Comparison of variability and change rate in tropospheric NO₂ column obtained from satellite products across China during 1997–2015

Figures & data

Figure 1. Temporal coverage of five satellite datasets (GOME/ERS-2 (H), SCIAMACHY (S), OMI (O), GOME-2/METOP_A (A) and GOME-2/METOP_B (B)) (left) and the spatial distribution of six regions (right) over China (the insert caption is Nanhai, China).

Table 1. Numbers and percentages of grids at five levels for each dataset within its temporal coverage.

NMV
D (L)	Number of grids					Percentage
	I (0)	II (≤12)	III (≤24)	IV (≤36)	V (≤48)	I (0) [%]	II (≤12) [%]	III (≤24) [%]	IV (≤36)	V (≤48)
A (84)	666	12,480	14,987	15,348^a	15,359	4.34	81.25	97.57	99.92%^a	99.99%
B (36)	743	14,646^a	15,359	−	−	4.84	95.35^a	99.99	−	−
H (72)	933	8101	12,785	14,524^a	15,013	6.08	52.81	83.35	94.69%^a	97.87%
O (132)	10098	34,738	44,354	53,039	57,692^a	16.45	56.58	72.24	86.38%	93.96%^a
S (108)	253	5553	9499	12,340	14,377^a	1.65	36.15	61.84	80.34%	93.60%^a

^aDenotes the selected threshold for each dataset, NMV represents the number of missing value at five levels, D (L) represents dataset and the temporal coverage of the dataset (unit: month), the percentage refers to the ratio of the number of grids at each level divided by the total number of grids. GOME/ERS-2 (H), SCIAMACHY (S), OMI (O), GOME-2/METOP_A (A) and GOME-2/METOP_B (B).

Figure 2. Temporal series of annual and seasonal average TNCs from five datasets over China during 1997–2015, the bar charts represent the multiple-year average TNC, and the curves represent the temporal series of average annual TNC. The black (grey) solid mark represents the highest level (abrupt reduction) of annual TNC. Unit: 10¹⁵ molecules cm⁻². GOME/ERS-2 (H), SCIAMACHY (S), OMI (O), GOME-2/METOP_A (A) and GOME-2/METOP_B (B).

Figure 3. Temporal series of annual TNCs from five datasets in six regions (NC, North China; SE, Southeast; NE, Northeast; SW, Southwest; NW, Northwest; TP, Tibetan Plateau) during 1997–2015. The curves represent the temporal series of average annual TNCs, and the bar charts represent the multiple-year average annual TNC from each dataset in six regions. The solid mark represents the highest level (abrupt reduction) of the annual TNC (Unit: 10¹⁵ molecules cm⁻²). GOME/ERS-2 (H), SCIAMACHY (S), OMI (O), GOME-2/METOP_A (A) and GOME-2/METOP_B (B).

Figure 4. Temporal series of seasonal TNCs from five datasets in six regions from 1997 to 2015, it represents temporal series of seasonal TNCs in NC, SE, NE, SW, NW and TP from left to right, and regional TNCs in spring, summer, autumn and winter from the upper to the bottom panel. The black (grey) solid circle represents the highest level (abrupt reduction) of the seasonal TNCs. Unit: 10¹⁵ molecules cm⁻². GOME/ERS-2 (H), SCIAMACHY (S), OMI (O), GOME-2/METOP_A (A) and GOME-2/METOP_B (B).

Figure 5. Spatial maps of changing rate and corresponding significance levels of TNC over China, unit of the trend: 10¹⁵ molecules cm⁻² yr⁻¹, Sig represents the significance level. GOME/ERS-2 (H), SCIAMACHY (S), OMI (O), GOME-2/METOP_A (A) and GOME-2/METOP_B (B).

Figure 6. Scatter plots and regression results between four pairs of datasets based on the multiple-year average annual TNC at the country level, the vertical axis (y) is for TNC from O, O, S, S and the horizontal axis (x) is for TNC from A, B, A, O, respectively. Unit: 10¹⁵ molecules cm⁻². GOME/ERS-2 (H), SCIAMACHY (S), OMI (O), GOME-2/METOP_A (A) and GOME-2/METOP_B (B).

Table 2. Parameters of regression between different datasets in six regions.

Region	O&A (2007–2013)				O&B (2013–2015)				S&A (2007–2011)				S&O (2005–2011)
	a	r^2	|D|	N	a	r^2	|D|	N	a	r^2	|D|	N	a	r^2	|D|	N
NC	0.56	0.89	617	693	0.56	0.85	658	677	1.20	0.96	214	624	1.85	0.93	702	662
SE	0.65	0.93	196	691	0.77	0.89	160	689	1.22	0.97	26	686	1.59	0.97	168	684
NE	0.70	0.95	93	653	0.76	0.83	92	513	1.23	0.95	10	513	1.63	0.96	114	490
NW	0.54	0.90	24	598	0.57	0.88	36	519	1.04	0.93	4	517	1.41	0.88	6	508
SW	0.77	0.87	38	688	0.80	0.84	59	670	0.82	0.93	54	569	0.95	0.90	15	564
TP	1.00	0.82	14	675	0.77	0.7	9	641	1.02	0.87	6	378	0.95	0.80	4	667

Note: a, r², |D| and N represent coefficients, r-square of linear regression, the absolute gap (unit:10¹⁵ molecules cm⁻²) and the number of sampling points. GOME/ERS-2 (H), SCIAMACHY (S), OMI (O), GOME-2/METOP_A (A) and GOME-2/METOP_B (B).

Figure 7. Spatial distribution of numbers of missing values on grid scale within each dataset under the predefined rule. GOME/ERS-2 (H), SCIAMACHY (S), OMI (O), GOME-2/METOP_A (A) and GOME-2/METOP_B (B).

Figure 8. Comparison of temporal series of the annual TNC calculated under different grid-selection rule: (a), grids with less than 12-month missing values are selected, (b), grids are selected by the rule described in Section 2.2 and (c) all grids are included. The solid mark represents the highest level of annual TNC. Unit: 10¹⁵ molecules cm⁻². GOME/ERS-2 (H), SCIAMACHY (S), OMI (O), GOME-2/METOP_A (A) and GOME-2/METOP_B (B).