Hu2015_TAAC
ORIGINAL PAPER
Assessing surface albedo change and its induced radiation budget under rapid urbanization with Landsat and GLASS data
Yonghong Hu&Gensuo Jia&Christine Pohl&
Xiaoxuan Zhang&John van Genderen
Received:16June2014/Accepted:19January2015
#Springer-Verlag Wien2015
Abstract Radiative forcing(RF)induced by land use(mainly surface albedo)change is still not well understood in climate change science,especially the effects of changes in urban albedo due to rapid urbanization on the urban radiation bud-get.In this study,a modified RF derivation approach based on Landsat images was used to quantify changes in the solar radiation budget induced by variations in surface albedo in Beijing from2001to2009.Field radiation records from a Beijing meteorological station were used to identify changes in RF at the local level.There has been rapid urban expansion over the last decade,with the urban land area increasing at about3.3%annually from2001to2009.This has modified three-dimensional urban surface properties,resulting in lower albedo due to complex building configurations of urban cen-ters and higher albedo on flat surfaces of suburban areas and cropland.There was greater solar radiation(6.93×108W)in the urban center in2009than in2001.However,large crop-land and urban fringe areas caused less solar radiation absorp-tion.RF increased with distance from the urban center(less than14km)and with greater urbanization,with the greatest value being0.41W/m2.The solar radiation budget in urban areas was believed to be mainly influenced by urban structural changes in the horizontal and vertical directions.Overall,the results presented herein indicate that cumulative urbanization impacts on the natural radiation budget could evolve into an important driver of local climate change.
1Introduction
Urbanization has changed land surface properties and resulted in alteration of biogeochemical cycles and possible climate feedback on local,regional,and global scales(Foley et al. 2005;Kalnay and Cai2003).As centers of economic devel-opment and population growth,urban areas consume large amounts of natural resources from surrounding areas,thereby affecting land use and land cover change well beyond their geographic boundaries(Grimm et al.2008).Urban expansion and decreasing natural(forest)or anthropogenic(cropland) surfaces are generally considered important factors influenc-ing land surface characterization.Furthermore,these process-es may strongly modify land-atmosphere interactions within a region,thereby affecting the local or regional climate(Ezber et al.2007;Kalnay and Cai2003).
Radiative forcing has been used as a measure of the effects of a factor on the balance of incoming and outgo-ing energy in land-atmosphere systems to assess
the Y.Hu(*)
:X.Zhang
Key Laboratory of Digital Earth Science,Institute of Remote sensing
and Digital Earth,Chinese Academy of Sciences,Beijing100094,
China
e-mail:huyh01@https://www.sodocs.net/doc/983472833.html,
X.Zhang
e-mail:zhangxiaoxuanzxx@https://www.sodocs.net/doc/983472833.html,
G.Jia
Key Laboratory of Regional Climate-Environment for East Asia,
Institute of Atmospheric Physics,Chinese Academy of Sciences,
Beijing100029,China
e-mail:jiong@https://www.sodocs.net/doc/983472833.html,
C.Pohl
Institute of Geospatial Science and Technology,Universiti Teknologi
Malaysia,81310Johor Bahru,Malaysia
e-mail:c.pohl@utm.my
X.Zhang
College of Geosciences,China University of Petroleum,
Qingdao266580,China
J.van Genderen
Geospatial Information Science Research Centre,Universiti Putra
Malaysia,43400Serdang,Malaysia
e-mail:genderen@itc.nl
Theor Appl Climatol
DOI10.1007/s00704-015-1385-2
relative influence of various human activities on climate change(Forster et al.2007).Land use change,which in-volves aerosol release and surface albedo change,is one of the principal human activities that can perturb the ra-diative balance of the Earth(Myhre and Myhre2003). Many studies have been conducted to investigate the ef-fects of biomass burning and aerosol change in altering atmospheric radiative forcing(RF)through direct or indi-rect radiation absorption,scattering,and emission.Studies of the influence of surface albedo on RF change based on changes in the shortwave radiation budget have mainly focused on potential natural vegetation(PNV)replace-ment by cropland at the global scale(Chung et al.2010; Forster et al.2007).Anthropogenic RF time series from pre-industrial times through to2010have indicated a de-crease(?0.07W/m2)of RF in response to surface albedo change,primarily owing to mid-latitude agricultural vari-ations in the northern hemisphere(Skeie et al.2011). Variations in radiative forcing in different vegetation datasets range from?0.6to0.5W/m2and have a nonlin-ear relationship with vegetation and surface albedo chang-es(Myhre and Myhre2003).Radiative forcing induced by deforestation and desertification on a global scale has in-fluenced climate change.With more than half the Earth’s population living in cities,urbanization has become an important component of global change.Indeed,urbaniza-tion deeply alters the land surface properties of cities and their surroundings,in addition to modifying albedo and its climate feedback(Grimm et al.2008;Kondo et al.2001; Menon et al.2010).
Feedback of urban albedo change on the land radiation budget has been used to quantify the effects of increasing outgoing radiation on climate change.Increasing surface al-bedo by changing the material comprising the urban canopy has been proposed as a method of mitigating urban heat islands or improving urban energy savings(Akbari et al. 2009;Jacobson and Ten Hoeve2012).Increasing urban albe-do by0.01causes an RF of?1.27W/m2,which can be achieved by replacing existing roofs with white ones in trop-ical and temperate urban areas,and application of such chang-es would be equivalent to a one-time offset of44Gt of emitted CO2(Akbari et al.2009).The negative RF caused by increas-ing the urban canopy albedo was reported to be?1.63W/m2 during summer in global land(Menon et al.2010).Climate models indicate that,in the long term,increasing urban albedo would lead to global cooling by0.01to0.07K based on moderate resolution imaging spectroradiometer(MODIS) and the Global Rural–Urban Mapping Project urban land covers(Akbari et al.2012).Simulation studies have also shown cooling effects of albedo on air temperature at local or regional scales(Jacobson and Ten Hoeve2012).Most stud-ies examined this effect through model simulation,rather than by exploring the actual urbanization effect on RF change using in situ measurements(V an Curen2012).Urbanization is a complex surface transformation of land cover types that alters inner urban structures,surface properties in urban fringe areas,and ecological or service functions in urban surround-ings(Seto et al.2012).Changes in urban albedo during ex-pansion primarily occur in spatially heterogeneous areas,in contrast to the homogenous changes that occur in models. Moreover,cooling induced by traditional urban canopy mate-rial and its replacement vary worldwide owing to cultural and economic development levels.Defining and quantifying feed-back of changes in urban albedo in response to urbanization could furnish more valuable case information for model stud-ies and practical advice for local governments dealing with urban energy saving and urban planning.
China has undergone rapid urbanization over the past two decades,with the urbanization rate increasing from26.4%in 1990to47.5%in2010.Few other countries have experienced such a large change over a short period.Urbanization,espe-cially small town development and conversion of large areas of urban surroundings to mixtures of urban–rural land,has resulted in urban patches with many surface fragments of nat-ural vegetation,cropland,and impervious land.In this study, Beijing was selected to examine the effects of urbanization on the regional radiation budget.The specific goals of this study were to(1)examine urbanization and related urban albedo change during the last decade using Landsat images and(2) quantify RF changes induced by urban albedo change.The data were then used to evaluate the feedback of urbanization on local or regional radiation budgets and urban climate effects.
2Study area
With a land area of1.6×104km2,Beijing(Fig.1)is one of the largest cities in China and has undergone rapid urbanization during the last few decades.The urban pop-ulation increased from7.98million in1990to16.86mil-lion in2010,giving an urbanization rate of86%. According to census statistics,the built-up area in Beijing has rapidly expanded at a rate of6.3%during the last two decades(Beijing Municipal Statistical Bureau 2010).As part of urbanization,city-wide ring roads were constructed to organize urban traffic and lead urban de-velopment.The region is influenced by the East Asian monsoon,with an average annual total solar radiation greater than450kJ/cm2.The hottest monthly air temper-ature exceeds26°C.Most mountains in the west and north are mainly covered by temperate deciduous broad-leaf https://www.sodocs.net/doc/983472833.html,rge plains have been developed in central and southern areas of the region at elevations less than 100m to meet the agricultural needs of the city(Fig.2).
Y.Hu et al.
3Data and methods 3.1Data
Landsat TM images from 31August 2001and 22September 2009(Path 123,Row 32)from the Institute of Remote Sensing and Digital Earth and United States Geological Survey were used to examine urban land use and surface al-bedo change.Surface albedo products from 2001to 2010were provided by the Global Land Surface Satellites (GLASS)project.With accuracy similar to MODIS albedo products,GLASS albedo products are gapless and continuous datasets with a temporal resolution of 8days and spatial res-olution of 1km (Liu et al.2013).Here,the averaged surface albedo derived from GLASS datasets was used to provide the average conditions of surface albedo from 2001to 2010to examine albedo change under urbanization and analyze its influence on the urban radiation budget.Additionally,
albedo
Fig.1Beijing mega city (left figure )and its topography variation (right figure ).The false color figure (left )is the composite of band 4,3,2from Landsat TM images in Sep 2009,and six ring roads of Beijing (closed curve in solid line in two figures)showed the development of urban expansion in
Beijing
Fig.2Comparison of urban land use a derived from Landsat image and high resolution true color image b from Google Earth in 2009.These test regions locate at near woodland,little town,urban center,and suburban,respectively.The land use type in red color in upper figures is urban and built-up,while yellow color is natural or planted vegetation,including forest,cropland,and grassland
Albedo and radiation change under urbanization
derived from Landsat images provided additional details in 30-m resolution for2001and2010,which could be used to identify variations in albedo in areas that have undergone dif-ferent levels of urbanization,as well as the changes in the surface radiation budget caused by these variations.
3.2Urbanization intensity derivation
A supervised classification method(Maximum Likelihood Classifier)was used to estimate urban and built-up land,as well as cropland,woodland,water,and other land in the study area based on Landsat TM images.According to visual iden-tification and Google Earth surveys,training samples (2001:10,086training pixels,2009:8162training pixels)were selected as input for the classifier of land use identification. The normalized difference vegetation index(NDVI)proc-essed from spectral reflectance of bands3and4(Y uan and Bauer2007)was introduced to improve spectral information during classification.To achieve optimal urban land use re-sults,cyclic validation of land use classification was conduct-ed to ensure the accuracy of land use classification.Four test regions(1×1km)were used to check the accuracy of that classification(Fig.2).If the average accuracy of the classifi-cation was less than90%,the training samples were checked for consistency between the training samples and classifica-tion results for each land cover type.Samples with inconsis-tent land cover types were removed,and the classification was repeated until the results attained the target accuracy.
Urbanization rate is usually quantified by the ratio of urban to total population and the percentage of urban to total land area(Hu and Jia2010).Changes in urbanization rate from 2001to2009were quantified based on the urbanization inten-sity(UI):
UI?U i?U j;e1T
where U i and U j are fractional covers of urban land in periods i and j,respectively.In this study,U i and U j are percentages of urban or built-up land use relative to all land use types in2001 and2009.
3.3Surface albedo derivation
3.3.1Albedo derived from Landsat images
Surface albedo indicates the percentage of solar radiation reflected by a body or surface to the amount incident upon it,which is considered a critical variable in climate change analysis and modeling(Roesch and Roeckner2006;Winton 2006).With about99%of energy occurring in the visible and infrared spectral region,shortwave radiation is the main type of solar radiation affecting climate.Shortwave albedo greatly alters the land surface energy balance,which can be detected by remote sensing methods(Csiszar and Gutman1999). Several algorithms have been proposed for derivation of broadband albedo using different band combinations of Landsat TM/ETM+(Liang2003;Liang2001).We used the direct retrieval method(Liang2003)to derive surface albedo and identify its changes over the last decade.This method is considered easier and more accurate for estimating surface albedo than physical inversion methods with limited observa-tions under normal conditions(Liang et al.2005).Those methods link top of atmosphere(TOA)directional reflectance with surface broadband albedo for a variety of surface and atmospheric conditions using radiative transfer model simula-tions such as atmospheric conditions,aerosol optical thick-ness,solar zenith angle,and azimuth angle.In our study,sur-face albedo in2001and2009was derived from Landsat im-ages using a direct retrieval algorithm for the study area to examine its spatial albedo change under rapid urbanization.
3.3.2Albedo from GLASS albedo products
To enable comparison of albedo results from Landsat im-ages,GLASS land surface products were used to derive surface albedo in September(Julian days249,257,265, and273)from2001to2010.The means and standard deviations of surface albedo were calculated according to dataset quality control information.Based on UI infor-mation(Fig.3),we further quantified the differences in albedo trends from2001to2010in areas of high,mod-erate,and low urbanization.
3.4Radiative forcing calculation
3.4.1Radiative forcing calculation based on station measurements
To investigate changes in the radiation budget induced by urban surface albedo at the local scale,meteorological records were used to examine the RF(Van Curen2012).First,the absorption of incident shortwave radiation was calculated ac-cording to the ratio between surface incident shortwave radi-ation from meteorological station data and theoretical TOA solar radiation.The latter radiation is a function of solar out-put,field latitude,and geometry of the Earth and Sun.This value can be derived using the following equation(Liou 2002):
S?S0
r0
r
2Z
?H
H
cosθ0
dh
2π
;e4T
where S0is the solar constant at mean Earth-Sun distance r0;θ0is the solar zenith angle;r is the Earth-Sun distance, which varies throughout the year according to the
Y.Hu et al.
elliptical orbit;h is the hour angle;and H is a half day (i.e.,from sunrise to noon or noon to sunset).
Second,the mean RF of outgoing shortwave radiation gen-erated by a specified increase of surface albedo is calculated by
RF ?0:01?SW ?f ;
e5T
where SW is the surface incident shortwave radiation.For comparison with remote sensing results,an increase in albedo of 0.01was considered the baseline change.Daily radiation records of the Beijing meteorological sta-tion in 2001and 2009were used to examine the mean annual RF change during the last decade.Because climate records respond sensitively to land surface changes in response to urbanization (He et al.2013),surface albedo change from 2001to 2009around the Beijing station at a spatial scale of 1×1km was used to examine the relation-ship between RF and albedo.
3.4.2Radiative forcing calculation based on remote sensing images
In a previous study,RF induced by surface albedo change was quantified according to an Earth energy budget ob-tained by satellite observations (Akbari et al.2009).While that study used the results of the global energy budget from RF calculations,the calculations were mod-ified with the latest global energy budget information in the present study (Trenberth et al.2009).According to a method proposed by Akbari et al.,RF per unit area can be quantified as follows:
1.Percentage of radiation absorbed by the atmosphere 341?79eT0f t23f =1?f eT?78
e2T
where (341–79)×f is the absorption of incident shortwave radiation (W/m 2),23f /(1?f )is the absorption of reflected solar radiation (W/m 2),and f is the percentage of radiation absorbed by the atmosphere for downward and reflected radiation,which is solved from the above equation as 26.6%.This value is very close to the fraction of radiation absorbed by the atmo-sphere (26%)reported by Akbari et al.(2009).2.Radiative forcing of surface albedo change 262?1?f eT=100?1:41;
e3T
where 262×(1?f )is the incident shortwave radiation for heating land surfaces and 1.41W/m 2is the RF from a surface reflectance change of 0.01.If surface albedo increases by 0.01,the RF is ?1.41W/m 2.This value was determined based on the annual average global insolation and cloud cover.To further identify RF values in Beijing,the annual downward radiation calculated from daily solar radiation measurements records at Beijing station was used to determine the average change in RF.The results revealed that the change was ap-proaching ?1.27W/m 2,which is in the range of ?1.27to ?1.63W/m 2for RF estimation on a large scale (Akbari et al.2009;Menon et al.2010).Therefore,we combined this value with changes in local urban albedo derived from Landsat im-ages to estimate the regional impact of urbanization on the radiation
budget.
Fig.3Urban land use change from 2001and 2009in Beijing mega https://www.sodocs.net/doc/983472833.html,nd use properties in 2001(a )and 2009(b )were
derived from Landsat TM images.Urbanization level was divided into three different levels (c )according to land use change
Albedo and radiation change under urbanization
4Results
4.1Urbanization trends
To examine the impact of urbanization on surface properties, we quantified urban development in Beijing at two spatial scales,the urban center(inner5th Ring Road)and the greater Beijing area.In general,trends regarding land use types in greater Beijing differed greatly(Fig.3).Urban or built-up land rapidly increased by3.3%annually from2001to2009, whereas vegetation cover declined by2.4%annually.The surface water area also decreased by about30%over the last 10years.The spatial pattern of urban land use change over the period indicates that the principal urbanization was around the 5th Ring Road in the southern,northern,and eastern regions. Satellite towns also rapidly developed in the Changping, Shunyi,Tongzhou,and Daxing districts(Fig.3).
We further examined land use change from the urban center to suburban areas(mainly the inner6th Ring Road)to evaluate the impact of urbanization on inner urban structures or land-scapes.The results indicated that the percentage of urban land in the urban center reached90%.The ring roads of Beijing have provided convenient urban transportation and signifi-cantly altered urban structures,resulting in the development of important business centers and extensive urban settlements. Here,the level of urbanization was defined as the percentage of urban land in specific zones,and changes in this level were examined in specific zones divided by ring road boundaries from2001to2009.As expected,there were obvious differ-ences in urban land cover from the urban center to rural re-gions,with fractional cover of urban land increasing by0.3, 0.6,1.3,4.7,and20%in the inner zones of the2nd,2nd–3rd, 3rd–4th,4th–5th,and5th–6th ring roads,respectively.These findings indicate that suburban or urban fringes are the areas undergoing the fastest change,with existing land being con-verted to urban land.
4.2Urban albedo change
GLASS surface albedo was used to explore statistical trends in albedo change under urbanization from2001to2010,and the albedo derived from Landsat images gave relatively detailed spatial patterns during urbanization.Decreased trends of sur-face albedo were detected in the urban center(Fig.4b)upon analysis of time series of the GLASS albedo,with surface albedo in2001being0.008higher than in2010.The spatial pattern of the10-year average surface albedo indicated that it was spatially heterogeneous,and there was a general pattern of higher albedo in cropland in the southeast portion of the study area.Lower albedo was observed in the urban center and northwest woodland area,except for the area surrounding Guanting Reservoir(Fig.4a).High albedo regions of cropland were found at100m a.s.l.(Figs.1,3,and4).To investigate albedo response to urbanization,we further quantified and compared differences in the average albedo in Beijing on a gradient from the urban center to natural surfaces in2001and 2009(Table1).Changes in urban albedo were influenced by urbanization.Theoretically,this influence involved two pro-cesses:(1)decreasing albedo from increased solar radiation absorption in dense building areas,and(2)increasing albedo from flat urban surfaces or urban canopy building materials of high reflectance.In general,surface albedo in2001and2009 increased from the inner2nd to the6th Ring Road,which indicated that the urban core area of dense buildings had lower solar radiation reflectance than other regions.For example,we estimated the inner2nd Ring Road average albedo to be0.120 and that of the5th–6th Ring Road to be0.145in2009. Statistical analysis of the relationship between urban albedo and urban land use was carried out for2009(Fig.4d).Except for other land(mainly bareland),urban areas had the highest surface reflectance(average albedo=0.144),whereas water bodies captured more solar radiation than other land use types, with an average albedo of0.069.
4.3Relationship between urbanization and surface radiation budget
RF indicates the solar radiation budget of the Earth-atmosphere system.Positive forcing(more incoming energy) warms the system,whereas negative forcing(more outgoing energy)cools it.Here,the relationship between urbanization and RF was quantified by Landsat data to examine urban climate effects.
4.3.1Surface radiation budget variation for three urbanization levels
Three different urbanization levels were selected according to land use change over the last decade(Fig.3)to examine RF variation under different urbanization intensities.These in-cluded the urban center(high level,urban land=90%in 2009),main cropland area(moderate level,urban land= 42%in2009),and main woodland area(low level,urban land=5%in2009).To investigate the mean changes in the radiation budget related to urbanization,heterogeneous urban distributions in different sub-areas were not included.We ex-amined the change in the surface radiation budget induced by changes in albedo relative to the baseline RF of?1.27W/m2 for three sub-areas using the GLASS datasets(Table2). Generally,changes in RF produced by variations in albedo decreased from high to low urbanization levels.A positive RF of1.016W/m2was detected in high-level regions owing to a decrease in urban albedo of?0.008from2001to2010, which generated a radiation absorption increase of6.93×108W.A reduced positive RF of0.902W/m2was detected in the moderate-level area,and further analysis showed that
Y.Hu et al.
RF reached 0.41W/m 2.An increased radiation absorption of 3.72×109W was detected in the low-level area,possibly be-cause of woodland management protection policies that have been implemented in recent decades.Theoretically,urban al-bedo decline influenced by high urbanization levels may cap-ture more solar radiation than that used to heat urban surfaces,which results in urban heat islands.In contrast to urban areas,such energy absorption in forest areas may be used during
physiological activities of vegetation,such as transpiration and photosynthesis.
4.3.2Relationship between RF and UI
To understand the relationship between urbanization and RF change,an experiment was conducted to quantify changes in RF across urbanization levels in the study area.A point at Tiananmen Square (longitude 116.391321°,latitude 39.903219°)in the urban center was used to create buffer regions every kilometer from the urban center to rural areas.These areas contained varying urbanization levels and RF values.RF variation with distance and UI were divided into two stages (Fig.5):(1)RF trends of linear increase from ?0.35to 0.40W/m 2were detected through linear regression (r 2=0.96)along the urban –rural direction for 1to 14km (around the 5th Ring Road).UI increased from 0.1to 4.2%in this highly developed urban area (urban land percentage ≥95%)in 2009,indicating increased urban building density and more complex urban canopy structure.(2)RF decreased from 0.36to 0.21W/m 2as distance increased from 15to 30km.The UI
Table 1Averaged albedo variation in different urban regions of Beijing from Landsat data Urban regions
20092001Albedo
Std Albedo Std 2nd ring road 0.12040.020.11970.01892–3rd ring road 0.12390.02270.12560.02153–4th ring road 0.13260.02370.13740.02304–5th ring road 0.14150.02380.14390.02215–6th ring road
0.1446
0.0247
0.1457
0.229
Fig.4Spatial-temporal change of surface albedo in the study area.a Averaged surface albedo and b mean albedo change in September from 2001to 2010from GLASS datasets;c spatial pattern of urban albedo and d its variation over different land cover types from Landsat data in 2009
Albedo and radiation change under urbanization
in this region increased from 5.5to 16.4%,which showed a decrease in urban land percentage of 64%in 2009.These findings indicate that the surrounding regions were rapidly covered by urban land during urbanization.The new buildings had lower density and complexity than in the urban center,which may have led to less radiation absorption.
5Discussion
5.1Urban center captures more solar radiation than urban fringe
Changes in urban albedo and urban radiation flux in response to various urban building properties have been detected in earlier studies.Simulation studies have shown that increasing building height and decreasing the uniformity of building height distribution greatly modifies surface albedo.A mean albedo decrease of greater than 0.15as building height in-creased from 5to 50m for the same building coverage has been reported (Aida 1982;Kondo et al.2001).Moreover,it is known that irregular urban structures capture over 20%more
solar radiation than flat surfaces (Aida 1982;Kondo et al.2001).Theoretically,because solar radiation absorption by dense buildings occurs through multiple reflections,tall build-ings and high-density urban centers usually capture more solar radiation than flat land surfaces such as cropland,grassland,or city squares.This partially explains why there was lower sur-face albedo in the Beijing urban center than in its surroundings (Fig.4).
Three 1×1km samples (Fig.6)from 2009were chosen to compare the urban albedo of different urban canopy configu-rations.These included dense building areas,sparse urban building areas,and flat urban squares.Samples a,b,and c in the 1×1km grid were at the Beijing Olympic Sports Center (center latitude 39.996994°,center longitude 116.390059°),a large residential community near the 3rd Ring Road (center latitude 39.965058°,center longitude 116.411731°),and Beijing Capital International Airport (center latitude:40.078240°,center longitude:116.596886°).These locations had building percentages of 8,92,and 36,respectively.As expected,dense building areas had lower surface albedos (0.12)than sparse building areas (0.15)and flat surfaces (0.2).Surface albedo was altered by building material.
Table 2
Radiation budget induced by surface albedo under different urbanization level from 2001to 2010based on GLASS albedo products
Urbanization level
High level Moderate level Low level Year
200120102001201020012010Average albedo
0.13720.1292
0.14910.142
0.12490.1187
Average albedo change from 2001to 2010?0.008?0.0071?0.0062Average RF caused by albedo change (W/m2) 1.0160.9020.787Radiation absorbed by land surface (W)
6.93×108
5.56×109
3.72×10
9
Fig.5Relationship between change of fractional urban cover from 2001to 2009and its induced radiative forcing.The left figure shows RF change along different distance and fractional urban cover.The right figure
shows the distance in identifying RF change from urban center to rural area with 1-km gap
Y .Hu et al.
White urban canopy surfaces usually had higher albedo,re-ducing radiation capture and mitigating the urban heat island effect (Jacobson and Ten Hoeve 2012;Prado and Ferreira 2005).
According to land use,albedo change and radiation budget analysis,variations in RF resulting from surface changes during urbanization can be controlled by two processes:(1)downward RF trends with increasing flat surfaces in urban fringe areas during natural surface or cropland replacement by urban land owing to increased surface reflectance;and (2)upward RF trends with more complex urban 3D structure configurations.Such changes in surface properties can enhance captured solar energy after multiple reflection of incident radiation.Therefore,radiation budget variations induced by surface changes are a synthetic effect of urban expansion and infrastruc-ture improvement,which intensifies urban heat islands and modifies urban energy consumption,especially in megacities under rapid urbanization such as Beijing.5.2Test case for RF in urban fringe areas using field measurements
RF produced by changes in surface albedo was used in global radiation budgets to examine the cooling effect of white roofs in urban areas (Akbari et al.2009).In the present study,satellite-based surface albedo changes were used to examine the overall information regarding RF change,and less work for testing the variation of RF on a local scale (Van Curen 2012).To investigate the reliability of RF trends on a local scale,field measurements from the Beijing meteorological station were used to derive RF in 2001and 2009.This station is in an urban fringe area (Fig.1),which we considered a typical site for investigating RF change because of its rapid urbanization from 2001to 2009.Representative land cover types in a 1×1km area surrounding the station were investi-gated.A large amount of cropland (42%)was found in 2001(Fig.7b ),whereas urban land accounted for about 80%of the total in 2009(Fig.7e,d ).During the rapid urbanization in recent decades,cropland was replaced by the 5th Ring Road and many buildings.Furthermore,the Landsat TM images indicated that surface albedo decreased by 0.01from 2001to 2009(Fig.7c,f ).
Changes in RF induced by an albedo increase/decrease of 0.01at the Beijing station were used to examine differ-ences in RF from 2001to 2009(Table 3).Because the station is a typical site with increasingly complex building configurations and declining albedo (Fig.7),the annual mean RF in 2009(1.656W/m 2)was greater than in 2001(1.47W/m 2).These findings indicate that more solar radi-ation was captured by land surfaces in 2009(Table 3),which concurs with our deduction from remote sensing data.RF in 2001and 2009showed an apparent seasonal change,with greater values being observed in summer than winter.Most monthly RFs in 2009were greater than those in 2001.The annual and monthly mean RF change from 2001to 2009indicated that radiation absorption by urban surfaces may alter local climate during
urbanization.
Fig.6Urban albedo variation in three typical urban surface (sparse building (a ),dense building (b ),and flat surface (c ))in 2009and its relative urban canopy
configuration in 1×1km samples.Figures in Column 1,2,and 3indicate overall surface
properties,surface albedo change,and filed urban canopy configuration
Albedo and radiation change under urbanization
5.3Implications of urban radiation budget alteration by urban albedo
According to Solomon et al.(2007),radiative forcing mea-sures the influence a factor has on the balance of incoming and outgoing energy in the Earth-atmosphere system.Therefore,the TOA radiation budget as modified by surface albedo change could be considered RF.This budget has been used to quantify the global radiation budget as affected by albedo change from a greater number of white roofs,using average radiative transfer conditions (Akbari et al.2009).Rapid urbanization significantly modifies surface albedo,
and can be used to quantify radiation budgets by employing RF definitions based on an assumption of relatively stable radiative transfer conditions.Therefore,we used the latest average Earth radiation budget of Trenberth et al.(2009)to quantify RF.Here,the RF values based on remote sensing data generally addressed average radiation budgets,which were supported by changes in surface albedo.Therefore,compara-ble albedo results in different urbanization level are important,and the similar spectral reflectance characterization in urban regions made this calculation reasonable (Zha et al.2003).RF values derived from field measurements of radiation compo-nents gave local RF trends under urbanization and induced changes in surface albedo.RF differences from varying urban-ization levels were considered an indication that the radiation budget is influenced by urban land cover and urban albedo changes.The climate zone of California was selected as a representative spatial scale for identifying RF by Van Curen (2012),and the results of our study substantiate their conclu-sion that urbanization information at local scales accounted for RF.
The latest IPCC report,AR5,stated low confidence regard-ing RF modified by surface albedo,with a range of ?0.15±0.1W/m 2.RF in urban areas is also influenced by greenhouse gases,aerosols,and cloud cover,which are usually defined as climate effects at large scales or in long time series (Akbari et al.2009;Menon et al.2010;Skeie et al.2011).Urbanization is considered a synthesis process in the modification of surface properties.Rapid changes in land cover (surface albedo)dur-ing urbanization comprise one of the most important drivers of surface properties and land-atmosphere interaction (Seto et
al.
Fig.7Surface character of test case site in 1×1km surroundings of Beijing meteorological station.Figure (a )and (d )are false color image composited from Landsat TM band 4,3,and 2in 2001and 2009,
respectively,b and e are building configurations from Google Earth and c and f showed the relative albedo change
Table 3Mean RF (W/m2)induced by
decreased surface albedo by 0.01at Beijing station
Month 2001200910.3730.48420.6670.7453 1.638 1.5204 2.025 2.4665 3.511 3.3776 2.003 3.9477 2.455 2.5098 2.134 1.8149 1.385 1.163100.597 1.006110.4770.43612
0.3790.358Annual RF
1.470
1.656
Y .Hu et al.
2012).RF is usually a driver of urban climate change.With the exception of warming trends from climate background, urbanization has been considered a key contributor to air tem-perature warming in the study area(Ren et al.2008;Wang et al.2012;He et al.2013).Therefore,it is important to iden-tify the contributions of the two drivers to understand urban climate effects.Changes in surface albedo may contribute greatly to the urban radiation budget,which is used to under-stand the mechanism of urban climate effects.More reason-able land-atmosphere transfer information and integrated ra-diation data from model simulations and remote sensing ob-servations may improve our understanding of RF changes during future urbanization.
Rapid urbanization in Beijing has converted many urban fringe areas to urban settlements during recent decades(Fig.2; Beijing Municipal Statistical Bureau2010).Given the urban-ization strategy of the Chinese government,this evolution will continue in most cities,especially in small towns.Owing to rapid population increase and business development,these policies will substantially alter land use and land cover in urban surroundings or create satellite towns.This will result in cumulative radiation budgets and urban heat island effects in many cities,which may further influence regional climate change.
6Conclusion
RF induced by surface albedo changes were derived from remote sensing and field measurements to examine radiation budgets under urbanization in Beijing from2001to2009.Our results indicate that Beijing underwent rapid urbanization dur-ing the last decade.This changed the surface albedo,which altered the urban radiation budget and urban climate.There were positive relationships between RF and urbanization level in urban centers.The urban canopy configuration modified the urban radiation budget by reducing albedo in these areas. We also found a higher positive RF in urban centers than in suburban areas,with flat surfaces in cropland and urban–rural fringe areas.Moreover,the Beijing meteorological station data indicated that positive RF was strongly related to urban cano-py configuration and surface albedo.
Integrated analysis of RF,urbanization,and changes in surface albedo demonstrated complex urban radiation budgets under different urbanization levels and relative surface https://www.sodocs.net/doc/983472833.html,ually,anthropogenic heat discharge and radiation absorption by urban buildings increase surface temperature. Data regarding these processes will facilitate analysis of urban heat islands and enable better understanding of their influence on local or regional climates.The results presented herein provide valuable information that will facilitate urban heat island mitigation and urban planning.Radiation inputs in ur-ban regions are important drivers of the urban heat island effect.Therefore,their quantification will improve our under-standing of urbanization impacts on climate change and the contribution of human activities to regional climate warming in China.
Acknowledgments This research was supported by the CAS Strategic Priority Research Program(XDA05090200),the National Natural Sci-ence foundation of China(41405064),and CMA/Henan Key Laboratory of Agrometeorological Support and Applied Technique(AMF201406). References
Aida M(1982)Urban albedo as a function of the urban structure—a model experiment.Boundary-Layer Meteorol23:405–413 Akbari H,Menon S,Rosenfeld A(2009)Global cooling:increasing world-wide urban albedos to offset CO2.Clim Change94:275–286.doi:10.1007/s10584-008-9515-9
Akbari H,Matthews HD,Seto D(2012)The long-term effect of increas-ing the albedo of urban areas.Environ Res Lett7:024004
Beijing Municipal Statistical Bureau(2010)Beijing’s statistical yearbook (Chinese),2010th edn.Beijing Municipal Statistical Bureau,Beijing Chung CE,Ramanthan V,Carmichael G,Tang Y,Adhikary B,Leung LR, Qian Y(2010)Anthropogenic aerosol radiative forcing in Asia de-rived from regional models with atmospheric and aerosol data as-similation.Atmos Chem Phys10:6007–6024.doi:10.5194/acp-10-6007-2010
Csiszar I,Gutman G(1999)Mapping global land surface albedo from NOAA A VHRR.J Geophys Res Atmos104:6215–6228.doi:10.
1029/1998JD200090
Ezber Y,Lutfi Sen O,Kindap T,Karaca M(2007)Climatic effects of urbanization in Istanbul:a statistical and modeling analysis.Int J Climatol27:667–679.doi:10.1002/joc.1420
Foley JA,DeFries R,Asner GP,Barford C,Bonan G,Carpenter SR, Chapin FS,Coe MT,Daily GC,Gibbs HK,Helkowski JH, Holloway T,Howard EA,Kucharik CJ,Monfreda C,Patz JA, Prentice IC,Ramankutty N,Snyder PK(2005)Global consequences of land use.Science309:570–574.doi:10.1126/science.1111772 Forster P et al(2007)Changes in atmospheric constituents and in radia-tive forcing.Cambridge University Press,Cambridge
Grimm NB,Faeth SH,Golubiewski NE,Redman CL,Wu J,Bai X, Briggs JM(2008)Global change and the ecology of cities.
Science319:756–760.doi:10.1126/science.1150195
He Y,Jia G,Hu Y,Zhou Z(2013)Detecting urban warming signals in climate records.Adv Atmos Sci30:1143–1153.doi:10.1007/ s00376-012-2135-3
Hu Y,Jia G(2010)Influence of land use change on urban heat island derived from multi-sensor data.Int J Climatol30:1382–1395.doi:
10.1002/joc.1984
Jacobson MZ,Ten Hoeve JE(2012)Effects of urban surfaces and white roofs on global and regional climate.J Climate25:1028–1044.doi:
10.1175/JCLI-D-11-00032.1
Kalnay E,Cai M(2003)Impact of urbanization and land-use change on climate.Nature423:528–531
Kondo A,Ueno M,Kaga A,Yamaguchi K(2001)The influence of urban canopy configuration on urban albedo.Boundary-Layer Meteorol 100:225–242.doi:10.1023/A:1019243326464
Liang SL(2001)Narrowband to broadband conversions of land surface albedo I:Algorithms.Remote Sens Environ76:213–238.doi:10.
1016/S0034-4257(00)00205-4
Liang SL(2003)A direct algorithm for estimating land surface broadband albedos from MODIS imagery.IEEE Trans Geosci Remote41:136–145.doi:10.1109/TGRS.2002.807751
Albedo and radiation change under urbanization
Liang SL,Stroeve J,Box JE(2005)Mapping daily snow/ice shortwave broadband albedo from Moderate Resolution Imaging Spectroradiometer(MODIS):the improved direct retrieval algo-rithm and validation with Greenland in situ measurement.J Geophys Res Atmos110:D10109.doi:10.1029/2004JD005493 Liou KN(2002)An introduction to atmospheric radiation,vol84.
International Geophysics.Academic Press,Los Angeles
Liu Q,Wang L,Qu Y,Liu N,Liu S,Tang H,Liang S(2013)Preliminary evaluation of the long-term GLASS albedo product.Int J Dig Earth 6:69–95.doi:10.1080/17538947.2013.804601
Menon S,Akbari H,Mahanama S,Sednev I,Levinson R(2010) Radiative forcing and temperature response to changes in urban albedos and associated CO2offsets.Environ Res Lett5:014005 Myhre G,Myhre A(2003)Uncertainties in radiative forcing due to sur-face albedo changes caused by land-use changes.J Climate16: 1511–1524
Prado RTA,Ferreira FL(2005)Measurement of albedo and analysis of its influence the surface temperature of building roof materials.Energy Build37:295–300
Ren G,Zhou Y,Chu Z,Zhou J,Zhang A,Guo J,Liu X(2008) Urbanization effects on observed surface air temperature trends in north china.J Climate21:1333–1348.doi:10.1175/2007JCLI1348.1 Roesch A,Roeckner E(2006)Assessment of snow cover and surface albedo in the ECHAM5general circulation model.J Climate19: 3828–3843.doi:10.1175/JCLI3825.1
Seto KC et al(2012)Urban land teleconnections and sustainability.Proc Natl Acad Sci109:7687–7692.doi:10.1073/pnas.1117622109Skeie RB,Berntsen TK,Myhre G,Tanaka K,Kvalevag MM,Hoyle CR (2011)Anthropogenic radiative forcing time series from pre-industrial times until2010.Atmos Chem Phys11:11827.doi:10.
5194/acp-11-11827-2011
Solomon S et al(2007)Contribution of working group i to the fourth assessment report of the intergovernmental panel on climate change.
United Kingdom,Cambridge
Trenberth KE,Fasullo JT,Kiehl J(2009)Earth’s global energy budget.Bull Am Meteorol Soc90:311–323.doi:10.1175/ 2008BAMS2634.1
Van Curen R(2012)The radiative forcing benefits of B cool roof^con-struction in California:quantifying the climate impacts of building albedo modification.Clim Change112:1071–1083.doi:10.1007/ s10584-011-0250-2
Wang J,Feng J,Yan Z,Hu Y,Jia G(2012)Nested high-resolution model-ing of the impact of urbanization on regional climate in three vast urban agglomerations in China.J Geophys Res Atmos117:D21103.
doi:10.1029/2012JD018226
Winton M(2006)Surface albedo feedback estimates for the AR4climate models.J Climate19:359–365.doi:10.1175/JCLI3624.1
Yuan F,Bauer ME(2007)Comparison of impervious surface area and normalized difference vegetation index as indicators of surface ur-ban heat island effects in Landsat imagery.Remote Sens Environ 106:375–386
Zha Y,Gao J,Ni S(2003)Use of normalized difference built-up index in automatically mapping urban areas from TM imagery.Int J Remote Sens24:583–594.doi:10.1080/01431160304987
Y.Hu et al.