返回 相似
资源描述：
ARTICLES https//doi.org/10.1038/s41560-019-0412-4 1 Institute for Atmospheric and Climate Science, ETH Zürich, Zürich, Switzerland. 2 Institute for Biodiversity and Ecosystem Dynamics, Faculty of Science, University of Amsterdam, Amsterdam, the Netherlands. 3 Climate Policy Group, Institute for Environmental Decisions, ETH Zürich, Zürich, Switzerland. 4 National Meteorological Information Centre, China Meteorological Administration, Beijing, China. 5 Van ’t Hoff Institute for Molecular Sciences, Faculty of Science, University of Amsterdam, Amsterdam, the Netherlands. 6 Energy Transition Studies, Energy Research Centre of the Netherlands, Netherlands Organisation for Applied Scientific Research ECN-TNO, Amsterdam, the Netherlands. 7 School of Advanced International Studies SAIS Europe, Johns Hopkins University, Bologna, Italy. *e-mail bart.sweertsenv.ethz.ch O ver the past two decades, solar photovoltaic PV electric- ity generation capacity has grown exponentially world- wide. Between 2000 and 2017, worldwide installed capacity increased from 4 to 385 GW 1 , consistently exceeding expectations 2,3 . China in particular is investing heavily in PV, increasing installed capacity from less than 1 GW in 2010 to 130 GW by the end of 2017 4 . In 2017, China accounted for over half of global PV capac- ity additions 5 . Having surpassed the 2020 PV development target of 110 GW, China is well on track to realize its goal of reaching 400 GW of installed PV capacity by 2030, to meet its commitment to the Paris Agreement of obtaining 20 of primary energy from renewable energy sources 6 . Solar radiation is often assumed to be constant over multiple years, but there is strong evidence for substantial multidecadal variations, referred to as ‘global dimming and brightening’ 7,8 . As a result of variations primarily in cloud characteristics and atmo- spheric aerosol concentrations, incoming radiation is scattered and absorbed, which modifies surface solar radiation 9,10 . In rapidly developing and heavily polluted regions, such as China, increasing anthropogenic aerosol emissions are considered a key cause of sub- stantial dimming 8 . Until recently, the extent and cause of observed radiation trends in China remained unclear due to uncertainties over data quality resulting from changes in the use of instruments and observational schedules 8,11,12 . However, Yang etuni00A0 al. 13 homog- enized a dataset covering 119 measurement stations by referenc- ing radiation observations to more numerous sunshine duration data from nearby stations. The resulting dataset showed consistent dimming across China between 1958 and 2005 −24 ± 1 Wm −2 ± s.d., followed by a period of brightening between 2005 and 2016 7 ± 2 Wm −2 . China is actively pursuing strategies to decrease air pollution, and shows signs of planning to phase out heavily pol- luting coal as its major energy source 14 . Thus, developing a deeper understanding of possible changes in surface solar radiation and their impact on a rapidly expanding PV sector is becoming increas- ingly important. Li etuni00A0al. 15 analysed satellite-derived solar radiation data to find a substantial aerosol-induced reduction of solar radia- tion in China, with large impacts particularly over eastern China. However, the satellite data used in that study were limited to 2003– 2015 and the aerosol effect was estimated using an atmospheric transfer model instead of being derived directly from measure- ments. Furthermore, the approach did not differentiate between anthropogenic and non-anthropogenic aerosols, which may lead to an overestimation of the effect of air pollution on surface radiation. Here, we take a different approach by applying a PV electricity generation model, as described below, to the homogenized ground- measured radiation dataset of Yang etuni00A0al. 13 , covering 1960–2015. We compute capacity factors CFs, defined as the ratio between a PV panel’s actual power output and maximum power output, deter- mined under laboratory conditions see Methods. We implicitly consider the effect of weather conditions as the simulated module’s temperature increases as a function of radiation levels, which leads to lower efficiencies. Annual average CFs of commercial PV systems generally range between 0.1 and 0.35 depending on surface radia- tion conditions and PV panel type 16 . We run experiments for three different panel settings horizontal fixed plane HOR, fixed plane with optimal tilt FIX and one-axis horizontal tracking 1AX. These experiments cover PV systems ranging from residential PV panels under suboptimal tilt to more optimized utility-scale PV Estimation of losses in solar energy production from air pollution in China since 1960 using surface radiation data Bart Sweertshairspace hairspace 1,2 *, Stefan Pfenningerhairspace hairspace 3 , Su Yang 1,4 , Doris Folini 1 , Bob vanuni00A0der Zwaan 5,6,7 and Martin Wild 1 China is the largest worldwide consumer of solar photovoltaic PV electricity, with 130thinspaceGW of installed capacity as of 2017. China’s PV capacity is expected to reach at least 400thinspaceGW by 2030, to provide 10 of its primary energy. However, anthro- pogenic aerosol emissions and changes in cloud cover affect solar radiation in China. Here, we use observational radiation data from 119 stations across China to show that the PV potential decreased on average by 11–15 between 1960 and 2015. The relationship between observed surface radiation and emissions of sulfur dioxide and black carbon suggests that strict air pollution control measures, combined with reduced fossil fuel consumption, would allow surface radiation to increase. We find that reverting back to 1960s radiation levels in China could yield a 12–13 increase in electricity generation, equivalent to an additional 14thinspaceTWh produced with 2016 PV capacities, and 51–74thinspaceTWh with the expected 2030 capacities. The corresponding economic benefits could amount to US1.9thinspacebillion in 2016 and US4.6–6.7thinspacebillion in 2030. NATURE ENERGY | www.nature.com/natureenergy ARTICLES NATURE ENERGY installations. We compute five-year mean baseline 1961–1965 and dimmed 2011–2015 CF scenarios to analyse the impact of changes in PV resources over 50 years in China. We explore the impact of historic changes in solar radiation by analysing CF value variations for the FIX experiment between 1960 and 2015 on the national and provincial level. We investigate the influence of the HOR and 1AX panel settings on these results and discuss the possible implications of historically observed radiation changes for present and future PV electricity generation. We find that air pollution accumulation since 1960 in China has decreased solar energy potential by up to 13, corresponding to a loss of 14 TWh of electricity in 2016. CFs 1960–2015 We calculate CFs at each of the 119 sites by feeding hourly solar radi- ation data derived from monthly mean records into the Global Solar Energy Estimator 17 GSEE and aggregating CF output to 5° 5° grid cell time series using the first difference method FDM 18 . We take spatially weighed averages over relevant grid cells to compute values for each province in China, as well as a nation-wide average see Methods for a more detailed description of the PV model setup and CF aggregation. Figure 1 shows that nation-wide averaged CFs in China decreased consistently starting in 1965. Between 1965 and 2008, mean CFs decreased by 12 from 0.162 to 0.142. After 2008, an increasing trend appears, but by 2015, CFs were still 10 lower than in 1965. Between 1960 and 2015, a negative change is observed in all months, with a strong reduction of 0.02–0.04 in the winter months November to February, and a weaker reduction of 0.01–0.02 in the summer months June to August Supplementary Fig. 1. Intra- annual variation of CFs increases through time Supplementary Fig. 2, potentially exacerbating intermittency issues inherent to the use of solar energy. The strong reduction in CFs observed in winter may be due to higher residential aerosol emissions as a result of bio- mass and fossil fuel-powered heating 19,20 . The five-year average baseline scenario 1961–1965 shows a spatial pattern of higher CFs in the northern and north-western provinces, with a maximum of 0.22 in Tibet, and lower values in the south-eastern provinces, with a minimum of 0.09 in Chongqing Fig. 2a. South-eastern provinces are characterized by, on aver- age, around 50 lower CFs than north-western provinces. Higher cloud cover over the humid subtropical south-eastern provinces than over the arid highlands of the north-western provinces is the primary cause for the difference in the geographical distribution of solar radiation. Additionally, regions with high altitudes, such as the Tibetan Plateau, receive more surface solar radiation as a result of a smaller total air column as shown for our dataset in Supplementary Fig. 3 21 . The overall distribution of CFs under dimmed 2011–2015 con- ditions is similar to that of the baseline scenario Fig. 2b. However, CFs are lower in 27 out of 31 provinces, with the largest absolute change observed in the heavily polluted eastern and southern prov- inces, as well as in Tibet and Qinghai Fig. 2c. The large change observed in Tibet and Qinghaitwo provinces largely located on the Tibetan Plateauis unexpected, as it is one of the least densely populated regions of China and contains limited industrial areas. However, as a result of the logarithmic relationship between aerosol optical depth AOD and aerosol radiative forcing 10 , regions with low background AOD, such as the Tibetan Plateau, may experience large changes in radiation due to relatively small changes in atmo- spheric aerosol concentrations. A comprehensive study analysing observations, reanalyses and global climate model ensemble simu- lations found that large-scale transported as well as locally emitted aerosols are a plausible cause for radiation reductions observed between 1960 and 2005 on the Tibetan Plateau 22 . Also, the absolute radiative effect of increasing AOD is proportional to the initial radi- ation levels. We find that percentage changes in CF between 1960 and 2015 are in fact largest in the densely populated and heavily pol- luted south-eastern regions Supplementary Fig. 4. The estimated changes increase the difference between the high-radiation north- western and low-radiation south-eastern provinces. Compared with the baseline, provincial CFs decreased by 20–28 in the five most heavily affected provinces. Four provinces show a small positive change 0 to 4. Robust linear regression trends computed for CF time series between 1960 and 2015 were found to be significant in all provinces with a negative change, but only in one out of four provinces with a positive change Supplementary Fig. 5. Sensitivity of results to panel settings Small-scale distributed PV systems are usually mounted at a fixed angle. The optimum mounting angle changes with latitude, but for residential applications, the angle is often constrained by the roof- top angle or building regulations. PV systems with one- or two-axis tracking, which more optimally utilize direct radiation by minimiz- ing the incidence angle of sunlight relative to the panel normal, are more frequently used in utility-scale solar farms. By the end of 2016, small-scale distributed PV accounted for 13 of installed PV capac- ity in China 23 . This share grew substantially in 2017, when distrib- uted capacity additions rose to 19 GW, accounting for 38 of newly installed PV capacity 24 . It is expected that the share of distributed PV will continue to rise between now and 2030 13 . Optimally tilted fixed-angle and tracking systems yield the larg- est efficiency gains in high-radiation areas in western and northern China Fig. 3a. We find that, for provinces in these regions, the FIX and 1AX experiments result in 13–21 and 47–54 higher yields, respectively, compared with HOR. In contrast, yields in south- ern and eastern China show no significant increase for FIX and a 14–25 increase for 1AX compared with HOR, due to a higher share of diffuse radiation in these regions. As a result, a large share of Chinese utility-scale solar parks is located in the high-radiation north-western provinces, such as Xinjiang, Qinghai and Gansu, while distributed PV is more evenly distributed across China 23 . Absorption and scattering of solar radiation by aerosols and clouds decrease the fraction of direct radiation and increase the fraction of diffuse radiation 25 as shown for our dataset in Supplementary Fig. 6. As the efficiency gains of PV panels equipped with tracking systems result from more effective use of direct radiation, they are more strongly affected in absolute terms and percentage wise by decreasing solar radiation resources than fixed panels are. Compared with 1961–1965 means, average CFs for 2011–2015 are 9 HOR, 11 FIX and 15 1AX lower Fig. 3b. The observed decrease in CFs for FIX is, on average, roughly half −11 as large as the benefit of switching from FIX to 1AX 24 under dimmed radiation conditions. The pattern of change is similar for all three experiments, with two exceptions Fig. 3b. First, in the most north-western province of Xinjiang, 0.170 0.165 0.160 0.155 0.150 0.145 0.140 0.135 1960 1970 1980 1990 2000 Mean 25th–75th percentiles 10th–90th percentiles 2010 Capacity f actor Year Fig. 1 | Changes in national CFs from 1960–2015 in China. National averages are based on CFs determined at 119 stations, which provide the essential solar radiation input data. National CFs were computed by taking the average value over all 5°thinspacethinspace5° FDM grid boxes covering China nthinspacethinspace51. NATURE ENERGY | www.nature.com/natureenergy ARTICLES NATURE ENERGY CF 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22 CF –0.035 –0.025 –0.015 –0.005 0.005 0.015 0.025 0.035 a b c Fig. 2 | Provincial five-year mean CFs in China. a–c, Mean CFs in Chinese provinces for 1961–1965 a; baseline, 2011–2015 b; dimmed and the difference between 1961–1965 and 2011–2015 c. Values were computed by feeding synthetic daily radiation profiles based on historical monthly radiation data into the GSEE PV electricity generation model and aggregating CF output to FDM time series before taking spatially weighed averages for each province. Taiwan is not included as data are only collected from mainland China. 0.08 –0.06 –0.04 –0.02 0 CF CF 0.02 0.04 0.06 0.10 0.12 0.14 0.16 0.18 0.20 0.22 0.24 0.26 0.28 0.30 a b c d e f Fig. 3 | Historic CFs and absolute change over the past 50thinspaceyears on the provincial level. a–c, 1961–1965 baseline CFs for HOR a, FIX b and 1AX panels c. d–f, Change in CFs between 1961–1965 baseline and 2011–2015 dimmed conditions for HOR d, FIX e and 1AX panels f. Taiwan is no
点击查看更多>>

京ICP备10028102号-1
电信与信息服务业务许可证：京ICP证120154号

地址：北京市大兴区亦庄经济开发区经海三路
天通泰科技金融谷 C座 16层 邮编：102600