文档库 最新最全的文档下载
当前位置:文档库 › Characteristics of aerosol optical and meteorological three dust events (2005–2010) Beijing

Characteristics of aerosol optical and meteorological three dust events (2005–2010) Beijing

Characteristics of aerosol optical and meteorological three dust events (2005–2010) Beijing
Characteristics of aerosol optical and meteorological three dust events (2005–2010) Beijing

Characteristics of aerosol optical properties and meteorological parameters during three major dust events (2005–2010)over Beijing,China

Chunxiang Cao a ,b ,Sheng Zheng a ,c ,Ramesh P.Singh d ,?

a

State Key Laboratory of Remote Sensing Science,Jointly Sponsored by the Institute of Remote Sensing and Digital Earth of Chinese Academy of Sciences and Beijing Normal University,Datun Road,Chaoyang District,Beijing 100101,PR China b

Center for Applications of Spatial Information Technologies in Public Health,Datun Road,Chaoyang District,Beijing 100101,PR China c

University of Chinese Academy of Sciences,No.19A,Yuquan Road,Shijingshan District,Beijing 100049,PR China d

School of Earth and Environmental Sciences,Schmid College of Science and Technology,Chapman University,One University Drive,Orange,CA 92866,USA

a r t i c l e i n f o a

b s t r a

c t

Article history:

Received 11May 2014

Received in revised form 12July 2014Accepted 17July 2014

Available online 30July 2014Multi-satellite sensors are capable of monitoring transport and characteristics of dust storms and changes in atmospheric parameters along their transport.The present paper discusses aerosol optical properties and meteorological parameters during major dust storm events occurred in the period 2005–2010over Beijing,China.The back trajectory model shows that the dust is transported from the Inner Mongolia and Mongolia arid regions to Beijing.High aerosol optical depth (AOD)at the wavelength 675nm and low ?ngstr?m exponent (AE)values in the wavelength 440–870nm are observed during dusty days.The aerosol size distribution (ASD)in coarse mode shows a large increase in the volume during dusty days.The single scattering albedo (SSA)increases with higher wavelength on dusty days,and is generally found to be higher compared to the days prior to and after the dust events,indicating the presence of high concentrations of scattering particles due to dust storm events.The physico-chemical properties of aerosols during dusty and non dusty days show distinct characteristics as reflected from the changes in the real and imaginary parts of refractive index (RI).In addition,the CO volume mixing ratio (COVMR)from Atmospheric Infrared Sounder (AIRS)shows a pronounced decrease on dusty days,while the H 2O mass mixing ratio (H 2OMMR)shows enhanced signal.Furthermore,enhanced level of water vapor (WV)using Moderate Resolution Imaging Spectroradiometer (MODIS)data is also observed in and around Beijing over the dust storms track.

?2014Elsevier B.V.All rights reserved.

Keyword:Dust storm Air pollution AERONET MODIS AIRS Beijing

1.Introduction

Since late 20th century,northern parts of China have experienced frequent dust storms (Chen,2001;Zhou and Zhang,2003)that affect the climate,visibility and air quality.The dust events observed in China are originated from Gobi and sand deserts of widespread arid and semiarid regions,which cover 13.6%area of the country (Zhu et al.,1986).Dust storms occur annually during spring season (March –May),and large

quantities of aeolian dust are transported to downwind regions.As a result,they are not only accumulated in the nearby region,forming the Loess Plateau (Liu,1985),but also undergo long-range transport depositing in the North Pacific Ocean (Duce et al.,1980;Shaw,1980).Dust storms with strong winds affect air quality that has serious health threat to the people living in the affected regions (Nickling and Brazel,1984;Littmann,1991;Swap et al.,1996).Episodically,the dust from Mongolia and northern China is transported across the North Pacific basin,reaching North America and beyond (Yu et al.,2008).Dusts influence the air quality,optical properties of aerosols,meteorological and atmospheric parameters,hydrologic cycle,

Atmospheric Research 150(2014)129–142

?Corresponding author.Tel.:+17142892057;fax:+17142892047.E-mail address:rsingh@https://www.wendangku.net/doc/341489177.html, (R.P.

Singh).

https://www.wendangku.net/doc/341489177.html,/10.1016/j.atmosres.2014.07.0220169-8095/?2014Elsevier B.V.All rights reserved.

Contents lists available at ScienceDirect

Atmospheric Research

j o u r n a l h o me p a g e :ww w.e l s e v i e r.c o m /l o c a t e /a t m o

s

monsoon system and climate as they are capable of changing radiative characteristics of the atmosphere(Tegen et al.,2004; Lau et al.,2006;Prasad and Singh,2007;Gautam et al.,2009; Singh,2014).In addition,dust storms have also been shown to influence the vertical structure of trace gases such as carbon monoxide(CO)(Bhattacharjee et al.,2007).

Beijing,the capital of China,is one of the most populous cities in the world with a population of20million(as in2011) covering an area of16,800km2.It is located at the foothills of Yan Mountains and Taihang Mountains,in the North China Plain(Chen et al.,2009).Dust storms,affecting Beijing, generally originate in the primary source regions of Mongolia and Inner Mongolia(Zhang et al.,2012).Beijing has a typical continental monsoon climate with four distinct seasons;during spring season(March–May)dust events occur every year affecting weather conditions,for example,the significant visibility reduction,and the significant PM10concentrations increase(Wang et al.,2008),that have long-term implications on human health.

Recently,ground-based Aerosol Robotic Network(AERONET) (Holben et al.,1998)and satellite data(Moderate Resolution Imaging Spectroradiometer(MODIS),Multiangle Imaging Spectroradiometer(MISR),and Ozone Monitoring Instrument (OMI))have been widely used to study aerosol properties and changes in optical/radiative properties due to dust events. Spaceborne lidar systems like Cloud–Aerosol Lidar and Infrared Pathfinder Satellite Observations(CALIPSO)have been used to understand the vertical variability of aerosols(Winker et al., 2007).Using satellite and ground observations,efforts have been made to study dust characteristics,its transport,short and long term impacts on climate around the world.The analysis of aerosol optical properties at different locations in the world shows robust differentiation in both the magnitude and spectral dependence of the absorption for desert dust,biomass burning, urban-industrial,and marine aerosols(Dubovik et al.,2002).In East Asia,maximum aerosol optical depth(AOD)together with low?ngstr?m exponents(AEs)is generally observed during spring season due to dust storm activities(Eck et al.,2005;Kim et al.,2008).In northern China,large variations in AOD and AE have been reported under typical background,and for environ-ments subjected to floating dust,and dust storm weather conditions.Occasionally,AE decreases to even zero or negative value during intense dust events(Xia et al.,2005;Cheng et al., 2006).Furthermore,the AOD and AE demonstrate contrary trends during all storm stages(pre-dust storm,dust storm,and post-dust storm),with the AOD indicating an obvious“Valley–Peak–Valley”pattern of variations,while AE demonstrate a “Peak–Valley–Peak”pattern(Xin et al.,2010).In Beijing,the AOD,AE,and single scattering albedo(SSA)show a distinct variation during dust and haze and foggy days(Yu et al.,2011). In general,higher AOD and lower AE are found over Beijing during dust events where dust is the major contributor of coarse mode particles to the net aerosol loading(Yu et al.,2013).Dust storms have also been associated with deterioration of regional air quality.For example,during the dust event of16–20April 2006over Beijing,a pronounced increase in hourly PM10 concentrations was observed which are up to1200μg/m3 (Wang et al.,2010),and during dust event of26April–3May, 2012,the air pollution index(API)increased up to473μg/m3 (Shen et al.,2013).The previous dust studies in Beijing mainly focus on one dust storm event or changes of few parameters acquired from ground-based measurements or satellite data.The comprehensive analysis of aerosol optical properties and meteorological parameters during dust storm events using ground and satellite data will help to find out dust impact on weather and climate at or above the surface level,and help to develop policy to reduce human health impacts of dust.

In the present study,we have studied the characteristics of aerosol and meteorological parameters using AERONET and satellite data during major dust storm events(2005–2010) over Beijing.We mainly focus on the changes in aerosol optical properties(AOD,AE,aerosol size distribution—ASD,refractive index—RI,and SSA from AERONET),API from ground based stations,and other parameters(water vapor—WV from MODIS, CO volume mixing ratio—COVMR,and H2O mass mixing ratio—H2OMMR from Atmospheric Infrared Sounder—AIRS).The Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT)model is used to determine the source of air mass and dust tracks.

2.Data and methods

In this study,we have considered data from ground based stations,MODIS,AIRS,and AERONET,and have analyzed the source of dust storms over Beijing using NOAA HYSPLIT model. The API simplifies the concentrations of several air pollutants to characterize air pollution level and air quality status in several levels(Zheng et al.,2014).It is related to the health impacts (Table1).The API data(as of the year2010)over Beijing is taken from the website of China National Environmental Monitoring Centre(CNEMC)(https://www.wendangku.net/doc/341489177.html,/citystatus/airMap.jsp).

The AERONET is a network of ground-based sun photometers to measure aerosol optical properties and validate satellite retrievals of aerosol optical properties(Holben et al.,1998). Numerous aerosol parameters(e.g.size distribution,complex refractive index—RI,phase function,SSA,AE,water vapor, spectral and broad-band fluxes)are available at different locations around the world from the AERONET network.In

Table1

API and health implications(Ministry of Environmental Protection of the People's Republic of China,2008).

API Air pollution level Health implications

0–50Excellent No health implications

51–100Good No health implications

101–200Lightly polluted Slight irritations may occur,individuals with breathing or heart problems should reduce outdoor exercise.

201–300Moderately polluted Healthy people will be noticeably affected.People with breathing or heart problems will experience reduced endurance in activities.These individuals and elders should remain indoors and restrict activities.

300+Severely polluted Healthy people will experience reduced endurance in activities.There may be strong irritations and symptoms and may trigger other illnesses.Elders and the sick should remain indoors and avoid exercise.Healthy individuals should avoid

outdoor activities.

130 C.Cao et al./Atmospheric Research150(2014)129–142

addition,the output provides measurement errors that are used for the assessment of the retrieval quality of aerosol parameters. The AE is inversely related to the average size of the particles in the aerosol,the smaller the particles,the larger the exponent (?ngstr?m,1961).The volume particle size distribution,aerosol size distribution(ASD),dV/d ln R(μm3/μm2)in the size range 0.05–15μm is retrieved from the spectral Sun and sky radiance data using the Dubovik and King approach(Dubovik and King, 2000).RI is an important optical property that represents the nature of aerosols.The real n(λ)(1.33≤n(λ)≤1.6)and imaginary k(λ)(0.0005≤k(λ)≤0.5)parts of the complex RI are retrieved for the wavelengths corresponding to sky radiance measurement,and provide information about the scattering and absorbing nature of aerosols.Higher values of real part of refractive index n(λ)represent scattering types of aerosol and higher values of imaginary refractive index k(λ)represent absorbing type of aerosols(Bohren and Huffman,1983;Sinyuk et al.,2003).The amount of light absorbed by each particle is measured by the SSA—the ratio between the light extinction due to scattering alone and the total light extinction from both scattering and absorption(Dubovik et al.,2000).We have used version1,level1.5of AERONET data since version2,level2 AERONET data including ASD,SSA,etc.,are not available soon after dust storm events of recent years.

We have adopted daily MODIS Level3(MOD08_D3) gridded product(water vapor(Infrared retrieval)for total column)with spatial resolution of1°.In addition,we have used AIRS products(COVMR,H2OMMR)having spatial resolution of 1°.The MODIS sensor onboard two NASA satellites,Aqua and Terra.The AIRS sensor onboard NASA's Aqua satellite.Both MODIS and AIRS products were downloaded from NASA Giovanni tools(https://www.wendangku.net/doc/341489177.html,/giovanni).As MODIS Level3products and AIRS products are sorted into 1°×1°cells,the central area of Beijing(longitude115.4°E–117.5°E,latitude39.4°N–41.0°N),is studied from116°to 117°E and from40°to41°N.The uncertainty in MODIS derived water vapor is in the range5–10%(Gao and Kaufman,2003;Prasad and Singh,2009).Preliminary comparisons to in-situ aircraft profiles indicate AIRS CO retrievals are approaching the15%accuracy target set by pre-launch simulations(McMillan et al.,2005).The water vapor retrievals from AIRS are in very good agreement with the radiosonde observations.The root mean square(RMS) difference is close to the expected goal accuracies,better than15%in2km layers for the water vapor in the troposphere(Divakarla et al.,2006;Gettelman et al.,2006).

The HYSPLIT model was used to determine the back trajectory of air masses(Draxler and Rolph,2013;Rolph, 2013).It has wide range of simulations related to atmospheric transport and dispersion of pollutants and hazardous materials, and deposition to the surface,e.g.volcanic ash,radioactive material,dust,air pollutants,etc.The model is run interactively through the website(https://www.wendangku.net/doc/341489177.html,/hypub-bin/ trajtype.pl?runtype=archive)of NOAA ARL(Air Resources Laboratory)by giving necessary input parameters.

3.Dust storm in Beijing

In China,according to(CCMB(China Central Meteorological Bureau),1979;Lin et al.,2011;Qian et al.,2002)four types of dust events are generally considered based on the visibility. Four types of dusts are classified as:

?Dust haze:Dust particles float up from the ground through winds and are suspended homogeneously in the air,the surface visibility may be about10km or less.

?Blowing dust:when dust particles are above ground by strong winds compared to those under the dust haze con-dition,visibility is reduced in the range1–10km.

?Dust storms:The dust particles are strongly transported into the air from the surface by storm events or turbulent winds, and the visibility is reduced to1km or less.

?Dust devils:The horizontal visibility is less than0.5km during such dust storm events.

There are several dust events during2005to2010in Beijing, here we have only considered dusts(dust storms)having visibility less than1km.Three dust storm events are identified for detailed study which occurred during April27–28,2005; April16–18,2006;and March20,22,2010(http://baike.baidu. com/view/3381041.htm;)(Li and Zhang,2012).In addition,we have acquired MODIS images(band combination1-4-3)prior to,during,and after dust storm events,clearly showing these dust events(Fig.1).The dust appears in beige color,and are clearly seen on April28and29,2005;April17,2006;and March 20,2010.We have considered April27–29,2005;April16–18, 2006;and March20,22,2010as dusty days,and studied the changes in aerosol and meteorological parameters during dusty and non-dusty days.

Desertified lands in China are mainly distributed in the arid, semiarid,and parts of sub-humid regions of the north of China, including the Inner Mongolia,Ningxia,Xinjiang,Tibet,etc. (Wang and Zhu,2001;Wang et al.,2002).In Inner Mongolia, located in the Northern China,there are five major deserts and five major sand lands,approximately0.1416million km2and 0.1255million km2,respectively.The five major desert regions are Badain Jaran,Tengger,Ulan Buh,Kubuqi,and Bayan Ondor. The five major sand regions include Mu Us,Hunshandake, Ujimqin,Horqin,and Hulunbeier(Dong and Ya,2004),desert and sand regions are the major sources for dust and sand storms.

The back trajectory(Fig.2)clearly shows the source and dust track,48h before reaching Beijing.Increasing the time of trajectories may also enhance uncertainty due to non-availability of dense network of meteorological observatories. As it can be seen from back trajectory of air mass at three heights,500,1500,and3000m,the source of air mass over Beijing is originated from Inner Mongolia and the border of China and Mongolia regions(Yin et al.,2007).The track of dust storms reaching Beijing is either from west,northwest and north.Mongolian cyclone along with cold front is the main sources for the changes in weather conditions.

4.AOD,?ngstr?m Exponent(AE)and air pollution index (API)

Fig.3a,b shows daily variations of AOD and AE,respectively, over Beijing during major dust storm events in the years2005–2010.The red column refers to AOD and the blue dot represents AE.There are two negative values of AE in Fig.3a,?0.010on March10,2006and?0.034on April23,2006.Negative values of AE are characteristics during the dust outbreak episode, indicating the presence of large dust particles.Similar AE values

131

C.Cao et al./Atmospheric Research150(2014)129–142

were also reported by Singh et al.(2005),Hamonou et al.(1999)and Prasad and Singh (2007)during dust events over Delhi,Sahara and Kanpur,respectively.A contrast in AE values during dust and non-dusty days are clearly seen.AE values near zero or slightly negative (Eck et al.,1999;O'Neill et al.,2001)are characteristics of desert dusts when the UV to mid visible AE (380to 500nm)is larger compared to the visible to near-infrared.This is characteristics of bimodal size distributions at relatively low optical depths with fine mode particles dominat-ing the wavelength dependence at short wavelengths and coarse mode particles dominating the wavelength dependence at longer wavelengths (Eck et al.,2005).Negative values of AE

are

Fig.1.MODIS images (band combination 1-4-3)prior to,during,and after dust storm events.The white color refers to clouds,and the beige color represents the dust.

132 C.Cao et al./Atmospheric Research 150(2014)129–142

also found by Xia et al.(2004)in Dunhuang,China and Xin et al.(2005)in Tengger Desert.During dust storm events,high AOD with low AE is observed on April 17,2006and March 22,2010.Conversely,during the days prior to and after dust storm events,low AOD with high AE is observed,obviously on April 15and 20,2006,as well as March 17and 25,2010.The detailed AOD and AE values for these three dust storm events are listed in Table 2.Eck et al.(2005)and Kim et al.(2008)also found similar results in East Asia.The AE show a minimum during spring time associated with dust storm activities (Eck et al.,2005).The AOD (AE)during the spring Asian dust period was found to be about 40–44%higher (39–56%lower)than the optical depths reported during the spring non-Asian dust period (Kim et al.,2008).Wang et al.(2010)found the dust plumes seen in the satellite true

color

Fig.1(continued.)

133

C.Cao et al./Atmospheric Research 150(2014)129–142

images are commonly associated with increasing AOD and small AE values,implying an increase in the amount of coarse particles.Fig.4shows the two different modes of AERONET AOD at 550nm,fine mode (0.05b r b 0.6μm)in blue column and coarse mode (0.6b r b 15μm),where r is the radius of particles.It is found that during dusty days,coarse mode AOD

increase

Fig.2.HYSPLIT 2days back trajectory starting from Beijing showing major sources of dust

storms.

Fig.3.Daily variation of AOD and ?ngstr?m Exponent in Beijing,the red column refers to AOD at 675nm,and the blue dot represents ?ngstr?m Exponent at 440–870nm.(a)from 2005to 2012,(b)during dusty days (April 27–29,2005;April 16–18,2006;and March 20,22,2010).

134 C.Cao et al./Atmospheric Research 150(2014)129–142

dramatically,indicating the presence of large particles due to dust events.Fig.5shows the API value during dust events March 5–April 5,2010,on March 20,API was found to be highest representing poor air https://www.wendangku.net/doc/341489177.html,bing the health implications of API in Table 1,the dust storms have serious impact on human health.

5.Characteristics of dust storms over Beijing 5.1.Aerosol size distribution (ASD)

Fig.6shows the aerosol size distribution (ASD)in the range 0.05–15μm,and the red line represents the dust storm events.The ASD shows bimodal nature during dust storm events,with maxima peaks in the ranges 0.5–10μm (coarse mode)and 0.05–0.5μm (fine mode),clearly indicating dusty days and the days prior to and after dust storm events.The ASD in coarse mode shows a large increase in the volume during dusty days.Moreover,the maxima ASD peak and corresponding radius in coarse mode show distinction between dusty days and the days prior to and after dust storm events.Table 3shows volume of aerosol size distribution of different radius of the aerosol particles,the peak of ASD (dV/d ln R,μm 3/μm 2)during dusty days in coarse mode is 0.8982, 1.9176, 1.2333and their

corresponding radius (μm)is 2.9400,2.2407,and 5.0613for the dusty days on April 29,2005,April 17,2006,and March 20,2010,respectively.During the days prior to and after dust storm events,the maxima peak of ASD (dV/d ln R,μm 3/μm 2)is about 0.158,0.149,0.079,0.046,and 0.018and the corre-sponding radius (μm)is 2.9400,3.8575,2.9400,0.1944,and 0.1129,respectively.Kim et al.(2008)also found the coarse mode particles are dominant during spring Asian dust period in East Asia.The large increase in the volume of ASD in coarse mode implies high mineral dust loading,and the mineral dust is the major component of aerosols.The mineral dust is mainly from the source of dust storms.Additionally,due to strong winds during dust events,local ground dust particles are strongly transported into the air from the surface.5.2.Single Scattering Albedo (SSA)

Fig.7shows the daily variation of SSA at four wavelengths (440,675,870and 1020nm)during major dust events in Beijing during 2005–2010.The SSA increases during dusty days at higher wavelengths (675,870,1020nm).Table 4shows SSA values during dust storm events,the average SSA at wave-lengths 675,870and 1020nm is found to be 0.89,0.96and 0.92on April 28,2005,April 17,2006and March 20,2010,respectively,much greater than that of 0.88and 0.80on non-dusty days April 15,2006and March 17,2010,respectively.The large average SSA indicates the presence of scattering and larger size particles and also mixing of dust and anthropogenic aerosols.Prasad and Singh (2007)have found that SSA increases with the wavelength during dust storm events over the Indo-Gangetic (IG)Plains.Yu et al.(2013)have found increasing SSA trend with wavelengths,from 0.891to 0.947at four wavelengths for dusty days in Beijing,we have also found similar results (Table 4).The average SSA at four wavelengths for the three dusty-days varies in the range 0.836–0.936,the average SSA is 0.836at 440nm for dusty days,lower compared to the observed value of 0.93at Kanpur during the 2002pre-monsoon season (Singh et al.,2004).This difference in SSA over Beijing and Kanpur is likely due to strong aerosol mixing,dust

Table 2

AOD and ?ngstr?m Exponent from AERONET for dusty days (April 27–29,2005;April 16–18,2006;and March 20,22,2010),as well as the days prior to and after dust storm events.Dust storm event Date AOD (675nm)AE (440–870nm)12005-04-29 1.2960.2512005-05-020.4130.6662

2006-04-150.2740.7002006-04-17 3.4580.1242006-04-200.280 1.1003

2010-03-170.190 1.3162010-03-22 3.169?0.1192010-03-25

0.084

1.135

Fig.4.Daily AERONET AOD at 500nm during dusty days (April 27–29,2005;April 16–18,2006;and March 20,22,2010),the blue column represents the fine mode,and the red column refers to the coarse mode.

135

C.Cao et al./Atmospheric Research 150(2014)129–142

with anthropogenic emissions (biomass burning smoke or anthropogenic activities),which can increase aerosol absorp-tion.Dubovik et al.(2002)have shown the spectral differences of SSA (ΔSSA,SSA 1022nm ?SSA 441nm )greater than 0.05for mineral dust aerosols.We have also observed similar results,i.e.values of ΔSSA during dust events with the results obtained by Dubovik et al.(2002).The ΔSSA is found to be 0.108,0.094and 0.096on April 28,2005,April 17,2006and March 20,2010,respectively.

5.3.Refractive Index (RI)

Fig.8shows daily variations of real and imaginary parts of refractive index (RI)during major dust events.Table 5clearly shows the distinction of RI during dusty days and the days prior to and after dust events.The real n (λ)and imaginary k(λ)parts show contrasting spectral behavior,n (λ)at higher wave-lengths is close to 1.53found from several models (K?pke et al.,1997;Shettle and Fenn,1979),indicating the impact of dust on the optical properties.In addition,n (λ)under moderate to strong dust event varies in the range 1.51±0.07at 440nm in Mediterranean region near Turkey,1.55±0.03at Bahrain –Persian Gulf,1.56±0.03at Solar-Vil.–Saudi Arabia,and 1.48±0.05at Cape Verde (Dubovik et al.,2002).n (λ)shows a high value greater than 1.5in the wavelength range 441–673nm,whereas at the higher wavelengths small value of n (λ)less than 1.5is observed for the IG Plains (Prasad and Singh,2007).n (λ)shows a similar high value greater than 1.5in all the four wavelengths in Beijing during dusty days,this is because the transport of dusts from source to local region as a result the surface area of coarse particles increase compared to the fine particles enhancing scattering.On April 29,2005,n (λ)

shows

Fig.5.Air pollution index during major dust storm events (March 20,22,

2010).

Fig.6.Aerosol size distribution from AERONET during dusty days (April 27–29,2005;April 16–18,2006;and March 20,22,2010).

136 C.Cao et al./Atmospheric Research 150(2014)129–142

small decrease with the increasing wavelength,and pro-nounced decrease on April17,2006.Low n(λ)during dusty days is likely due to high RH and resultant hygroscopic growth, similar to the conditions found over Goddard Space Flight Center(Dubovik et al.,2002).

The imaginary part of refractive index k(λ)is found to be low(b0.008)at every wavelength during dusty days compared with the days prior to and after dust storm events,showing the lower absorption of the dust.This is also supported by the higher SSA which means lower absorption during dusty days (Fig.7).The decrease in k(λ)shows dominance of mineral dust aerosols during the dust storms and is likely to be attributed to the decrease in fraction of anthropogenic aerosols.In addition, k(λ)decreases with the increasing wavelength.This is similar to the findings of Dey et al.(2004)and Prasad and Singh (2007).k(λ)values are less than0.0045during dust events over the IG Plains,and the small decrease at higher wave-lengths is a characteristic of mineral dust(Dey et al.,2004; Prasad and Singh,2007).

5.4.CO volume mixing ratio(COVMR)

The vertical profiles of COVMR during the days prior to and after dust events(solid lines)show maximum value at pressure levels in the range407–618hPa(Fig.9).During dusty days (dotted lines),the COVMR has maximum value at surface level (905hPa).In addition,the COVMR values during dusty days in 2005and2010increase with the increasing air pressure,which

Table3

Aerosol size distribution during major dust storm events.

Aerosol size distribution

Radius(μm)2005-04-282005-05-022006-04-152006-04-172006-04-202010-03-172010-03-202010-03-25 0.05000.00100.00080.00060.00500.00040.00040.00190.0003

0.06560.00460.00460.00490.01820.00250.00180.00510.0024

0.08610.00970.01350.01660.03750.00800.00650.00760.0095

0.11290.00960.02040.02560.04750.01370.01910.00680.0179

0.14820.00590.01840.02150.04730.01380.03980.00420.0165

0.19440.00320.01290.01310.05060.01000.04630.00230.0094

0.25510.00230.00940.00780.06870.00690.02730.00150.0047

0.33470.00270.00910.00620.10930.00590.01240.00130.0029

0.43920.00510.01200.00740.16380.00700.00750.00190.0025

0.57620.01250.01950.01220.20880.01080.00720.00380.0029

0.75610.02880.03200.02290.26880.01810.00940.00930.0040

0.99200.05210.04860.04100.44760.02810.01290.02330.0054

1.30160.07610.06950.06120.91540.03890.01630.05190.0069

1.70780.10470.09750.0753 1.58440.05060.02040.10190.0088

2.24070.15190.13200.0918 1.91760.06410.02650.19180.0115

2.94000.23190.15790.1206 1.74790.07690.03400.37730.0147

3.85750.32890.14850.1490 1.16450.07960.03910.75570.0168

5.06130.35680.10020.13800.55950.06260.0353 1.23330.0150

6.64070.23420.04610.07960.19910.03290.0228 1.13970.0096

8.71310.07410.01430.02580.05520.01080.01010.42310.0043 11.43230.00980.00300.00450.01290.00220.00310.05170.0013

15.00000.00050.00040.00040.00280.00030.00070.0020

0.0003

Fig.7.Total mode single scattering albedo from AERONET during dusty days(April27–29,2005;April16–18,2006;and March20,22,2010).137

C.Cao et al./Atmospheric Research150(2014)129–142

are different from the days prior to and after dust events.With the onset of the dust storm events,the COVMR is obviously lower at the pressure levels 407–802hPa,due to strong winds during dust storm events.

5.5.H 2O mass mixing ratio (H 2OMMR)and water vapor (WV)Kim et al.(2004)have observed an enhancement of the water vapor mixing ratio within the dust layer compared to the air above and below the dust layer.The increased water vapor in

the dust layer is also found to affect the radiative property of the dust layer (Kim et al.,2004).A large portion of water vapor within the Asian dust layer (ADL)is enhanced over the edge of a highland and the plains in China (Yoon et al.,2006).Fig.10shows vertical profiles of H 2OMMR during dust storm events.An enhancement in the H 2OMMR is observed during dusty days at pressure levels between 600and 850hPa in 2005,700–1000hPa in 2006,and 300–700hPa in 2010.The possible source of water vapor at corresponding pressure levels could be due to advection of ground surface moisture during dust storm events.In addition,airborne dust particles could cause significant radiative heating at shorter wavelengths and cooling at long wavelengths,which in turn influence the thermodynamics and temperature profile in the atmosphere (Won et al.,2004).The change of temperature profile will cause the change of water vapor profile in the atmosphere.

Besides the vertical profiles of H 2OMMR,we have analyzed the total WV column (infrared retrieval)derived from MODIS data in Fig.11.Enhanced WV total column is found in and around Beijing along the dust track compared to the day prior to the dust storm event (April 25,2005).In the northwest of Beijing,middle of Inner Mongolia,the WV content increase from 0.0–0.5to 0.5–1.0cm.An increase in the total WV column

Table 4

Total mode single scattering albedo during major dust storm events.SSA (Total)Days

44067587010202005-04-280.79620.87810.89890.90452005-05-020.80570.81550.82420.83032006-04-150.88510.88400.87080.87432006-04-170.87230.95690.96300.96672006-04-200.88170.86340.83470.82662010-03-170.78010.79690.79300.79732010-03-200.83990.90200.92320.93542010-03-25

0.9069

0.8841

0.8717

0.8695

Fig.8.Real and imaginary parts of refractive index from AERONET during dusty days (April 27–29,2005;April 16–18,2006;and March 20,22,2010).

138 C.Cao et al./Atmospheric Research 150(2014)129–142

is found in the south of Beijing,from 0.0–1.0cm to 1.5–3.5cm,similar enhancement in total column of WV during 2005dust event is also observed by Prasad and Singh (2007)over the IG Plains.

6.Conclusions

The results discussed in this paper shows characteristics of aerosol and meteorological parameters during major dust storm events (2005–2010)over Beijing,China,using ground-based AERONET and satellite data (MODIS and AIRS).The results show pronounced changes in water vapor column,meteorological parameters and aerosol optical properties during dust days as compared to the days prior to and after dust storm events.

AOD is observed much higher during dusty days,while AE shows decrease during dusty days.High AOD with corresponding low AE are characteristic of dust.Low AE is also associated with the presence of coarse dust particles,which is also supported from the ASD values.The ASD shows bimodal distribution during dust storm events,and the ASD in coarse mode shows a large increase in the volume during dusty days.

The optical properties of aerosols,SSA and RI,show changes during dust storm events.The SSA increases during dusty days at higher wavelengths.The SSA value at 440nm is lower compared to that observed at Kanpur (Singh et al.,2004).This difference in SSA shows strong aerosol mixing which is attributed to the mixing of dusts with the pollutants from biomass burning and anthropogenic activities.The real part of complex RI,n (λ),shows high values (N 1.5)in all the four

Table 5

Refractive index during major dust storm events.

RI(real)at wavelength (nm)

RI(imaginary)at wavelength (nm)Days

440

675

870

1020

440

675

870

1020

2005-04-28 1.6000 1.6000 1.6000 1.60000.00800.00400.00390.00412005-05-02 1.6000 1.6000 1.6000 1.60000.01790.01430.01380.01402006-04-15 1.5713 1.5834 1.5909 1.59150.00740.00680.00810.00822006-04-17 1.6000 1.5772 1.5655 1.54270.00540.00190.00190.00192006-04-20 1.5371 1.5545 1.5651 1.56730.01400.01360.01660.01762010-03-17 1.3939 1.3980 1.4112 1.42960.03410.02250.02000.01902010-03-20 1.6000 1.6000 1.6000 1.60000.00250.00170.00160.00162010-03-25

1.4920

1.5154

1.5323

1.5464

0.0104

0.0108

0.0110

0.0109

Fig.9.CO volume mixing ratio from AIRS during major dust storm events.Dotted lines indicate profile during dust storm events,while solid lines are prior to and after the events.

139

C.Cao et al./Atmospheric Research 150(2014)129–142

wavelengths during dusty days.The imaginary part of complex RI,k(λ),decreases with the wavelength,and shows low value at every wavelength during dusty days.In addition to aerosols,API shows higher values during dusty days due to poor air quality.

With the onset of the dust storm events,the COVMR and H 2OMMR from AIRS show contrasting differences during dust storm events.The COVMR is obviously lower at pressure levels between 407–802hPa,while the H 2OMMR is higher around 700hPa.In addition,the total WV column from MODIS data shows an enhancement during dusty days.Due to dust storm event in the year 2005,enhanced level of WV is observed in and around Beijing over the dust storms track.A contrast difference in aerosol and meteorological parameters is observed during dusty and the days prior to and after dust storm events.In addition,compared with previous studies focusing on one dust event or few parameters in Beijing,aerosol optical properties and meteorological parameters during three major dust events (2005–2010)are analyzed in this study.An extensive database of aerosol and meteorological parameters during dust

storm

Fig.10.H 2O mass mixing ratio from AIRS during major dust storm events.Dotted lines indicate profile during dust storm events,while solid lines are prior and after the

events.

Fig.11.Water vapor content from MODIS during dusty day and the day prior to the dust storm event.The left figure represents April 25,2005,and the right figure stands for the dusty day on April 29,2005.The blue track refers to 2days back trajectory of dust starting from Beijing on April 27,2005.

140 C.Cao et al./Atmospheric Research 150(2014)129–142

events from AERONET and satellite will be of a great use in quantitative evaluation of the dust impact on the regional weather and climate.

Acknowledgment

This paper was supported by the National High Technology Research and Development Program of China(863Program) (Grant no.2013AA12A302);National Natural Science Founda-tion of China(Grant no.41171330).Concerning AERONET data used in this paper,we thank Hongbin Chen and Philippe Goloub for theirs efforts in establishing and maintaining Beijing site.We are grateful to China National Environmental Monitoring Centre for making API available,and NASA Giovanni team for making satellite data available.The MODIS and AIRS data used in this effort were acquired as part of the activities of NASA's Science Mission Directorate,and are archived and distributed by the Goddard Earth Sciences(GES)Data and Information Services Center(DISC).We gratefully acknowledge the NOAA Air Resources Laboratory(ARL)for the provision of the HYSPLIT transport and dispersion model and/or READY website(http:// https://www.wendangku.net/doc/341489177.html,)used in the present study.We are grateful to the two anonymous reviewers for their comments and suggestions which have helped us in improving the earlier version of the manuscript.

References

?ngstr?m,A.,1961.Techniques of determinig the turbidity of the atmosphere.

Tellus13(2),214–223.

Bhattacharjee,P.S.,Prasad,A.K.,Kafatos,M.,Singh,R.P.,2007.Influence of a dust storm on carbon monoxide and water vapor over the Indo-Gangetic Plains.

J.Geophys.Res.Atmos.1984–2012112(D18).

Bohren, C.F.,Huffman, D.R.,1983.Absorption and scattering by a sphere.

Absorption and Scattering of Light by Small Particlespp.82–129.

CCMB(China Central Meteorological Bureau),1979.Standard on Weather Observations.Meteorological Press,Beijing(In Chinese).

Chen,G.T.,2001.History of strong dust storms in Beijing and ecological environmental change in nearby regions.J.Desert Res.21(4),402–407(in Chinese).

Chen,Y.,Zhao,C.S.,Zhang,Q.,Deng,Z.Z.,Huang,M.Y.,Ma,X.C.,2009.Aircraft study of mountain chimney effect of Beijing,China.J.Geophys.Res.Atmos.

1984–2012114(D8).

Cheng,T.,Wang,H.,Xu,Y.,Li,H.,Tian,L.,2006.Climatology of aerosol optical properties in northern China.Atmos.Environ.40(8),1495–1509.

Dey,S.,Tripathi,S.N.,Singh,R.P.,Holben,B.,2004.Influence of dust storms on the aerosol optical properties over the Indo-Gangetic basin.J.Geophys.

Res.Atmos.1984–2012109(D20).

Divakarla,M.G.,Barnet,C.D.,Goldberg,M.D.,McMillin,L.M.,Maddy,E.,Wolf,W., Zhow,L.H.,Liu,X.P.,2006.Validation of AIRS temperature and water vapor retrievals with matched radiosonde measurements and forecasts.J.

Geophys.Res.Atmos.1984–2012111(D09S15).

Dong,J.L.,Ya,J.,2004.Deserts and desertification land in Inner Mongolia.Inn.

Mong.For.Investig.Des.27(1),3–7(in Chinese).

Draxler,R.R.,Rolph,G.D.,2013.HYSPLIT(HYbrid Single-Particle Lagrangian Integrated Trajectory)Model access via NOAA ARL READY(Website (https://www.wendangku.net/doc/341489177.html,/HYSPLIT.php))NOAA Air Resources Laboratory, College Park,MD.

Dubovik,O.,King,M.D.,2000.A flexible inversion algorithm for retrieval of aerosol optical properties from Sun and sky radiance measurements.J.

Geophys.Res.Atmos.1984–2012105(D16),20673–20696.

Dubovik,O.,Smirnov,A.,Holben,B.N.,King,M.D.,Kaufman,Y.J.,Eck,T.F., Slutsker,I.,2000.Accuracy assessments of aerosol optical properties retrieved from Aerosol Robotic Network(AERONET)Sun and sky radiance measurements.J.Geophys.Res.Atmos.1984–2012105(D8), 9791–9806.

Dubovik,O.,Holben,B.,Eck,T.F.,Smirnov,A.,Kaufman,Y.J.,King,M.D.,Tanré,D., Slutsker,I.,2002.Variability of absorption and optical properties of key aerosol types observed in worldwide locations.J.Atmos.Sci.59(3),590–608.Duce,R.A.,Unni,C.K.,Ray,B.J.,Prospero,J.M.,Merrill,J.T.,1980.Long-range atmospheric transport of soil dust from Asia to the tropical North Pacific—temporal variability.Science209(4464),1522–1524.

Eck,T.,Holben,B.,Reid,J.,Dubovik,O.,Smirnov,A.,O'neill,N.,Slutsker,I.,Kinne,S.

,1999.Wavelength dependence of the optical depth of biomass burning, urban,and desert dust aerosols.J.Geophys.Res.Atmos.1984–2012104 (D24),31333–31349.

Eck,T.,Holben,B.,Dubovik,O.,Smirnov,A.,Goloub,P.,Chen,H.,Chatenet,B., Gomes,L.,Zhang,X.Y.,Tsay,S.C.,2005.Columnar aerosol optical properties at AERONET sites in central eastern Asia and aerosol transport to the tropical mid‐Pacific.J.Geophys.Res.Atmos.1984–2012110(D6).

Gao, B.C.,Kaufman,Y.J.,2003.Water vapor retrievals using Moderate Resolution Imaging Spectroradiometer(MODIS)near‐infrared channels.

J.Geophys.Res.Atmos.1984–2012108(D13).

Gautam,R.,Liu,Z.Y.,Singh,R.P.,Hsu,N.,2009.Two contrasting dust-dominant periods over India observed from MODIS and CALIPSO data.Geophys.Res.

Lett.36(Article Number:L06813,2009).

Gettelman,A.,Collins,W.D.,Fetzer,E.J.,Eldering,A.,Irion,F.W.,Duffy,P.B., Bala,G.,2006.Climatology of upper-tropospheric relative humidity from the Atmospheric Infrared Sounder and implications for climate.J.

Clim.19,6104–6121.

Hamonou,E.,Chazette,P.,Balis,D.,Dulac,F.,Schneider,X.,Galani,E.,Ancellet,G., Papayannis,A.,1999.Characterization of the vertical structure of Saharan dust export to the Mediterranean basin.J.Geophys.Res.Atmos.1984–2012104 (D18),22257–22270.

Holben,B.N.,Eck,T.F.,Slutsker,I.,Tanre,D.,Buis,J.P.,Setzer,A.,Vermote,E., Reagan,J.A.,Kaufman,Y.J.,Nakajima,T.,1998.AERONET—a federated instrument network and data archive for aerosol characterization.Remote Sens.Environ.66(1),1–16.

Kim,S.W.,Yoon,S.C.,Jefferson,A.,Won,J.G.,Dutton,E.G.,Ogren,J.A.,Anderson,T.L., 2004.Observation of enhanced water vapor in Asian dust layer and its effect on atmospheric radiative heating rates.Geophys.Res.Lett.31,L18113. Kim,S.W.,Yoon,S.C.,Kim,J.,2008.Columnar Asian dust particle properties observed by sun/sky radiometers from2000to2006in Korea.Atmos.

Environ.42(3),492–504.

K?pke,P.,Hess,M.,Schult,I.,Shettle,E.P.,1997.Global aerosol data set.Max-Planck-Institut für Meteorologie Hamburg,Germany.

Lau,K.M.,Kim,M.K.,Kim,K.M.,https://www.wendangku.net/doc/341489177.html,n summer monsoon anomalies induced by aerosol direct forcing:the role of the Tibetan Plateau.Clim.

Dyn.26,855–864.https://www.wendangku.net/doc/341489177.html,/10.1007/s00382-006-0114-z(2006). Li,X.L.,Zhang,H.S.,2012.Micrometeorological characteristics during severe dust storms in spring of the Beijing area in2010.Clim.Environ.Res.17(4), 400–408(in Chinese).

Lin,T.H.,Christina Hsu,N.,Tsay,S.C.,Huang,S.J.,https://www.wendangku.net/doc/341489177.html,n dust weather categorization with satellite and surface observations.Int.J.Remote Sens.

32(1),153–170.

Littmann,T.,1991.Dust storm frequency in Asia:climatic control and variability.Int.J.Climatol.11(4),393–412.

Liu,T.S.,1985.Loess and the Environment.China Ocean Press,Beijing,China(in Chinese).

McMillan,W.W.,Barnet,C.,Strow,L.,Chahine,M.T.,McCourt,M.L.,Warner,J.X., Novelli,P.C.,Korontzi,S.,Maddy,E.S.,Datta,S.,2005.Daily global maps of carbon monoxide from NASA's Atmospheric Infrared Sounder.Geophys.

Res.Lett.32,L11801.

Ministry of Environmental Protection of the People's Republic of China,2008.

Technical requirements for urban ambient air quality daily report and forecast(in Chinese).

Nickling,W.G.,Brazel,A.J.,1984.Temporal and spatial characteristics of Arizona dust storms(1965–1980).J.Climatol.4(6),645–660.

O'Neill,N.,Eck,T.,Holben,B.,Smirnov,A.,Dubovik,O.,Royer,A.,2001.Bimodal size distribution influences on the variation of Angstrom derivatives in spectral and optical depth space.J.Geophys.Res.Atmos.1984–2012106 (D9),9787–9806.

Prasad,A.K.,Singh,R.P.,2007.Changes in aerosol parameters during major dust storm events(2001–2005)over the Indo-Gangetic Plains using AERONET and MODIS data.J.Geophys.Res.Atmos.1984–2012112(D9).

Prasad,A.K.,Singh,R.P.,2009.Validation of MODIS Terra,AIRS,NCEP/DOE AMIP-II Reanalysis-2,and AERONET Sun photometer derived integrated precipitable water vapor using ground-based GPS receivers over India.J.

Geophys.Res.Atmos.1984–2012114(D05107).

Qian,W.H.,Quan,L.S.,Shi,S.Y.,2002.Variations of the dust storm in China and its climatic control.J.Clim.15(10),1216–1229.

Rolph,G.D.,2013.Real-time Environmental Applications and Display sYstem (READY)(Website(https://www.wendangku.net/doc/341489177.html,))NOAA Air Resources Laboratory,College Park,MD.

Shaw,G.E.,1980.Transport of Asian desert aerosol to the Hawaiian Islands.J.

Appl.Meteorol.19(11),1254–1259.

Shen,X.X.,Liu,C.S.,Shi,R.H.,Bai,K.X.,Wang,C.,Gao,W.,2013.Preliminary study of a dust event over Beijing by using satellite data and ground-based measurements.Proc.SPIE8869,Remote Sensing and Modeling

141

C.Cao et al./Atmospheric Research150(2014)129–142

of Ecosystems for Sustainability X,886912(pp.886912-886912-886911).

Shettle,E.P.,Fenn,R.W.,1979.Models for the aerosols of the lower atmosphere and the effects of humidity variations on their optical properties.DTIC Document.

Singh,R.P.,2014.Dust storms and their influence on atmospheric parameters over the Indo-Gangetic plains(book chapter2)In:Sundaresan,J.,et al.

(Eds.),Geospatial Technologies and Climate Change.,10,pp.21–35. Singh,R.P.,Dey,S.,Tripathi,S.N.,Tare,V.,Holben,B.,2004.Variability of aerosol parameters over Kanpur,northern India.J.Geophys.Res.109(D23),D23206. Singh,S.,Nath,S.,Kohli,R.,Singh,R.,2005.Aerosols over Delhi during pre-monsoon months:characteristics and effects on surface radiation forcing.

Geophys.Res.Lett.32(13).

Sinyuk,A.,Torres,O.,Dubovik,O.,https://www.wendangku.net/doc/341489177.html,bined use of satellite and surface observations to infer the imaginary part of refractive index of Saharan dust.Geophys.Res.Lett.30(2),1081.

Swap,R.,Ulanski,S.,Cobbett,M.,Garstang,M.,1996.Temporal and spatial characteristics of Saharan dust outbreaks.J.Geophys.Res.Atmos.

1984–2012101(D2),4205–4220.

Tegen,I.,Werner,M.,Harrison,S.,Kohfeld,K.,2004.Relative importance of climate and land use in determining present and future global soil dust emission.Geophys.Res.Lett.31,L05105.

Wang,T.,Zhu,Z.D.,2001.Studies on the sandy desertification in China.Chin.J.

Eco-Agric.9(2),7–12(in Chinese).

Wang,T.,Zhu,Z.D.,Wu,W.,2002.Sandy desertification in the north of China.

Sci.China Ser.D Earth Sci.45(1),23–34.

Wang,Y.Q.,Zhang,X.Y.,Gong,S.L.,Zhou,C.H.,Hu,X.Q.,Liu,H.L.,Niu,T.,Yang,Y.Q., 2008.Surface observation of sand and dust storm in East Asia and its application in CUACE/Dust.Atmos.Chem.Phys.8(3),545–553.

Wang,W.J.,Cheng,T.T.,Zhang,R.J.,Jia,X.,Han,Z.W.,Zhang,X.L.,Xu,X.F.,Li,D.P., 2010.Insights into an Asian dust event sweeping Beijing during April 2006:particle chemical composition,boundary layer structure,and radiative forcing.J.Geophys.Res.Atmos.1984–2012115(D18). Winker,D.M.,Hunt,W.H.,McGill,M.J.,2007.Initial performance assessment of CALIOP.Geophys.Res.Lett.34(19).

Won,J.G.,Yoon,S.C.,Kim,S.W.,Jefferson,A.,Dutton,E.G.,Holben,B.,2004.

Estimation of direct radiative forcing of Asian dust aerosols with sun/sky radiometer and lidar measurements at Gosan,Korea.J.Meteorol.Soc.Jpn.

82(1),115–130.Xia,X.A.,Chen,H.B.,Wang,P.C.,2004.Aerosol properties in a Chinese semiarid region.Atmos.Environ.38(27),4571–4581.

Xia,X.A.,Chen,H.B.,Wang,P.C.,Zong,X.M.,Qiu,J.H.,Gouloub,P.,2005.Aerosol properties and their spatial and temporal variations over North China in spring2001.Tellus B57(1),28–39.

Xin,J.,Wang,S.,Wang,Y.,Yuan,J.,Zhang,W.,Sun,Y.,2005.Optical properties and size distribution of dust aerosols over the Tengger Desert in Northern China.Atmos.Environ.39(32),5971–5978.

Xin,J.Y.,Du,W.P.,Wang,Y.S.,Gao,Q.X.,Li,Z.Q.,Wang,M.X.,2010.Aerosol optical properties affected by a strong dust storm over central and northern China.

Adv.Atmos.Sci.27(3),562–574.

Yin,X.H.,Shi,S.Y.,Zhang,M.Y.,Li,J.,2007.Change characteristic of Beijing dust weather and its sand-dust source areas.Plateau Meteorol.26(5), 1039–1044(in Chinese).

Yoon,S.C.,Kim,S.W.,Kim,J.,Sohn,B.J.,Jefferson,A.,Choi,S.J.,Cha,D.H.,Lee,D.K., Anderson,T.L.,Doherty,S.J.,2006.Enhanced water vapor in Asian dust layer:entrainment processes and implication for aerosol optical proper-ties.Atmos.Environ.40(13),2409–2421.

Yu,H.,Remer,L.A.,Chin,M.,Bian,H.,Kleidman,R.G.,Diehl,T.,2008.A satellite-based assessment of transpacific transport of pollution aerosol.J.Geophys.

Res.Atmos.1984–2012113(D14).

Yu,X.N.,Zhu,B.,Yin,Y.,Yang,J.,Li,Y.W.,Bu,X.L.,2011.A comparative analysis of aerosol properties in dust and haze-fog days in a Chinese urban region.

Atmos.Res.99(2),241–247.

Yu,X.N.,Shi,C.Z.,Ma,J.,Zhu,B.,Li,M.,Wang,J.,Yang,S.Y.,Kang,N.,2013.Aerosol optical properties during firework,biomass burning and dust episodes in Beijing.Atmos.Environ.81,475–484.

Zhang,Q.J.,Laurent,B.,Velay-Lasry,F.,Ngo,R.,Derognat,C.,Marticorena,B., Albergel,A.,2012.An air quality forecasting system in Beijing—application to the study of dust storm events in China in May2008.J.Environ.Sci.24

(1),102–111.

Zheng,S.,Cao,C.X.,Singh,R.P.,https://www.wendangku.net/doc/341489177.html,parison of ground based indices(API and AQI)with satellite based aerosol products.Sci.Total Environ.488, 398–412.

Zhou,Z.J.,Zhang,G.C.,2003.Typical severe dust storms in northern China during1954–2002.Chin.Sci.Bull.48(21),2366–2370.

Zhu,Z.D.,Liu,S.,Wu,Z.,Di,X.M.,1986.Deserts in China.Institute of Desert Research,Chinese Academy of Sciences,Lanzhou(132).

142 C.Cao et al./Atmospheric Research150(2014)129–142

比较PageRank算法和HITS算法的优缺点

题目:请比较PageRank算法和HITS算法的优缺点,除此之外,请再介绍2种用于搜索引擎检索结果的排序算法,并举例说明。 答: 1998年,Sergey Brin和Lawrence Page[1]提出了PageRank算法。该算法基于“从许多优质的网页链接过来的网页,必定还是优质网页”的回归关系,来判定网页的重要性。该算法认为从网页A导向网页B的链接可以看作是页面A对页面B的支持投票,根据这个投票数来判断页面的重要性。当然,不仅仅只看投票数,还要对投票的页面进行重要性分析,越是重要的页面所投票的评价也就越高。根据这样的分析,得到了高评价的重要页面会被给予较高的PageRank值,在检索结果内的名次也会提高。PageRank是基于对“使用复杂的算法而得到的链接构造”的分析,从而得出的各网页本身的特性。 HITS 算法是由康奈尔大学( Cornell University ) 的JonKleinberg 博士于1998 年首先提出。Kleinberg认为既然搜索是开始于用户的检索提问,那么每个页面的重要性也就依赖于用户的检索提问。他将用户检索提问分为如下三种:特指主题检索提问(specific queries,也称窄主题检索提问)、泛指主题检索提问(Broad-topic queries,也称宽主题检索提问)和相似网页检索提问(Similar-page queries)。HITS 算法专注于改善泛指主题检索的结果。 Kleinberg将网页(或网站)分为两类,即hubs和authorities,而且每个页面也有两个级别,即hubs(中心级别)和authorities(权威级别)。Authorities 是具有较高价值的网页,依赖于指向它的页面;hubs为指向较多authorities的网页,依赖于它指向的页面。HITS算法的目标就是通过迭代计算得到针对某个检索提问的排名最高的authority的网页。 通常HITS算法是作用在一定范围的,例如一个以程序开发为主题的网页,指向另一个以程序开发为主题的网页,则另一个网页的重要性就可能比较高,但是指向另一个购物类的网页则不一定。在限定范围之后根据网页的出度和入度建立一个矩阵,通过矩阵的迭代运算和定义收敛的阈值不断对两个向量authority 和hub值进行更新直至收敛。 从上面的分析可见,PageRank算法和HITS算法都是基于链接分析的搜索引擎排序算法,并且在算法中两者都利用了特征向量作为理论基础和收敛性依据。

pagerank算法实验报告

PageRank算法实验报告 一、算法介绍 PageRank是Google专有的算法,用于衡量特定网页相对于搜索引擎索引中的其他网页而言的重要程度。它由Larry Page 和Sergey Brin在20世纪90年代后期发明。PageRank实现了将链接价值概念作为排名因素。 PageRank的核心思想有2点: 1.如果一个网页被很多其他网页链接到的话说明这个网页比较重要,也就是pagerank值会相对较高; 2.如果一个pagerank值很高的网页链接到一个其他的网页,那么被链接到的网页的pagerank值会相应地因此而提高。 若页面表示有向图的顶点,有向边表示链接,w(i,j)=1表示页面i存在指向页面j的超链接,否则w(i,j)=0。如果页面A存在指向其他页面的超链接,就将A 的PageRank的份额平均地分给其所指向的所有页面,一次类推。虽然PageRank 会一直传递,但总的来说PageRank的计算是收敛的。 实际应用中可以采用幂法来计算PageRank,假如总共有m个页面,计算如公式所示: r=A*x 其中A=d*P+(1-d)*(e*e'/m) r表示当前迭代后的PageRank,它是一个m行的列向量,x是所有页面的PageRank初始值。 P由有向图的邻接矩阵变化而来,P'为邻接矩阵的每个元素除以每行元素之和得到。 e是m行的元素都为1的列向量。 二、算法代码实现

三、心得体会 在完成算法的过程中,我有以下几点体会: 1、在动手实现的过程中,先将算法的思想和思路理解清楚,对于后续动手实现 有很大帮助。 2、在实现之前,对于每步要做什么要有概念,然后对于不会实现的部分代码先 查找相应的用法,在进行整体编写。 3、在实现算法后,在寻找数据验证算法的过程中比较困难。作为初学者,对于 数据量大的数据的处理存在难度,但数据量的数据很难寻找,所以难以进行实例分析。

PageRank算法的核心思想

如何理解网页和网页之间的关系,特别是怎么从这些关系中提取网页中除文字以外的其他特性。这部分的一些核心算法曾是提高搜索引擎质量的重要推进力量。另外,我们这周要分享的算法也适用于其他能够把信息用结点与结点关系来表达的信息网络。 今天,我们先看一看用图来表达网页与网页之间的关系,并且计算网页重要性的经典算法:PageRank。 PageRank 的简要历史 时至今日,谢尔盖·布林(Sergey Brin)和拉里·佩奇(Larry Page)作为Google 这一雄厚科技帝国的创始人,已经耳熟能详。但在1995 年,他们两人还都是在斯坦福大学计算机系苦读的博士生。那个年代,互联网方兴未艾。雅虎作为信息时代的第一代巨人诞生了,布林和佩奇都希望能够创立属于自己的搜索引擎。1998 年夏天,两个人都暂时离开斯坦福大学的博士生项目,转而全职投入到Google 的研发工作中。他们把整个项目的一个总结发表在了1998 年的万维网国际会议上(WWW7,the seventh international conference on World Wide Web)(见参考文献[1])。这是PageRank 算法的第一次完整表述。 PageRank 一经提出就在学术界引起了很大反响,各类变形以及对PageRank 的各种解释和分析层出不穷。在这之后很长的一段时间里,PageRank 几乎成了网页链接分析的代名词。给你推荐一篇参考文献[2],作为进一步深入了解的阅读资料。

PageRank 的基本原理 我在这里先介绍一下PageRank 的最基本形式,这也是布林和佩奇最早发表PageRank 时的思路。 首先,我们来看一下每一个网页的周边结构。每一个网页都有一个“输出链接”(Outlink)的集合。这里,输出链接指的是从当前网页出发所指向的其他页面。比如,从页面A 有一个链接到页面B。那么B 就是A 的输出链接。根据这个定义,可以同样定义“输入链接”(Inlink),指的就是指向当前页面的其他页面。比如,页面C 指向页面A,那么C 就是A 的输入链接。 有了输入链接和输出链接的概念后,下面我们来定义一个页面的PageRank。我们假定每一个页面都有一个值,叫作PageRank,来衡量这个页面的重要程度。这个值是这么定义的,当前页面I 的PageRank 值,是I 的所有输入链接PageRank 值的加权和。 那么,权重是多少呢?对于I 的某一个输入链接J,假设其有N 个输出链接,那么这个权重就是N 分之一。也就是说,J 把自己的PageRank 的N 分之一分给I。从这个意义上来看,I 的PageRank,就是其所有输入链接把他们自身的PageRank 按照他们各自输出链接的比例分配给I。谁的输出链接多,谁分配的就少一些;反之,谁的输出链接少,谁分配的就多一些。这是一个非常形象直观的定义。

相关文档