A new snow cover mapping algorithm for Chinese geostationary meteorological satellite FY-4A AGRI data

Figures & data

Figure 1. Study area and study data in this study.

Figure 2. Flowchart of the snow cover mapping algorithm for FY-4A AGRI data.

Table 1. Number of training and validation samples for each spatial and temporal division.

Region	Time period 1 (10:45, 11:00, 11:15)	Time period 2 (12:00, 13:00)	Time period 3 (13:45, 14:00, 14:15)
Area 1	10631/4557	30290/12982	12105/5189
Area 2	5604/2403	10029/4299	6856/2939
Area 3	8593/3683	18476/7876	9601/4115
Area 4	11361/4870	23503/10073	10227/4383
Area 5	3255/1395	5009/2148	3027/1298
Area 6	5078/2177	8332/3571	5319/2288

Note: the number before slash is for training sample and the number after slash is for validation sample.

Table 2. Criteria and thresholds for detection of cloud with high confidence.

Criteria	Region	P1	P2	P3	P4	P5	P6	P7
R1.61 > P1 * NDFSI + P2	Area 1	−0.2	0.55	235	240	−2.5	−30	5
BT7.1 < P3	Area 2	−0.2	0.55	235	243	−2.0	−25	0
BT10.7 < P4	Area 3	−0.2	0.55	233	240	−2.0	−20	0
BT12.0 – BT10.7 < P5	Area 4	−0.2	0.45	237	240	−2.5	−30	5
BT12.0 – BT3.8 < P6	Area 5	−0.2	0.40	243	255	−2.0	−20	0
BT8.5 – BT10.7 > P7	Area 6	−0.2	0.35	245	265	−2.0	−15	0

Table 3. Confusion matrix for FY-4A snow cover data versus reference data.

	FY-4A data
Reference data	Snow	Snow-free
Snow	True Positive (TP)	False Negative (FN)
Snow-free	False Positive (FP)	True Negative (TN)

Figure 3. Snow mapping results from FY-4A AGRI on December 5, 2019 at different times (a)-(h) and (i) the fusion result. Three MODIS snow cover products, including MOD10A1 (NDSI > 0.1), NIEER MODIS-only SC and NIEER MODIS-CGF SC, are also displayed.

Figure 4. Average daily cloud coverage with one standard deviation for different snow cover products during the period of 2018–2023.

Figure 5. Evaluation metrics for different snow cover products validated using in-situ snow depth data.

Figure 6. The accuracy of the FY-4A fusion snow cover product varies at each ground snow-depth measurement station for (a) overall accuracy (OA), (b) F-score (FS), (c) underestimation error (UE) and (d) overestimation error (OE)

Table 4. Evaluation metrics of FY-4A fusion snow cover product versus Landsat-8 OLI data.

	Number of OLI images	OA (%)	FS (%)	UE (%)	OE (%)
Area 1	15	90.70	93.49	1.44	7.85
Area 2	10	88.63	91.87	0.97	10.39
Area 3	10	89.13	94.25	0.03	10.86
Area 4	20	88.35	87.48	2.15	9.50
Area 5	6	88.59	90.78	4.94	6.47
Area 6	2	86.42	67.14	6.24	7.34
All	63	89.05	91.65	1.73	9.22

Figure 7. Comparison of different snow cover products for three cases: (a) eastern Inner Mongolia, January 5 2019; (b) Central China, February16, 2020; (c) Northern Xinjiang, March 22, 2019.

Figure 8. Daily snow cover extent over China extracted from different snow cover products during the period of 2019–2023.

Figure 9. Scatterplots of snow cover extent between the FY-4A fusion SC product and the NIEER MODIS SC products for (a) the NIEER MODIS-only SC product and (b) the NIEER MODIS-CGF SC product over the entire year and the snow season.

Figure 10. Time series of evaluation metrics of the FY-4A fusion snow cover product compared to other MODIS snow cover products on a weekly basis for (a) overall accuracy; (b) F-score; (c) overestimation error; and (d) underestimated error.

Table 5. Comparison of evaluation metrics for FY-4A fusion snow cover product with/without small divisions versus the NIEER MODIS-only SC product.

Region	FY-4A fusion SC with small divisions				FY-4A fusion SC without small divisions
	OA (%)	FS (%)	UE (%)	OE (%)	OA (%)	FS (%)	UE (%)	OE (%)
Area 1	95.47	86.78	1.80	2.74	95.07	86.62	1.31	3.62
Area 2	96.62	91.45	0.30	3.08	97.18	92.91	0.38	2.44
Area 3	92.84	91.31	3.94	3.23	92.74	91.01	4.76	2.50
Area 4	92.32	80.93	2.64	5.04	91.54	79.25	2.52	5.94
Area 5	98.77	75.10	0.60	0.63	98.48	73.31	0.65	0.87
Area 6	99.71	45.16	0.17	0.13	99.64	36.56	0.20	0.15
All	95.57	87.06	1.71	2.72	95.27	86.50	1.72	3.01

Data availability statement

The data that support the findings of this study are available from the corresponding author upon request.