- Email: [email protected]
Using land use change trajectories to quantify the effects of urbanization on urban heat island
Available online at www.sciencedirect.com
ScienceDirect Advances in Space Research 53 (2014) 463–473 www.elsevier.com/locate/asr
Using land use change trajectories to quantify the effects of urbanization on urban heat island Huihui Feng a,b,c,1, Xiaofeng Zhao a,b,⇑, Feng Chen a,b,2, Lichun Wu d,3 a
Key Lab of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, Xiamen 361021, China b Xiamen Key Lab of Urban Metabolism, Xiamen 361021, China c Nanjing Institute of Geography and Limnology, Chinese Academy of Sciences, Nanjing 210008, China d School of Resources and Safety Engineering, Central South University, Changsha 410083, China Received 19 June 2013; received in revised form 30 October 2013; accepted 14 November 2013 Available online 22 November 2013
Abstract This paper proposed a quantitative method of land use change trajectory, which means the succession among different land use types across time, to examine the effects of urbanization on an urban heat island (UHI). To accomplish this, multi-temporal images from Landsat 5 Thematic Mapper (TM) and Landsat 7 Enhanced Thematic Mapper Plus (ETM+) of Xiamen City in China from 1987 to 2007 were selected. First, the land use change trajectories were extracted based on the classified images from different years. Then the land surface temperatures (LST) were retrieved and the magnitudes of the UHI were evaluated using the UHI intensity (UHII) indicator. Finally, the indices of the contribution to UHI intensity (CUHII) were constructed and calculated to quantify the effects of each land use change trajectory on the UHI during urbanization. The results demonstrated that the land use change trajectories and CUHII are effective and useful in quantifying the effects of urbanization on UHI. In Xiamen City, a total of 2218 land use change trajectories were identified and 530 of them were the existing urban or urbanization trajectories. The UHII presents a trend of continuous increase from 0.83 °C in 1987 to 2.14 °C in 2007. With respect to the effects of urban growth on UHI, the contribution of existing urban area to UHI decreased during urbanization. Prior to 2007, the existing urban area of trajectory NO. 44444 had the most significant effect on UHI with the greatest CUHII, while the value has decreased from 55.00% in 1987 to 13.03% in 2007 because of the addition of new urbanized area. In 2007, the greatest CUHII was replaced by a trajectory from farmland to built-up area (NO. 22224) with the CUHII of 21.98%, followed by the existing urban area of trajectory NO. 44444 with the CUHII of 13.03%. These results provide not only a new methodology to assess the environmental effects of urbanization, but also decision-supports for the planning and management of cities. Ó 2013 COSPAR. Published by Elsevier Ltd. All rights reserved. Keywords: Urbanization; Land use change trajectory; Urban heat island; Contribution to UHI intensity (CUHII); Xiamen
1. Introduction Urbanization profoundly influences the ecosystem of an urban area. One of the best-documented examples of this,
⇑ Corresponding author. at: Key Lab of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, Xiamen 361021, China. Tel.: +86 592 6190672; fax: +86 592 6190977. E-mail addresses: [email protected] (H. Feng), [email protected] (X. Zhao), [email protected] (F. Chen), [email protected] (L. Wu). 1 Tel.: 86 25 86882167; fax: 86 25 57714759. 2 Tel.: Tel.: 86 592 6190672; fax: 86 592 6190977. 3 Tel.: 86 25 86882167; fax: 86 25 57714759.
is the urban heat island (UHI) effect, which refers to the higher temperatures found in urban areas compared with the surrounding rural area (Grimm et al., 2008). UHI not only influences the living environment (Konopacki and Akbari, 2002), but also increases energy consumption (Kolokotroni et al., 2012), and even harms human health (Changnon et al., 1996). Recently, a great deal of research has been carried out on the causes and impacts of UHIs, and their qualitative and quantitative characteristics have been documented (Rosenfeld et al., 1998; Rizwan et al., 2008; Imhoff et al., 2010). UHI was observed through air temperature at the early stage. For example, Howard (1818) firstly proposed the
0273-1177/$36.00 Ó 2013 COSPAR. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.asr.2013.11.028
464
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
phenomenon of UHI when he observed a higher the air temperature of urban area than the surrounding rural area of London. Jusuf et al. (2007) used the qualitative and quantitative methods to evaluate the influence of land use on the UHI from air temperature in Singapore. In recent decades, since assessment of land surface temperature (LST) from satellite data was first proposed by Rao (1972), this remote sensing method has been widely used to analyze UHI characteristics on a regional and global spatial scale (Gallo et al., 1993; Weng, 2001; Sobrino and Romaguera, 2004; Zhao et al., 2010b; Schwarz et al., 2011; Weng et al., 2011). LST is strongly related to the surface properties and much previous research focused on the relationship between the LST and land use in mono-temporal (Weng et al., 2004), across different seasons (Gallo et al., 1993; Li et al., 2011) or different years (Chen et al., 2006; Yuan and Bauer, 2007; Zhang et al., 2009). The results of this research effort indicated a positive correlation between the LST and impervious land use, and a negative correlation with the vegetation fraction. Furthermore, the effects of urbanization on the UHI were also analyzed. For example, Gallo et al. (1996) concluded that the transformation from rural to urban land use can impact the trend in temperature in a similar manner to that which would be expected under an enhanced greenhouse warming scenario. Amiri et al. (2009) examined the relationship between the temporal dynamics of LST and land use through the method of temperature vegetation index (TVX) space. They found that the LST trajectory in TVX space was reflected by the shift from the high vegetation – low LST fraction, to the low vegetation – high LST areas during the process of urbanization. Zhou et al. (2011) showed that the impact of urbanization on the UHI can be mitigated by balancing land use composition through optimizing land use configuration. However, the quantitative relationship between urbanization and its effects on the UHI is still ambiguous. The environmental response due to urbanization associated with different land use changes (e.g., from farmland to built-up areas and from forest to built-up areas.), is highly variable because of the different properties of each land use type (Houghton and Goodale, 2004; Pauleit et al., 2005). Therefore, it is essential to quantify the effects of different land use change during urbanization on the UHI, to understand fully the characteristics of the UHI. This study used land use change trajectory as a method to quantify the effects of different land use change types on the UHI during urbanization. A trajectory was defined as the succession areas of land use types, through this method, the urban areas could be divided into several parts according to the transformation of other land use types to built-up area (Lambin, 1997; Mertens and Lambin, 2000; Petit and Lambin, 2001). For example, Liu and Zhou (2005) simulated the process of urban growth through the land use change trajectory in Beijing. Zhou et al. (2008) analyzed the spatial pattern analysis of land cover change trajectories in Tarm Basin, northwest China. Wang et al.
(2012, 2013) illustrated the spatial patterns and the driving forces of the land cover change trajectories in the Xihe watershed of the Loess Plateau, China. The results of researches could support the methodology of this research. After the retrieval of LST and the calculation of the UHII, land use change trajectories were extracted and their CUHII calculated. In this way, the quantitative relationship between urbanization and its effects on the UHI was analyzed in the context of urbanization in Xiamen City. The results not only provide a useful tool for quantifying the effects of urbanization, but also support the planning and management of cities. 2. Study area and data pre-processing 2.1. Study area The city of Xiamen, which has experienced rapid urbanization in the past three decades, is selected as the study area for this research. It locates at 24°25’–24°55’ N and 117°53’–117°28’ E, and is situated on the southeast coast of Fujian Province and at the estuary of Jiulong River. It comprises Xiamen Island, Gulang Island, and the coastal part to the north of the Jiulong River (Fig. 1). It has an area of more than 1565 km2, a sea area of 390 km2, and in 2007, a population of 2.43 million with nearly 70% of which in urban areas (Xiamen Bureau of Statistics, 2008). In 1980, Xiamen Special Economic Zone was established. Since then, Xiamen has experienced rapid urbanization, which has led to significant environmental and ecological effects (Zhao et al., 2010a). Meanwhile, UHI in Xiamen has become more and more obvious (Xu and Chen, 2004; Zhao et al., 2010a,b; Huang and Huang, 2011). 2.2. Data resources and pre-processing Images acquired from the Landsat 5 Thematic Mapper (TM) and Landsat 7 Enhanced Thematic Mapper Plus (ETM+) sensors (Jan 17, 1987; Jan 15, 1992; Jan 12, 1997; Jan 2, 2002, and Jan 8, 2007) were used in this research. The multi-images were co-registered to the same coordinate system of UTM/WGS84 based on the image of 1987. Moreover, the images were then re-sampled using the nearest neighbor algorithm with a spatial resolution of 30 m for all bands. To remove the atmospheric influence, the MODTRAN4-based FLAASH module in the software of ENVI4.7 was adopted to correct atmospheric errors. Its parameters included the information about the sensor and scene, atmosphere and aerosol model, and the atmosphere correction model. For the study area of this paper, the scene center location, sensor altitude, sensor type, flight date and time were obtained from the Landsat TM/ ETM+ header book, the average elevation was 0.05 km, the atmospheric model and aerosol model were Mid-Latitude Summer (MLS) and urban, and the aerosol retrieval was adopted the K-T method. Finally, to analyze the land use change, all the individual images were classified
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
465
Fig. 1. The study area.
through the Support Vector Machine (SVM) method. According to current Chinese land use classification, five major land use types of forest, farmland, water, built-up area and barren land were extracted. These types were numbered from 1 to 5 for their digital forms, respectively. To assess the classification accuracy, high resolution historical images from Google Earth were used as the reference layers, and five hundred random sample points were generated for comparing classification data and reference data. 3. Methods 3.1. Land use change trajectory To generate the land use trajectories of urbanization, all the classified images are first integrated in GIS using raster format with ArcGISe 9.3 software, and then calculated through the overlay analysis, as in the method of Zhou et al. (2008). The equation is: Trac ¼ Cl1987 10000 þ Cl1992 1000 þ Cl1997 100 þ Cl2002 10 þ Cl2007
ð1Þ
where Cl1987, Cl1992, Cl1997, Cl2002, and Cl2007 are the land use classified data of the years 1987, 1992, 1997, 2002, and 2007, respectively. A trajectory code contains two meanings. Firstly, it describes the succession among different land cover types. Furthermore, it implies the time of succession happened. For example, the trajectory code NO. 12345 refers to the transformation process of forest ! farmland ! water ! built-up area ! barren land across the study period. To assess the effects of urbanization on the UHI, only the trajectories of the existing urban and urbanization areas are considered. Therefore, two types of trajectories are identified in this research: the first is the existing urban trajectory, which means the area already built-up prior to the study period (trajectory code: No. 44444); and the other is the urbanization trajectories, which means those areas urbanized from other land use types to build-up area during the study (trajectory code: No. ****4, where * is the land use type other than built-up area).
3.2. LST retrieval The LST retrieval algorithm used in this paper can be divided into three steps (Artis and Carnahan, 1982; Weng et al., 2004; Chen et al., 2006). The first step is to calculate the spectral radiance Lb from the digital number of Band 6: Lb ¼ Lmin þ ðLmax Lmin Þ DN 6 =255
ð2Þ
where Lmax, Lmin are the maximum and minimum thermal radiation energy, which can be obtained from the sensor handbook. DN6 is digital number of Band 6. For Landsat 5 TM, Lmax = 15.600W/(m2 sr), Lmin = 1.238W/(m2 sr). For Landsat 7 ETM+, Lmax,61 = 17.040W/(m2 sr), Lmin,61 = 0 and Lmax,62 = 12.650W/(m2 sr), Lmin,62 = 3.2W/(m2 sr). Lmin,61, Lmax,61, Lmin,62 and Lmax,62 are the minimum and maximum thermal radiation energy of Band 61 and 62 for Landsat 7 ETM+. Next, spectral radiance is converted to brightness temperature T6: T6 ¼
K2 lnðK 1 =Lb þ 1Þ
where K1, K2 are constant. For Landsat 5 K1 = 607.76W/(m2 sr lm) and K2 = 1260.56K. Landsat 7 ETM+, K1 = 666.09W/(m2 sr lm) K2 = 1281.71K. Finally, the LST is calculated as below (Artis Carnahan, 1982): Ts ¼
T6 1 þ ðk T 6 =aÞ ln e
ð3Þ TM, For and and
ð4Þ
where k ¼ 11:5lm, a = 0.01438 mK, e is the land surface emissivity, which can be estimated through the method proposed by Sobrino et al. (2004). 3.3. The indicator of CUHII To quantify the effects of different types of land use change during urbanization on the UHI, the contribution
466
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
Fig. 2. The map of land uses classification from the year of 1987 to 2007.
to UHI intensity (CUHII) is built in this research. The definition of the contribution is similar to the concept used in the study of Cairns et al. (2000), which refers to a portion of the total effects of each of two or more causes on a geography phenomenon. The calculation processes of CUHII
for each stage of urbanization can be divided into three parts: First, the indicator of urban heat island intensity (UHII) is calculated to reflect the LST difference between the urban and rural areas. It is a well-known indicator of the UHI
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
467
(Kim and Jong-Jin Baik, 2002; Memon et al., 2009; Schwarz et al., 2011; Schwarz et al., 2012) and the equation is as follow:
Table 2 The statistics of top 21 trajectories. Trajectory
Count
Area ratio (%)
Accumulation (%)
UHII ¼ T 0urban T 0rural
22224 44444 22444 22244 24444 33334 22544 22524 12444 25444 22254 12224 14444 22554 11114 34444 33344 12244 54444 33354 22234
76624 45801 29337 21177 15395 12101 10045 9562 4716 4190 4042 3765 3750 2670 2471 2298 2158 2018 1916 1887 1788
21.91 13.09 8.39 6.05 4.4 3.46 2.87 2.73 1.35 1.2 1.16 1.08 1.07 0.76 0.71 0.66 0.62 0.58 0.55 0.54 0.51
21.91 35 43.39 49.44 53.84 57.3 60.17 62.9 64.25 65.45 66.61 67.69 68.76 69.52 70.23 70.89 71.51 72.09 72.64 73.18 73.69
ð5Þ
where T 0urban and T 0rural are the average LST of urban and rural areas, respectively. Then, the average LST of the urban area could be represented as follow: Pn 0 i S traj;i T traj;i 0 T urban ¼ ð6Þ S urban where Straj,i and Surban are areas of a specific urban trajectory i and total urban area for each stage, respectively, T 0traj;i and T 0urban are the corresponding average LST of each area, and n is the number of urban trajectories. Finally, substituting Eq. (6) into Eq. (5) means that Eq. (5) can be transformed to: Pn 0 i S traj;i T traj;i ¼1 ð7Þ S urban ðUHII þ T 0rural Þ
average LST of each trajectory are two key parameters in determining the value of the CUHII.
and the CUHII is defined as: CUHII i ¼
S traj;i T 0traj;i S urban ðUHII þ T 0rural Þ
S traj;i T 0traj;i ¼ 100% S urban T 0urban
4. Results and discussion ð8Þ
larger value of the CUHII indicates greater effect of the corresponding land use change trajectory on the UHII. Furthermore, it is also demonstrated that the area and
4.1. The extraction and analysis of land use change trajectory of urbanization The results of land use classification are shown in Fig. 2 and the overall accuracies are: 92.17%, 90.18%, 93.45%,
Table 1 Error matrix of the land use classification from 1987 to 2007. Year
Land cover
Producers accuracy (%)
Users accuracy (%)
Overall accuracy (%)
Kappa statistics
1987
Forest Farmland Water Built-up Barren land Forest Farmland Water Built-up Barren land Forest Farmland Water Built-up Barren land Forest Farmland Water Built-up Barren land Forest Farmland Water Built-up Barren land
92.90 89.47 98.11 86.67 50.00 80.75 93.97 98.11 90.48 50.00 84.40 97.85 97.74 91.18 90.00 92.18 93.90 97.37 78.95 75.00 94.42 88.18 97.46 85.37 80.00
93.41 91.40 93.69 76.47 100.00 94.89 85.39 94.55 86.36 100.00 97.54 87.92 99.24 93.94 81.82 95.38 87.50 97.37 88.24 100.00 96.37 88.18 97.46 88.61 54.55
92.17
0.8839
90.18
0.8545
93.45
0.9074
92.46
0.8959
91.95
0.8913
1992
1997
2002
2007
468
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
92.46%, and 91.95%, respectively. The Kappa indices are: 0.8839, 0.8545, 0.9074, 0.8959 and 0.8913, respectively (Tables 1–4). The statistical results show that during the entire study period, the built-up area expanded from 3.51% in 1987 to 14.70% in 2007, which is the greatest land use change type of the study area. Meanwhile, the greatest decrease from 37.67% to 22.94% occurred in farmland see Table 1. Land use change trajectories are extracted through the overlay analysis method in ArcGISe 9.3. A total of 2218 land use change trajectories are identified and 530 of them are the existing urban or urbanization trajectories. Furthermore, 21 of the 530 trajectories occupied an area greater than 0.5% with a total area of 73.69% (Table 2). Among these 21 trajectories, 13.09% are the existing urban area, 49.98% are from farmland, 5.28% from water, and 5.34% from forest and barren land. Fig. 3 shows the spatial patterns of the land use change trajectories. The areas with green, yellow, blue color series and purple refer to the transformation from forest, farmland, water and barren land to built-up area, respectively. The area colored red represents existing urban area prior to the study period. The figure clearly shows that the existing urban area of trajectory No. 44444 is located in the
southwest of Xiamen Island, while the urban growth areas are mainly located in the northeast part of the island and inland areas. 4.2. The results of LST retrieval The results of LST retrieved from different years are shown in Fig. 4, which shows significant LST differences. A further statistic of LST is shown in Fig. 5. LST over the whole study area varies because of the different climatological and environmental conditions across the study period, which causes difficulty in exploring the overall LST change. However, some meaningful information about the LST can be discovered when analyzing each land use type. To analyze the UHI in detail, land use classification maps are taken as mosaics to extract the corresponding LST. Fig. 6 shows the LST with different land use types during the study period. It can be seen that in the period before 2007 farmland exhibited the highest LST, which might be attributed to the reduction in latent heat flux and increase in sensible heat of farmland in winter (Jin et al., 2005; Rizwan et al., 2008; Imhoff et al., 2010; Schwarz et al., 2012). Following that, the built-up area
Fig. 3. The map of the urbanization trajectory.
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
469
Fig. 4. Thematic map of LST from 1987 to 2007.
presented the highest LST, which implied a strengthening of the UHI magnitude. It was related to industrial zones and large infrastructure constructed in coastal areas during the rapid course of urbanization, since both large impervious ground surfaces, large-sized and endothermic factory building roofs were the sources of these hot spots (Zhao
et al., 2010b). Forest was the coolest area during the study period as the high vegetable fraction could mitigate the LST (Weng et al., 2004; Rizwan et al., 2008; Onishi et al., 2010). To examine quantitatively the magnitude of the UHI, the UHII for different years is calculated using Eq. (5). It
470
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
4.3. The CUHII of land use change trajectory
Fig. 5. The retrieval results of LST (°C).
Fig. 6. The LST change of each land use type (°C).
Fig. 7. The values of UHII from the year of 1987 to 2007 (°C).
can be seen that the UHII presents a trend of continuous increase from 0.83 °C in 1987 to 2.14 °C in 2007. An exception was in 1992, when the UHII decreased compared with that of 1987. As shown in Fig. 7, the average LST and the standard deviation of the study area in 1992 were the least and greatest during the study period with values of 10.82 °C and 2.05 °C, respectively. Furthermore, in this year, water exhibited the highest LST, whereas it was almost the coolest type in other years. The reason for this might be attributed to the cold wave that occurred in the winter of 1992 (China Meteorological Data Sharing Service System, 2011); the difference of specific heat capacity among land use types would lead to various LST changes, and increase the standard deviation. Meanwhile, the specific heat capacity of water is much greater than other land use types, which would lead a relatively high LST.
To quantify the effects of different land use change types on the UHI, initially the land use change trajectories were taken as mosaics to extract the corresponding LST and the results are shown in Table 3. In 2007, all the top 21 trajectories were transformed into built-up area. The average LST and standard deviation of all the trajectories are: 15.57 °C and 0.67, respectively, which shows a minor difference among the trajectories. Then, the CUHIIs are calculated according to Eqs (7) and (8). The results are shown in Table 4, where the symbol “–” means that the trajectory has not transformed into built-up area in the corresponding year. Trajectory of No. 44444 is only one major part of the built-up area in 1987, thus the CUHII value in 1987 is less than 1. As for year 2007, the total CUHII of the trajectories list in Table 4 is 73.82%, which is still less than 1. That is because only the Top 21 trajectories are selected in this study, and CUHII of other trajectories have not taken into account. It can be seen that before 2002, the existing urban area of trajectory NO. 44444 had the most significant effect on the UHI with the greatest CUHII, but that this value has since decreased continuously. In 2007, the trajectory with the greatest CUHII was NO. 22224 (21.98%), followed by the existing urban area of trajectory NO. 44444 (13.03%). It can be concluded that urbanization plays a significant role on the UHI, which weakened the contribution of the existing urban area. The CUHII map of 2007 is shown in Fig. 8. The areas of high contribution are mainly located in the north of the inland, and the areas of low contribution are mainly located near the seaside of the inland and in the south of Xiamen Island. As defined in Eq. (8), the CUHII is dependent upon the area and average LST of each land use change trajectory. Because of minor differences among the average LST of land use change trajectories, the area of each trajectory plays a more important role on the CUHII. The larger the area of the land use change trajectory, the greater the impact on the CUHII. 5. Conclusion This paper quantified the effects of urbanization on the UHI through the method of land use change trajectory and the indices of CUHII in Xiamen, China. The method of land use change trajectory was developed to analyze the land use change types during urbanization and provide the spatial calculation units. The index of CUHII was built to quantify the effects of each land use change trajectory on the UHI during urbanization. The results demonstrated that: (1) The method of land use change trajectory and the indicator of CUHII provided a tool for quantifying the effects of urbanization on the UHI. Among the urbanization trajectories, the most significant transformation was from farmland to built-up area. (2) The study area presented a continuous increase of UHII from 0.83 °C in
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
471
Table 3 The average LSTs of land use change trajectories in different years (°C). Trajectory
1987–01–17
1992–01–15
1997–01–12
2002–01–02
2007–01–08
22224 44444 22444 22244 24444 33334 22544 22524 12444 25444 22254 12224 14444 22554 11114 34444 33344 12244 54444 33354 22234
19.19 18.55 19.04 19.06 18.89 17.10 19.23 19.30 18.23 19.17 19.34 18.58 18.26 19.45 17.21 17.91 16.91 18.37 17.80 17.46 18.74
12.03 11.42 11.94 12.05 11.48 11.39 12.24 12.35 11.54 11.05 12.31 11.59 11.03 12.33 9.00 11.32 11.28 11.67 11.02 10.63 11.93
15.88 15.16 15.31 15.71 15.23 13.82 14.71 14.85 14.94 15.22 15.97 15.49 14.76 14.57 13.48 14.96 13.71 15.51 14.94 13.18 15.08
18.74 17.99 18.08 18.06 17.99 16.26 18.34 18.70 17.66 18.22 17.78 18.42 17.51 17.75 16.29 17.87 17.24 17.71 17.81 17.19 17.01
15.71 15.58 15.81 15.79 15.70 14.91 16.46 16.35 15.27 16.11 16.24 15.54 14.90 16.15 13.30 15.54 15.68 15.42 15.45 15.37 15.65
Table 4 The CUHII of urbanization land use change trajectories (%). Trajectory
1987–1–17
1992–1–15
1997–1–12
2002–1–2
2007–1–8
22224 44444 22444 22244 24444 33334 22544 22524 12444 25444 22254 12224 14444 22554 11114 34444 33344 12244 54444 33354 22234
–* 55.00 – – – – – – – – – – – – – – – – – – –
– 38.99 – – 13.18 – – – – – – – 3.08 – – 1.94 – – 1.57 – –
– 27.57 17.84 – 9.31 – – – 2.80 2.53 – – 2.20 – – 1.37 – – 1.14 – –
– 21.77 14.01 10.11 7.32 – 4.87 – 2.20 2.02 – – 1.74 – – 1.09 0.98 0.94 0.90 – –
21.98 13.03 8.47 6.11 4.41 3.29 3.02 2.85 1.32 1.23 1.20 1.07 1.02 0.79 0.60 0.65 0.62 0.57 0.54 0.53 0.51
*
means land is not covered by urban in this year.
1987 to 2.14 °C in 2007; an exception was in 1992, which might be attributed to the cold wave that year (3) The maximum CUHII before 2002 was the existing urban area of trajectory NO. 44444, but the value decreased from 55.00% in 1987 to 13.03% in 2007. In 2007, it was replaced by trajectory NO. 22224 the contribution of which was 21.98%. The areas of the land use change trajectories played a significantly more important role on the CUHII, than the minor LST differences among them. Therefore, urbanization significantly influenced the UHII and weakened the CUHII of the existing urban area.
This research provides a new method to study quantitatively the effects of urbanization on the UHI based on the method of land use change trajectory. Two areas need to be addressed in future research. First, the classification accuracy was a key factor in the land use change trajectory method, especially in the long time-series analysis. The relationship between classification errors and the reliability of the research should be further discussed. Second, there were many other indicators to indicate UHI effects in addition to the UHII, such as magnitude, hot island area and range. (Schwarz et al., 2011; Schwarz et al., 2012). These
472
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
Fig. 8. The CUHII map of land use change trajectories in 2007.
diverse indicators and their similarities and differences should also be the subject of further study when used to analyze the UHI. Acknowledgments This study was supported by the National Natural Science Foundation of China (No. 41371392), the National Key Technology R&D Program (2012BAC21B03), the Knowledge Innovation Programs of Chinese Academy of Sciences (IUEQN-2012-01), and also the Natural Science Foundation of Fujian Province (2013J05062) and Ningbo City (2013A610180). References Amiri, R., Weng, Q., Alimohammadi, A., Alavipanah, S.K., 2009. Spatial–temporal dynamics of land surface temperature in relation to
fractional vegetation cover and land use/cover in the Tabriz urban area. Iran. Remote Sens. Environ. 113 (12), 2606–2617. Artis, D.A., Carnahan, W.H., 1982. Survey of emissivity variability in thermography of urban areas. Remote Sens. Environ. 12 (4), 313–329. Cairns, M.A., Haggerty, P.K., Alvarez, R., Jong, B.H.J.D., Olmsted, I., 2000. Tropical Mexico’s recent land-use change: a region’s contribution to the global carbon cycle. Ecol. Appl. 10 (5), 1426–1441. Changnon, S.A., Kunkel, K.E., Reinke, B.C., 1996. Impacts and responses to the 1995 heat wave: a call to action. Bull. Am. Meteorol. Soc. 77 (7), 1497–1506. Chen, X.-L., Zhao, H.-M., Li, P.-X., Yin, Z.-Y., 2006. Remote sensing image-based analysis of the relationship between urban heat island and land use/cover changes. Remote Sen. of Environ. 104 (2), 133–146. China Meteorological Data Sharing Service System. (2011). “http:// www.cma.gov.cn/2011qxfw/2011qsjgx/.”. Gallo, K.P., Easterling, D.R., Peterson, T.C., 1996. The influence of land use/land cover on climatological values of the Diurnal temperature range. J. Clim. 9, 2941–2944. Gallo, K.P., McNab, A.L., Karl, T.R., Brown, J.F., Hood, J.J., Tarpley, J.D., 1993. The Use of NOAA AVHRR Data for Assessment of the Urban Heat Island Effect. J. Appl. Meteorol. 32 (5), 899–908.
H. Feng et al. / Advances in Space Research 53 (2014) 463–473 Grimm, N.B., Faeth, S.H., Golubiewski, N.E., Redman, C.L., Wu, J., Bai, X., Briggs, M., 2008. Global change and the ecology of cities. Science 319, 756–760. Houghton, R.A., Goodale, C.L., 2004. Effects of land-use change on the carbon balance of terrestrial ecosystems. Geophys. Monogr. 153 (1–4), 85–98. Howard, L. 1818. The Climate of London: deduced from Meteorological observations, made at different places in the neighbourhood of the metropolis. London, W. Phillips, sold also by J. and A. Arch. Huang, L., Huang, J., (2011). Preliminary study on urban heat island effect in Xiamen during 2001–2007. 2011 International Conference on Remote Sensing, Environment and Transportation Engineering (RSETE): 3330–3333. Imhoff, M.L., Zhang, P., Wolfe, R.E., Bounoua, L., 2010. Remote sensing of the urban heat island effect across biomes in the continental USA. Remote Sens. Environ. 114 (3), 504–513. Jin, M., Dickinson, R.E., Zhang, D., 2005. The footprint of urban areas on global climate as characterized by MODIS. J. Clim. 18 (10), 1551– 1565. Jusuf, S.K., Wong, N.H., Hagen, E., et al., 2007. The influence of land use on the urban heat island in Singapore. Habitat Int. 31 (2), 232–242. Kim, Y.-H., Baik, Jong.-Jin., 2002. Maximum urban heat island intensity in Seoul. J. Appl. Meteorol. Climatol. 41, 651–659. Kolokotroni, M., Ren, X., Davies, M., Mavrogianni, A., 2012. London’s urban heat island: impact on current and future energy consumption in office buildings. Energy Build. 47, 302–311. Konopacki, S., Akbari, H., (2002). Energy savings for heat-island reduction strategies in Chicago and Houston (including updates for Baton Rouge, Sacramento, and Salt Lake City). L. B. N. Laboratory. Lambin, E.F., 1997. Modelling and monitoring land-cover change processes in tropical regions. Prog. Phys. Geogr. 21 (3), 375–393. Li, J., Song, C., Cao, L., Zhu, F., Meng, X., Wu, J., 2011. Impacts of landscape structure on surface urban heat islands: A case study of Shanghai. China Remote Sens. Environ. 115 (12), 3249–3263. Liu, H., Zhou, Q., 2005. Developing urban growth predictions from spatial indicators based on multi-temporal images. Comput. Environ. Urban Syst. 29, 580–594. Memon, R.A., Leung, D.Y.C., Liu, C.-H., 2009. An investigation of urban heat island intensity (UHII) as an indicator of urban heating. Atmos. Res. 94 (3), 491–500. Mertens, B., Lambin, E.F., 2000. Land-cover-change trajectories in Southern Cameroon. Ann. Assoc. Am. Geogr. 90 (3), 467–494. Onishi, A., Cao, X., Ito, T., Shi, F., Imura, H., 2010. Evaluating the potential for urban heat-island mitigation by greening parking lots. Urban For. Urban Greening 9 (4), 323–332. Pauleit, S., Ennos, R., Golding, Y., 2005. Modeling the environmental impacts of urban land use and land cover change—a study in Merseyside. UK. Landscape Urban Plan. 71 (2–4), 295–310. Petit, C.C., Lambin, E.F., 2001. Integration of multi-source remote sensing data for land cover change detection. Int. J. Geog. Inf. Sci. 15 (8), 785–803. Rao, P.K., 1972. Remote sensing of urban“heat islands”from an environmental satellite. Bull. Am. Meteorol. Soc. 53, 647–648. Rizwan, A.M., Dennis, L.Y.C., Liu, C., 2008. A review on the generation, determination and mitigation of urban heat island. J. Environ. Sci. 20 (1), 120–128.
473
Rosenfeld, A.H., Akbari, H., Romm, J.J., Pomerantz, M., 1998. Cool communities: strategies for heat island mitigation and smog reduction. Energy Build. 28 (1), 51–62. Schwarz, N., Lautenbach, S., Seppelt, R., 2011. Exploring indicators for quantifying surface urban heat islands of European cities with MODIS land surface temperatures. Remote Sens. Environ. 115 (12), 3175– 3186. Schwarz, N., Schlink, U., Franck, U., Großmannc, K., 2012. Relationship of landsurface and airtemperatures and its implications for quantifying urban heat island indicators–an application for the city of Leipzig (Germany). Ecol. Ind. 18, 693–704. Sobrino, J.A., Jime´nez-Mun˜oz, J.C., Paolini, L., 2004. Land surface temperature retrieval from LANDSAT TM 5. Remote Sens. Environ. 90 (4), 434–440. Sobrino, J.A., Romaguera, M., 2004. Land surface temperature retrieval from MSG1-SEVIRI data. Remote Sens. Environ. 92 (2), 247–254. Wang, D., Gong, J., Chen, L., Zhang, L., Song, Y., Yue, Y., 2012. Spatiotemporal pattern analysis of land use/cover change trajectories in Xihe watershed. Int. J. Appl. Earth Obs. Geoinf. 14 (1), 12–21. Wang, D., Gong, J., Chen, L., Zhang, L., Song, Y., Yue, Y., 2013. Comparative analysis of land use/cover change trajectories and their driving forces in two small watersheds in the western Loess Plateau of China. Obs. Geoinf. 21, 241–252. Weng, Q., 2001. A remote sensing–GIS evaluation of urban expansion and its impact on surface temperature in the Zhujiang Delta. China Int. J. Remote Sens. 22 (10), 1999–2014. Weng, Q., Lu, D., Schubring, J., 2004. Estimation of land surface temperature–vegetation abundance relationship for urban heat island studies. Remote Sens. Environ. 89 (4), 467–483. Weng, Q., Rajasekar, U., Hu, X., 2011. Modeling urban heat islands with multi-temporal ASTER images. IEEE Trans. Geosci. Remote Sens. 49 (10), 4080–4089. Xiamen Bureau of Statistics (2008). Year book of Xiamen special economic zone, Beijing:China Statistics Press (in Chinese). Xu, H., Chen, B., 2004. Remote sensing of the urban heat island and its changes in Xiamen City of SE China. J. Environ. Sci. 16 (2), 276-218. Yuan, F., Bauer, M.E., 2007. Comparison of impervious surface area and normalized difference vegetation index as indicators of surface urban heat island effects in Landsat imagery. Remote Sens. Environ. 106 (3), 375–386. Zhang, Y., Odeh, I.O.A., Han, C., 2009. Bi-temporal characterization of land surface temperature in relation to impervious surface area, NDVI and NDBI, using a sub-pixel image analysis. Int. J. Appl. Earth Obs. Geoinf. 11 (4), 256–264. Zhao, J., Dai, D., Lin, T., Tang, L., 2010a. Rapid urbanisation, ecological effects and sustainable city construction in Xiamen. Int. J. Sustainable Dev. World Ecol. 17 (4), 271–272. Zhao, X., Huang, J., Ye, H., Wang, K., Qiu, Q., 2010b. Spatiotemporal changes of the urban heat island of a coastal city in the context of urbanisation. Int. J. Sustainable Dev. World Ecol. 17 (4), 311–316. Zhou, Q., Li, B., Kurban, A., 2008. Spatial pattern analysis of land cover change trajectories in Tarm Basin, northwest China. Int. J. Remote Sens. 29 (19), 5495–5509. Zhou, W., Huang, G., Cadenasso, M.L., 2011. Does spatial configuration matter? understanding the effects of land cover pattern on land surface temperature in urban landscapes. Landscape Urban Plan. 102 (1), 54–63.
ScienceDirect Advances in Space Research 53 (2014) 463–473 www.elsevier.com/locate/asr
Using land use change trajectories to quantify the effects of urbanization on urban heat island Huihui Feng a,b,c,1, Xiaofeng Zhao a,b,⇑, Feng Chen a,b,2, Lichun Wu d,3 a
Key Lab of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, Xiamen 361021, China b Xiamen Key Lab of Urban Metabolism, Xiamen 361021, China c Nanjing Institute of Geography and Limnology, Chinese Academy of Sciences, Nanjing 210008, China d School of Resources and Safety Engineering, Central South University, Changsha 410083, China Received 19 June 2013; received in revised form 30 October 2013; accepted 14 November 2013 Available online 22 November 2013
Abstract This paper proposed a quantitative method of land use change trajectory, which means the succession among different land use types across time, to examine the effects of urbanization on an urban heat island (UHI). To accomplish this, multi-temporal images from Landsat 5 Thematic Mapper (TM) and Landsat 7 Enhanced Thematic Mapper Plus (ETM+) of Xiamen City in China from 1987 to 2007 were selected. First, the land use change trajectories were extracted based on the classified images from different years. Then the land surface temperatures (LST) were retrieved and the magnitudes of the UHI were evaluated using the UHI intensity (UHII) indicator. Finally, the indices of the contribution to UHI intensity (CUHII) were constructed and calculated to quantify the effects of each land use change trajectory on the UHI during urbanization. The results demonstrated that the land use change trajectories and CUHII are effective and useful in quantifying the effects of urbanization on UHI. In Xiamen City, a total of 2218 land use change trajectories were identified and 530 of them were the existing urban or urbanization trajectories. The UHII presents a trend of continuous increase from 0.83 °C in 1987 to 2.14 °C in 2007. With respect to the effects of urban growth on UHI, the contribution of existing urban area to UHI decreased during urbanization. Prior to 2007, the existing urban area of trajectory NO. 44444 had the most significant effect on UHI with the greatest CUHII, while the value has decreased from 55.00% in 1987 to 13.03% in 2007 because of the addition of new urbanized area. In 2007, the greatest CUHII was replaced by a trajectory from farmland to built-up area (NO. 22224) with the CUHII of 21.98%, followed by the existing urban area of trajectory NO. 44444 with the CUHII of 13.03%. These results provide not only a new methodology to assess the environmental effects of urbanization, but also decision-supports for the planning and management of cities. Ó 2013 COSPAR. Published by Elsevier Ltd. All rights reserved. Keywords: Urbanization; Land use change trajectory; Urban heat island; Contribution to UHI intensity (CUHII); Xiamen
1. Introduction Urbanization profoundly influences the ecosystem of an urban area. One of the best-documented examples of this,
⇑ Corresponding author. at: Key Lab of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, Xiamen 361021, China. Tel.: +86 592 6190672; fax: +86 592 6190977. E-mail addresses: [email protected] (H. Feng), [email protected] (X. Zhao), [email protected] (F. Chen), [email protected] (L. Wu). 1 Tel.: 86 25 86882167; fax: 86 25 57714759. 2 Tel.: Tel.: 86 592 6190672; fax: 86 592 6190977. 3 Tel.: 86 25 86882167; fax: 86 25 57714759.
is the urban heat island (UHI) effect, which refers to the higher temperatures found in urban areas compared with the surrounding rural area (Grimm et al., 2008). UHI not only influences the living environment (Konopacki and Akbari, 2002), but also increases energy consumption (Kolokotroni et al., 2012), and even harms human health (Changnon et al., 1996). Recently, a great deal of research has been carried out on the causes and impacts of UHIs, and their qualitative and quantitative characteristics have been documented (Rosenfeld et al., 1998; Rizwan et al., 2008; Imhoff et al., 2010). UHI was observed through air temperature at the early stage. For example, Howard (1818) firstly proposed the
0273-1177/$36.00 Ó 2013 COSPAR. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.asr.2013.11.028
464
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
phenomenon of UHI when he observed a higher the air temperature of urban area than the surrounding rural area of London. Jusuf et al. (2007) used the qualitative and quantitative methods to evaluate the influence of land use on the UHI from air temperature in Singapore. In recent decades, since assessment of land surface temperature (LST) from satellite data was first proposed by Rao (1972), this remote sensing method has been widely used to analyze UHI characteristics on a regional and global spatial scale (Gallo et al., 1993; Weng, 2001; Sobrino and Romaguera, 2004; Zhao et al., 2010b; Schwarz et al., 2011; Weng et al., 2011). LST is strongly related to the surface properties and much previous research focused on the relationship between the LST and land use in mono-temporal (Weng et al., 2004), across different seasons (Gallo et al., 1993; Li et al., 2011) or different years (Chen et al., 2006; Yuan and Bauer, 2007; Zhang et al., 2009). The results of this research effort indicated a positive correlation between the LST and impervious land use, and a negative correlation with the vegetation fraction. Furthermore, the effects of urbanization on the UHI were also analyzed. For example, Gallo et al. (1996) concluded that the transformation from rural to urban land use can impact the trend in temperature in a similar manner to that which would be expected under an enhanced greenhouse warming scenario. Amiri et al. (2009) examined the relationship between the temporal dynamics of LST and land use through the method of temperature vegetation index (TVX) space. They found that the LST trajectory in TVX space was reflected by the shift from the high vegetation – low LST fraction, to the low vegetation – high LST areas during the process of urbanization. Zhou et al. (2011) showed that the impact of urbanization on the UHI can be mitigated by balancing land use composition through optimizing land use configuration. However, the quantitative relationship between urbanization and its effects on the UHI is still ambiguous. The environmental response due to urbanization associated with different land use changes (e.g., from farmland to built-up areas and from forest to built-up areas.), is highly variable because of the different properties of each land use type (Houghton and Goodale, 2004; Pauleit et al., 2005). Therefore, it is essential to quantify the effects of different land use change during urbanization on the UHI, to understand fully the characteristics of the UHI. This study used land use change trajectory as a method to quantify the effects of different land use change types on the UHI during urbanization. A trajectory was defined as the succession areas of land use types, through this method, the urban areas could be divided into several parts according to the transformation of other land use types to built-up area (Lambin, 1997; Mertens and Lambin, 2000; Petit and Lambin, 2001). For example, Liu and Zhou (2005) simulated the process of urban growth through the land use change trajectory in Beijing. Zhou et al. (2008) analyzed the spatial pattern analysis of land cover change trajectories in Tarm Basin, northwest China. Wang et al.
(2012, 2013) illustrated the spatial patterns and the driving forces of the land cover change trajectories in the Xihe watershed of the Loess Plateau, China. The results of researches could support the methodology of this research. After the retrieval of LST and the calculation of the UHII, land use change trajectories were extracted and their CUHII calculated. In this way, the quantitative relationship between urbanization and its effects on the UHI was analyzed in the context of urbanization in Xiamen City. The results not only provide a useful tool for quantifying the effects of urbanization, but also support the planning and management of cities. 2. Study area and data pre-processing 2.1. Study area The city of Xiamen, which has experienced rapid urbanization in the past three decades, is selected as the study area for this research. It locates at 24°25’–24°55’ N and 117°53’–117°28’ E, and is situated on the southeast coast of Fujian Province and at the estuary of Jiulong River. It comprises Xiamen Island, Gulang Island, and the coastal part to the north of the Jiulong River (Fig. 1). It has an area of more than 1565 km2, a sea area of 390 km2, and in 2007, a population of 2.43 million with nearly 70% of which in urban areas (Xiamen Bureau of Statistics, 2008). In 1980, Xiamen Special Economic Zone was established. Since then, Xiamen has experienced rapid urbanization, which has led to significant environmental and ecological effects (Zhao et al., 2010a). Meanwhile, UHI in Xiamen has become more and more obvious (Xu and Chen, 2004; Zhao et al., 2010a,b; Huang and Huang, 2011). 2.2. Data resources and pre-processing Images acquired from the Landsat 5 Thematic Mapper (TM) and Landsat 7 Enhanced Thematic Mapper Plus (ETM+) sensors (Jan 17, 1987; Jan 15, 1992; Jan 12, 1997; Jan 2, 2002, and Jan 8, 2007) were used in this research. The multi-images were co-registered to the same coordinate system of UTM/WGS84 based on the image of 1987. Moreover, the images were then re-sampled using the nearest neighbor algorithm with a spatial resolution of 30 m for all bands. To remove the atmospheric influence, the MODTRAN4-based FLAASH module in the software of ENVI4.7 was adopted to correct atmospheric errors. Its parameters included the information about the sensor and scene, atmosphere and aerosol model, and the atmosphere correction model. For the study area of this paper, the scene center location, sensor altitude, sensor type, flight date and time were obtained from the Landsat TM/ ETM+ header book, the average elevation was 0.05 km, the atmospheric model and aerosol model were Mid-Latitude Summer (MLS) and urban, and the aerosol retrieval was adopted the K-T method. Finally, to analyze the land use change, all the individual images were classified
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
465
Fig. 1. The study area.
through the Support Vector Machine (SVM) method. According to current Chinese land use classification, five major land use types of forest, farmland, water, built-up area and barren land were extracted. These types were numbered from 1 to 5 for their digital forms, respectively. To assess the classification accuracy, high resolution historical images from Google Earth were used as the reference layers, and five hundred random sample points were generated for comparing classification data and reference data. 3. Methods 3.1. Land use change trajectory To generate the land use trajectories of urbanization, all the classified images are first integrated in GIS using raster format with ArcGISe 9.3 software, and then calculated through the overlay analysis, as in the method of Zhou et al. (2008). The equation is: Trac ¼ Cl1987 10000 þ Cl1992 1000 þ Cl1997 100 þ Cl2002 10 þ Cl2007
ð1Þ
where Cl1987, Cl1992, Cl1997, Cl2002, and Cl2007 are the land use classified data of the years 1987, 1992, 1997, 2002, and 2007, respectively. A trajectory code contains two meanings. Firstly, it describes the succession among different land cover types. Furthermore, it implies the time of succession happened. For example, the trajectory code NO. 12345 refers to the transformation process of forest ! farmland ! water ! built-up area ! barren land across the study period. To assess the effects of urbanization on the UHI, only the trajectories of the existing urban and urbanization areas are considered. Therefore, two types of trajectories are identified in this research: the first is the existing urban trajectory, which means the area already built-up prior to the study period (trajectory code: No. 44444); and the other is the urbanization trajectories, which means those areas urbanized from other land use types to build-up area during the study (trajectory code: No. ****4, where * is the land use type other than built-up area).
3.2. LST retrieval The LST retrieval algorithm used in this paper can be divided into three steps (Artis and Carnahan, 1982; Weng et al., 2004; Chen et al., 2006). The first step is to calculate the spectral radiance Lb from the digital number of Band 6: Lb ¼ Lmin þ ðLmax Lmin Þ DN 6 =255
ð2Þ
where Lmax, Lmin are the maximum and minimum thermal radiation energy, which can be obtained from the sensor handbook. DN6 is digital number of Band 6. For Landsat 5 TM, Lmax = 15.600W/(m2 sr), Lmin = 1.238W/(m2 sr). For Landsat 7 ETM+, Lmax,61 = 17.040W/(m2 sr), Lmin,61 = 0 and Lmax,62 = 12.650W/(m2 sr), Lmin,62 = 3.2W/(m2 sr). Lmin,61, Lmax,61, Lmin,62 and Lmax,62 are the minimum and maximum thermal radiation energy of Band 61 and 62 for Landsat 7 ETM+. Next, spectral radiance is converted to brightness temperature T6: T6 ¼
K2 lnðK 1 =Lb þ 1Þ
where K1, K2 are constant. For Landsat 5 K1 = 607.76W/(m2 sr lm) and K2 = 1260.56K. Landsat 7 ETM+, K1 = 666.09W/(m2 sr lm) K2 = 1281.71K. Finally, the LST is calculated as below (Artis Carnahan, 1982): Ts ¼
T6 1 þ ðk T 6 =aÞ ln e
ð3Þ TM, For and and
ð4Þ
where k ¼ 11:5lm, a = 0.01438 mK, e is the land surface emissivity, which can be estimated through the method proposed by Sobrino et al. (2004). 3.3. The indicator of CUHII To quantify the effects of different types of land use change during urbanization on the UHI, the contribution
466
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
Fig. 2. The map of land uses classification from the year of 1987 to 2007.
to UHI intensity (CUHII) is built in this research. The definition of the contribution is similar to the concept used in the study of Cairns et al. (2000), which refers to a portion of the total effects of each of two or more causes on a geography phenomenon. The calculation processes of CUHII
for each stage of urbanization can be divided into three parts: First, the indicator of urban heat island intensity (UHII) is calculated to reflect the LST difference between the urban and rural areas. It is a well-known indicator of the UHI
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
467
(Kim and Jong-Jin Baik, 2002; Memon et al., 2009; Schwarz et al., 2011; Schwarz et al., 2012) and the equation is as follow:
Table 2 The statistics of top 21 trajectories. Trajectory
Count
Area ratio (%)
Accumulation (%)
UHII ¼ T 0urban T 0rural
22224 44444 22444 22244 24444 33334 22544 22524 12444 25444 22254 12224 14444 22554 11114 34444 33344 12244 54444 33354 22234
76624 45801 29337 21177 15395 12101 10045 9562 4716 4190 4042 3765 3750 2670 2471 2298 2158 2018 1916 1887 1788
21.91 13.09 8.39 6.05 4.4 3.46 2.87 2.73 1.35 1.2 1.16 1.08 1.07 0.76 0.71 0.66 0.62 0.58 0.55 0.54 0.51
21.91 35 43.39 49.44 53.84 57.3 60.17 62.9 64.25 65.45 66.61 67.69 68.76 69.52 70.23 70.89 71.51 72.09 72.64 73.18 73.69
ð5Þ
where T 0urban and T 0rural are the average LST of urban and rural areas, respectively. Then, the average LST of the urban area could be represented as follow: Pn 0 i S traj;i T traj;i 0 T urban ¼ ð6Þ S urban where Straj,i and Surban are areas of a specific urban trajectory i and total urban area for each stage, respectively, T 0traj;i and T 0urban are the corresponding average LST of each area, and n is the number of urban trajectories. Finally, substituting Eq. (6) into Eq. (5) means that Eq. (5) can be transformed to: Pn 0 i S traj;i T traj;i ¼1 ð7Þ S urban ðUHII þ T 0rural Þ
average LST of each trajectory are two key parameters in determining the value of the CUHII.
and the CUHII is defined as: CUHII i ¼
S traj;i T 0traj;i S urban ðUHII þ T 0rural Þ
S traj;i T 0traj;i ¼ 100% S urban T 0urban
4. Results and discussion ð8Þ
larger value of the CUHII indicates greater effect of the corresponding land use change trajectory on the UHII. Furthermore, it is also demonstrated that the area and
4.1. The extraction and analysis of land use change trajectory of urbanization The results of land use classification are shown in Fig. 2 and the overall accuracies are: 92.17%, 90.18%, 93.45%,
Table 1 Error matrix of the land use classification from 1987 to 2007. Year
Land cover
Producers accuracy (%)
Users accuracy (%)
Overall accuracy (%)
Kappa statistics
1987
Forest Farmland Water Built-up Barren land Forest Farmland Water Built-up Barren land Forest Farmland Water Built-up Barren land Forest Farmland Water Built-up Barren land Forest Farmland Water Built-up Barren land
92.90 89.47 98.11 86.67 50.00 80.75 93.97 98.11 90.48 50.00 84.40 97.85 97.74 91.18 90.00 92.18 93.90 97.37 78.95 75.00 94.42 88.18 97.46 85.37 80.00
93.41 91.40 93.69 76.47 100.00 94.89 85.39 94.55 86.36 100.00 97.54 87.92 99.24 93.94 81.82 95.38 87.50 97.37 88.24 100.00 96.37 88.18 97.46 88.61 54.55
92.17
0.8839
90.18
0.8545
93.45
0.9074
92.46
0.8959
91.95
0.8913
1992
1997
2002
2007
468
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
92.46%, and 91.95%, respectively. The Kappa indices are: 0.8839, 0.8545, 0.9074, 0.8959 and 0.8913, respectively (Tables 1–4). The statistical results show that during the entire study period, the built-up area expanded from 3.51% in 1987 to 14.70% in 2007, which is the greatest land use change type of the study area. Meanwhile, the greatest decrease from 37.67% to 22.94% occurred in farmland see Table 1. Land use change trajectories are extracted through the overlay analysis method in ArcGISe 9.3. A total of 2218 land use change trajectories are identified and 530 of them are the existing urban or urbanization trajectories. Furthermore, 21 of the 530 trajectories occupied an area greater than 0.5% with a total area of 73.69% (Table 2). Among these 21 trajectories, 13.09% are the existing urban area, 49.98% are from farmland, 5.28% from water, and 5.34% from forest and barren land. Fig. 3 shows the spatial patterns of the land use change trajectories. The areas with green, yellow, blue color series and purple refer to the transformation from forest, farmland, water and barren land to built-up area, respectively. The area colored red represents existing urban area prior to the study period. The figure clearly shows that the existing urban area of trajectory No. 44444 is located in the
southwest of Xiamen Island, while the urban growth areas are mainly located in the northeast part of the island and inland areas. 4.2. The results of LST retrieval The results of LST retrieved from different years are shown in Fig. 4, which shows significant LST differences. A further statistic of LST is shown in Fig. 5. LST over the whole study area varies because of the different climatological and environmental conditions across the study period, which causes difficulty in exploring the overall LST change. However, some meaningful information about the LST can be discovered when analyzing each land use type. To analyze the UHI in detail, land use classification maps are taken as mosaics to extract the corresponding LST. Fig. 6 shows the LST with different land use types during the study period. It can be seen that in the period before 2007 farmland exhibited the highest LST, which might be attributed to the reduction in latent heat flux and increase in sensible heat of farmland in winter (Jin et al., 2005; Rizwan et al., 2008; Imhoff et al., 2010; Schwarz et al., 2012). Following that, the built-up area
Fig. 3. The map of the urbanization trajectory.
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
469
Fig. 4. Thematic map of LST from 1987 to 2007.
presented the highest LST, which implied a strengthening of the UHI magnitude. It was related to industrial zones and large infrastructure constructed in coastal areas during the rapid course of urbanization, since both large impervious ground surfaces, large-sized and endothermic factory building roofs were the sources of these hot spots (Zhao
et al., 2010b). Forest was the coolest area during the study period as the high vegetable fraction could mitigate the LST (Weng et al., 2004; Rizwan et al., 2008; Onishi et al., 2010). To examine quantitatively the magnitude of the UHI, the UHII for different years is calculated using Eq. (5). It
470
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
4.3. The CUHII of land use change trajectory
Fig. 5. The retrieval results of LST (°C).
Fig. 6. The LST change of each land use type (°C).
Fig. 7. The values of UHII from the year of 1987 to 2007 (°C).
can be seen that the UHII presents a trend of continuous increase from 0.83 °C in 1987 to 2.14 °C in 2007. An exception was in 1992, when the UHII decreased compared with that of 1987. As shown in Fig. 7, the average LST and the standard deviation of the study area in 1992 were the least and greatest during the study period with values of 10.82 °C and 2.05 °C, respectively. Furthermore, in this year, water exhibited the highest LST, whereas it was almost the coolest type in other years. The reason for this might be attributed to the cold wave that occurred in the winter of 1992 (China Meteorological Data Sharing Service System, 2011); the difference of specific heat capacity among land use types would lead to various LST changes, and increase the standard deviation. Meanwhile, the specific heat capacity of water is much greater than other land use types, which would lead a relatively high LST.
To quantify the effects of different land use change types on the UHI, initially the land use change trajectories were taken as mosaics to extract the corresponding LST and the results are shown in Table 3. In 2007, all the top 21 trajectories were transformed into built-up area. The average LST and standard deviation of all the trajectories are: 15.57 °C and 0.67, respectively, which shows a minor difference among the trajectories. Then, the CUHIIs are calculated according to Eqs (7) and (8). The results are shown in Table 4, where the symbol “–” means that the trajectory has not transformed into built-up area in the corresponding year. Trajectory of No. 44444 is only one major part of the built-up area in 1987, thus the CUHII value in 1987 is less than 1. As for year 2007, the total CUHII of the trajectories list in Table 4 is 73.82%, which is still less than 1. That is because only the Top 21 trajectories are selected in this study, and CUHII of other trajectories have not taken into account. It can be seen that before 2002, the existing urban area of trajectory NO. 44444 had the most significant effect on the UHI with the greatest CUHII, but that this value has since decreased continuously. In 2007, the trajectory with the greatest CUHII was NO. 22224 (21.98%), followed by the existing urban area of trajectory NO. 44444 (13.03%). It can be concluded that urbanization plays a significant role on the UHI, which weakened the contribution of the existing urban area. The CUHII map of 2007 is shown in Fig. 8. The areas of high contribution are mainly located in the north of the inland, and the areas of low contribution are mainly located near the seaside of the inland and in the south of Xiamen Island. As defined in Eq. (8), the CUHII is dependent upon the area and average LST of each land use change trajectory. Because of minor differences among the average LST of land use change trajectories, the area of each trajectory plays a more important role on the CUHII. The larger the area of the land use change trajectory, the greater the impact on the CUHII. 5. Conclusion This paper quantified the effects of urbanization on the UHI through the method of land use change trajectory and the indices of CUHII in Xiamen, China. The method of land use change trajectory was developed to analyze the land use change types during urbanization and provide the spatial calculation units. The index of CUHII was built to quantify the effects of each land use change trajectory on the UHI during urbanization. The results demonstrated that: (1) The method of land use change trajectory and the indicator of CUHII provided a tool for quantifying the effects of urbanization on the UHI. Among the urbanization trajectories, the most significant transformation was from farmland to built-up area. (2) The study area presented a continuous increase of UHII from 0.83 °C in
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
471
Table 3 The average LSTs of land use change trajectories in different years (°C). Trajectory
1987–01–17
1992–01–15
1997–01–12
2002–01–02
2007–01–08
22224 44444 22444 22244 24444 33334 22544 22524 12444 25444 22254 12224 14444 22554 11114 34444 33344 12244 54444 33354 22234
19.19 18.55 19.04 19.06 18.89 17.10 19.23 19.30 18.23 19.17 19.34 18.58 18.26 19.45 17.21 17.91 16.91 18.37 17.80 17.46 18.74
12.03 11.42 11.94 12.05 11.48 11.39 12.24 12.35 11.54 11.05 12.31 11.59 11.03 12.33 9.00 11.32 11.28 11.67 11.02 10.63 11.93
15.88 15.16 15.31 15.71 15.23 13.82 14.71 14.85 14.94 15.22 15.97 15.49 14.76 14.57 13.48 14.96 13.71 15.51 14.94 13.18 15.08
18.74 17.99 18.08 18.06 17.99 16.26 18.34 18.70 17.66 18.22 17.78 18.42 17.51 17.75 16.29 17.87 17.24 17.71 17.81 17.19 17.01
15.71 15.58 15.81 15.79 15.70 14.91 16.46 16.35 15.27 16.11 16.24 15.54 14.90 16.15 13.30 15.54 15.68 15.42 15.45 15.37 15.65
Table 4 The CUHII of urbanization land use change trajectories (%). Trajectory
1987–1–17
1992–1–15
1997–1–12
2002–1–2
2007–1–8
22224 44444 22444 22244 24444 33334 22544 22524 12444 25444 22254 12224 14444 22554 11114 34444 33344 12244 54444 33354 22234
–* 55.00 – – – – – – – – – – – – – – – – – – –
– 38.99 – – 13.18 – – – – – – – 3.08 – – 1.94 – – 1.57 – –
– 27.57 17.84 – 9.31 – – – 2.80 2.53 – – 2.20 – – 1.37 – – 1.14 – –
– 21.77 14.01 10.11 7.32 – 4.87 – 2.20 2.02 – – 1.74 – – 1.09 0.98 0.94 0.90 – –
21.98 13.03 8.47 6.11 4.41 3.29 3.02 2.85 1.32 1.23 1.20 1.07 1.02 0.79 0.60 0.65 0.62 0.57 0.54 0.53 0.51
*
means land is not covered by urban in this year.
1987 to 2.14 °C in 2007; an exception was in 1992, which might be attributed to the cold wave that year (3) The maximum CUHII before 2002 was the existing urban area of trajectory NO. 44444, but the value decreased from 55.00% in 1987 to 13.03% in 2007. In 2007, it was replaced by trajectory NO. 22224 the contribution of which was 21.98%. The areas of the land use change trajectories played a significantly more important role on the CUHII, than the minor LST differences among them. Therefore, urbanization significantly influenced the UHII and weakened the CUHII of the existing urban area.
This research provides a new method to study quantitatively the effects of urbanization on the UHI based on the method of land use change trajectory. Two areas need to be addressed in future research. First, the classification accuracy was a key factor in the land use change trajectory method, especially in the long time-series analysis. The relationship between classification errors and the reliability of the research should be further discussed. Second, there were many other indicators to indicate UHI effects in addition to the UHII, such as magnitude, hot island area and range. (Schwarz et al., 2011; Schwarz et al., 2012). These
472
H. Feng et al. / Advances in Space Research 53 (2014) 463–473
Fig. 8. The CUHII map of land use change trajectories in 2007.
diverse indicators and their similarities and differences should also be the subject of further study when used to analyze the UHI. Acknowledgments This study was supported by the National Natural Science Foundation of China (No. 41371392), the National Key Technology R&D Program (2012BAC21B03), the Knowledge Innovation Programs of Chinese Academy of Sciences (IUEQN-2012-01), and also the Natural Science Foundation of Fujian Province (2013J05062) and Ningbo City (2013A610180). References Amiri, R., Weng, Q., Alimohammadi, A., Alavipanah, S.K., 2009. Spatial–temporal dynamics of land surface temperature in relation to
fractional vegetation cover and land use/cover in the Tabriz urban area. Iran. Remote Sens. Environ. 113 (12), 2606–2617. Artis, D.A., Carnahan, W.H., 1982. Survey of emissivity variability in thermography of urban areas. Remote Sens. Environ. 12 (4), 313–329. Cairns, M.A., Haggerty, P.K., Alvarez, R., Jong, B.H.J.D., Olmsted, I., 2000. Tropical Mexico’s recent land-use change: a region’s contribution to the global carbon cycle. Ecol. Appl. 10 (5), 1426–1441. Changnon, S.A., Kunkel, K.E., Reinke, B.C., 1996. Impacts and responses to the 1995 heat wave: a call to action. Bull. Am. Meteorol. Soc. 77 (7), 1497–1506. Chen, X.-L., Zhao, H.-M., Li, P.-X., Yin, Z.-Y., 2006. Remote sensing image-based analysis of the relationship between urban heat island and land use/cover changes. Remote Sen. of Environ. 104 (2), 133–146. China Meteorological Data Sharing Service System. (2011). “http:// www.cma.gov.cn/2011qxfw/2011qsjgx/.”. Gallo, K.P., Easterling, D.R., Peterson, T.C., 1996. The influence of land use/land cover on climatological values of the Diurnal temperature range. J. Clim. 9, 2941–2944. Gallo, K.P., McNab, A.L., Karl, T.R., Brown, J.F., Hood, J.J., Tarpley, J.D., 1993. The Use of NOAA AVHRR Data for Assessment of the Urban Heat Island Effect. J. Appl. Meteorol. 32 (5), 899–908.
H. Feng et al. / Advances in Space Research 53 (2014) 463–473 Grimm, N.B., Faeth, S.H., Golubiewski, N.E., Redman, C.L., Wu, J., Bai, X., Briggs, M., 2008. Global change and the ecology of cities. Science 319, 756–760. Houghton, R.A., Goodale, C.L., 2004. Effects of land-use change on the carbon balance of terrestrial ecosystems. Geophys. Monogr. 153 (1–4), 85–98. Howard, L. 1818. The Climate of London: deduced from Meteorological observations, made at different places in the neighbourhood of the metropolis. London, W. Phillips, sold also by J. and A. Arch. Huang, L., Huang, J., (2011). Preliminary study on urban heat island effect in Xiamen during 2001–2007. 2011 International Conference on Remote Sensing, Environment and Transportation Engineering (RSETE): 3330–3333. Imhoff, M.L., Zhang, P., Wolfe, R.E., Bounoua, L., 2010. Remote sensing of the urban heat island effect across biomes in the continental USA. Remote Sens. Environ. 114 (3), 504–513. Jin, M., Dickinson, R.E., Zhang, D., 2005. The footprint of urban areas on global climate as characterized by MODIS. J. Clim. 18 (10), 1551– 1565. Jusuf, S.K., Wong, N.H., Hagen, E., et al., 2007. The influence of land use on the urban heat island in Singapore. Habitat Int. 31 (2), 232–242. Kim, Y.-H., Baik, Jong.-Jin., 2002. Maximum urban heat island intensity in Seoul. J. Appl. Meteorol. Climatol. 41, 651–659. Kolokotroni, M., Ren, X., Davies, M., Mavrogianni, A., 2012. London’s urban heat island: impact on current and future energy consumption in office buildings. Energy Build. 47, 302–311. Konopacki, S., Akbari, H., (2002). Energy savings for heat-island reduction strategies in Chicago and Houston (including updates for Baton Rouge, Sacramento, and Salt Lake City). L. B. N. Laboratory. Lambin, E.F., 1997. Modelling and monitoring land-cover change processes in tropical regions. Prog. Phys. Geogr. 21 (3), 375–393. Li, J., Song, C., Cao, L., Zhu, F., Meng, X., Wu, J., 2011. Impacts of landscape structure on surface urban heat islands: A case study of Shanghai. China Remote Sens. Environ. 115 (12), 3249–3263. Liu, H., Zhou, Q., 2005. Developing urban growth predictions from spatial indicators based on multi-temporal images. Comput. Environ. Urban Syst. 29, 580–594. Memon, R.A., Leung, D.Y.C., Liu, C.-H., 2009. An investigation of urban heat island intensity (UHII) as an indicator of urban heating. Atmos. Res. 94 (3), 491–500. Mertens, B., Lambin, E.F., 2000. Land-cover-change trajectories in Southern Cameroon. Ann. Assoc. Am. Geogr. 90 (3), 467–494. Onishi, A., Cao, X., Ito, T., Shi, F., Imura, H., 2010. Evaluating the potential for urban heat-island mitigation by greening parking lots. Urban For. Urban Greening 9 (4), 323–332. Pauleit, S., Ennos, R., Golding, Y., 2005. Modeling the environmental impacts of urban land use and land cover change—a study in Merseyside. UK. Landscape Urban Plan. 71 (2–4), 295–310. Petit, C.C., Lambin, E.F., 2001. Integration of multi-source remote sensing data for land cover change detection. Int. J. Geog. Inf. Sci. 15 (8), 785–803. Rao, P.K., 1972. Remote sensing of urban“heat islands”from an environmental satellite. Bull. Am. Meteorol. Soc. 53, 647–648. Rizwan, A.M., Dennis, L.Y.C., Liu, C., 2008. A review on the generation, determination and mitigation of urban heat island. J. Environ. Sci. 20 (1), 120–128.
473
Rosenfeld, A.H., Akbari, H., Romm, J.J., Pomerantz, M., 1998. Cool communities: strategies for heat island mitigation and smog reduction. Energy Build. 28 (1), 51–62. Schwarz, N., Lautenbach, S., Seppelt, R., 2011. Exploring indicators for quantifying surface urban heat islands of European cities with MODIS land surface temperatures. Remote Sens. Environ. 115 (12), 3175– 3186. Schwarz, N., Schlink, U., Franck, U., Großmannc, K., 2012. Relationship of landsurface and airtemperatures and its implications for quantifying urban heat island indicators–an application for the city of Leipzig (Germany). Ecol. Ind. 18, 693–704. Sobrino, J.A., Jime´nez-Mun˜oz, J.C., Paolini, L., 2004. Land surface temperature retrieval from LANDSAT TM 5. Remote Sens. Environ. 90 (4), 434–440. Sobrino, J.A., Romaguera, M., 2004. Land surface temperature retrieval from MSG1-SEVIRI data. Remote Sens. Environ. 92 (2), 247–254. Wang, D., Gong, J., Chen, L., Zhang, L., Song, Y., Yue, Y., 2012. Spatiotemporal pattern analysis of land use/cover change trajectories in Xihe watershed. Int. J. Appl. Earth Obs. Geoinf. 14 (1), 12–21. Wang, D., Gong, J., Chen, L., Zhang, L., Song, Y., Yue, Y., 2013. Comparative analysis of land use/cover change trajectories and their driving forces in two small watersheds in the western Loess Plateau of China. Obs. Geoinf. 21, 241–252. Weng, Q., 2001. A remote sensing–GIS evaluation of urban expansion and its impact on surface temperature in the Zhujiang Delta. China Int. J. Remote Sens. 22 (10), 1999–2014. Weng, Q., Lu, D., Schubring, J., 2004. Estimation of land surface temperature–vegetation abundance relationship for urban heat island studies. Remote Sens. Environ. 89 (4), 467–483. Weng, Q., Rajasekar, U., Hu, X., 2011. Modeling urban heat islands with multi-temporal ASTER images. IEEE Trans. Geosci. Remote Sens. 49 (10), 4080–4089. Xiamen Bureau of Statistics (2008). Year book of Xiamen special economic zone, Beijing:China Statistics Press (in Chinese). Xu, H., Chen, B., 2004. Remote sensing of the urban heat island and its changes in Xiamen City of SE China. J. Environ. Sci. 16 (2), 276-218. Yuan, F., Bauer, M.E., 2007. Comparison of impervious surface area and normalized difference vegetation index as indicators of surface urban heat island effects in Landsat imagery. Remote Sens. Environ. 106 (3), 375–386. Zhang, Y., Odeh, I.O.A., Han, C., 2009. Bi-temporal characterization of land surface temperature in relation to impervious surface area, NDVI and NDBI, using a sub-pixel image analysis. Int. J. Appl. Earth Obs. Geoinf. 11 (4), 256–264. Zhao, J., Dai, D., Lin, T., Tang, L., 2010a. Rapid urbanisation, ecological effects and sustainable city construction in Xiamen. Int. J. Sustainable Dev. World Ecol. 17 (4), 271–272. Zhao, X., Huang, J., Ye, H., Wang, K., Qiu, Q., 2010b. Spatiotemporal changes of the urban heat island of a coastal city in the context of urbanisation. Int. J. Sustainable Dev. World Ecol. 17 (4), 311–316. Zhou, Q., Li, B., Kurban, A., 2008. Spatial pattern analysis of land cover change trajectories in Tarm Basin, northwest China. Int. J. Remote Sens. 29 (19), 5495–5509. Zhou, W., Huang, G., Cadenasso, M.L., 2011. Does spatial configuration matter? understanding the effects of land cover pattern on land surface temperature in urban landscapes. Landscape Urban Plan. 102 (1), 54–63.