The reason is that the state adjusted the administrative division of Hefei during this period, resulting in a rapid increase in the population in this area. During the past 20 years, the national and provincial governments gradually approved the construction of Hefei Binhu New District, High-tech District, and Economic Development District, respectively; in addition, Hefei was identified as one of the National Comprehensive Science Centers. Hefei responded to the slogan of “Great Lake City, Innovation Highland” and accelerated the construction of the built-up area actively from 2000 to 2020.
The whole trend of the total population, GDP, and built-up area curve were similar to each other, and the regional change trend of vegetation area was opposite to them. To evaluate the correlation of these parameters quantitatively, the correlation coefficients between the four parameter curves were calculated, and the results are shown in Table 7. The results show that the absolute values of all correlation coefficients are greater than 0.85. Therefore, with the development of population and social economy, the construction land in Hefei expanded and the area of vegetation decreased, which made the land use types change greatly. Therefore, population growth and socio-economic development were two important factors to study the change in regional land cover structure.

4. Discussion

The LCC based on the satellite images has been widely studied in the past thirty years. Cities in China have been experienced high-speed development under the Chinese Government’s policy of reform and opening up in the past thirty years, especially in the past twenty years. It is worth noting that most previous studies focused on the metropolis, coastal, and developed cities, e.g., top 20 cities in GDP, Beijing, Shanghai, Guangzhou, and Nanjing, et al., yet seldom on the relatively small, inland, but fast developing cities. Hefei was out of the top 80 cities in GDP in China in 2000, which was a significantly underdeveloped city, yet speedily rose into the top 20 cities in GDP in China in 2020. Its GDP growth ranks first in all Chinese cities and Hefei is definitely deemed as a representative city and one of the most dynamic cities in economic development in China in the past twenty years. It is of great importance to study spatio-temporal variation in LCC and the balance between the high-speed economic development and environment protection for the city.
In addition, the fast urban expansion inevitably leads to a change in LCC in the suburbs and new town. The previous studies of LCC often reported variation in a whole city, yet seldom focused on the old town or core area. The density of the population and intensity of economic activity in the old town are often obviously higher than those in the suburb and new town. Although the variation in LCC in the old town is not as obvious as the new town, it is of significance to especially focus on the study of LCC in the old town. Therefore, we focused on the spatio-temporal variation in LCC in the city with the fastest economic growth in China from 2000 to 2020, from global and local viewpoints, respectively. The results in the study show that the spatial-temporal variations in land cover in the city from 2000 to 2020 are obvious. Yet, both the area and coverage of vegetation increased significantly in the old town of Hefei, although the vegetation in the whole city showed a decreasing trend, leading to the LST in the old town decreasing. The economic engagements in the old town also show a decreasing trend from the night light images, indicating the core of the city changed from the old town to new town. It is mainly due to the movement of administration centers (including the municipal government of the city and provincial government of Anhui) from the old town to the new town. As for the rapid improvements in urbanization, e.g., the increment in population and human activities, the urbanization issues would gradually appear in the old town, while urban development is limited by the land use in the old town. It is an effective way to solve the problem of movement of administration centers from the old town to the new town. Under fast urban development, this urbanization pattern is significant: (1) avoiding the serious urbanization issues (limited land and road, and over-concentrated human activities) in the old town; (2) great improvement in development of new town; (3) commendable enhancement in area and coverage of vegetation in the old town because of its decentration, maintaining the green space in the city. Therefore, both fast urbanization and improvements in the environment could simultaneously occur, and the relationship between economic development and environment protection is balanced. This is representative of urbanization patterns in China during the past twenty years.
The accuracy of land cover maps remains a great uncertainty [53]. In this study, based on the GEE platform, we used the rapid construction of sample datasets and massive and multisource remote sensing images. Spectral characteristic index, digital elevation dataset (SRTM), and nighttime light product serve as input variables used in the RF algorithm to achieve the rapid and autonomous interpretation of land cover results in 2000, 2006, 2011, 2016, and 2020. Although there are different characteristics among modern classification algorithms [54,55], the only RF algorithm was used in this study, while the results show that the method and platform have obvious advantages in large-scale land cover classification. The average accuracy of the OA reaches 93%, and the classification results are better than similar land cover products in detail. Compared with other data, SRTM and nighttime light data acquisition methods are suitable for large-scale land cover classification monitoring. Relevant studies on the impact of these two kinds of data on the accuracy of land cover classification have been reported by many researchers [56,57]. In this study, the impacts of SRTM, nighttime light product data, and four vegetation indexes on the accuracy of the land cover classification and the causes are discussed. The influence of these different datasets on the overall classification accuracy is shown in Table 8.
Table 8 shows that nearly all overall accuracies of the classification show slight downward trends after those datasets are removed. There is no obvious primary factor among those different kinds of datasets. Yet, after SRTM and nighttime light products are removed simultaneously, the overall accuracy of the classification obviously decreased. It indicates that both two data are useful for improvements in the classification accuracy, although the spatial resolution of the two data is relatively coarse (500 m and 90 m). In the case of multi-source data fusion, the classification accuracy is effectively improved. For future large-scale land cover classification monitoring studies, nighttime light and SRTM digital elevation data are suggested to be used as input parameters to improve the classification accuracy of the land cover.

5. Conclusions

As the fastest-growing city economy in China from 2000 to 2020, Hefei’s GDP growth rate ranks first among all cities in China. With the support of the powerful ability of the GEE platform in huge data processing and parallel computing, based on the constructed sample data and sufficient auxiliary data. Combined with the powerful RF algorithm, the 719 Landsat TM, ETM+, and OLI images are used to generate multi-year land cover products with strong timeliness, and the LCC in Hefei from 2000 to 2020 is studied based on the constructed sample data and sufficient auxiliary data with the powerful RF algorithm. We found that, generally, spatio-temporal variations in land cover in Hefei from 2000 to 2020 are obvious with excellent classification accuracies: the built-up area expanded from 419.72 km2 to 1530.20 km2, with a total growth rate of 264.58%, and the vegetation coverage decreased by 16.61% (1652.56 km2); locally, both the area and coverage of vegetation in the old town of Hefei increased significantly, although the high-speed economic development and urbanization inevitably caused a decrement in vegetation and water area in the suburb, and the fast economic growth led to the LST increasing in the new town yet decreasing in the old town due to the variation in vegetation coverage and the decentration of administration centers; the population and the social economy are two driving factors for the variation in LCC. Both the high-speed economic development and environmental improvement simultaneously occurred in Hefei from 2000 to 2020.

Author Contributions

Conceptualization, J.G. and L.T.; methodology, L.X.; software, L.X. and Y.W.; validation, L.X.; writing—original draft preparation, J.G. and L.X.; writing—review and editing, J.G., L.X. and L.T.; visualization, L.X. and Y.W.; supervision, J.G. and L.T.; project administration, J.G. and L.T. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the National Natural Science Foundation of China, grant number 41801234 and 41701383; Anhui Provincial Natural Science Foundation, grant number 2208085MD90; Fundamental Research Funds for the Central Universities under Grant JZ2022HGTB0253.

Data Availability Statement

The data that we used in this study can be requested by contacting the corresponding author.

Acknowledgments

We thank the Google Earth Engine platform and developers for their support. We also thank the journal editor and the anonymous reviewers for their useful comments and great efforts on this paper.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Cristina, G.; Joanne, C.W.; Michael, A.W. Optical remotely sensed time series data for land cover classification: A review. ISPRS-J. Photogramm. Remote Sens. 2016, 116, 55–72. [Google Scholar]
  2. Kavitha, A.V.; Srikrishna, A.; Satyanarayana, C. A Review on Detection of Land Use and Land Cover from an Optical Remote Sensing Image. IOP Conf. Ser. Mater. Sci. Eng. 2021, 1074, 12002. [Google Scholar] [CrossRef]
  3. Gorelick, N.; Hancher, M.; Dixon, M.; Ilyushchenko, S.; Thau, D.; Moore, R. Google Earth Engine: Planetary-scale geospatial analysis for everyone. Remote Sens. Environ. 2017, 202, 18–27. [Google Scholar] [CrossRef]
  4. Amani, M.; Ghorbanian, A.; Ahmadi, S.A.; Kakooei, M.; Brisco, B. Google Earth Engine Cloud Computing Platform for Remote Sensing Big Data Applications: A Comprehensive Review. Ieee J. Sel. Top. Appl. Earth Observ. Remote Sens. 2020, 13, 5326–5350. [Google Scholar] [CrossRef]
  5. Fu, D.J.; Xiao, H.; Su, F.Z.; Zhou, C.H.; Dong, J.W.; Zeng, Y.L.; Yan, K.; Li, S.W.; Wu, J.; Wu, W.Z.; et al. Remote sensing cloud computing platform development and Earth science application. J. Remote Sens. 2021, 25, 220–230. [Google Scholar]
  6. Xie, S.; Liu, L.; Zhang, X.; Yang, J.; Gao, Y. Automatic Land-Cover Mapping using Landsat Time-Series Data based on Google Earth Engine. Remote Sens. 2019, 11, 3023. [Google Scholar] [CrossRef] [Green Version]
  7. Kumar, L.; Mutanga, O. Google Earth Engine Applications Since Inception: Usage, Trends, and Potential. Remote Sens. 2018, 10, 1509. [Google Scholar] [CrossRef] [Green Version]
  8. Hansen, M.C.; Potapov, P.V.; Moore, R.; Hancher, M.; Turubanova, S.A.; Tyukavina, A.; Thau, D.; Stehman, S.V.; Goetz, S.J.; Loveland, T.R.; et al. High-Resolution Global Maps of 21st-Century Forest Cover Change. Science 2013, 342, 850–853. [Google Scholar] [CrossRef] [Green Version]
  9. Dong, J.; Xiao, X.; Menarguez, M.A.; Zhang, G.; Qin, Y.; Thau, D.; Biradar, C.; Iii, B.M. Mapping paddy rice planting area in northeastern Asia with Landsat 8 images, phenology-based algorithm and Google Earth Engine. Remote Sens. Environ. 2016. [Google Scholar] [CrossRef] [Green Version]
  10. Du, Z.; Yang, J.; Ou, C.; Zhang, T. Agricultural Land Abandonment and Retirement Mapping in the Northern China Crop-Pasture Band Using Temporal Consistency Check and Trajectory-Based Change Detection Approach. IEEE Trans. Geosci. Remote Sens. 2022, 60, 1–12. [Google Scholar] [CrossRef]
  11. Zhang, K.; Dong, X.; Liu, Z.; Gao, W.; Hu, Z.; Wu, G. Mapping Tidal Flats with Landsat 8 Images and Google Earth Engine: A Case Study of the China’s Eastern Coastal Zone circa 2015. Remote Sens. 2019, 11, 924. [Google Scholar] [CrossRef] [Green Version]
  12. Yang, W.; Qi, W.; Fang, J. Using Google Earth Engine to monitor co-seismic landslide recovery after the 2008 Wenchuan earthquake. Earth Surf. Dynam. Discuss. 2020. [Google Scholar] [CrossRef]
  13. DeVries, B.; Huang, C.; Armston, J.; Huang, W.; Jones, J.W.; Lang, M.W. Rapid and robust monitoring of flood events using Sentinel-1 and Landsat data on the Google Earth Engine. Remote Sens. Environ. 2020, 240, 111664. [Google Scholar] [CrossRef]
  14. Sazib, N.; Mladenova, I.; Bolten, J. Leveraging the Google Earth Engine for Drought Assessment Using Global Soil Moisture Data. Remote Sens. 2018, 10, 1265. [Google Scholar] [CrossRef] [Green Version]
  15. Shamshiri, R.; Eide, E.; Høyland, K.V. Spatio-temporal distribution of sea-ice thickness using a machine learning approach with Google Earth Engine and Sentinel-1 GRD data. Remote Sens. Environ. 2022, 270, 112851. [Google Scholar] [CrossRef]
  16. Qiao, Z.; Liu, L.; Qin, Y.; Xu, X.; Wang, B.; Liu, Z. The Impact of Urban Renewal on Land Surface Temperature Changes: A Case Study in the Main City of Guangzhou, China. Remote Sens. 2020, 12, 794. [Google Scholar] [CrossRef] [Green Version]
  17. Zhang, Y.; Hu, Q.; Zou, F. Spatio-Temporal Changes of Vegetation Net Primary Productivity and Its Driving Factors on the Qinghai-Tibetan Plateau from 2001 to 2017. Remote Sens. 2021, 13, 1566. [Google Scholar] [CrossRef]
  18. Liu, G.; Li, J.; Nie, P. Tracking the history of urban expansion in Guangzhou (China) during 1665–2017: Evidence from historical maps and remote sensing images. Land Use Pol. 2022, 112, 105773. [Google Scholar] [CrossRef]
  19. Tu, L.; Qin, Z.; Li, W.; Geng, J.; Yang, L.; Zhao, S.; Zhan, W.; Wang, F. Surface urban heat island effect and its relationship with urban expansion in Nanjing, China. J. Appl. Remote Sens. 2016, 10, 26037. [Google Scholar] [CrossRef] [Green Version]
  20. Nor, A.; Corstanje, R.; Harris, J.A.; Brewer, T. Impact of rapid urban expansion on green space structure. Ecol. Indic. 2017, 81, 274–284. [Google Scholar] [CrossRef]
  21. Zhang, Z.; Ling, Y.I.; Shunguang, H.U.; Liu, B.; Liu, F.; Zhao, X.; Wang, X.; Shi, L.; Jinyong, X.U.; Sisi, Y.U. Urban Expansion in China Based on Remote Sensing Technology:A Review. Chin. Geogr. Sci. 2018, 28, 727–743. [Google Scholar] [CrossRef] [Green Version]
  22. Gong, J.; Hu, Z.; Chen, W.; Liu, Y.; Wang, J. Urban expansion dynamics and modes in metropolitan Guangzhou, China. Land Use Pol. 2018, 72, 100–109. [Google Scholar] [CrossRef]
  23. Sun, Z.; Xu, R.; Du, W.; Wang, L.; Lu, D. High-Resolution Urban Land Mapping in China from Sentinel 1A/2 Imagery Based on Google Earth Engine. Remote Sens. 2019, 11, 752. [Google Scholar] [CrossRef] [Green Version]
  24. Sengupta, D.; Chen, R.S.; Meadows, M.E.; Choi, Y.R.; Banerjee, A.; Zilong, X. Mapping Trajectories of Coastal Land Reclamation in Nine Deltaic Megacities using Google Earth Engine. Remote Sens. 2019, 11, 2621. [Google Scholar] [CrossRef] [Green Version]
  25. Huang, H.; Chen, Y.; Clinton, N.; Wang, J.; Wang, X.; Liu, C.; Gong, P.; Yang, J.; Bai, Y.; Zheng, Y.; et al. Mapping major land cover dynamics in Beijing using all Landsat images in Google Earth Engine. Remote Sens. Environ. 2017, 202, 166–176. [Google Scholar] [CrossRef]
  26. Zhao, S. Rapid urban expansion and its ecological consequences in Shanghai, China. Front. Ecol. Environ. 2006, 4, 341–346. [Google Scholar] [CrossRef] [Green Version]
  27. Zhang, D.; Zhang, L. Land Cover Change in the Central Region of the Lower Yangtze River Based on Landsat Imagery and the Google Earth Engine: A Case Study in Nanjing, China. Sensors 2020, 20, 2091. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  28. Yao, Y.L.; Liu, P.X.; Chen, L.L.; Lu, L.M. Remote sensing analysis of urban expansion in Hefei in recent 30 years. Econ. Geogr. 2013, 33, 67–74. [Google Scholar]
  29. Wolanin, A.; Camps-Valls, G.; Gómez-Chova, L.; Mateo-García, G.; Christiaan, V.; Zhang, Y.; Guanter, L. Estimating crop primary productivity with Sentinel-2 and Landsat 8 using machine learning methods trained with radiative transfer simulations. Remote Sens. Environ. 2019, 225, 441–457. [Google Scholar] [CrossRef]
  30. Amaral, S.; Monteiro, A.; Camara, G.; Quintanilha, J.A. DMSP/OLS night-time light imagery for urban population estimates in the Brazilian Amazon. Int. J. Remote Sens. 2006, 27, 855–870. [Google Scholar] [CrossRef]
  31. Baugh, K.; Hsu, F.C.; Elvidge, C.D.; Zhizhin, M. Nighttime Lights Compositing Using the VIIRS Day-Night Band: Preliminary Results. Proc. Asia-Pac. Adv. Netw. 2013, 35, 70. [Google Scholar] [CrossRef] [Green Version]
  32. Chen, J.; Ban, Y.; Li, S. China: Open access to Earth land-cover map. Nature 2015, 514, 434. [Google Scholar]
  33. Zhang, X.; Liu, L.; Wu, C.; Chen, X.; Zhang, B. Development of a global 30-m impervious surface map using multi-source and multi-temporal remote sensing datasets with the Google Earth Engine platform. Earth Syst. Sci. Data 2020, 12, 1625–1648. [Google Scholar] [CrossRef]
  34. Zhang, X.; Liu, L.; Chen, X.; Xie, S.; Gao, Y. Fine Land-Cover Mapping in China Using Landsat Datacube and an Operational SPECLib-Based Approach. Remote Sens. 2019, 11, 1056. [Google Scholar] [CrossRef] [Green Version]
  35. Zhang, X.; Liu, L.; Chen, X.; Gao, Y.; Mi, J. GLC_FCS30: Global land-cover product with fine classification system at 30 m using time-series Landsat imagery. Earth Syst. Sci. Data 2020, 13, 2753–2776. [Google Scholar] [CrossRef]
  36. Hu, Y.; Dong, Y. An automatic approach for land-change detection and land updates based on integrated NDVI timing analysis and the CVAPS method with GEE support. ISPRS-J. Photogramm. Remote Sens. 2018, 146, 347–359. [Google Scholar] [CrossRef]
  37. Kussul, N.; Lavreniuk, M.; Skakun, S.; Shelestov, A. Deep Learning Classification of Land Cover and Crop Types Using Remote Sensing Data. IEEE Geosci. Remote Sens. Lett. 2017, 14, 778–782. [Google Scholar] [CrossRef]
  38. Wang, Q.; Blackburn, G.A.; Onojeghuo, A.O.; Dash, J.; Atkinson, P.M. Fusion of Landsat 8 OLI and Sentinel-2 MSI Data. IEEE Trans. Geosci. Remote Sens. 2017, 55, 3885–3899. [Google Scholar] [CrossRef] [Green Version]
  39. Qihao, W. Remote sensing of impervious surfaces in the urban areas: Requirements, methods, and trends. Remote Sens. Environ. 2011, 117, 34–49. [Google Scholar]
  40. Belgiu, M.; Drăguţ, L. Random forest in remote sensing: A review of applications and future directions. ISPRS-J. Photogramm. Remote Sens. 2016, 114, 24–31. [Google Scholar] [CrossRef]
  41. Tucker, C.J. Red and photographic infrared linear combinations for monitoring vegetation. Remote Sens. Environ. 1979, 8, 127–150. [Google Scholar] [CrossRef] [Green Version]
  42. McFEETERS, S.K. The use of the Normalized Difference Water Index (NDWI) in the delineation of open water features. Int. J. Remote Sens. 1996, 17, 1425–1432. [Google Scholar] [CrossRef]
  43. Xu, H. A Study on Information Extraction of Water Body with the Modified Normalized Difference Water Index(MNDWI). J. Remote Sens. 2005, 9, 589–595. [Google Scholar]
  44. Zha, Y.; Gao, J.; Ni, S. Use of normalized difference built-up index in automatically mapping urban areas from TM imagery. Int. J. Remote Sens. 2003, 24, 583–594. [Google Scholar] [CrossRef]
  45. Pekel, J.F.; Cottam, A.; Gorelick, N.; Belward, A.S. High-resolution mapping of global surface water and its long-term changes. Nature 2016, 540, 418–422. [Google Scholar] [CrossRef] [PubMed]
  46. Chen, J.; Chen, Y.H.; He, C.Y.; Shi, P.J. Sub-pixel Model for Vegetation Fraction Estimation based on Land Cover Classification. J. Remote Sens. 2001, 5, 416–422. [Google Scholar]
  47. Jacquin, A.; Sheeren, D.; Lacombe, J.P. Vegetation cover degradation assessment in Madagascar savanna based on trend analysis of MODIS NDVI time series. Int. J. Appl. Earth Obs. Geoinf. 2010, 12, S3–S10. [Google Scholar] [CrossRef] [Green Version]
  48. Gao, L.; Wang, X.; Johnson, B.A.; Tian, Q.; Wang, Y.; Verrelst, J.; Mu, X.; Gu, X. Remote sensing algorithms for estimation of fractional vegetation cover using pure vegetation index values: A review. ISPRS-J. Photogramm. Remote Sens. 2020, 159, 364–377. [Google Scholar] [CrossRef]
  49. Tu, L.; Qin, Z.; Yang, L.; Wang, F.; Geng, J.; Zhao, S. Identifying the Lambertian Property of Ground Surfaces in the Thermal Infrared Region via Field Experiments. Remote Sens. 2017, 9, 481. [Google Scholar] [CrossRef] [Green Version]
  50. Yi, S.; Li, L. Analysis on the Characteristics of Central Urban Area’s Spatial Morphology and Its Cause of Formation in Big Cities. In Advances in Intelligent Systems and Computing; Springer: Singapore, 2017; Volume 502, pp. 817–825. [Google Scholar]
  51. Liu, Z.; Zhan, W.; Bechtel, B.; Voogt, J.; Lai, J.; Chakraborty, T.; Wang, Z.; Li, M.; Huang, F.; Lee, X. Surface warming in global cities is substantially more rapid than in rural background areas. Commun. Earth Environ. 2022, 3, 219. [Google Scholar] [CrossRef]
  52. Zhou, S.; Zhu, Q.; Guo, Q. Variation of land-use structure in Nanjing over the last decade. Soils 2005, 37, 394–399. [Google Scholar]
  53. Thinh, T.V.; Duong, P.C.; Nasahara, K.N.; Tadono, T. How Does Land Use/Land Cover Map’s Accuracy Depend on Number of Classification Classes? Sola 2019, 15, 28–31. [Google Scholar] [CrossRef] [Green Version]
  54. Dabija, A.; Kluczek, M.; Zagajewski, B.; Raczko, E.; Kycko, M.; Al-Sulttani, A.H.; Tardà, A.; Pineda, L.; Corbera, J. Comparison of Support Vector Machines and Random Forests for Corine Land Cover Mapping. Remote Sens. 2021, 13, 777. [Google Scholar] [CrossRef]
  55. Olofsson, P.; Foody, G.M.; Herold, M.; Stehman, S.V.; Woodcock, C.E.; Wulder, M.A. Good practices for estimating area and assessing accuracy of land change. Remote Sens. Environ. 2014, 148, 42–57. [Google Scholar] [CrossRef]
  56. Li, X.; Li, D.; Xu, H.; Wu, C. Intercalibration between DMSP/OLS and VIIRS night-time light images to evaluate city light dynamics of Syria’s major human settlement during Syrian Civil War. Int. J. Remote Sens. 2017, 38, 5934–5951. [Google Scholar] [CrossRef]
  57. Li, X.; Zhou, Y. Urban mapping using DMSP/OLS stable night-time light: A review. Int. J. Remote Sens. 2017, 38, 6030–6046. [Google Scholar] [CrossRef]
Remotesensing 15 01604 g001 550
Figure 1. Spatial position of the study area.
Figure 1. Spatial position of the study area.
Remotesensing 15 01604 g001
Remotesensing 15 01604 g002 550
Figure 2. Flowchart of land cover mapping and mechanism analysis. Land cover mapping includes sample selection, basic data, and RF classification.
Figure 2. Flowchart of land cover mapping and mechanism analysis. Land cover mapping includes sample selection, basic data, and RF classification.
Remotesensing 15 01604 g002
Remotesensing 15 01604 g003 550
Figure 3. The number of covered images and sample points in the study area.
Figure 3. The number of covered images and sample points in the study area.
Remotesensing 15 01604 g003
Remotesensing 15 01604 g004 550
Figure 4. Comparison of different land cover products in 2020. (A) Landsat annual TOA false-color composition in 2020 (RGB: B5, B4, B3), (B) GLC_ FCS30, (C) GlobeLand30, and (D) this product. Comparison of different kinds of land cover types from 1 to 8.
Figure 4. Comparison of different land cover products in 2020. (A) Landsat annual TOA false-color composition in 2020 (RGB: B5, B4, B3), (B) GLC_ FCS30, (C) GlobeLand30, and (D) this product. Comparison of different kinds of land cover types from 1 to 8.
Remotesensing 15 01604 g004
Remotesensing 15 01604 g005 550
Figure 5. Hefei land cover classification results in the five periods (1999–2001, 2005–2007, 2010–2012, 2015–2017, and 2019–2020) from 2000 to 2020.
Figure 5. Hefei land cover classification results in the five periods (1999–2001, 2005–2007, 2010–2012, 2015–2017, and 2019–2020) from 2000 to 2020.
Remotesensing 15 01604 g005
Remotesensing 15 01604 g006 550
Figure 6. Spatio-temporal distribution variation in FVC in Hefei from 2000 to 2020: FVC values in 2000, FVC values in 2020, and FVC changes between 2000 and 2020 from global and local viewpoints, respectively.
Figure 6. Spatio-temporal distribution variation in FVC in Hefei from 2000 to 2020: FVC values in 2000, FVC values in 2020, and FVC changes between 2000 and 2020 from global and local viewpoints, respectively.
Remotesensing 15 01604 g006
Remotesensing 15 01604 g007 550
Figure 7. Spatiotemporal distribution variation in water body in Hefei from 2000 to 2020.
Figure 7. Spatiotemporal distribution variation in water body in Hefei from 2000 to 2020.
Remotesensing 15 01604 g007
Remotesensing 15 01604 g008 550
Figure 8. Change and characteristics of built-up area in Hefei from 2000 to 2020.
Figure 8. Change and characteristics of built-up area in Hefei from 2000 to 2020.
Remotesensing 15 01604 g008
Remotesensing 15 01604 g009 550
Figure 9. Comparison of the annual mean LST in Hefei from 2000 to 2020.
Figure 9. Comparison of the annual mean LST in Hefei from 2000 to 2020.
Remotesensing 15 01604 g009
Remotesensing 15 01604 g010 550
Figure 10. Comparison of the night in Hefei from 2000 to 2020 from the global and local viewpoints, respectively.
Figure 10. Comparison of the night in Hefei from 2000 to 2020 from the global and local viewpoints, respectively.
Remotesensing 15 01604 g010
Remotesensing 15 01604 g011 550
Figure 11. Change chart of four parameters in Hefei area from 2000 to 2020.
Figure 11. Change chart of four parameters in Hefei area from 2000 to 2020.
Remotesensing 15 01604 g011
Table
Table 1. GDP growth rate among major cities in China from 2000 to 2020 (unit: ×108 CNY).
Table 1. GDP growth rate among major cities in China from 2000 to 2020 (unit: ×108 CNY).
City20002020Growth Rate (%)City20002020Growth Rate (%)
Hefei32510,0452991Yangzhou47260481181
Dongguan49096501869Xiamen50263841172
Changsha65612,1421751Hangzhou138316,1061065
Shenzhen166527,6701562Qingdao115012,400978
Zhengzhou73812,0031526Ningbo117612,408955
Chongqing159025,0021473Guangzhou237625,019953
Beijing247936,1021356Tianjin163914,083759
Xi’an68910,0201354Shanghai455138,700750
Nanjing102114,8171351Dalian11117030533
Chengdu131317,7161249Shijiazhuang10035935492
Suzhou154120,1701201Shenyang11196571487
Wuhan120715,6161194Ha’erbin10035183417
Table
Table 2. Datasets used in this study.
Table 2. Datasets used in this study.
DataYearSpatial ResolutionTemporal ResolutionDescriptionData Sources
Landsat *2000–
2020		30 m16 daysMulti-bands for classificationhttp://landsat.usgs.gov/ (accessed on 1 January 2020)
MCD12Q1.006 *2001–
2019		500 m1 yearLand cover products for comparisonhttps://lpdaac.usgs.gov/products/mcd12q1v006/ (accessed on 1 January 2020)
SRTM3 *200030 m-DEM for classificationhttp://www2.jpl.nasa.gov/srtm/ (accessed on 1 January 2020)
MOD11A12000–
2020		1 km1 dayLST for analysishttps://lpdaac.usgs.gov/products/mod11a1v061/ (accessed on 1 January 2020)
DMSP/OLS *2001–
2013		30 arc seconds1 yearNighttime satellite images for classification and analysishttps://ngdc.noaa.gov/eog/dmsp/ (accessed on 1 January 2020)
NPP/VIIRS *2012–
2020		15 arc seconds1 monthNighttime satellite images for classification and analysishttps://ngdc.noaa.gov/eog/viirs/ (accessed on 1 January 2020)
GlobeLand30202030 m-Land cover products for comparisonhttp://www.globeland30.com/ (accessed on 1 January 2020)
GLC_FCS30202030 m-Land cover products for comparisonhttp://data.casearth.cn/sdo/detail/5fbc7904819aec1ea2dd7061 (accessed on 1 January 2020)
Note that year(s) represents the temporal range of the datasets used; * represents data available online (https://earthengine.google.com (accessed on 1 January 2020)).
Table
Table 3. The selected Landsat image collections for the study area.
Table 3. The selected Landsat image collections for the study area.
Satellite SensorThree-Year PeriodDate FrameNumber
Landsat 51999–20011 March to 15 June and 16 July to 31 November153
Landsat 52005–20071 March to 15 June and 16 July to 31 November154
Landsat 72010–20121 March to 15 June and 16 July to 31 November151
Landsat 82015–20161 March to 15 June and 16 July to 31 November155
Landsat 82019–20201 March to 15 June and 16 July to 31 November106
Table
Table 4. Accuracy evaluation of land cover classification results in the five periods: 1999–2001, 2005–2007, 2010–2012, 2015–2017, and 2019–2020.
Table 4. Accuracy evaluation of land cover classification results in the five periods: 1999–2001, 2005–2007, 2010–2012, 2015–2017, and 2019–2020.
AccuracyYear PeriodBarrenBuilt-Up AreaWaterVegetation
Producer’s1999–20010.740.550.980.98
2005–20070.810.870.950.94
2010–20120.930.880.960.90
2015–20170.940.790.980.93
2019–20200.930.910.980.94
User’s1999–20010.890.840.980.91
2005–20070.930.900.960.90
2010–20120.820.9010.91
2015–20170.900.880.980.91
2019–20200.990.920.960.89
Overall1999–20010.92
2005–20070.92
2010–20120.91
2015–20170.92
2019–20200.94
Kappa1999–20010.85
2005–20070.87
2010–20120.88
2015–20170.89
2019–20200.92
Table
Table 5. The transfer matrix of land cover types in Hefei from 2000 to 2020 (unit: km2).
Table 5. The transfer matrix of land cover types in Hefei from 2000 to 2020 (unit: km2).
Land Cover Type2020
BarrenBuiltup_AreaWaterVegetationTotal Area
2000Barren9.322.340.364.1316.15
Builtup_area3.85286.978.38117.19416.39
Water1.8519.20988.9832.241042.27
Vegetation142.891279.92426.978062.259912.03
Total Area157.911588.431424.698215.8111,386.84
Table
Table 6. Variation in LST in the three different regions of Hefei from 2000 to 2020.
Table 6. Variation in LST in the three different regions of Hefei from 2000 to 2020.
NAMEMINMAXRANGEMEANSTD
Old town−3.503.036.54−1.011.34
New town−3.817.7411.550.661.41
Suburb−3.592.996.58−0.430.81
Table
Table 7. Correlation coefficient matrix of four parametric curves.
Table 7. Correlation coefficient matrix of four parametric curves.
ParameterTotal PopulationGDPBuilt-Up AreaVegetation Area
Total Population1
GDP0.8951
Built-up Area0.9220.9481
Vegetation Area−0.904−0.968−0.9951
Table
Table 8. Influence of several kinds of datasets on the overall accuracy of classification.
Table 8. Influence of several kinds of datasets on the overall accuracy of classification.
Year20002006201120162020
Original overall accuracy0.920.920.920.920.94
After removing nighttime light product 0.90 0.91 0.91 0.92 0.88
After removing SRTM 0.90 0.86 0.89 0.85 0.83
After removing the above two datasets 0.89 0.88 0.85 0.78 0.77
After removing NDVI 0.90 0.90 0.92 0.91 0.92
After removing NDWI 0.89 0.91 0.92 0.90 0.91
After removing MNDWI 0.90 0.91 0.91 0.92 0.92
After removing NDBI 0.88 0.91 0.91 0.92 0.91
Blue numbers show the decrements in overall accuracy.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Geng, J.; Xu, L.; Wang, Y.; Tu, L. Study of Land Cover Change in the City with the Fastest Economic Growth in China (Hefei) from 2000 to 2020 Based on Google Earth Engine Platform. Remote Sens. 2023, 15, 1604. https://doi.org/10.3390/rs15061604

AMA Style

Geng J, Xu L, Wang Y, Tu L. Study of Land Cover Change in the City with the Fastest Economic Growth in China (Hefei) from 2000 to 2020 Based on Google Earth Engine Platform. Remote Sensing. 2023; 15(6):1604. https://doi.org/10.3390/rs15061604

Chicago/Turabian Style

Geng, Jun, Lichen Xu, Yuping Wang, and Lili Tu. 2023. "Study of Land Cover Change in the City with the Fastest Economic Growth in China (Hefei) from 2000 to 2020 Based on Google Earth Engine Platform" Remote Sensing 15, no. 6: 1604. https://doi.org/10.3390/rs15061604

APA Style

Geng, J., Xu, L., Wang, Y., & Tu, L. (2023). Study of Land Cover Change in the City with the Fastest Economic Growth in China (Hefei) from 2000 to 2020 Based on Google Earth Engine Platform. Remote Sensing, 15(6), 1604. https://doi.org/10.3390/rs15061604

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop