2.1 Description of pollution control scenarios
For the residential combustion of straw and firewood alternative scheme in the control scenarios design of this study, the method of building a household biogas digester per household in rural areas of Henan Province was adopted based on the government’s current policy of encouraging rural household biogas digesters plus the clean advantages of biogas energy, so farmers could use biogas from household biogas digesters to replace the traditional straw burning for cooking (straw burned for cooking is converted into biogas, and the household firewood is used in the same way as before). The specific calculation method of biogas production is shown in section 2.3.3 and attached Fig.A.1. It has been calculated that this kind of control scenarios produced a total of about 5 billion cubic meters of biogas in 2016, which is 500 million cubic meters more than that in the end of 2020 in Henan’s 13th Five-Year Plan, so the biogas production under the control scenarios is basically in line with government planning. For the straw open burning alternative scheme, based on the “one village, one factory” straw use model designed by Tsinghua University (Building Energy Efficiency Research Center, Tsinghua University., 2016), each village establishes a factory of biomass briquette fuel, the factory produces biomass briquette fuel for use in local farm households, and traditional heating equipment in farm households is replaced by small biomass heaters. The production of biomass briquette fuel in each region(18 cities in Henan Province)is based on the ratio of 30%, 50% and 70% of the remaining straw volume which is used for open burning in control scenarios to convert to biomass briquette fuel, and the remaining 70%, 50% and 30% of the straw is still burned in the open.
In general, the control scenarios of this study is straws used for cooking in rural areas were uniformly replaced by biogas for farmers to use, and open burning of straw is converted it into biomass briquette fuel for farmers to use on a proportional basis.
In this study, special characters were set to represent different scenarios. The B represents the base scenario; the C1 represents that straws used for cooking in rural areas were uniformly replaced by biogas, and the amount of straw open burning is converted into biomass briquette fuel at a rate of 30%; C2 represents that straws used for cooking in rural areas were uniformly replaced by biogas, and the amount of straw open burning is converted into biomass briquette fuel at a rate of 50%; C3 represents that straws used for cooking in rural areas were uniformly replaced by biogas, and the amount of straw open burning is converted into biomass briquette fuel at a rate of 70%. Seeing the attached Table A.3-A.12 for detailed results of emissions inventory in base scenario and control scenarios.
2.2 Air Quality Model Setup
Numerical modeling of air quality is an important tool for atmospheric environment research. In this study, the impact of biomass burning to air quality and the environmental benefit of alternative comprehensive straw utilization were evaluated numerically by the WRF-CMAQ model system.
The simulation period of this study uses May, June and October in 2016 to represent the crop harvest period in Henan Province. Since the initial conditions in CMAQ are the clean atmospheric components of the model species, the start date of the model run is set seven days prior to the first day of the simulation period in order to reduce the influence of initial conditions on the simulation results.
The CMAQ model simulations were designed by using three layers of grid nesting, namely D01, D02 and D03 and with resolutions of 36 km, 12 km and 4 km respectively. The number of simulated domain grids in D01 is 160 x 130, which covered the area of the most of China, Japan, Korea, Mongolia and parts of Southeast Asia. The number of simulated domain grids in D02 is 140 x 140, which covered the area of the most of Southeast China. The number of simulated domain grids in D03 is 160 x 160, which covered the area of all Henan Province. The WRF simulation domain and CMAQ simulation domain is shown in Fig. 1.
The meteorological conditions of this study were generated by Weather Research Forecast model (WRFv4.1.3). The WRF model system is vertically divided into 43 layers and has different parameterization schemes for various physicochemical processes in the atmosphere. The operational parameterization scheme of the WRF model in this study includes: Thompson graupel scheme for the microphysical process; the YSU scheme for the boundary layer process; the Monin-Obukhov scheme for the near-surface layer; the RRTM scheme for the long-wave radiation; the Goddard shortwave scheme for the short-wave radiation and the Grell-Devenyi ensemble scheme for the cumulus convection. The topographic and surface type data used by the WRF are based on MODIS land use type data, and fnl global re-analysis information from the National Center for Environmental Prediction (NCEP) in the United States was used as meteorological input data for weather field simulations.
The CMAQ model enables the quantitative calculation of complex chemical reaction processes by setting the atmospheric chemical mechanism. Two types of CMAQ models was used in this study, one of which is the traditional CMAQv5.3.1 model and the other is an improved Source-orientated model based on the CMAQv5.3.1 version, both of which will be mentioned below. The SAPRC99 chemical reaction mechanism was used for both types of CMAQ in this study, which including 78 chemical species and 211 reactions.
The Source-orientated air quality model system was used to analyze the impact of open burning of straw plus residential combustion of straw and firewood in Henan Province on the contribution to air pollution during the crop harvest period. Source-orientated air quality model use marker tracking to achieve source apportionment of pollution emissions ( Ying et al., 2004), and pollutant emission source contributions are assessed by calculating the source labeling of emissions rather than by direct changes in pollutant emissions (Ying and Kleeman, 2006). Source-orientated air quality model used reactive tracers to track the physical processes of emission, transport, reaction, removal of target pollutants and their associated primary precursors and intermediates in the atmosphere. Source apportionment of particulate matter is achieved by “tagging” pollutants emitted from specific regions and categories of pollutant emission sources, and pollutants with each source tag were tracked throughout the science procedures of air quality models (Ying and Krishnan, 2010; Zhang et al., 2012).
Particulate tracer component has been used to track the spreading of numerous PM components from each sector and nominated as ATCR plus unique source tag. For instance, ATCR_X1 was used to represent the primary particulate matter tracer generated by open burning of straw in each region of Henan Province ; ATCR_X2 was used to represent the primary particulate matter tracer generated by residential combustion of straw and firewood in each region of Henan Province ; ATCR_X3 was used to represent the primary particulate matter tracer generated by other remaining emission source in each region of Henan Province; ATCR_H represents tracers of primary particulate matter emitted by Henan Province; ATCR_W represents tracers of total amount of primary particulate matter in the atmosphere within Henan Province (ATCR_W includes tracers of primary particulate matter emitted within Henan Province and tracers of primary particulate matter transported from outside Henan Province). ATCR_R represents tracers of primary particulate matter emitted by each region of Henan Province; ATCR_T represents tracers of primary particulate matter emitted by all regions of Henan Province.
The primary particulate of PM2.5 components from open burning of straw plus residential combustion of straw and firewood in Henan Province extracted in this study include aerosols in the Aitken nuclei mold (including organic carbon, non-combustible material, inorganic carbon, chlorine ions and other material components) and aerosols in the accumulation mode (including organic carbon, non-combustible material, inorganic carbon, sodium ions, chlorine ions, metal compounds and other material components). The primary particulate matter is calculated according to the following formula, as shown in (1), (2), (3), where P represents the total amount of primary particulate matter in the atmosphere in May, June and October in each region of Henan Province; P1 represents primary particulate matter emissions from open burning of straw in each region of Henan Province, P2 represents primary particulate matter emissions from residential combustion of straw and firewood in each region of Henan Province, and P3 represents primary particulate matter emissions from the remaining emission source in each region of Henan Province.
P1 = Pⅹ\(\frac{\text{A}\text{T}\text{C}\text{R}\_\text{H}}{\text{A}\text{T}\text{C}\text{R}\_\text{W}}\) ⅹ\(\frac{\text{A}\text{T}\text{C}\text{R}\_\text{R}}{\text{A}\text{T}\text{C}\text{R}\_\text{T}}\)ⅹ\(\frac{\text{A}\text{T}\text{C}\text{R}\_\text{X}1}{\text{A}\text{T}\text{C}\text{R}\_\text{X}1+\text{A}\text{T}\text{C}\text{R}\_\text{X}2+\text{A}\text{T}\text{C}\text{R}\_\text{X}3}\) (1)
P2 = Pⅹ\(\frac{\text{A}\text{T}\text{C}\text{R}\_\text{H}}{\text{A}\text{T}\text{C}\text{R}\_\text{W}}\) ⅹ\(\frac{\text{A}\text{T}\text{C}\text{R}\_\text{R}}{\text{A}\text{T}\text{C}\text{R}\_\text{T}}\)ⅹ\(\frac{\text{A}\text{T}\text{C}\text{R}\_\text{X}2}{\text{A}\text{T}\text{C}\text{R}\_\text{X}1+\text{A}\text{T}\text{C}\text{R}\_\text{X}2+\text{A}\text{T}\text{C}\text{R}\_\text{X}3}\) (2)
P3 = Pⅹ\(\frac{\text{A}\text{T}\text{C}\text{R}\_\text{H}}{\text{A}\text{T}\text{C}\text{R}\_\text{W}}\) ⅹ\(\frac{\text{A}\text{T}\text{C}\text{R}\_\text{R}}{\text{A}\text{T}\text{C}\text{R}\_\text{T}}\)ⅹ\(\frac{\text{A}\text{T}\text{C}\text{R}\_\text{X}3}{\text{A}\text{T}\text{C}\text{R}\_\text{X}1+\text{A}\text{T}\text{C}\text{R}\_\text{X}2+\text{A}\text{T}\text{C}\text{R}\_\text{X}3}\) (3)
Traditional air quality models can reproduce the effects of different control measures on the atmospheric environment and provide a scientific basis for the development of relevant pollution control policies. In this study, different scenarios are simulated separately through using CMAQ models. The pollutant reduction in the control scenarios is analyzed by comparing the difference of the simulated values of atmospheric pollutants between the base scenario and control scenarios.
In this study, the simulation results were compared to observation data, and the NMB(Normalized Mean Bias), NME (Normalized Mean Error) and R (Correlation coefficient) were calculated as shown in equations (4), (5), and (6), respectively. Where, Cm is the daily average concentration value of pollutants simulated on the i-th day; C0 is the daily average concentration of pollutants observed on the i-th day;‾Cm is the average value of the simulated concentration values for each day of the evaluation period;‾C0 is the average value of the observed concentration values for each day of the evaluation period; N is the total number of days in the evaluation period. The performance of the CMAQ model will be discussed below.
2.3 Emissions
This study, two sets of gridded emission inventories were established to drive the air quality model: (1) base scenario emission inventories and (2) alternative control scenarios emission inventories with different proportion of replacing. According to the described in section 2.1, the base emission scenario emission inventories includes open burning of straw and residential combustion of agroforestry biomass (The residential combustion of agroforestry biomass includes residential burning of straw and residential combustion of fuelwood) in 18 cities in Henan Province. The inventories of control scenarios include biogas combustion, biomass briquette fuel combustion and residual straw open burning (residual straw is the amount of straw in the field that has not been converted to biomass briquette fuel in control scenarios). The calculation method of each emission inventory is described as below. Inventory of anthropogenic emissions source other than open burning of straw and residential combustion of agroforestry biomass in Henan Province were used from Bai et al (Bai et al., 2020). The REAS2 emission inventory (Kurokawa et al., 2013) and MEIC emission inventory (Li et al., 2017) were used for anthropogenic source emission in simulated areas outside Henan Province. Natural emission sources in simulated areas was estimated by using the MEGAN model (Liang et al., 2020). The calculation method of pollutant emission inventories in this study use a bottom-up emission factor approach, as shown in the formula (7):
Ei =∑(AjⅹEFi,j) / 1000 (7)
where subscripts i and j represent the type of pollutant and the type of burning source, respectively; Ei the annual emissions of a certain pollutant, t; Aj is the annual activity level of a certain burned source, t; EFi,j is the emission factor of a certain pollutant in a certain burned source (g /kg).
The reasonableness of emission factors has a significant impact on the accuracy of
emission inventory calculations. The emission factors of this paper mainly refers to the inventory compilation guidelines issued by the Ministry of Environmental Protection and the statistical results of Zhou et al (Zhou et al., 2017). The emission factors involved in this article are listed in Table A.2.
2.3.1 Open and residential burning of straw
The amount of straw burning in open air and the amount of straw used to residential combustion is calculated by formula (8)
Ai,k = Pi,k × Nk × Ri,k ×Dk × CEk (8)
where subscripts i and k represent the region (The region is 18 cities in Henan Province) and the main crop type (rice, wheat, maize, beans, peanuts, oilseed rape and cotton), respectively; Ai,k is the amount of each crop straw burning in open air or the amount of residential combustion of each crop straw in each region, t; Pi;k is the amount of crop-specific yield in each region ,t; Nk is the straw-to-product ratio of each straw type; Ri;k is the percentage of open burning of straw or the percentage of residential combustion of straw in each region; Dk is the dry matter fraction of each straw type; CEk is the combustion efficiency of each type of straw burning in open air or residential combustion of straw.
The annual crop yield Pi;k comes from the Henan Statistical Yearbook 2017 (HNBS, 2017) .The value of Ri;k for the proportion of open burning of straw and residential straw combustion in each region is important for the calculation of the emissions inventory. Ma et al. (Ma et al,2018) used a large number of questionnaires to calculate the proportion of straw open burning and straw residential combustion in 18 municipalities in Henan Province during the 2014 period, which can reflect the actual situation in Henan relatively very well. Otherwise, there are few research statistics on the proportion of straw open burning and straw residential combustion at municipal level in Henan Province. Therefore, the proportion of straw open burning and straw residential combustion in Henan Province in 2016 was studied through using Ma’s research results. The crop straw-to-product ratio Nk comes from the technical guide for the preparation of air pollutant emission inventory of biomass combustion. The combustion efficiency of straw open burning plus straw residential combustion and straw dry matter fraction use the statistical results of He et al. (He et al., 2015). Table 1 shows comprehensive value of each crop type.
Table 1
Comprehensive value of each crop type
Crop
|
Rice
|
Wheat
|
Corn
|
Beans
|
Peanut
|
Rapeseed
|
Cotton
|
Ri;k (open burning,%)
|
13.8
|
8.1
|
1.1
|
1.3
|
0.7
|
48.3
|
0.1
|
Ri;k (residential combustion,%)
|
5.9
|
4
|
9.2
|
50.4
|
8.5
|
26.3
|
86.2
|
CEk ( open burning,%)
|
0.93
|
0.92
|
0.92
|
0.68
|
0.82
|
0.9
|
0.9
|
CEk (residential combustion,%)
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Dk (%)
|
0.89
|
0.89
|
0.87
|
0.91
|
0.94
|
0.83
|
0.83
|
Nk (%)
|
1.323
|
1.718
|
1.269
|
1.5
|
1.5
|
1.5
|
1.5
|
Hk(%)
|
0.4
|
0.45
|
0.5
|
0.4
|
0.4
|
0.4
|
0.6
|
2.3.2 Fuelwood residential combustion
Due to the difficulties in obtaining the consumption data of residential fuelwood directly, only the total amount of consumption data for fuelwood of Henan Province in 2011 was counted in China Energy Statistical Yearbook 2012 (CESY, 2012), therefore, this part of the calculation of residential firewood combustion is based on the total amount of firewood in 2011 as the original data, and the total amount of firewood combustion in Henan Province in 2016 was estimated by using sown area data in 2011 and 2016 which from the “Henan Provincial Statistical Yearbook 2017” as reference factors. Finally, the amount of fuelwood combustion in each region was obtained by using the rural population of each region in Henan Statistical Yearbook 2017 as the distribution factor.
2.3.3 Biogas production in the control scenarios
For the straw residential combustion alternative scheme in the control scenarios, a residential biogas digester is constructed for each household, that is, farmers use the biogas produced by the household biogas digester to replace the traditional straw combustion for cooking. Biogas production in each region is calculated according to the following formula.
Ai = Zi,k×Dkⅹηk (9)
where subscripts i and k represent the region and the main crop type; Ai is the source activity level (t), that is, the amount of biogas production in each region; Zi,k is the amount of straw used to convert biogas in each region; Dk is the dry matter rate of different straw type, and is shown in Table 1 above. ηk is the straw to biogas conversion rate of different straw type. The straw to biogas conversion rate for each straw type (except cotton) were derived from Zhang et al (Zhang et al., 2014), and the straw to biogas conversion rate for cotton were comprehensive calculation from Zhang et al (Zhang et al., 2014). and Cui et al. (Cui et al., 2013). The straw to biogas conversion rate for each straw type is show in Table 1 for details.
2.3.4 Production of biomass briquette fuel and the amount of open burning of residual straw
According to the calculation, after the residential straw consumption in each region is converted into biogas according to formula (6), the calorific value of biogas after conversion in each region cannot reach the calorific value of residential straw consumption before conversion. Therefore, a portion of the open burning of straw is used to convert biogas to ensure that the calorific value of the biogas after conversion can reach the calorific value of the residential combustion of straw before conversion (Therefore, the relevant calculation of biogas production in control scenarios is performed firstly to ensure the rationality of the control scenarios, and then the calculation of biomass briquette fuel is performed). The detailed material accounting process for control scenarios is shown in Fig.A.1, and the corresponding values for the material accounting process are shown in Table A.1.
As for the scheme of biomass briquette fuel instead of the open burning of straw, based on the “one village, one factory” model of straw utilization mentioned in section 2.1, the biomass briquette fuel processing factory will be established in each village, the produced biomass briquette fuel is used for heating in winter for local farm households, and traditional heating equipment in farm households is replaced by small biomass heaters. The amount of straw used for open burning in control scenarios in each region will be converted into biomass briquette fuel at the ratio of 30%, 50% and 70%, the remaining corresponding 70%, 50% and 30% of the straw remains open burning. The calculation method of the activity level of biomass briquette fuel uses formula (10).
Ai = Gi × η (10)
where the subscript i represent the region; Ai is the activity level ,that is the production of biomass briquette fuel in each region; Gj is the amount of straw used for open burning in control scenarios in each region; and η is the proportion of straw used to cover biomass briquette fuel, that is, 30%, 50% and 70%. The calculation method of remaining 70%, 50% and 30% of straw open burning emission inventory are still according to the method described in Section 2.3.1.
2.4 Spatial distribution, Temporal distribution and Speciation of NMVOCs and PM2.5
The spatial allocation of the emission sources for each scenario in this study was completed by ARCGIS. The method of spatial distribution of pollutants was mentioned in the supplementary material. Figure 2 shows the spatial distribution of PM2.5 emitted from different emission sources. The spatial distribution of NOX and VOCs from different emission sources is shown in the Fig.A.2 and Fig.A.3, respectively.
The types of pollutant emission sources in the anthropogenic source emission inventory of Henan Province were divided into 15 major categories, namely power combustion, industrial combustion, residential combustion, catering, on-road mobile, non-road mobile, dust source, industrial process, industrial solvents, residential solvents, fuel evaporation from service stations, fuel evaporation from refineries, agriculture, residential combustion of straw and firewood (RB), open burning of straw (OB). The temporal allocation factors for RB and OB was mentioned in the supplementary material.
In addition, speciation of NMVOCs and PM2.5 was described in the supplementary material.