- Email: [email protected]
The aerosol effect on direct normal irradiance in Europe under clear skies
Renewable Energy 68 (2014) 475e484
Contents lists available at ScienceDirect
Renewable Energy journal homepage: www.elsevier.com/locate/renene
The aerosol effect on direct normal irradiance in Europe under clear skies E. Nikitidou*, A. Kazantzidis, V. Salamalikis Laboratory of Atmospheric Physics, Physics Department, University of Patras, 26500 Rio, Greece
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 August 2013 Accepted 12 February 2014 Available online 12 March 2014
The effect of spatial and temporal variability of aerosol optical depth (AOD) on direct normal irradiance (DNI) under clear skies is studied, with the synergetic use of satellite and ground-based data as well as calculations from a radiative transfer model. The area of interest is Europe; data from May to September during 13 years (2000e2012) are analyzed. The aerosol effect on DNI is high in areas influenced by desert dust intrusions and intense anthropogenic activities, such as the Mediterranean basin and the Po Valley in Italy. In May, the attenuation of DNI from aerosols, over these areas, can reach values up to 35% and 45% respectively, which corresponds to 4 and 6 kWh m2 per day. In most areas, even for periods with lower values of AOD, the attenuation of DNI is found to be around 20%, which corresponds to about 2 e3 kWh m2 less received DNI per day, compared to the corresponding value on an aerosol clean day. However, the DNI has increased during the recent years, due to the decreasing tendency of AOD over most areas of Europe. The increase is around 6e12%, which corresponds to an amount of 0.5e1.25 more kWh m2 received per day, compared to a clean day. The percentage differences of daily DNI from the corresponding monthly climatological value reveals that day-to-day differences (due to AOD changes) from the monthly mean, by 20%, can occur. The significance of the aerosol changes in Europe reveals the necessity for near real-time measurements or forecasts of AOD when reliable estimations of DNI are required. Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: Aerosols Direct normal irradiance DNI Satellite estimates
1. Introduction Aerosols are one of the most important constituents in the atmosphere that affect the incoming solar radiation, either directly through absorbing and scattering processes or indirectly by changing the optical properties and lifetime of clouds. Even though aerosols have been in the center of scientific research for years; their high spatiotemporal variability and complex atmospheric interactions induce challenges in the determination of their radiative forcing, which still holds high uncertainties [1]. Many studies focus on the spatiotemporal distribution of aerosol properties in Europe. Chubarova [2] studied the seasonal distribution, using data from the MODIS instrument and the AERONET network. The highest aerosol loads were observed in south and southeastern parts of Europe during the warmer months of the year. Over these areas, permanent high values of the Ångström a
* Corresponding author. Tel.: þ30 2610996079; fax: þ30 2610997989. E-mail addresses: [email protected] (E. Nikitidou), [email protected] (A. Kazantzidis), [email protected] (V. Salamalikis). http://dx.doi.org/10.1016/j.renene.2014.02.034 0960-1481/Ó 2014 Elsevier Ltd. All rights reserved.
coefficient were found indicating the presence of anthropogenic aerosols. Marmer and Langmann [3] studied the interannual variability of the aerosol distribution over Europe, using a regional atmospheric chemistry model. They highlighted the role of meteorological conditions which can induce a variability of up to 100% in the monthly mean of the aerosol load. The weekly variability of aerosols in Europe presents the lowest values during the weekend due to the decrease in transport and other anthropogenic activities [4e6]. As a result of legislations regarding atmospheric pollution, negative trends in aerosol amounts have been observed during the last years in many areas of Europe, the so-called “brightening effect” [7e11]. In order to complement the ground measurements and account for their limited spatial coverage, satellites have been employed for Earth’s and atmosphere’s observation. Satellites can provide global coverage and produce long-term datasets. The MODerate resolution Imaging Spectroradiometer (MODIS), on board the NASA Terra satellite, has been in operation since February 2000. The sunsynchronous satellite provides global coverage every 1e2 days and acquires data in 36 spectral bands (http://modis.gsfc.nasa.gov/) [12]. The Collection 005 (C005) is the dataset of MODIS products
476
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
provided by the version 5.2 of the algorithm, which has replaced older versions [13,14]. The prelaunch expected error in AOD retrievals over land is (0.05 þ 0.15 AOD), while AOD is estimated to within this expected error in more than 60% and 72% of the cases over ocean and over land respectively [15]. The MODIS data have been extensively validated against ground-based measurements [16e19] and used to calculate, amongst others, the aerosol trends and events [7,20,21], as well as the surface shortwave, ultraviolet [22e24] and visible solar irradiance [25]. The accurate knowledge of the amount of solar irradiance reaching a surface on the Earth and its temporal variability is essential for the efficient performance of solar power applications. The most important constituent of solar radiation is the Direct Normal Irradiance (DNI), which is the radiation received by a surface perpendicular to the direction of the sun. Although systems measuring the Global Horizontal Irradiance (GHI) are abundant, DNI measurements are not so common and these data are usually estimated indirectly with the use of radiative transfer or decomposition models [26e28]. Under clear skies, aerosols become the dominant factor that affect the intensity of solar irradiance reaching the ground. It has been shown that the variability in DNI due to aerosols is more important than the one induced in GHI [29], while the uncertainty in its calculation is dominated by uncertainties in the aerosol optical properties [30]. Suri et al. [31] studied the DNI over Europe as this is provided by 5 different datasets. The annual sum was found to reach the highest values in areas of the Mediterranean, Southern and Central Spain, Portugal, Sicily, Sardinia and Provence. Gueymard [29] used AOD measurements from 180 stations of the AErosol RObotic NETwork (AERONET) over the world to study the variability of AOD and its effect on DNI and GHI. He concluded that some areas experience a high variability that makes resource assessments potentially too optimistic for bankability if based only on limited data series. In this study, the effect of AOD on DNI under clear skies is studied, with the synergetic use of satellite data from the MODIS instrument, complementary data from the AERONET network and calculations from a radiative transfer model. The area of interest is Europe and data from a 13-year period (2000e2012) are analyzed.
The DNI temporal and spatial variability due to aerosols is examined, as well as the increase induced by the observed decreasing tendency of AOD over Europe. The uncertainties of DNI estimations induced by the use of aerosol climatological values are investigated. 2. Data and methodology The AOD at 550 nm is taken from the MODIS Terra daily Level-3 data (Collection 5.1) which have a spatial resolution of 1 1 (100 100 km). Terra was launched in December 1999 and started providing MODIS data in March 2000. The MODIS AOD at 550 nm, from the Terra satellite is used in this study from the beginning of operation (March 1st, 2000) until the end of 2012. The AOD wavelength dependence, described by the Ångström-a exponent, is provided by MODIS but is not used in this study since it still presents considerable errors when compared with ground-based measurements [14,18]. For the Ångström-a exponent, the level 2.0 climatological data from AERONET were used. The level 2.0 data are pre- and post-field calibrated, automatically cloud-screened and manually checked. The monthly climatological values of Ångstrom a exponent (440e 870 nm) were chosen for 37 stations in Europe, which present data availability longer than 3 years during the period 2000e2013. These data were then interpolated for the region of our study in order to obtain the spatial resolution of MODIS (1 1 ). The interpolation was performed using the method of Ordinary Kriging, combined with a linear semi-variogram model. The AERONET sites used in this study are shown in Fig. 1. The area examined (latitude: 29.5 N e 59.5 N, longitude: 9.5 W e 38.5 E) covers Europe, with the exception of the northern latitudes of Scandinavia, the northern coasts of Africa and the Mediterranean coasts of Middle East. In order to estimate the DNI on the surface, the radiative transfer model SBDART was used [32], which is included in the LibRadtran package [33] (www.libradtran.org). Typical vertical profiles for the basic atmospheric gases, pressure and temperature, for the midlatitudes, were used [34]. The surface albedo was set at 0.2 for the shortwave (SW) range (280e3000 nm) [35]. The aerosol vertical profile is described by Shettle [36]. The AOD is described by the
Fig. 1. AERONET stations used for the retrieval of the gridded monthly averages of Ångström a exponent.
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
Ångström-a exponent and b coefficient, the latter being calculated by the equation:
b ¼ s*la
(1)
where s is the MODIS AOD at l ¼ 550 nm and a the AERONET Ångström exponent. The solver of the radiative transfer equation is the discrete ordinate algorithm with 6 streams considered and the sphericity of the Earth is taken into account. The AOD provided by MODIS Terra was considered to be constant during the day. For every satellite pixel and day, the clear-sky DNI was calculated every 15 min, from sunrise to sunset, as a function of the solar zenith angle, and the Ångström-a and b parameters. A correction for the Earth e Sun distance was subsequently applied. The DNI was then calculated by integrating the 15min values over the day length. The DNI is also calculated for aerosol-free conditions in order to derive the absolute and percentage (%) differences in the received DNI between the measured and the aerosol-free conditions:
Diff ¼ DNI DNIclear Diffð%Þ ¼
DNI DNIclear *100 DNIclear
(2) (3)
where DNIclear is the calculated DNI under aerosol-free conditions. These values were calculated by running the model with the same conditions as before, but with no aerosols present. The latitudes lower than 34 are not depicted in the maps of the following figures, as the satellite provides few data during the investigated period due to the high surface albedo of the deserts in these areas. During the 13-year period examined, only months which had data availability higher than 40% (more than 12 days per month) were considered. As a result, only data during the period MayeSeptember of each year were analyzed, since the data availability during the rest of the months is not adequate due to the dominance of thick cloudiness. 3. Results 3.1. Seasonal variability of the Ångström-a exponent The monthly Ångström-a exponent climatological values, taken from the selected AERONET stations in Europe, were interpolated to match the MODIS spatial resolution (1 1 ). The monthly Ångström a exponent maps for Europe during the MayeSeptember period are presented in Fig. 2. The lowest values (around 0.8) are presented during May in Southern Europe, Mediterranean and the northern coasts of Africa and constitute and evidence for the frequent desert-dust intrusions during this period. The influence of desert dust is very important in July and August too, for Southern Spain and the western coasts of Africa but is limited in Eastern Mediterranean. The highest values of the Ångström-a exponent (1.4e1.8) are observed during summer in Central-Eastern Europe and the Balkans, indicating the presence of fine mode particles, which persist during this period due to the reduction of precipitation. In September, these values decrease further and moderate values are observed in continental Europe, while maritime areas are described by values around 0.8e1.1. 3.2. DNI changes due to the AOD variability 3.2.1. The aerosol effect on DNI The absolute and percentage differences of DNI under clear skies have been calculated for every day and every satellite pixel during
477
the examined period. The monthly average values of the percentage difference over the 13-year period are presented in Fig. 3. In Mediterranean, the highest percentage differences are observed in May reaching values down to 35%. This corresponds to an attenuation of about 4 kWh m2 per day. The highest reductions during this month are evident around Po Valley, with differences down to 45% (6 kWh m2 per day). Lower values of 25 to 15% (3 to 2 kWh m2 per day) are revealed in Spain, Southern France, Eastern and Northern Europe, while Central Europe and the southern parts of Ireland and Britain present an attenuation of around 30 to 25% (4 to 3.5 kWh m2 per day). The data availability at the northern parts of Ireland and Britain was lower than 40% so the results for these areas were considered not significant for this study. In June, the percentage difference is slightly lower than that over the Mediterranean basin and Northern Europe but moderately increases in areas of central Europe. In Northern France, the Low Countries, most parts of Germany and Poland, the attenuation of DNI is around 30e35%, while it remains high around the Po Valley and the far western part of the Mediterranean basin. The situation over continental Europe is similar in July, with the exception of a slight increase of the aerosol effect in North-Eastern regions. In the Mediterranean, the effect of aerosols on DNI decreases over Portugal and Po Valley (around 10 and 30% respectively) but becomes stronger over the western coasts of Africa, where percentage difference of up to 40% (5 kWh m2 per day) are found. In August, the aerosol effect on DNI is around 25 to 30% (3 to 3.5 kWh m2 per day) over most parts of Mediterranean, Central and Eastern Europe and slightly lower over France and Spain, where differences of around 2 kWh m2 per day are calculated. The effect is stronger in South-Western Mediterranean and in the interior areas of Turkey. In September, the percentage differences decrease in Central and Eastern Europe, while there is an increase in the North Sea (25 to 35%). The importance of the aerosol effect on DNI is clear, as even for months with low aerosol loads, the attenuation of DNI is around 20%, which corresponds to about 2 to 3 kWh m2 per day. This effect becomes even more significant at areas burdened with aerosols throughout most time of the year, such as the industrial region of Northern Italy. 3.2.2. Annual differences of DNI For each year, the average values of modeled DNI have been calculated over the period MayeSeptember. Fig. 4 presents the percentage differences, for every year, between these mean values and the average over the period MayeSeptember for all years. There is an increasing tendency in the calculated differences. During the first years of the period examined, the differences in DNI are negative; they reach low values, around zero, in the middle of the period and then increase during the last years, indicating a positive trend in the received DNI due to the reduction of AOD (“brightening” effect). This “brightening” effect, concerning the European area that is studied here, has been identified in various other scientific studies [37e40]. During the years 2000e2003, the DNI received during the MayeSeptember period is lower than the average derived from the entire 13-year period, for most regions of Europe. Central Europe received lower DNI around 4e10% during 2000 and 2001, which corresponds to 1 to 0.25 kWh m2 per day. However, positive differences (2e4%) are revealed in some areas of Northern Europe. In 2002, Eastern Europe and the Balkans present the highest negative difference of DNI, between 8 and 18% (0.75e1.75 kWh m2 per day) compared to the average; this probably may be due to excessive wildfires in this area [41,42]. In 2003, low negative values ( 10 to 4%, or 1 to 0.25 kWh m2 per day) are observed in West-Central Europe and Western
478
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
Fig. 2. Monthly average maps of the Ångström a exponent over Europe.
Mediterranean due to exceptionally hot and dry conditions and large-scale intense forest fires [43e45]. In 2004, small increases start to appear, which reach þ2 to þ8%, in 2005, at latitudes between 50 and 55 N. From 2006 to 2008, the differences are small, around 4%, which corresponds to an aerosol effect on DNI less than 0.5 kWh m2 per day. However, increasing values of DNI appear during the last years (2009e2012). In 2012, 6e12% more DNI is received on most areas of Central and Eastern Europe, the Balkans and the Italian peninsula. These positive differences correspond to an increase of 0.5e1.25 kWh m2 in received DNI per day. These differences are further studied in the next section by averaging them over specific regions of Europe. 3.2.3. DNI and AOD differences per area The above mentioned differences in DNI for each satellite pixel were then averaged over each of the five different areas in Europe. These areas (Fig. 5) are Central-Western Mediterranean (34.5 N e 44.5 N, 9.5 W e 18.5 E), Eastern Mediterranean-Black
Sea (32.5 N e 46.5 N, 18.5 E e 36.5 E), Central-Western Europe (44.5 N e 54.5 N, 4.5 W e 14.5 E), Eastern Europe (44.5 N e 55.5 N, 14.5 E e 38.5 E) and Northern Europe (54.5 N e 59.5 N, 5.5 E e 29.5 E). The percentage differences of the annual mean AOD and DNI values, from the climatological values over the whole period (2000e2012) were calculated for each area and for the MayeSeptember period (Fig. 6). The annual mean AOD values (Fig. 6a) are higher than the 13-year average, until 2003, after which a decrease starts, which becomes more evident during the end of the period. During 2010e2012, the AOD in CentralWestern Europe and Central-Western Mediterranean is lower by 15e17% (w0.04) than the 13-year average, while in Eastern Europe and Eastern Mediterranean a similar decrease is revealed only in 2011 and 2012. With the exception of year 2002, where strong positive differences are observed, the area of Northern Europe has the lowest variability during the examined period. The corresponding percentage differences in DNI are presented in Fig. 6b. Over all regions, the decrease during the beginning of the
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
479
Fig. 3. Monthly mean values of the aerosol effect (%) on DNI. Only the period MayeSeptember is presented, which fulfills the data availability requirements.
examined period and the increase at the end is evident. With the exception of Northern Europe, all areas received around 3% (0.3 kWh m2 per day) less DNI during 2000 and 2001. In 2002, the highest decrease is revealed in Eastern Europe (w10%, 0.9 kWh m2 per day). The differences are moderate for all regions in the following years but increased DNI values are found after 2009. During the end of the examined period, an increase in the received DNI of 4e6% (0.4e0.6 kWh m2 per day) can be observed in all regions, except Northern Europe. Fig. 7 presents the percentage differences of the daily DNI from the corresponding monthly climatological value, calculated from the daily DNI values during the entire time period. The positive trend in DNI differences is evident in all areas, when looking at the daily differences as well, except in Northern Europe. Additionally, in all areas, day-to-day percentage differences of DNI from the monthly climatological value, due to changes in AOD of up to 20%, are revealed. The high negative differences, which were observed in Eastern Europe in 2002, can now be seen that occur during the
period of mid August e mid September. During this time, the differences can reach values higher than 40%. It is during the same period of 2002 that high differences in Northern Europe appear as well. 3.3. DNI from MODIS In this section, the monthly climatological MODIS AOD values, derived from the 13-year period, have been used as input data to the SBDART model, in order to estimate the DNI values. The goal is to examine the differences between these DNI values and the monthly ones, calculated from the available daily data during the examined period. The analysis is performed for the months May to September and the results are averaged over each of the 5 areas described before. Fig. 8 shows the differences between the DNI derived from the monthly MODIS-AOD climatology and the corresponding monthly mean DNI calculated from all the daily AOD data during the period of the study. The derived differences are negative
480
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
Fig. 4. Differences (%) between the mean DNI for each year (2000e2012) and the average for the 13-year period. Only data for MayeSeptember are considered.
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
481
(0.4 kWh m2 per day). In Central-Western Mediterranean the underestimation is around 2.5e3%, which corresponds to 0.21e 0.28 kWh m2 per day, except in May when it reaches 0.33 kWh m2 per day. In May, the highest differences, for each region, are found, close to and larger than 0.3 kWh m2 per day, except for Eastern Europe, which has a slightly higher underestimation in August (0.37 kWh m2 per day). These differences are smaller than half a kWh m2 per day on average, but still have to be taken into account if someone chooses to use the satellite climatology for modeling the DNI received by a solar power farm. 4. Conclusions
Fig. 5. Europe divided into 5 areas of interest: Central-Western Mediterranean (34.5 N e 44.5 N, 9.5 W e 18.5 E), Eastern Mediterranean-Black Sea (32.5 N e 46.5 N, 18.5 E e 36.5 E), Central-Western Europe (44.5 N e 54.5 N, 4.5 W e 14.5 E), Eastern Europe (44.5 N e 55.5 N, 14.5 E e 38.5 E) and Northern Europe (54.5 N e 59.5 N, 5.5 E e 29.5 E).
for all months and all regions, showing that when the MODIS-AOD monthly climatology is used to estimate the DNI, there is an underestimation of a few percent, as opposed to the monthly DNI values calculated from the daily MODIS data. The lowest differences are observed in Northern Europe in June and July, around 0.16 kWh m2 per day, which corresponds to an underestimation of 1.5%. Northern Europe has higher differences for the rest of the months and the maximum difference for the area is found in May
The effect of aerosols on DNI in Europe was calculated for clearsky conditions on a daily basis, using data from the MODIS instrument, on board the NASA Terra satellite, and complementary data of Ångström a exponent from 37 AERONET stations. The period of the study was from the beginning of the satellite operation (March 2000) until the end of 2012. The DNI was estimated, using the SBDART radiative transfer model, with a spatial resolution of 1 1 ; the corresponding aerosol effect was derived for every satellite pixel. The monthly climatological values of the Ångström-a exponent, taken from AERONET, were interpolated to the MODIS spatial resolution. The lowest values, around 0.6, were observed in spring across the Mediterranean coasts and were evident of the intrusion of desert-dust aerosols over this area. Permanent high values, indicating the presence of fine-mode aerosols, were observed over many areas of Central Europe, most notably the Po Valley. The calculated percentage (%) difference was high in areas influenced by desert-dust intrusions and intense anthropogenic
Fig. 6. Difference (%) between the mean AOD (6a) and DNI (6b) value for each year (2000e2012) and the average for the 13-year period for the 5 areas of interest.
482
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
Fig. 7. Difference (%) between the daily DNI and the corresponding monthly mean, for each of the 5 areas of interest.
activities, such as the Mediterranean basin and the Po Valley. The attenuation of DNI due to aerosols, over these areas, can reach values up to 35% and 45% and corresponds to 4 and 6 kWh m2 per day, respectively. In June and July, a significant decrease in DNI values (30e35%) due to aerosols was calculated over the Low Countries, Germany and Poland. In most areas, even for periods with lower AOD values, the attenuation of DNI was found to be
around 20%, which corresponds to values about 2e3 kWh m2 lower than that received per day. The MODIS-AOD values and the differences in DNI were averaged over each of the 5 different areas in Europe: Central-Western Mediterranean, Eastern Mediterranean-Black Sea, Central-Western, Eastern and Northern Europe. During 2010e2012, the AOD in Central-Western Europe and Central-Western Mediterranean was
Fig. 8. Difference (%) between the DNI derived from the monthly AOD climatology and the corresponding monthly mean DNI calculated from the available AOD daily data during the period 2000e2012 and over each of the 5 areas of interest.
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
lower by 15e17% (w0.04) than the 13-year average, while in Eastern Europe and Eastern Mediterranean a similar decrease was found only in 2011 and 2012. When compared to the climatological values of DNI for the 2000e2012 period, it was shown that the decreasing tendency of the AOD values resulted in a DNI change of 10%. The percentage differences of the daily DNI from the corresponding climatological monthly value showed that, apart from the positive trend in the DNI differences in all areas, day-to-day differences from the monthly mean of up to 20% can occur. The effect of calculating the DNI from the monthly AOD climatology instead of the monthly DNI mean calculated from all the daily AOD data was investigated. According to the results, the DNI was found to be underestimated in the range 0.16e0.37 kWh m2 during the Maye September period. DNI is of vital importance to solar power systems; detailed information of its spatiotemporal variations is often required, rather than for GHI, a parameter that is more usually examined. A detailed study of the changes in DNI over Europe due to the variability of AOD, the DNI “brightening” effect and the uncertainties induced from the use of an aerosol monthly climatology instead measurements was provided. The decreasing tendency of AOD over Europe caused significant changes in DNI. The increased energy production from solar power plants in most areas of the European continent set the need for additional studies about the DNI variability. This study highlighted the need for near real-time measurements or reliable forecasts of AOD for the estimation of DNI and emphasized to the use of AOD climatological data for Europe.
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
Acknowledgments The authors would like to thank the principal investigators of the AERONET stations (listed in Fig. 1) for providing the data. The MODIS teams are acknowledged for providing the Collection 5 aerosol products from the Terra and Aqua satellites. The LibRadtran team (www.libradtran.org) is acknowledged for providing the model algorithm. This study was mainly conducted and funded by project “Hellenic Network of Solar Energy” (HNSE, 09SYN-32-778), funded by the General Secretariat for Research and Technology, Greek Ministry of Education, Lifelong Learning and Religious Affairs. References [1] Intergovernmental Panel on Climate Change (IPCC). Climate change 2007, working group i report: “The physical science basis”, summary for policymakers. Paris; February 2007. [2] Chubarova NY. Seasonal distribution of aerosol properties over Europe and their impact on UV irradiance. Atmos Meas Tech 2009;2:593e608. [3] Marmer E, Langmann B. Aerosol modeling over Europe: 1. Interannual variability of aerosol distribution. J Geophys Res 2007;112(D23S15). http:// dx.doi.org/10.1029/2006JD008113. [4] Bäumer D, Rinke R, Vogel B. Weekly periodicities of aerosol optical thickness over Central Europe e evidence of an anthropogenic direct aerosol effect. Atmos Chem Phys 2008;8:83e90. http://dx.doi.org/10.5194/acp-8-83-2008. [5] Xia X, Eck TF, Holben BN, Phillippe G, Chen H. Analysis of the weekly cycle of aerosol optical depth using AERONET and MODIS data. J Geophys Res 2008;113(D14217). http://dx.doi.org/10.1029/2007JD009604. [6] Georgoulias AK, Kourtidis KA. On the aerosol weekly cycle spatiotemporal variability over Europe. Atmos Chem Phys 2011;11:4611e32. [7] Papadimas CD, Hatzianastassiou N, Mihalopoulos N, Querol X, Vardavas I. Spatial and temporal variability in aerosol properties over the Mediterranean basin based on 6-year (2000e2006) MODIS data. J Geophys Res 2008;113(D11205). http://dx.doi.org/10.1029/2007JD009189. [8] Karnieli A, Derimian Y, Indoitu R, Panov N, Levy RC, Remer LA, et al. Temporal trend in anthropogenic sulfur aerosol transport from central and eastern Europe to Israel. J Geophys Res 2009;114(D00D19). http://dx.doi.org/10.1029/ 2009JD011870. [9] Streets DG, Yan F, Chin M, Diehl T, Mahowald N, Schultz M, et al. Anthropogenic and natural contributions to regional trends in aerosol optical depth,
[22]
[23] [24]
[25]
[26]
[27]
[28]
[29] [30]
[31]
[32]
[33]
[34]
[35] [36]
483
1980e2006. J Geophys Res 2009;114(D00D18). http://dx.doi.org/10.1029/ 2008JD011624. Koukouli ME, Kazadzis S, Amiridis V, Ichoku C, Balis DS, Bais AF. Signs of a negative trend in the MODIS aerosol optical depth over the Southern Balkans. Atmos Environ 2010;44:1219e28. Yoon J, von Hoyningen-Huene W, Vountas M, Burrows J. Analysis of linear long-term trend of aerosol optical thickness derived from SeaWiFS using BAER over Europe and South China. Atmos Chem Phys 2011;11:12149e67. Remer LA, Kaufman YJ, Tanré D, Mattoo S, Chu DA, Martins JV, et al. The MODIS aerosol algorithm, products and validation. J Atmos Sci 2005;62(4): 947e73. Levy RC, Remer LA, Dubovik O. Global aerosol optical properties and application 15 to moderate resolution imaging spectroradiometer aerosol retrieval over land. J Geophys Res 2007a;112(D13210). http://dx.doi.org/10.1029/ 2006JD007815. Levy RC, Remer LA, Mattoo S, Vermote EF, Kaufman YJ. Second-generation operational algorithm: retrieval of aerosol properties over land from inversion of moderate resolution imaging spectroradiometer spectral reflectance. J Geophys Res 2007b;112(D13211):20. http://dx.doi.org/10.1029/2006JD007811. Remer LA, Kleidman RG, Levy RC, Kaufman YJ, Tanré D, Mattoo S, et al. Global aerosol climatology from the MODIS satellite sensors. J Geophys Res 2008;113(D14S07). http://dx.doi.org/10.1029/2007JD009661. Chu DA, Kaufman YJ, Ichoku C, Remer LA, Tanré D, Holben BN. Validation of MODIS aerosol optical depth retrieval over land. Geophys Res Lett 2002;29(12). http://dx.doi.org/10.1029/2001GL013205. Santese M, De Tomasi F, Perrone MR. Moderate resolution imaging spectroradiometer (MODIS) and aerosol robotic network (AERONET) retrievals during dust outbreaks over the mediterranean. J Geophys Res 2007;112(D18201). http://dx.doi.org/10.1029/2007JD008482. Levy RC, Remer LA, Kleidman RG, Mattoo S, Ichoku C, Kahn R, et al. Global evaluation of the collection 5 MODIS dark-target aerosol products over land. Atmos Chem Phys 2010;10:10399e420. De Meij A, Lelieveld J. Evaluating aerosol optical properties observed by ground-based and satellite remote sensing over the Mediterranean and the Middle East in 2006. Atmos Res 2011;99:415e33. Gkikas A, Hatzianastassiou N, Mihalopoulos N. Aerosol events in the broader Mediterranean basin based on 7-year (2000-2007) MODIS C005 data. Ann Geophys 2009;27:3509e22. Zhang J, Reid JS. A decadal regional and global trend analysis of the aerosol optical depth using a data-assimilation grade over-water MODIS and Level 2 MISR aerosol products. Atmos Chem Phys 2010;10:10949e63. Bisht G, Venturini V, Islam S, Jiang L. Estimation of the net radiation using MODIS (Moderate Resolution Imaging Spectroradiometer) data for clear sky days. Remote Sens Environ 2005;97:52e67. Tang B, Li Z-L, Zhang R. A direct method for estimating net surface shortwave radiation from MODIS data. Remote Sens Environ 2006;103:115e26. Huang G, Ma M, Liang S, Liu S, Li X. A LUT-based approach to estimate surface solar irradiance by combining MODIS and MTSAT data. J Geophys Res 2011;116(D22201). http://dx.doi.org/10.1029/2011JD016120. Nikitidou E, Kazantzidis A. On the differences of ultraviolet and visible irradiance calculations in the Mediterranean basin due to model e and satellite e derived climatologies of aerosol optical properties. Int J Climatol; 2012. http:// dx.doi.org/10.1002/joc.3638. Battles FJ, Rubio MA, Tovar J, Olmo FJ, Alados-Arboledas L. Empirical modeling of hourly direct irradiance by means of hourly global irradiance. Energy 2000;25:675e88. López G, Rubio MA, Batlles FJ. Estimation of hourly direct normal from measured global solar irradiance in Spain. Renew Energy 2000;21: 175e86. Janjai S, Sricharoen K, Pattarapanitchai S. Semi-empirical models for the estimation of clear sky solar global and direct normal irradiances in the tropics. Appl Energy 2011;88:4749e55. Gueymard CA. Temporal variability in direct and global irradiance at various time scales as affected by aerosols. Sol Energy 2012;86:3544e53. Halthore RN, Schwartz SE, Michalsky JJ, Anderson GP, Ferrare RA, Holben BN, et al. Comparison of model estimated and measured direct-normal solar irradiance. J Geophys Res 1997;102(D25):29991e30002. Suri M, Remund J, Cebecauer T, Hoyer-Klick C, Dumortier D, Huld T, et al. In: SolarPACES conference “Concentrated solar power and Chemical Energy systems”, 15e18 Sep. 2009, Berlin, Germany; 2009. Ricchiazzi P, Yang S, Gautier C, Sowle D. SBDART: a research and Teaching software tool for plane-parallel radiative transfer in the Earth’s atmosphere. Bull Am Meteorol Soc 1998;79:2101e14. Mayer B, Kylling A. Technical note: the libRadtran software package for radiative transfer calculations e description and examples of use. Atmos Chem Phys 2005;5:1855e77. Anderson G, Glough S, Kneizys F, Chetwynd J, Shettle E. AFGL atmospheric constituent profiles (0e120km). Technical Report AFGL-TR-86-0110. MA 01736: AFGL (OPI), Hanson AFB; 1986. Csiszar I, Gutman G. Mapping global land surface albedo from NOAA AVHRR. J Geophys Res 1999;104(D6):6215e28. Shettle EP. Models of aerosols, clouds and precipitation for atmospheric propagation studies. In: Paper Presented at Conference on Atmospheric Propagation in the UV, Visible, IR and MM-Region and Related System Aspects, NATO Adv. Group for Aerosp. Res. And Dev., Copenhagen; 1989.
484
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
[37] Wild M, Gilgen H, Roesch A, Ohmura A, Long CN, Dutton EG, et al. From dimming to brightening: decadal changes in surface solar radiation. Science 2005;308:847e50. [38] Abakumova GM, Gorbarenko EV, Nezval EI, Shilovtseva OA. Fifty years of actinometrical measurements in Moscow. Int J Remote Sens 2008;29: 2629e65. [39] Ruckstuhl C, Philipona R, Behrens K, Coen MC, Dürr B, Heimo A, et al. Aerosol and cloud effects on solar brightening and the recent rapid warming. Geophys Res Lett 2008;35:L12708. http://dx.doi.org/10.1029/ 2008GL034228. [40] Russak V. Changes in solar radiation and their influence on temperature trend in Estonia (1955-2007). J Geophys Res 2009;114:D00D01. http://dx.doi.org/ 10.1029/2008JD010613. [41] Ovadneiviate_ J, Kvietkus K, Marsalka A. 2002 summer fires in Lithuania: impact on the Vilnius city air quality and the inhabitants health. Sci Total Environ 2006;356(1e3):11e21.
[42] Hänniken OO, Salonen RO, Koistinen K, Lanki T, Barregard L, Jantunen M. Population exposure to fine particles and estimated excess mortality in Finland from an East European wildfire episode. J Expo Sci Environ Epidemiol 2009;19:414e22. [43] Pace G, Meloni D, di Sara A. Forest fire aerosol over the Mediterranean basin during summer 2003. J Geophys Res 2005;110(D21). http://dx.doi.org/ 10.1029/2005JD005986. [44] García-Herréra R, Diaz J, Trigo RM, Luterbacher J, Fischer AM. A review of the eurpean summet heatwave of 2003. Crit Rev Environ Sci Technol 2010;40(4): 267e306. [45] Péré JC, Mallet M, Pont V, Bessagnet B. Impact of aerosol direct radiative forcing on the radiative budget, surface heat fluxes, and atmospheric dynamics during the heat wave of summer 2003 over western Europe: a modeling study. J Geophys Res 2011;116(D23). http://dx.doi.org/10.1029/ 2011JD016240.
Contents lists available at ScienceDirect
Renewable Energy journal homepage: www.elsevier.com/locate/renene
The aerosol effect on direct normal irradiance in Europe under clear skies E. Nikitidou*, A. Kazantzidis, V. Salamalikis Laboratory of Atmospheric Physics, Physics Department, University of Patras, 26500 Rio, Greece
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 August 2013 Accepted 12 February 2014 Available online 12 March 2014
The effect of spatial and temporal variability of aerosol optical depth (AOD) on direct normal irradiance (DNI) under clear skies is studied, with the synergetic use of satellite and ground-based data as well as calculations from a radiative transfer model. The area of interest is Europe; data from May to September during 13 years (2000e2012) are analyzed. The aerosol effect on DNI is high in areas influenced by desert dust intrusions and intense anthropogenic activities, such as the Mediterranean basin and the Po Valley in Italy. In May, the attenuation of DNI from aerosols, over these areas, can reach values up to 35% and 45% respectively, which corresponds to 4 and 6 kWh m2 per day. In most areas, even for periods with lower values of AOD, the attenuation of DNI is found to be around 20%, which corresponds to about 2 e3 kWh m2 less received DNI per day, compared to the corresponding value on an aerosol clean day. However, the DNI has increased during the recent years, due to the decreasing tendency of AOD over most areas of Europe. The increase is around 6e12%, which corresponds to an amount of 0.5e1.25 more kWh m2 received per day, compared to a clean day. The percentage differences of daily DNI from the corresponding monthly climatological value reveals that day-to-day differences (due to AOD changes) from the monthly mean, by 20%, can occur. The significance of the aerosol changes in Europe reveals the necessity for near real-time measurements or forecasts of AOD when reliable estimations of DNI are required. Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: Aerosols Direct normal irradiance DNI Satellite estimates
1. Introduction Aerosols are one of the most important constituents in the atmosphere that affect the incoming solar radiation, either directly through absorbing and scattering processes or indirectly by changing the optical properties and lifetime of clouds. Even though aerosols have been in the center of scientific research for years; their high spatiotemporal variability and complex atmospheric interactions induce challenges in the determination of their radiative forcing, which still holds high uncertainties [1]. Many studies focus on the spatiotemporal distribution of aerosol properties in Europe. Chubarova [2] studied the seasonal distribution, using data from the MODIS instrument and the AERONET network. The highest aerosol loads were observed in south and southeastern parts of Europe during the warmer months of the year. Over these areas, permanent high values of the Ångström a
* Corresponding author. Tel.: þ30 2610996079; fax: þ30 2610997989. E-mail addresses: [email protected] (E. Nikitidou), [email protected] (A. Kazantzidis), [email protected] (V. Salamalikis). http://dx.doi.org/10.1016/j.renene.2014.02.034 0960-1481/Ó 2014 Elsevier Ltd. All rights reserved.
coefficient were found indicating the presence of anthropogenic aerosols. Marmer and Langmann [3] studied the interannual variability of the aerosol distribution over Europe, using a regional atmospheric chemistry model. They highlighted the role of meteorological conditions which can induce a variability of up to 100% in the monthly mean of the aerosol load. The weekly variability of aerosols in Europe presents the lowest values during the weekend due to the decrease in transport and other anthropogenic activities [4e6]. As a result of legislations regarding atmospheric pollution, negative trends in aerosol amounts have been observed during the last years in many areas of Europe, the so-called “brightening effect” [7e11]. In order to complement the ground measurements and account for their limited spatial coverage, satellites have been employed for Earth’s and atmosphere’s observation. Satellites can provide global coverage and produce long-term datasets. The MODerate resolution Imaging Spectroradiometer (MODIS), on board the NASA Terra satellite, has been in operation since February 2000. The sunsynchronous satellite provides global coverage every 1e2 days and acquires data in 36 spectral bands (http://modis.gsfc.nasa.gov/) [12]. The Collection 005 (C005) is the dataset of MODIS products
476
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
provided by the version 5.2 of the algorithm, which has replaced older versions [13,14]. The prelaunch expected error in AOD retrievals over land is (0.05 þ 0.15 AOD), while AOD is estimated to within this expected error in more than 60% and 72% of the cases over ocean and over land respectively [15]. The MODIS data have been extensively validated against ground-based measurements [16e19] and used to calculate, amongst others, the aerosol trends and events [7,20,21], as well as the surface shortwave, ultraviolet [22e24] and visible solar irradiance [25]. The accurate knowledge of the amount of solar irradiance reaching a surface on the Earth and its temporal variability is essential for the efficient performance of solar power applications. The most important constituent of solar radiation is the Direct Normal Irradiance (DNI), which is the radiation received by a surface perpendicular to the direction of the sun. Although systems measuring the Global Horizontal Irradiance (GHI) are abundant, DNI measurements are not so common and these data are usually estimated indirectly with the use of radiative transfer or decomposition models [26e28]. Under clear skies, aerosols become the dominant factor that affect the intensity of solar irradiance reaching the ground. It has been shown that the variability in DNI due to aerosols is more important than the one induced in GHI [29], while the uncertainty in its calculation is dominated by uncertainties in the aerosol optical properties [30]. Suri et al. [31] studied the DNI over Europe as this is provided by 5 different datasets. The annual sum was found to reach the highest values in areas of the Mediterranean, Southern and Central Spain, Portugal, Sicily, Sardinia and Provence. Gueymard [29] used AOD measurements from 180 stations of the AErosol RObotic NETwork (AERONET) over the world to study the variability of AOD and its effect on DNI and GHI. He concluded that some areas experience a high variability that makes resource assessments potentially too optimistic for bankability if based only on limited data series. In this study, the effect of AOD on DNI under clear skies is studied, with the synergetic use of satellite data from the MODIS instrument, complementary data from the AERONET network and calculations from a radiative transfer model. The area of interest is Europe and data from a 13-year period (2000e2012) are analyzed.
The DNI temporal and spatial variability due to aerosols is examined, as well as the increase induced by the observed decreasing tendency of AOD over Europe. The uncertainties of DNI estimations induced by the use of aerosol climatological values are investigated. 2. Data and methodology The AOD at 550 nm is taken from the MODIS Terra daily Level-3 data (Collection 5.1) which have a spatial resolution of 1 1 (100 100 km). Terra was launched in December 1999 and started providing MODIS data in March 2000. The MODIS AOD at 550 nm, from the Terra satellite is used in this study from the beginning of operation (March 1st, 2000) until the end of 2012. The AOD wavelength dependence, described by the Ångström-a exponent, is provided by MODIS but is not used in this study since it still presents considerable errors when compared with ground-based measurements [14,18]. For the Ångström-a exponent, the level 2.0 climatological data from AERONET were used. The level 2.0 data are pre- and post-field calibrated, automatically cloud-screened and manually checked. The monthly climatological values of Ångstrom a exponent (440e 870 nm) were chosen for 37 stations in Europe, which present data availability longer than 3 years during the period 2000e2013. These data were then interpolated for the region of our study in order to obtain the spatial resolution of MODIS (1 1 ). The interpolation was performed using the method of Ordinary Kriging, combined with a linear semi-variogram model. The AERONET sites used in this study are shown in Fig. 1. The area examined (latitude: 29.5 N e 59.5 N, longitude: 9.5 W e 38.5 E) covers Europe, with the exception of the northern latitudes of Scandinavia, the northern coasts of Africa and the Mediterranean coasts of Middle East. In order to estimate the DNI on the surface, the radiative transfer model SBDART was used [32], which is included in the LibRadtran package [33] (www.libradtran.org). Typical vertical profiles for the basic atmospheric gases, pressure and temperature, for the midlatitudes, were used [34]. The surface albedo was set at 0.2 for the shortwave (SW) range (280e3000 nm) [35]. The aerosol vertical profile is described by Shettle [36]. The AOD is described by the
Fig. 1. AERONET stations used for the retrieval of the gridded monthly averages of Ångström a exponent.
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
Ångström-a exponent and b coefficient, the latter being calculated by the equation:
b ¼ s*la
(1)
where s is the MODIS AOD at l ¼ 550 nm and a the AERONET Ångström exponent. The solver of the radiative transfer equation is the discrete ordinate algorithm with 6 streams considered and the sphericity of the Earth is taken into account. The AOD provided by MODIS Terra was considered to be constant during the day. For every satellite pixel and day, the clear-sky DNI was calculated every 15 min, from sunrise to sunset, as a function of the solar zenith angle, and the Ångström-a and b parameters. A correction for the Earth e Sun distance was subsequently applied. The DNI was then calculated by integrating the 15min values over the day length. The DNI is also calculated for aerosol-free conditions in order to derive the absolute and percentage (%) differences in the received DNI between the measured and the aerosol-free conditions:
Diff ¼ DNI DNIclear Diffð%Þ ¼
DNI DNIclear *100 DNIclear
(2) (3)
where DNIclear is the calculated DNI under aerosol-free conditions. These values were calculated by running the model with the same conditions as before, but with no aerosols present. The latitudes lower than 34 are not depicted in the maps of the following figures, as the satellite provides few data during the investigated period due to the high surface albedo of the deserts in these areas. During the 13-year period examined, only months which had data availability higher than 40% (more than 12 days per month) were considered. As a result, only data during the period MayeSeptember of each year were analyzed, since the data availability during the rest of the months is not adequate due to the dominance of thick cloudiness. 3. Results 3.1. Seasonal variability of the Ångström-a exponent The monthly Ångström-a exponent climatological values, taken from the selected AERONET stations in Europe, were interpolated to match the MODIS spatial resolution (1 1 ). The monthly Ångström a exponent maps for Europe during the MayeSeptember period are presented in Fig. 2. The lowest values (around 0.8) are presented during May in Southern Europe, Mediterranean and the northern coasts of Africa and constitute and evidence for the frequent desert-dust intrusions during this period. The influence of desert dust is very important in July and August too, for Southern Spain and the western coasts of Africa but is limited in Eastern Mediterranean. The highest values of the Ångström-a exponent (1.4e1.8) are observed during summer in Central-Eastern Europe and the Balkans, indicating the presence of fine mode particles, which persist during this period due to the reduction of precipitation. In September, these values decrease further and moderate values are observed in continental Europe, while maritime areas are described by values around 0.8e1.1. 3.2. DNI changes due to the AOD variability 3.2.1. The aerosol effect on DNI The absolute and percentage differences of DNI under clear skies have been calculated for every day and every satellite pixel during
477
the examined period. The monthly average values of the percentage difference over the 13-year period are presented in Fig. 3. In Mediterranean, the highest percentage differences are observed in May reaching values down to 35%. This corresponds to an attenuation of about 4 kWh m2 per day. The highest reductions during this month are evident around Po Valley, with differences down to 45% (6 kWh m2 per day). Lower values of 25 to 15% (3 to 2 kWh m2 per day) are revealed in Spain, Southern France, Eastern and Northern Europe, while Central Europe and the southern parts of Ireland and Britain present an attenuation of around 30 to 25% (4 to 3.5 kWh m2 per day). The data availability at the northern parts of Ireland and Britain was lower than 40% so the results for these areas were considered not significant for this study. In June, the percentage difference is slightly lower than that over the Mediterranean basin and Northern Europe but moderately increases in areas of central Europe. In Northern France, the Low Countries, most parts of Germany and Poland, the attenuation of DNI is around 30e35%, while it remains high around the Po Valley and the far western part of the Mediterranean basin. The situation over continental Europe is similar in July, with the exception of a slight increase of the aerosol effect in North-Eastern regions. In the Mediterranean, the effect of aerosols on DNI decreases over Portugal and Po Valley (around 10 and 30% respectively) but becomes stronger over the western coasts of Africa, where percentage difference of up to 40% (5 kWh m2 per day) are found. In August, the aerosol effect on DNI is around 25 to 30% (3 to 3.5 kWh m2 per day) over most parts of Mediterranean, Central and Eastern Europe and slightly lower over France and Spain, where differences of around 2 kWh m2 per day are calculated. The effect is stronger in South-Western Mediterranean and in the interior areas of Turkey. In September, the percentage differences decrease in Central and Eastern Europe, while there is an increase in the North Sea (25 to 35%). The importance of the aerosol effect on DNI is clear, as even for months with low aerosol loads, the attenuation of DNI is around 20%, which corresponds to about 2 to 3 kWh m2 per day. This effect becomes even more significant at areas burdened with aerosols throughout most time of the year, such as the industrial region of Northern Italy. 3.2.2. Annual differences of DNI For each year, the average values of modeled DNI have been calculated over the period MayeSeptember. Fig. 4 presents the percentage differences, for every year, between these mean values and the average over the period MayeSeptember for all years. There is an increasing tendency in the calculated differences. During the first years of the period examined, the differences in DNI are negative; they reach low values, around zero, in the middle of the period and then increase during the last years, indicating a positive trend in the received DNI due to the reduction of AOD (“brightening” effect). This “brightening” effect, concerning the European area that is studied here, has been identified in various other scientific studies [37e40]. During the years 2000e2003, the DNI received during the MayeSeptember period is lower than the average derived from the entire 13-year period, for most regions of Europe. Central Europe received lower DNI around 4e10% during 2000 and 2001, which corresponds to 1 to 0.25 kWh m2 per day. However, positive differences (2e4%) are revealed in some areas of Northern Europe. In 2002, Eastern Europe and the Balkans present the highest negative difference of DNI, between 8 and 18% (0.75e1.75 kWh m2 per day) compared to the average; this probably may be due to excessive wildfires in this area [41,42]. In 2003, low negative values ( 10 to 4%, or 1 to 0.25 kWh m2 per day) are observed in West-Central Europe and Western
478
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
Fig. 2. Monthly average maps of the Ångström a exponent over Europe.
Mediterranean due to exceptionally hot and dry conditions and large-scale intense forest fires [43e45]. In 2004, small increases start to appear, which reach þ2 to þ8%, in 2005, at latitudes between 50 and 55 N. From 2006 to 2008, the differences are small, around 4%, which corresponds to an aerosol effect on DNI less than 0.5 kWh m2 per day. However, increasing values of DNI appear during the last years (2009e2012). In 2012, 6e12% more DNI is received on most areas of Central and Eastern Europe, the Balkans and the Italian peninsula. These positive differences correspond to an increase of 0.5e1.25 kWh m2 in received DNI per day. These differences are further studied in the next section by averaging them over specific regions of Europe. 3.2.3. DNI and AOD differences per area The above mentioned differences in DNI for each satellite pixel were then averaged over each of the five different areas in Europe. These areas (Fig. 5) are Central-Western Mediterranean (34.5 N e 44.5 N, 9.5 W e 18.5 E), Eastern Mediterranean-Black
Sea (32.5 N e 46.5 N, 18.5 E e 36.5 E), Central-Western Europe (44.5 N e 54.5 N, 4.5 W e 14.5 E), Eastern Europe (44.5 N e 55.5 N, 14.5 E e 38.5 E) and Northern Europe (54.5 N e 59.5 N, 5.5 E e 29.5 E). The percentage differences of the annual mean AOD and DNI values, from the climatological values over the whole period (2000e2012) were calculated for each area and for the MayeSeptember period (Fig. 6). The annual mean AOD values (Fig. 6a) are higher than the 13-year average, until 2003, after which a decrease starts, which becomes more evident during the end of the period. During 2010e2012, the AOD in CentralWestern Europe and Central-Western Mediterranean is lower by 15e17% (w0.04) than the 13-year average, while in Eastern Europe and Eastern Mediterranean a similar decrease is revealed only in 2011 and 2012. With the exception of year 2002, where strong positive differences are observed, the area of Northern Europe has the lowest variability during the examined period. The corresponding percentage differences in DNI are presented in Fig. 6b. Over all regions, the decrease during the beginning of the
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
479
Fig. 3. Monthly mean values of the aerosol effect (%) on DNI. Only the period MayeSeptember is presented, which fulfills the data availability requirements.
examined period and the increase at the end is evident. With the exception of Northern Europe, all areas received around 3% (0.3 kWh m2 per day) less DNI during 2000 and 2001. In 2002, the highest decrease is revealed in Eastern Europe (w10%, 0.9 kWh m2 per day). The differences are moderate for all regions in the following years but increased DNI values are found after 2009. During the end of the examined period, an increase in the received DNI of 4e6% (0.4e0.6 kWh m2 per day) can be observed in all regions, except Northern Europe. Fig. 7 presents the percentage differences of the daily DNI from the corresponding monthly climatological value, calculated from the daily DNI values during the entire time period. The positive trend in DNI differences is evident in all areas, when looking at the daily differences as well, except in Northern Europe. Additionally, in all areas, day-to-day percentage differences of DNI from the monthly climatological value, due to changes in AOD of up to 20%, are revealed. The high negative differences, which were observed in Eastern Europe in 2002, can now be seen that occur during the
period of mid August e mid September. During this time, the differences can reach values higher than 40%. It is during the same period of 2002 that high differences in Northern Europe appear as well. 3.3. DNI from MODIS In this section, the monthly climatological MODIS AOD values, derived from the 13-year period, have been used as input data to the SBDART model, in order to estimate the DNI values. The goal is to examine the differences between these DNI values and the monthly ones, calculated from the available daily data during the examined period. The analysis is performed for the months May to September and the results are averaged over each of the 5 areas described before. Fig. 8 shows the differences between the DNI derived from the monthly MODIS-AOD climatology and the corresponding monthly mean DNI calculated from all the daily AOD data during the period of the study. The derived differences are negative
480
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
Fig. 4. Differences (%) between the mean DNI for each year (2000e2012) and the average for the 13-year period. Only data for MayeSeptember are considered.
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
481
(0.4 kWh m2 per day). In Central-Western Mediterranean the underestimation is around 2.5e3%, which corresponds to 0.21e 0.28 kWh m2 per day, except in May when it reaches 0.33 kWh m2 per day. In May, the highest differences, for each region, are found, close to and larger than 0.3 kWh m2 per day, except for Eastern Europe, which has a slightly higher underestimation in August (0.37 kWh m2 per day). These differences are smaller than half a kWh m2 per day on average, but still have to be taken into account if someone chooses to use the satellite climatology for modeling the DNI received by a solar power farm. 4. Conclusions
Fig. 5. Europe divided into 5 areas of interest: Central-Western Mediterranean (34.5 N e 44.5 N, 9.5 W e 18.5 E), Eastern Mediterranean-Black Sea (32.5 N e 46.5 N, 18.5 E e 36.5 E), Central-Western Europe (44.5 N e 54.5 N, 4.5 W e 14.5 E), Eastern Europe (44.5 N e 55.5 N, 14.5 E e 38.5 E) and Northern Europe (54.5 N e 59.5 N, 5.5 E e 29.5 E).
for all months and all regions, showing that when the MODIS-AOD monthly climatology is used to estimate the DNI, there is an underestimation of a few percent, as opposed to the monthly DNI values calculated from the daily MODIS data. The lowest differences are observed in Northern Europe in June and July, around 0.16 kWh m2 per day, which corresponds to an underestimation of 1.5%. Northern Europe has higher differences for the rest of the months and the maximum difference for the area is found in May
The effect of aerosols on DNI in Europe was calculated for clearsky conditions on a daily basis, using data from the MODIS instrument, on board the NASA Terra satellite, and complementary data of Ångström a exponent from 37 AERONET stations. The period of the study was from the beginning of the satellite operation (March 2000) until the end of 2012. The DNI was estimated, using the SBDART radiative transfer model, with a spatial resolution of 1 1 ; the corresponding aerosol effect was derived for every satellite pixel. The monthly climatological values of the Ångström-a exponent, taken from AERONET, were interpolated to the MODIS spatial resolution. The lowest values, around 0.6, were observed in spring across the Mediterranean coasts and were evident of the intrusion of desert-dust aerosols over this area. Permanent high values, indicating the presence of fine-mode aerosols, were observed over many areas of Central Europe, most notably the Po Valley. The calculated percentage (%) difference was high in areas influenced by desert-dust intrusions and intense anthropogenic
Fig. 6. Difference (%) between the mean AOD (6a) and DNI (6b) value for each year (2000e2012) and the average for the 13-year period for the 5 areas of interest.
482
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
Fig. 7. Difference (%) between the daily DNI and the corresponding monthly mean, for each of the 5 areas of interest.
activities, such as the Mediterranean basin and the Po Valley. The attenuation of DNI due to aerosols, over these areas, can reach values up to 35% and 45% and corresponds to 4 and 6 kWh m2 per day, respectively. In June and July, a significant decrease in DNI values (30e35%) due to aerosols was calculated over the Low Countries, Germany and Poland. In most areas, even for periods with lower AOD values, the attenuation of DNI was found to be
around 20%, which corresponds to values about 2e3 kWh m2 lower than that received per day. The MODIS-AOD values and the differences in DNI were averaged over each of the 5 different areas in Europe: Central-Western Mediterranean, Eastern Mediterranean-Black Sea, Central-Western, Eastern and Northern Europe. During 2010e2012, the AOD in Central-Western Europe and Central-Western Mediterranean was
Fig. 8. Difference (%) between the DNI derived from the monthly AOD climatology and the corresponding monthly mean DNI calculated from the available AOD daily data during the period 2000e2012 and over each of the 5 areas of interest.
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
lower by 15e17% (w0.04) than the 13-year average, while in Eastern Europe and Eastern Mediterranean a similar decrease was found only in 2011 and 2012. When compared to the climatological values of DNI for the 2000e2012 period, it was shown that the decreasing tendency of the AOD values resulted in a DNI change of 10%. The percentage differences of the daily DNI from the corresponding climatological monthly value showed that, apart from the positive trend in the DNI differences in all areas, day-to-day differences from the monthly mean of up to 20% can occur. The effect of calculating the DNI from the monthly AOD climatology instead of the monthly DNI mean calculated from all the daily AOD data was investigated. According to the results, the DNI was found to be underestimated in the range 0.16e0.37 kWh m2 during the Maye September period. DNI is of vital importance to solar power systems; detailed information of its spatiotemporal variations is often required, rather than for GHI, a parameter that is more usually examined. A detailed study of the changes in DNI over Europe due to the variability of AOD, the DNI “brightening” effect and the uncertainties induced from the use of an aerosol monthly climatology instead measurements was provided. The decreasing tendency of AOD over Europe caused significant changes in DNI. The increased energy production from solar power plants in most areas of the European continent set the need for additional studies about the DNI variability. This study highlighted the need for near real-time measurements or reliable forecasts of AOD for the estimation of DNI and emphasized to the use of AOD climatological data for Europe.
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
Acknowledgments The authors would like to thank the principal investigators of the AERONET stations (listed in Fig. 1) for providing the data. The MODIS teams are acknowledged for providing the Collection 5 aerosol products from the Terra and Aqua satellites. The LibRadtran team (www.libradtran.org) is acknowledged for providing the model algorithm. This study was mainly conducted and funded by project “Hellenic Network of Solar Energy” (HNSE, 09SYN-32-778), funded by the General Secretariat for Research and Technology, Greek Ministry of Education, Lifelong Learning and Religious Affairs. References [1] Intergovernmental Panel on Climate Change (IPCC). Climate change 2007, working group i report: “The physical science basis”, summary for policymakers. Paris; February 2007. [2] Chubarova NY. Seasonal distribution of aerosol properties over Europe and their impact on UV irradiance. Atmos Meas Tech 2009;2:593e608. [3] Marmer E, Langmann B. Aerosol modeling over Europe: 1. Interannual variability of aerosol distribution. J Geophys Res 2007;112(D23S15). http:// dx.doi.org/10.1029/2006JD008113. [4] Bäumer D, Rinke R, Vogel B. Weekly periodicities of aerosol optical thickness over Central Europe e evidence of an anthropogenic direct aerosol effect. Atmos Chem Phys 2008;8:83e90. http://dx.doi.org/10.5194/acp-8-83-2008. [5] Xia X, Eck TF, Holben BN, Phillippe G, Chen H. Analysis of the weekly cycle of aerosol optical depth using AERONET and MODIS data. J Geophys Res 2008;113(D14217). http://dx.doi.org/10.1029/2007JD009604. [6] Georgoulias AK, Kourtidis KA. On the aerosol weekly cycle spatiotemporal variability over Europe. Atmos Chem Phys 2011;11:4611e32. [7] Papadimas CD, Hatzianastassiou N, Mihalopoulos N, Querol X, Vardavas I. Spatial and temporal variability in aerosol properties over the Mediterranean basin based on 6-year (2000e2006) MODIS data. J Geophys Res 2008;113(D11205). http://dx.doi.org/10.1029/2007JD009189. [8] Karnieli A, Derimian Y, Indoitu R, Panov N, Levy RC, Remer LA, et al. Temporal trend in anthropogenic sulfur aerosol transport from central and eastern Europe to Israel. J Geophys Res 2009;114(D00D19). http://dx.doi.org/10.1029/ 2009JD011870. [9] Streets DG, Yan F, Chin M, Diehl T, Mahowald N, Schultz M, et al. Anthropogenic and natural contributions to regional trends in aerosol optical depth,
[22]
[23] [24]
[25]
[26]
[27]
[28]
[29] [30]
[31]
[32]
[33]
[34]
[35] [36]
483
1980e2006. J Geophys Res 2009;114(D00D18). http://dx.doi.org/10.1029/ 2008JD011624. Koukouli ME, Kazadzis S, Amiridis V, Ichoku C, Balis DS, Bais AF. Signs of a negative trend in the MODIS aerosol optical depth over the Southern Balkans. Atmos Environ 2010;44:1219e28. Yoon J, von Hoyningen-Huene W, Vountas M, Burrows J. Analysis of linear long-term trend of aerosol optical thickness derived from SeaWiFS using BAER over Europe and South China. Atmos Chem Phys 2011;11:12149e67. Remer LA, Kaufman YJ, Tanré D, Mattoo S, Chu DA, Martins JV, et al. The MODIS aerosol algorithm, products and validation. J Atmos Sci 2005;62(4): 947e73. Levy RC, Remer LA, Dubovik O. Global aerosol optical properties and application 15 to moderate resolution imaging spectroradiometer aerosol retrieval over land. J Geophys Res 2007a;112(D13210). http://dx.doi.org/10.1029/ 2006JD007815. Levy RC, Remer LA, Mattoo S, Vermote EF, Kaufman YJ. Second-generation operational algorithm: retrieval of aerosol properties over land from inversion of moderate resolution imaging spectroradiometer spectral reflectance. J Geophys Res 2007b;112(D13211):20. http://dx.doi.org/10.1029/2006JD007811. Remer LA, Kleidman RG, Levy RC, Kaufman YJ, Tanré D, Mattoo S, et al. Global aerosol climatology from the MODIS satellite sensors. J Geophys Res 2008;113(D14S07). http://dx.doi.org/10.1029/2007JD009661. Chu DA, Kaufman YJ, Ichoku C, Remer LA, Tanré D, Holben BN. Validation of MODIS aerosol optical depth retrieval over land. Geophys Res Lett 2002;29(12). http://dx.doi.org/10.1029/2001GL013205. Santese M, De Tomasi F, Perrone MR. Moderate resolution imaging spectroradiometer (MODIS) and aerosol robotic network (AERONET) retrievals during dust outbreaks over the mediterranean. J Geophys Res 2007;112(D18201). http://dx.doi.org/10.1029/2007JD008482. Levy RC, Remer LA, Kleidman RG, Mattoo S, Ichoku C, Kahn R, et al. Global evaluation of the collection 5 MODIS dark-target aerosol products over land. Atmos Chem Phys 2010;10:10399e420. De Meij A, Lelieveld J. Evaluating aerosol optical properties observed by ground-based and satellite remote sensing over the Mediterranean and the Middle East in 2006. Atmos Res 2011;99:415e33. Gkikas A, Hatzianastassiou N, Mihalopoulos N. Aerosol events in the broader Mediterranean basin based on 7-year (2000-2007) MODIS C005 data. Ann Geophys 2009;27:3509e22. Zhang J, Reid JS. A decadal regional and global trend analysis of the aerosol optical depth using a data-assimilation grade over-water MODIS and Level 2 MISR aerosol products. Atmos Chem Phys 2010;10:10949e63. Bisht G, Venturini V, Islam S, Jiang L. Estimation of the net radiation using MODIS (Moderate Resolution Imaging Spectroradiometer) data for clear sky days. Remote Sens Environ 2005;97:52e67. Tang B, Li Z-L, Zhang R. A direct method for estimating net surface shortwave radiation from MODIS data. Remote Sens Environ 2006;103:115e26. Huang G, Ma M, Liang S, Liu S, Li X. A LUT-based approach to estimate surface solar irradiance by combining MODIS and MTSAT data. J Geophys Res 2011;116(D22201). http://dx.doi.org/10.1029/2011JD016120. Nikitidou E, Kazantzidis A. On the differences of ultraviolet and visible irradiance calculations in the Mediterranean basin due to model e and satellite e derived climatologies of aerosol optical properties. Int J Climatol; 2012. http:// dx.doi.org/10.1002/joc.3638. Battles FJ, Rubio MA, Tovar J, Olmo FJ, Alados-Arboledas L. Empirical modeling of hourly direct irradiance by means of hourly global irradiance. Energy 2000;25:675e88. López G, Rubio MA, Batlles FJ. Estimation of hourly direct normal from measured global solar irradiance in Spain. Renew Energy 2000;21: 175e86. Janjai S, Sricharoen K, Pattarapanitchai S. Semi-empirical models for the estimation of clear sky solar global and direct normal irradiances in the tropics. Appl Energy 2011;88:4749e55. Gueymard CA. Temporal variability in direct and global irradiance at various time scales as affected by aerosols. Sol Energy 2012;86:3544e53. Halthore RN, Schwartz SE, Michalsky JJ, Anderson GP, Ferrare RA, Holben BN, et al. Comparison of model estimated and measured direct-normal solar irradiance. J Geophys Res 1997;102(D25):29991e30002. Suri M, Remund J, Cebecauer T, Hoyer-Klick C, Dumortier D, Huld T, et al. In: SolarPACES conference “Concentrated solar power and Chemical Energy systems”, 15e18 Sep. 2009, Berlin, Germany; 2009. Ricchiazzi P, Yang S, Gautier C, Sowle D. SBDART: a research and Teaching software tool for plane-parallel radiative transfer in the Earth’s atmosphere. Bull Am Meteorol Soc 1998;79:2101e14. Mayer B, Kylling A. Technical note: the libRadtran software package for radiative transfer calculations e description and examples of use. Atmos Chem Phys 2005;5:1855e77. Anderson G, Glough S, Kneizys F, Chetwynd J, Shettle E. AFGL atmospheric constituent profiles (0e120km). Technical Report AFGL-TR-86-0110. MA 01736: AFGL (OPI), Hanson AFB; 1986. Csiszar I, Gutman G. Mapping global land surface albedo from NOAA AVHRR. J Geophys Res 1999;104(D6):6215e28. Shettle EP. Models of aerosols, clouds and precipitation for atmospheric propagation studies. In: Paper Presented at Conference on Atmospheric Propagation in the UV, Visible, IR and MM-Region and Related System Aspects, NATO Adv. Group for Aerosp. Res. And Dev., Copenhagen; 1989.
484
E. Nikitidou et al. / Renewable Energy 68 (2014) 475e484
[37] Wild M, Gilgen H, Roesch A, Ohmura A, Long CN, Dutton EG, et al. From dimming to brightening: decadal changes in surface solar radiation. Science 2005;308:847e50. [38] Abakumova GM, Gorbarenko EV, Nezval EI, Shilovtseva OA. Fifty years of actinometrical measurements in Moscow. Int J Remote Sens 2008;29: 2629e65. [39] Ruckstuhl C, Philipona R, Behrens K, Coen MC, Dürr B, Heimo A, et al. Aerosol and cloud effects on solar brightening and the recent rapid warming. Geophys Res Lett 2008;35:L12708. http://dx.doi.org/10.1029/ 2008GL034228. [40] Russak V. Changes in solar radiation and their influence on temperature trend in Estonia (1955-2007). J Geophys Res 2009;114:D00D01. http://dx.doi.org/ 10.1029/2008JD010613. [41] Ovadneiviate_ J, Kvietkus K, Marsalka A. 2002 summer fires in Lithuania: impact on the Vilnius city air quality and the inhabitants health. Sci Total Environ 2006;356(1e3):11e21.
[42] Hänniken OO, Salonen RO, Koistinen K, Lanki T, Barregard L, Jantunen M. Population exposure to fine particles and estimated excess mortality in Finland from an East European wildfire episode. J Expo Sci Environ Epidemiol 2009;19:414e22. [43] Pace G, Meloni D, di Sara A. Forest fire aerosol over the Mediterranean basin during summer 2003. J Geophys Res 2005;110(D21). http://dx.doi.org/ 10.1029/2005JD005986. [44] García-Herréra R, Diaz J, Trigo RM, Luterbacher J, Fischer AM. A review of the eurpean summet heatwave of 2003. Crit Rev Environ Sci Technol 2010;40(4): 267e306. [45] Péré JC, Mallet M, Pont V, Bessagnet B. Impact of aerosol direct radiative forcing on the radiative budget, surface heat fluxes, and atmospheric dynamics during the heat wave of summer 2003 over western Europe: a modeling study. J Geophys Res 2011;116(D23). http://dx.doi.org/10.1029/ 2011JD016240.