Blue skies over China: The effect of pollution-control on solar power generation and revenues

Air pollution is the single most important environmental health risk, causing about 7 million premature deaths annually worldwide. China is the world’s largest emitter of anthropogenic air pollutants, which causes major negative health consequences. The Chinese government has implemented several policies to reduce air pollution, with success in some but far from all sectors. In addition to the health benefits, reducing air pollution will have side-benefits, such as an increase in the electricity generated by the solar photovoltaic panels via an increase in surface solar irradiance through a reduction of haze and aerosol-impacted clouds. We use the global aerosol-climate model ECHAM6-HAM2 with the bottom-up emissions inventory from the Community Emission Data System and quantify the geographically specific increases in generation and economic revenue to the Chinese solar photovoltaic fleet as a result of reducing or eliminating air pollution from the energy, industrial, transport, and residential and commercial sectors. We find that by 2040, the gains will be substantial: the projected solar photovoltaic fleet would produce between 85–158 TWh/year of additional power in clean compared to polluted air, generating US$6.9–10.1 billion of additional annual revenues in the solar photovoltaic sector alone. Furthermore, we quantify the cost of adopting best-practice emission standards in all sectors and find that the revenue gains from the increased solar photovoltaic generation could offset up to about 13–17% of the costs of strong air pollution control measures designed to reach near-zero emissions in all sectors. Hence, reducing air pollution in China will not only have clear health benefits, but the side-effect of increased solar power generation would also offset a sizeable share of the costs of air pollution control measures.

respectively [2]. When comparing the total generating costs of a near-zero emissions plant to those of a natural gas combined-cycle plant, the generating costs of a natural gas combinedcycle unit in the province of Zhejiang is US 9.36 ¢/kWh, while those for Zhoushan No.4 in the same province is US 3.1 ¢/kWh -about one-third as much [2].
Retrofitting coal-fired power plants with the above-described combination of emission-removal technologies to reach near-zero emissions will increase electricity generation costs by US$ 0.1-0.16 ¢/kWh [2]. Because the thermal electricity generation (~90% coal-fired) in China in 2014 was 4,222 TWh [3], which accounts for the plants that generate electricity for sale to third parties and for the electricity used in plants for their own purposes, the cost of retrofitting the thermal plants is between US$4.2 and 6.7 billion/year.
We assume that the costs of retrofitting the combustion process for heat generation from electric boilers plants and the combustion processes in transformation processes with the same combination of air pollution control systems as is used for electricity generation are proportional to the level of SO2 emissions and thus amount to US$1.5-2.3 billion/year and US$1. 5-2.4 billion/year, respectively.
Hence, the total retrofitting cost in the energy sector, i.e., electricity and heat generation plants and transformation processes, amounts to US$7.2-11.4 billion/year.

Industrial combustion and industrial processes
Industrial boilers provide heat or process steam to meet the needs of the facilities in which they are installed. These facilities can be parts of the iron and steel industry, the chemical and petrochemical industry, the non-ferrous metals and non-metallic minerals industry, etc. While emissions from coal-fired power plants and coal-fired industrial boilers are affected by a number of variables such as coal type and composition and the type of combustion technology, the emission control technologies used to limit emissions from stack gases are essentially the same [4]. Hence, we assume that coal-fired industrial boilers are retrofitted with the same combination of air pollution control systems as is used in coal-fired power plants and that the cost is proportional to the level of SO2 emissions and thus amount to US$12.0-19.1 billion/year.
It is also possible to replace the use of coal in industrial boilers with another, less-polluting fuel, such as natural gas. In many cases, converting coal-fired boilers to gas-fired boilers can be profitable because the changes to the equipment are likely to be less expensive than installing air pollution control equipment; also, the use of natural gas would lead to lower emission characteristics.
China's natural gas production is rising at a fast pace but not fast enough to meet the demand required by the government to clean the country's air. China has the world's largest reserves of shale natural gas, and much of it could be recovered if cost were not a limitation [5]. The boilers and stoves used in the residential and commercial sectors are difficult to retrofit with effective pollution control equipment because of the small scale and age of the units, and it is also difficult to ensure that these units operate correctly. Hence, we assume that the residential and commercial sectors will switch to natural gas, while the industrial boilers will be equipped with the same pollution control equipment as is used in coal-fired power plants.
We discuss the industrial processes in the main text in the section Clean-air policies and their cost.

Road transport and domestic navigation
Gasoline and diesel fuels contain sulfur because it is a natural component of crude oil. The sulfur content in the fuel is the most important parameter affecting the introduction of measures to limit end-of-the-pipe emissions: A fuel with a sulfur content of ~50 ppm (parts per million) allows for the use of diesel particulate filters with an efficiency of ~50% and for the selective catalytic reduction (SCR) of NOx with an efficiency of ~80%. In contrast, a fuel with near-zero sulfur content of ~10 ppm enables huge advances in fuel-efficient vehicle design and advanced control technology because it allows for the use of NOx absorbers with an efficiency of over 90%, which enables engine designs with higher fuel efficiency and particulate filters that achieve an efficiency close to 100%, thereby emitting ~99% less PM2.5 than uncontrolled vehicles [6].
The costs of reducing sulfur content in the fuel depend on the state of existing refineries, current fuel quality, and emissions standards but such costs can be divided into two types: the cost associated with fuel production and the cost associated with vehicle emission control technologies. Estimates of the costs associated with fuel production accounts for upfront refinery investment, such as capital equipment upgrades, and direct operating costs, such as catalysts and chemicals [7]. The costs of upgrading China's refineries to produce near-zero sulfur 10 ppm gasoline and diesel fuels are US 0.7¢ and US 1.7¢ per liter, respectively [8], which is comparable to international experiences and equivalent to 0.6-1.5% of the pump price.
This translates into a total investment requirement of US$4.3 billion/year after accounting for upfront refinery investments, such as capital equipment upgrades, and direct operating costs, such as catalysts and chemicals [7].
Estimates of the cost for the introduction of advanced emission control technologies in vehicles account for the additional costs to manufacturers for equipping these vehicles with advanced emission control technologies to meet international best-practice standards, i.e., the adoption of the China 6 standard in gasoline and diesel vehicles.  China has also about 10,500 coastal vessels, including small passenger ships, fishing boats, etc., and 147,200 river vessels [12]. Most of the fishing boats and small passenger ships already operate with low-sulfur fuel oil due to their limited sizes and hence limited engine capacities.
Overall, low oil prices favor solutions with the lowest capex, i.e., MGO, while high oil prices favor solutions with a higher capex, i.e., scrubbers or liquefied natural gas (LNG). Under stricter international emission standards, the demand for scrubbers may increase, and the costs may go down as production scales. Also, the price of HSFO is expected to fall sharply when the cap set by the IMO comes into force in 2020, while the price of ULSFs is expected to dramatically increase. Thus, the use of scrubbers may be the most cost-effective way for larger ships to comply with the sulfur limit.
Emissions from the transport sector originate not only from road transport and navigation, but also from the combustion processes in rail transport. Emissions from rail transport account for 3% of SO2 emissions from the transport sector, or 1% of total SO2 emissions from all sectors.
Thus, because emissions from rail transport are not significant, we exclude rail transport when estimating the cost of reaching near-zero emissions. Hence, the total cost of reducing emissions in the transport sector due to road transport and domestic navigation is US$15.3 billion/year.

Residential and commercial sector
Households and businesses in China burned 119 million tons of coal in 2014, which accounted for 4% of total national consumption [3]. Both the boilers employed to heat residential and commercial buildings and stoves used for cooking lack effective pollution control equipment and retrofitting them with such equipment is difficult because of the age and small scale of the units, as well as the difficulty of ensuring that these units operate correctly. Also, the cost of retrofitting the units may be too high for poor owners.
To significantly improve air quality at the urban level, the burning of coal and other pollution sources, such as wood, biomass, and waste, can be replaced with natural gas or propane. Also, it is possible to switch to electrical appliances fed with electricity from low-emission sources, though the prior method is the most efficient way because the equipment is already in place.
To reduce air pollution levels, cities and villages in Northern China and, more notoriously, the nation's capital, Beijing, have started to replace coal-fired residential heating and cooking with gas-powered stoves and boilers. Table B shows that to replace coal in residential and commercial uses with natural gas or propane, China would need to procure an additional 93.86 billion m 3 of natural gas, which would represent a 70% increase over the total of 159.32 billion m 3 of natural gas consumed in 2014 [3]. In our calculations, we assume that the switch is made to a single fuel, i.e., natural gas, because in East Asia the costs per unit of energy of natural gas and propane are similar.
We also assume that one energy unit of natural gas can substitute for one energy unit of coal in household and commercial users; which is a conservative assumption because the new natural gas heating devices are likely to be more efficient than old coal-fired boilers.  billion/year. We acknowledge the volatility of the price of natural gas, which can influence the results. The import prices for natural gas via pipeline and LNG during 2015-2016 were rather stable, but in previous years, the prices had been more than 50% higher. We do not estimate the investment cost of converting a coal-fired boiler to a natural gas-fired boiler, because of the lack of available data on the capacity of installed residential and commercial boilers and the cost of replacing those boilers. It is likely that the costs of converting boilers will represent a small share of the total costs and savings from shifting from coal to natural gas or propane. Also, switching from coal to gas involves the construction of natural gas distribution networks, pipelines, and household connection facilities, the prices of which are uncertain. Hence, we acknowledge these uncertainties and exclude these estimates from this analysis.   Figure 4 and Table 6 and [23] Figure 3 and Table 5,

Policies to control and reduce air pollution
China has been implementing policies to control air pollution for more than three decades. The plan introduced, for the first time, politically binding targets for SO2, NOx, and CO2 intensity [31]. The emission standards became stricter and were comparable to those in Europe and the United States. The Emission Standard for Air Pollutants from Thermal Power Plants limited SO2 emission concentrations for new and existing coal-fired plants to 100 mg/m 3 and 200 mg/m 3 , respectively, except in some provinces where the coal used to fuel the plant has a high sulfur content. There, higher emission limits were allowed [31]. For key regions of pollution control, which account for more than 66% of China's GDP, the limit was 50 mg/m 3 .
In 2013, a winter-long episode of severe haze over many provinces and cites in eastern China became worldwide news. The concentration of PM2.5 in Beijing, for instance, was 40 times the limit recommended by the World Health Organization [32].

Counterfactual scenarios
The counterfactual scenarios represent the emissions of the energy sector if no pollution control policies had been introduced since 2006 and the emissions of the industrial sector if the same pollution control policies as used for the energy sector had been applied to industry as well. We estimate emissions in the counterfactual scenarios for SO2, BC, and OC, we do so in two steps.
First, we estimate the emissions factor for the energy sector for the year before the policies to control air pollutants from coal-fired power plants became effective. We calculate the emissions factor as the emission estimates divided by the energy consumption at a regional level for the 22 provinces, 4 municipalities and 5 autonomous regions for 2005; see Equation (A1). We use province-level data on Fuel use for power generation as a driver of emissions from the energy sector [3]. Because the Chinese National Bureau of Statistics does not report energy consumption data for Tibet but it does report thermal power generation, we use the last as a driver for the energy sector. We convert thermal power generation (TWh) into the units given for fuel use (million tonnes of coal equivalent, Mtce), multiplying TWh by the yearly average heat rates for fossil fuel-fired power plants depending on their average fuel consumption [3].
This assumption does not significantly influence the results, because thermal power generation in Tibet is very low.  We calculate the emissions for the Industry counterfactual scenario as described in Equation (A3). We use province-level data on Industrial sector end use [3] as a driver of emissions from industry. The data on Industrial sector end use are given as coal, petroleum, and electricity consumption. We convert the coal and petroleum consumption given in Mt into Mtce, using the same procedure as described for the energy sector.
where Eindustry (I, j) is the emissions (Mt) in province i for year j, E ' industry is the actual emissions (Mt) in province i for year j, Epower (I,j) is from Equation (A2), and E'power (I,j) is the actual emissions (Mt) in province i for year j.    Area III Areas other than areas in Resource Areas I-II, including Tibet autonomous region Note: Regional solar resources as classified in class I, II or III in descending order of solar endowment.   and transport sectors (d). Data and material from [15,[17][18][19]. Data and material from [15,[17][18][19].

Effect of potential clean-air policies for SO2 on surface solar irradiance
The solar radiation benefits of eliminating all current, actual emissions will be greater than if the focus were on eliminating SO2 emissions only: eliminating all emissions in the energy sector increases surface radiation by up to 6 W/m 2 , as compared to the increase of 2.4 W/m 2 seen when eliminating only SO2 (Fig Fa-d).  [15,[17][18][19].
The solar radiation benefits of eliminating all aerosol emissions in the energy sector compared to eliminating in the same sector only SO2 emissions (3.6 W/m 2 ) are greater than for the energy and industrial sectors combined (2.4 W/m 2 ). This non-linear increase in surface solar radiation is due to the relative amount of SO2 emissions per sector compared to total emissions. As seen in Table D, the ratio of SO2 emissions to total emissions in the energy and industrial sectors combined (0.98) is larger than in the energy sector alone (0.82). Thus, for the energy and industrial sectors combined, eliminating only SO2 emissions is closer to eliminating all aerosol emissions than for the energy sector alone. Also, there is no one-to-one correspondence between the ratio of "SO2 emissions to all emissions" and of "SO2 induced irradiance reductions to irradiance reductions from all emissions" because the model computes effects of, e.g., aerosol mixing, aerosol cloud interactions, and different transport and deposition properties.    Revenues discounted to the present using a discount rate of 5% and 8%. Sector-specific annual costs are averages of a low and a high cost scenario, for a break-d own of sub-sector-specific costs and uncertainty ranges see Table I. Data and material from [15,[17][18][19]. and for a feed-in tariff that reduces over time as the national PV system cost reduces following a technological learning rate of 20% starting in 2017, i.e., the year of the last available feed-in tariffs, and for a feed-in tariff without technological learning, i.e., equal to the feed-in tariffs in 2017. Revenues discounted to the present using a discount rate of 5% and 8%. Sector-specific annual costs are averages of a low and a high cost scenario, for a break-down of sub-sector-specific costs and uncertainty ranges see Table I. Data and material from [15,[17][18][19].