Solar global ultraviolet and broadband global radiant fluxes and their relationships with aerosol optical depth at New Delhi

Authors


Abstract

The solar global ultraviolet (GUV) and broadband global (G) radiation flux obtained on a horizontal plane in Delhi, during April 2010 to March 2011 have been used to investigate the temporal variability of these radiations and their ratio, fraction of UV (FUV). For the first time the clearness index (KT) has been estimated over Delhi and its variability during different months of the year and season has been studied in detail. The impact of atmospheric aerosols on KT has also been studied. It has been found that for every unit increase in aerosol optical depth (AOD) at 340 nm, KT decreases by 0.06. A strong anti-correlation with correlation coefficient − 0.75 is observed between AOD and KT. On the basis of our field experience and observations at Delhi it is found that for highly cloudy and overcast conditions 0⩽KT⩽0.15, for partial cloudy or hazy conditions 0.15⩽KT⩽0.21 and KT > 0.21 for clear sky conditions. In addition, during foggy days in winter we have found KT values lying in the range 0.12–0.18 at Delhi. The day-time daily-averaged fluxes GUV and G varied in the range 0.15–1.23 MJ m−2 and 3.36–27.02 MJ m−2, respectively. The GUV and G showed similar pattern during the year except for the wet season when the FUV increased possibly due to an increase in water vapour concentration. Copyright © 2012 Royal Meteorological Society

1. Introduction

Solar radiation is the ultimate source of energy that drives the climate system on Earth. The ultraviolet (UV) and the broadband (or shortwave) radiation (that includes visible and near infrared) of the solar spectrum are of utmost importance to all life forms on earth. The importance of these radiations on biological systems (Wharton and Cockerell, 1998; Grant, 2002), photochemistry (Dickerson et al., 1997; Kondo, 2004), human health (Diffey,1991; Lu et al., 1996), radiation budget (Ramanathan et al., 2001; Zhang et al., 2004), agriculture (Caldwell et al., 1995; Zhao et al., 2005) and ecosystem (Franklin and Forster, 1997; Hansson and Hylander, 2009) have been extensively studied world over. It is also well known that the amount of these radiation vary extensively on spatial as well as temporal scales. These variations become more complex to understand due to the presence of aerosols, clouds, water vapour and other gases in the atmosphere. Nonetheless, the measurement of solar irradiance is essential for studying the atmospheric phenomena, large-scale weather analysis and prediction, climate change studies etc. The measurement of geographical distribution of irradiance and its changes with time is a necessary requirement not only in weather and climate studies but also in agricultural practice and food production, hydrology, ecology and energy development programmes etc. (see review by Okogbue et al., 2009).

A wide range of measurements of surface UV irradiance is available worldwide (Fioletov et al., 2001; Lam et al., 2002; Canada et al., 2003; Ogunjobi and Kim, 2004; Perrin et al., 2005; Jacovides et al., 2006; Hu et al., 2007; Kim et al., 2008; Xia et al., 2008); however, only a few have been reported from the Indian region (Panicker et al., 2009). This is in spite of the fact that there are a large number of aerosols measurements currently being undertaken in the country and it has already been established that the aerosols attenuate more UV irradiance than ozone does under the cloudless conditions, particularly, for wavelengths longer than 315 nm (Iqbal, 1983; Koronakis et al., 2002; Kalashnikova et al., 2007). A few short-term measurements of UV irradiance in the pristine high altitude regions of Himalayas have also shown UV measurements getting affected by the atmospheric aerosols (Singh and Singh, 2004; Srivastava et al., 2006). The effect of aerosols on the fraction of UV (FUV)radiation to global broadband shortwave radiation has also been observed by some researchers recently (Xia et al., 2008; Jacovides et al., 2009; Bo et al., 2010). In spite of its significance, ground-based instrumentation systems to monitor UV radiation and the FUV routinely and objectively have been established only at few locations in the country.

Similarly, there are not many surface observations of global broadband solar radiation in the country and as a consequence there is a lack of data on clearness index as well. The clearness index (KT), which is the ratio of global to extraterrestrial radiation is a general indicator of the transmission of radiation through the atmosphere at a particular location. A low value of KT indicates lower global solar radiation at surface, usually due to clouds or aerosols, whereas, a high value means higher radiation at the surface. KT is also an important parameter for design, development and application of solar energy collection, conversion and conservation devices (Coppolino, 1994; Ali et al., 2003; Okogbue et al., 2009). In this study, for the first time we have estimated the clearness index at Delhi for one complete year (2010 to 2011) and also studied the impact of aerosols on KT. The variation of KT during different season has also been studied in detail. On the basis of the clearness index we have also classified the different sky conditions. In addition, we have also presented the solar UV radiant flux, broadband global radiation measurements and the radiometric FUV during the year and studied its monthly and seasonal variation over Delhi.

2. Data utilized

2.1. Measurements

The radiometric measurements used in this study were collected at the National Physical Laboratory in New Delhi (28.38°N, 77.10°E, 235 m AMSL) during the time period 1 April 2010 to 31 March 2011. The database consists of hourly broadband global (G) and global UV (GUV) radiant fluxes measured on a horizontal plane. Broadband global radiant flux, G, (total bandwidth, 285–2800 nm) was measured using the Kipp & Zonen CMP-21 model pyranometer (Delft, The Netherlands). The GUV (total bandwidth, 280–400 nm) radiant flux was measured with the Kipp & Zonen CUV-4 model radiometer. The bandwidth at 50% point for CMP-21 is 310–2800 nm and that for the CUV-4 instrument it is 305–385 nm. Broadband global radiant flux measurements have an estimated experimental error of 3%, whereas GUV radiant flux measurements have experimental errors < 10%. All these instruments are new and have been installed after proper calibration, certified from the company. The flux data were recorded in W/m2 for every 2 min, 24 h a day during the entire observation period.

For measuring aerosol optical depth (AOD), Microtops Sunphotometer—make Solar Light Company, USA, has been used. The instrument gives column AOD at five wavelengths from UV to near infrared (340, 500, 675, 870 and 1020 nm) with full width at half maximum of ± 2 to 10 nm. The instrument is periodically calibrated by the Solar Light Company, USA, every year to ensure the stability and reliability of AOD measurements. Besides calibration issues, error in the Microtops measurements can also be caused by filter degradation, temperature effects and poor pointing at the sun (Morys et al., 2001). The pointing accuracy of the optical filters towards the sun is very important for accurate measurement of atmospheric constituents. The instrument was operated using a stable wooden platform with a precision elevation–azimuth assembly, so as to strictly maintain the pointing accuracy of optical filters. The pointing accuracy of the instrument is better than 0.1° and long-term stability of the filter is better than 0.1 nm year−1. The accuracy of measurements for precision and consistency of the Microtops are discussed in detail by Srivastava et al. (2006) and maximum uncertainty in the AOD measurements was found to be about 1.8% at 500 nm.

The meteorological parameters (relative humidity, wind speed, wind direction, temperature and rainfall) have been recorded with the help of meteorological sensors, co-located with the other instruments. These sensors are made by Campbell Scientific Group, Canada. Wind sensors model is 05 103 wind monitor, temperature and relative humidity (RH) probe, model HMP 45 C and rainfall measurement is done by using rain gauges, model TE 525 M-L.

2.2. Definition of clearness index

Solar radiation outside the earth's atmosphere is called extraterrestrial radiation. The ratio of the global solar radiation measured at the surface to the extraterrestrial radiation is called the clearness index (KT). KT is very useful in characterizing the sky conditions over a particular location (Okogbue et al., 2002) and in estimating the diffuse radiation (Okogbue et al., 2009). It can also be used to analyse the effects of aerosols, water vapour and clouds on the radiation (Hu et al., 2007).

The daily (or monthly) clearness index (KT) may be calculated using daily (or monthly) averaged global solar radiation measured on the horizontal surface (H) and the daily (or monthly) average extraterrestrial radiation (H0) as KT = H/H0 (Angstrom, 1924; Page, 1961; Duffie and Beckman, 1991; A-Hinai and Al-Alawi, 1995). Where the extraterrestrial H0 radiation is estimated as (Iqbal, 1983; Ulgen and Hepbasli, 2002):

equation image(1)
equation image(2)

where, Gsc is the solar constant equal to 1367 W/m2, ϕ is the latitude of the study site, δ is the solar declination angle, ωs, the sunrise hour angle and nday is the day of the year starting from the first of January as 1. The solar declination and sunshine hour angle can be computed by using the following equations as given by Duffie and Beckman (1991):

equation image(3)
equation image(4)

Before reaching the surface of the earth, radiation from the sun is attenuated by the atmospheric molecules, aerosols particles etc. So KT values depend on the location and the time of year. In this study, we have estimated it for the city of Delhi during April 2010 to March 2011.

2.3. Site characteristics

Delhi is among the ten most polluted cities in the world and the second largest Indian mega city with an average population growth rate of 3.85% per year (Srivastava et al., 2008). Although the vehicles are the biggest contributor to the ambient total suspended particulate matter level, significant contributions from other sources such as industries, roadside dust, trans-boundary migration, power plants and local sources have also been observed (Srivastava and Jain, 2007). The area around Delhi falls on the border zone lying between the rich rain-washed Ganges plains to the east and the semi-arid tracts of Rajasthan about 160 km to the west and southwest. The soil around Delhi is dusty and it has a semi-arid climate with extreme weather conditions. During peak summers, temperature rises even beyond 45 °C and in winter temperatures may fall to below 2 °C. In the pre-monsoon period (April to June), frequent dust storms from western and north to western desert regions cause large scale loading of aerosols over Delhi (Singh et al., 2005; Prasad et al., 2007; Pandithurai et al., 2008). The rainy season is from July to September when the climate sometimes becomes very humid. In this study, the annual data have been classified into four seasons, namely spring (February to March), summer (April to June), monsoon (July to September) and winter (October to January) on the basis of the existing meteorological conditions. The experimental site is located in the central Delhi and there are no major industries located nearby.

The meteorological conditions of Delhi from April 2010 to March 2011, the observation period, have been plotted in Figure 1. It represents the annual variation of meteorological parameters like temperature, RH, wind speed, wind direction and total rainfall plotted as monthly means. The vertical bars denote the 1σ standard deviations from the monthly mean. Temperature peaks during summer months with highest average value in the month of May (36 °C) and the lowest in January (14 °C). Average daily temperature rises even beyond 40 °C in summers and it falls below 7 °C in winter months. Average monthly temperature decreased gradually from October to reach a minimum in January and again started rising till April. Most of the rains occurred during July, August and September. August observed highest rainfall of 303.4 mm. The total rainfall during the observation period was 870.9 mm. In monsoon season (July to September), RH was maximum (average ∼75%) with peak value during September. On the other hand, April was the driest with average RH being only about 16%. The monthly average wind speed during the observation was 1.97 ms−1. It ranged between 0.53 and 4.58 ms−1 with the highest wind speed observed during the month of June. The wind directions shown in figure have been measured in degree clockwise from north. The winds are mostly from westerly or south-westerly direction except for the monsoon season when it is blowing from the East and South East direction. Such variations in different meteorological parameters has a strong influence on the aerosols properties, UV radiation as well as broadband solar radiation observed over Delhi, as will be described later.

Figure 1.

Variation of monthly averaged, temperature, relative humidity, wind speed, wind direction (clockwise with respect to North) and the total monthly rainfall during April 2010 to March 2011 over Delhi

3. Results and discussion

3.1. Daily G and UV radiant fluxes

Figure 2(a) shows the day-time (07:00 to 19:00 h) daily average variation of solar broadband global radiation flux (G) along with the UV radiation flux (GUV) as observed on a horizontal plane during the period of measurements over Delhi. The data were recorded at every 2 min of interval and hourly values of radiation were derived by integrating the data every hour. The G and GUV radiation fluxes show similar pattern throughout the observation. The day-time daily-averaged GUV flux varied between 0.15 and 1.23 MJ m−2 with an average annual value of 0.67 ± 0.24 MJ m−2. Similarly, the daily-averaged G varied in the range 3.36–27.02 MJ m−2 with average annual value 15.81 ± 5.47 MJ m−2.

Figure 2.

(a) Annual pattern of daily-averaged (07:00 to 19:00 h) broadband global radiation flux (G) and ultraviolet radiation flux (GUV) (b) Contour plot of hourly variation of broadband global radiation flux (G) and (c) global ultraviolet radiation flux (GUV) during day time at Delhi during 1 April 2010 to 31 March 2011

In order to show the hourly variation of G and GUV during day-time covering the entire period of observations, we have plotted the contours of G and GUV flux in Figure 2(b) and (c), respectively. On the x-axis, the day 1 denotes the starting date of observation, 1 April 2010. The annual pattern of daily G and GUV radiation shows that the values are generally low during winters as compared to those during summers. However, the minimum values of both G and GUV radiation are observed on the cloudy days during monsoon season. The maximum values were observed during summer months for both, GUV radiation and G. During monsoon, the values were slightly higher than the winter months but lesser than the spring months. The GUV flux shows an increase during monsoon but the broadband flux shows a relative decrease in its value. The increase in GUV flux and the corresponding decrease in global G flux during monsoon season may be due to the increased atmospheric water vapour level during this period of high RH. RH can abate the long wavelengths radiation remarkably well through absorption process, leaving the spectral UV portion unaltered. This process results in an effective reduction of broadband solar radiation reaching the ground. Water vapour absorption on the long wavelength (infrared wave band) does not affect the spectral GUV radiation, thus leaving the short wave radiation unchanged. As a result UV fraction increased as RH increases. Similar behaviour of UV fraction has also been reported earlier by Fiester et al. (1992) and Martinez-Lozano et al. (1999).

3.2. Monthly mean hourly variation of UV fraction and KT

The KT has been calculated every hour daily using the method described in Section 2.2. In order to do this we have integrated the hourly data of GUV radiation and broadband global radiation, G. Figure 3(a) shows the diurnal variation of monthly means of hourly integrated clearness index for four different months representing the four different seasons (May-summer, August-monsoon, December-winter and March-spring seasons). Along with the KT values we have also plotted the corresponding UV radiation, broadband radiation and UV fraction values in Figure 3. As expected, the clearness index has very low values during the sunrise and sunset times. The solar radiation received at the earth surface during the hours close to the sunrise and sunset is mainly of diffuse radiation. During the local noon period, when solar zenith angle is minimum, the clearness index reaches to its maximum value. It ranges from 0.24 to 0.58 at 13:00 h. Its monthly average value during the day (7:00 to 19:00 h) is minimum during August (0.14) and maximum in March (0.31). Season wise, the KT value is maximum during spring, followed by winter, summer and monsoon. It is noticed that the KT values are almost three times in March as compared to that during August whereas the values during May and December are comparable. However, it is important to know that the decrease in KT values during May is mainly due to high aerosol loading and in December it is due to the effect of fog and haze.

Figure 3.

Diurnal variation of monthly means of (a) clearness index, (b) fraction of UV (FUV), (c) ultraviolet radiation flux (GUV) and (d) broadband global radiation (G) for 4 months representing different seasons (May-summer, August-monsoon, December-winter and March-spring seasons)

In Figure 3(b), the radiometric ratio or fraction of GUV to broadband global radiation, denoted as FUV, has been plotted for the four representative months in the year. Where FUV = GUV/G, expressed as a percentage. It is interesting to notice that the FUV for all the months shown are nearly similar (in the range 3.2–4.3%) except for the monsoon month of August when it is significantly high (5.5%). This enhancement may be due to increased humidity-induced absorption at longer wavelengths, leading finally to higher GUV to G ratios. Hourly integrated daily-averaged GUV and G has also been plotted for the typical months representing the four seasons in Figure 3(c) and (d), respectively. Clearness index, GUV and G show the expected bell-shaped distribution. All these three parameters reach maximum during local noon time at 13:00 h and gradually decrease to lowest values during sunrise and sunset time. The monthly averaged, hourly GUV is found to be maximum in the dry summer month of May with a peak value of 0.117 MJ m−2 (average 0.066 MJ m−2) and minimum in the winter month of December with a peak value 0.064 MJ m−2 (average 0.029 MJ m−2). Similarly, the value for the global broadband radiation, G, is also maximum during May at a peak value of 2.886 MJ m−2 (average 1.687 MJ m−2) and minimum during December with maximum value 1.754 MJ m−2 (average 0.804 MJ m−2). It can be noticed that the reduction in the peak value of GUV radiation from summer (May) to monsoon (August) is about 55% whereas the reduction in the corresponding peak value of G is more than 60%. This additional decrease in G is likely due to the increased absorption due to water vapour and clouds, particularly in the near infrared region, which is comparatively less absorbing in the UV range (Martinez-Lozano et al., 1994; Jacovides et al., 2006). This is the reason why FUV has a high value during the monsoon month of August. The range of FUV at Delhi may be compared with several previously reported FUV values worldwide, for example, Al-Aruri et al. (1998) reported FUV in the range from 4.2% (December) to 5.2% (August) in Kuwait, Elhadidy et al. (1990) reported range of 2.1–4.5% in Dhahran, Hu et al. (2007) showed FUV varied between 3.0 and 5.0% in different parts of China.

A complete feature of hourly KT values during during the whole experimental period is displayed in Figure 4, top panel. The day 1 in the x-axis corresponds to 1 April 2010. The contour clearly shows enhanced values during mid-day and highest values during the month of April 2011. The low KT values during the monsoon may also be noticed. The corresponding contour plot for the FUV has been shown in the bottom panel of Figure 4. It can be noticed clearly that during the monsoon period, the FUV values are quite high compared to all other months, as also described earlier. A comparison of the two contours clearly indicates that during monsoon period even though the KT value is low, the UVG is quite high.

Figure 4.

Contour plot of clearness index (KT) (top) and FUV (bottom) during 1 April 2010 to 31 March 2011 at Delhi

3.3. Monthly average of daily UV fraction, G, UV and KT

Figure 5 shows the monthly mean of daily-averaged GUV and G along with the fraction FUV and KT values. The monthly averaged flux values showed more or less similar pattern as daily values shown earlier (Figure 2(a)). The monthly values however give a better picture of the annual variability in the flux and its ratio FUV. The average flux values for both, GUV and G were found to be minimum in December which gradually increased starting January and seemed to be reaching a maximum during the month of March. The monthly average GUV during day time varied between 0.383 MJ m−2 (minimum during December) and 0.858 MJ m−2 (maximum during May) whereas G varied between 10.650 MJ m−2 (minimum during December) and 22.006 MJ m−2 (maximum during May) throughout the year. The contribution of FUV in G varied from 3.0 to 5.4% on a monthly basis. Although the percentage contribution of UV peaked in August and showed minimum during December, it showed enhanced values during the monsoon months of July to September. This also coincides with the high average RH values during these months which lied in the range 68–75% (to make it more clear, RH has been re-plotted with the UV fraction in the middle section of Figure 5). During this time there is steep decrease in G, while the value for FUV does not decrease by the same order. This is attributed to the fact that water vapour can abate the long wave radiation substantially through absorption process leaving the short wave spectral radiation (GUV) unaffected. An increase in RH may lead to the larger extinction of broadband radiation G in comparison to that of the GUV, as a result the higher values of FUV during these months.

Figure 5.

Comparative monthly average of daily FUV and clearness index (top), FUV and RH (middle) and global ultraviolet radiation (GUV) and broadband global radiation (G) (bottom) over Delhi

Similar effect of RH on FUV has also been reported from the measurements taken in China by Hu et al. (2007), on the island of Cyprus by Jacovides et al. (2006) and by Canada et al. (2003) over Spain. At Spain, the value of FUV lied in the range 4.4–5.6% at Valencia and 3.9–4.5% at Cordoba (Canada et al., 2003). The annual average value of FUV at Delhi was found to be 4.2 ± 0.7% which is approximately similar to the values at Taoyuan, China (Latitude 28° 55'N) having FUV 4.5 ± 0.7% (Hu et al., 2007). In northern China plain it is 3.85% due to heavy aerosol loading. The attenuation of UV by aerosols is generally considerably larger than that in the whole short wave range because aerosol extinction generally decreases with wavelength (Xia et al., 2008).

The transmission of radiation through the atmosphere is affected by aerosols, water vapour, gases and clouds. KT is a general indicator of all the extinction (scattering + absorption) occurring in the atmosphere due to these factors (Liu and Jordan, 1960; Elhadidy et al., 1990; Hu et al., 2007). Top panel in Figure 5 shows the monthly averaged clearness index during daytime. It is maximum during March (0.29) and minimum during August (0.15). The monthly KT values are found to be anti-correlated with the FUV values. This is possibly due to the combined effects of variable aerosols, water vapour and clouds during different months in the year. The monthly average of hourly integrated broadband solar radiation and the hourly integrated UV radiation obtained daily on a horizontal surface over Delhi for 1 year (April 2010 to Mar 2011) have been shown in tabular form in Table I and Table II respectively. The present findings may be compared against earlier studies worldwide; for example, Canada et al. (2003) for Spain and Jacovides et al. (2006) for the Cyprus island.

Table 1. Monthly averaged hourly integrated broadband global radiation on a horizontal surface (MJ m−2) over Delhi
Time (h)April 2010May 2010June 2010July 2010August 2010September 2010October 2010November 2010December 2010January 2011February 2011March 2011
07:000.1550.3080.3180.2160.1300.0750.0440.0000.0000.0000.0300.000
08:000.6770.8440.7630.5950.4440.3020.4200.0950.0600.1590.3560.095
09:001.3311.4931.3541.0160.8280.7791.0070.4440.4290.6590.9830.444
10:002.0032.1441.8751.4451.2071.1971.5870.8940.9271.2311.6220.894
11:002.5572.5782.3131.8161.6231.5722.0171.3231.3921.7542.1731.323
12:002.9062.8412.6231.9881.7851.8392.2771.6141.7462.0072.5331.614
13:002.9622.8862.7071.9711.8122.0172.3041.7541.9212.1742.6241.754
14:002.7732.7102.5241.8861.8661.9362.1101.6201.8362.0582.5061.620
15:002.4202.3242.1151.7351.5571.5051.7061.3691.5371.7002.2041.369
16:001.8511.8251.5671.5941.1521.0751.1560.8991.0901.2541.6790.899
17:001.2041.1961.1681.0120.8700.6800.5680.3880.5230.7411.0540.388
18:000.5050.6370.6660.6000.4430.2650.0990.0470.0850.2140.4280.047
19:000.0660.1390.1950.1830.1030.0350.0000.0000.0000.0000.0220.000
Table 2. Monthly averaged hourly integrated ultraviolet radiation flux on a horizontal surface (MJ m−2) over Delhi
Time (h)April 2010May 2010June 2010July 2010August 2010September 2010October 2010November 2010December 2010January 2011February 2011March 2011
07:000.0070.0120.0120.0100.0070.0050.0030.0000.0000.0000.0000.002
08:000.0250.0300.0290.0280.0230.0170.0170.0090.0040.0030.0070.016
09:000.0500.0560.0530.0490.0430.0390.0380.0240.0170.0170.0270.040
10:000.0780.0840.0760.0700.0620.0590.0620.0430.0320.0350.0500.068
11:000.1040.1040.0980.0880.0830.0800.0800.0570.0470.0540.0730.095
12:000.1200.1160.1120.0970.0920.0930.0910.0640.0580.0670.0840.111
13:000.1220.1170.1170.0970.0950.1020.0930.0660.0640.0740.0920.117
14:000.1120.1070.1090.0930.0970.0970.0840.0590.0580.0690.0870.111
15:000.0960.0910.0890.0850.0800.0750.0660.0450.0480.0570.0700.094
16:000.0700.0680.0620.0760.0590.0510.0430.0270.0310.0390.0500.069
17:000.0420.0430.0440.0470.0420.0310.0210.0120.0140.0190.0280.040
18:000.0180.0220.0240.0260.0210.0120.0050.0010.0020.0040.0090.016
19:000.0030.0060.0080.0080.0050.0020.0000.0000.0000.0000.0000.002

3.4. KT versus AOD

The aerosols are known to attenuate the solar radiation through scattering and absorption. The scattering efficiency of aerosols depends upon the size distribution or the real part of the refractive index whereas the absorption efficiency depends upon the imaginary part of the refractive index. The finer particles have greater extinction effect on shorter wavelength as compared to the longer ones. In order to study the effects of aerosols we need to measure the column AOD in the atmosphere. AOD is a measure of the total extinction (scattering + absorption) of solar radiation in the atmosphere. By measuring AOD at different wavelengths we can parameterize the effective size of the aerosols in the atmosphere. The relation between the AOD and the wavelength can be best described by Angstrom formula τ(λ) = βλ−α where τ is the AOD, λ is the wavelength in µm (Angstrom, 1964), α, called the Angstrom exponent, and β, is the Angstrom turbidity coefficient. α is a rough indicator of the size distribution of the aerosols particles in the column while β represents the aerosol loading in the atmosphere, which is also the AOD at λ = 1 µm (Eck et al., 1999; Reid et al., 1999, Soni et al., 2011). β is an essential parameter in the estimation of the expected maximum performance of a solar installation, in the calculation of the maximum potential in air conditioning installations and in the determination of the level of photosynthetic activity efficacy (Pinazo et al., 1995). A low α value indicates the abundance of coarse-mode particles and high α values indicates the dominance of fine mode particles (Dubovik et al., 2002; Singh et al., 2006). Several research papers have reported about the AOD and the aerosol size distribution of Delhi region during different season and attributed them to various natural and anthropogenic sources (Singh et al., 2010; Soni et al., 2010).

In this case, AOD measurements were done at 340, 500, 675, 870 and 1020 nm using MICROTOPS-II sunphotometer during clear sky conditions. The clear sky conditions mean the cloudless days or the intervals of the day that were cloud free. The data were taken nearly every half an hour interval during the day time 09:00 to 17:00 h on cloud-free days. The observations were taken only when the sun was completely visible so that the field of view remains cloud free during the observations. More details about the microtops observations are given elsewhere (Srivastava et al., 2006; Singh et al., 2010). The α was retrieved by least square fit on a log–log plot scale of the observed AOD versus wavelength following the Angstrom power law for variation of AOD with wavelength. In this case, α and β have been estimated using AOD measurements at 340, 500, 675 and 870 nm as the AOD at 1020 nm may be contaminated due to the water vapour absorption.

Figure 6 shows the monthly averaged AOD measured at 500 nm and α and β parameters estimated at Delhi during the observation. A total of 117 d of clear sky conditions were observed. The daily-averaged AOD at 500 nm, α and β values were averaged monthly to be plotted in the figure. The vertical bars indicate the 1σ standard deviation from the mean values. It is apparent from the figure that the AOD values are high during the summer months of May (0.78 ± 0.28) and June (0.82 ± 0.30) and again during October (0.89 ± 0.32) and November (0.88 ± 0.33) of winter months. High value in May and June is due to the enhancement in natural desert dust aerosols from the western and southwestern region whereas during October and November it is mainly due to enhanced anthropogenic pollution during the festival month of Diwali (Singh et al., 2010). The average monthly α values during the months of November, December, January and February are relatively high ranging from 0.95 to 1.0, further indicating the dominance of fine mode aerosol particles. On the other hand, the monsoon months of July and August also show higher values of α suggesting that coarse-mode aerosol particles have been removed mostly by precipitation; that is, fine mode aerosols particles remain in the atmosphere. The month of September has the lowest AOD values (0.51 ± 0.16) with moderate α and β values. There are substantial variations in AOD, α and β throughout the year.

Figure 6.

Monthly variations of AOD (500 nm), Angstrom exponent α (340–870 nm) and Angstrom turbidity coefficient β calculated from observation at Delhi from April 2010 to March 2011 during clear sky days

3.5. UV fraction versus AOD

By comparing the Figures 5 and 6 it has been noticed that generally when AOD is high the UV fraction is low. This may be because an increased AOD leads to a larger extinction of GUV radiation than G. Similar results have also been reported by Hu et al. (2007) earlier. Similarly, when α is high clearness index is low and the vice versa. On the other hand, the low AOD during monsoon does not show an increase in clearness index due to the possible effects of cloud and enhanced water vapour in the atmosphere. As the scattering efficiency depends on the size of aerosols, the increase of finer particles in winter months, indicated by high α and moderate AOD values (Figure 6), led to a larger extinction of UV radiation which resulted in the comparatively lower UV fraction. In order to see the effect of AOD on KT we have plotted the daily clearness index variation with respect to the corresponding daily average AOD in Figure 7, upper panel. A strong anti-correlation with correlation coefficient of the order of − 0.75 is observed between AOD at 340 nm and KT indicating a decrease in KT with increasing AOD. For every unit increase in AOD, KT decreases by 0.06. The high AOD observed over Delhi throughout the year is one of the reasons of relatively low KT values observed.

Figure 7.

Variation of clearness index (KT) with AOD at 340 nm (top) and variation of FUV with AOD at 500 nm (bottom) at Delhi during clear sky conditions

The lower panel in Figure 7 shows the variation of FUV for the corresponding variation in AOD at 500 nm. This shows the direct effect of aerosol on the FUV measurements at Delhi. The AOD at 500 nm is chosen here as it is directly related to the atmospheric turbidity levels. A negative linear correlation between FUV and the AOD at 500 nm of the order of − 0.51 can be noticed. For every unit increase in AOD at 500 nm the average FUV is found to decrease by ∼0.7%. Similar regression analysis between FUV and the AOD at 340 nm (not shown in figure) also shows a decrease in FUV of the order of 0.5% for every unit increase in AOD. Bo et al. (2010) have studied the relationship between UV radiation and AOD in China in detail and have found strong negative linear correlation between the UV fraction and AOD, which was stronger than that between UV radiation and AOD. In their observation the monthly average hourly FUV varied from 2.9 to 4.6% at different locations. Similarly, Jacovides et al. (2009) have also found similar negative correlations separately between UVA and UVB fractions versus AOD at an eastern Mediterranean site Cyprus.

3.6. Sky conditions and frequency distribution of KT

The sky condition is classified by using the clearness index which is a widely used index since it depends only on solar irradiation measurements as a measured parameter (Muneer, 1995, 1998; Li et al., 2004). High clearness index indicates high global solar radiation dominated by the direct radiation while low clearness index indicates low global solar radiation which may be usually due to cloudy sky or high aerosol loading giving rise to high portion of diffuse radiation components. Although, there are no definite KT values to define the sky conditions, different researches have adopted different values for KT depending upon their locations, prevailing atmospheric conditions and field experiences. For example, Reindl et al. (1990) proposed KT > 0.6 for clear sky and KT < 0.2 for cloudy sky at various European and North American locations; Kuye and Jagtap (1992) have used KT > 0.65 for very clear sky and 0.15⩽KT⩽0.35 for cloudy skies at Port Harcourt, Nigeria; Li et al. (2004) have used 0⩽KT⩽0.15 for overcast, 0.15⩽KT⩽0.7 for partly cloudy and KT > 0.7 for clear sky at Hong Kong. Recently, Okogbue et al. (2009) classified 0⩽KT⩽0.15 for overcast sky, 0.15⩽KT⩽0.6 for partially cloudy and 0.6⩽KT⩽∞ to define clear sky conditions at a tropical station Ile-Ife, Nigeria. For this study we have found 0⩽KT⩽0.15 for highly cloudy and overcast conditions, 0.15⩽KT⩽0.21 for partial cloudy or hazy conditions and KT > 0.21 for clear sky conditions based on our field experience and observations. In addition, during foggy days in winter we have found KT values lying in the range 0.12–0.18 at Delhi.

Figure 8 shows the frequency distribution (frequency of occurrence) and cumulative frequency for every 0.03 interval of the daily clearness index obtained during 1 April 2010 to 31 March 2011 at Delhi. It shows that during most of the year (about 50%) daily KT values lies between 0.18 and 0.24 for the annual distribution with marked peak value in the interval 0.21–0.24. About 40% of the days in the year, the daily KT values are below 0.18 and only during the remaining 10% of the days it goes beyond 0.24. To see the seasonal effect we have divided the whole study period into four different seasons—spring (February to March), summer (April to June), monsoon (July to September) and winter (October to January). Figure 9 shows the frequency distribution of daily clearness index during these seasons. In summer season KT values peak in the range 0.21–0.24 with most of the values (∼55%) lying above 0.21. The spring season on the other hand shows peak KT value in the range 0.24–0.27 and a significant high frequency (∼50%) lying above 0.24. Moreover, during summer season the maximum KT value reach only upto 0. 27 whereas during spring the KT values reach as high as 0.45. This denotes that the spring season is relatively more clear than the summer as far as sun shine is concerned. This is mainly due to high aerosols present in the atmosphere during summer season (Singh et al., 2005, 2010), as also indicated by high AOD observed during summer (Figure 6). On the other hand, during monsoon, most of the days (∼75%) KT lies below 0.18 value indicating significantly cloudy sky conditions. During winters, about 30% of time KT is below 0.18 and peaks in the range 0.18–0.24 with the highest frequency (∼65%). It goes beyond 0.24 only rarely (∼5%) during winter.

Figure 8.

Frequency of occurrence along with cumulative frequency of daily clearness index (KT) from April 2010 to March 2011 over Delhi

Figure 9.

Frequency of occurrence along with cumulative frequency of daily clearness index (KT) for different seasons during April 2010 to March 2011 over Delhi

4. Conclusions

The integrated hourly and daily values of UV (300–385 nm) and broadband (310–2800 nm) solar radiations were measured at Delhi and analysed to see their monthly and seasonal variations. In addition, we have also estimated the clearness index, for the first time over Delhi, and studied not only their variability during the day and during different seasons but also tried to study the impact of AOD and relative humidity on KT. The following conclusions may be drawn from this study:

  • 1.The day-time daily-averaged UV radiation flux (GUV) varied between 0.15–1.23 MJ m−2 (annual average 0.67 ± 0.24 MJ m−2) and G varied in the range 3.36–27.02 MJ m−2 (annual average 15.81 ± 5.47 MJ m−2). Both the fluxes showed similar pattern during the year except for the wet season when the UV fraction increases possibly due to the increase in water vapour concentration.
  • 2.The daily clearness index KT value at Delhi was found to lie between 0.18 and 0.24 during most of the year. It was found to be maximum during spring, followed by winter, summer and monsoon seasons. Its monthly average value during the day (7:00 to 19:00 h) was minimum in August (0.14) and maximum in March (0.31). The KT was found to be anti-correlated with the UV fraction, and the RH.
  • 3.From this study we have found that at Delhi during highly cloudy or overcast conditions 0⩽KT⩽0.15, during partial cloudy or hazy conditions 0.15⩽KT⩽0.21 and for clear sky conditions KT > 0.21. Further, during foggy days in winter the KT values were found lying in the range 0.12–0.18.
  • 4.A strong anti-correlation with correlation coefficient − 0.75 is observed between AOD at 340 nm and KT indicating a decrease in KT with increasing AOD. For every unit increase in AOD, KT decreases by 0.06.
  • 5.The correlation between the daily radiometric ratio (FUV) versus AOD at 500 nm was found to be weak and negative, further suggesting that the FUV decreases by 0.66% when AOD increases by 0.1.

Acknowledgements

The author Tarannum Bano is thankful to Council for Scientific and Industrial Research (CSIR) for her Research Associateship being provided during this work. A part of this work was also sponsored by the Indian Council for Medical Research (ICMR). Authors are also thankful to the two anonymous referees for their valuable comments and suggestions.

Ancillary