Synthesis of the ICESat/ICESat-2 and CryoSat-2 observations to reconstruct time series of lake level

Figures & data

Figure 1. Geographical distribution of the sample lakes in this study.

Table 1. Satellite basic parameters.

Mission	Sensor (type)	Data record	Repeat cycle(day)	Along-track space (m)	Footprint Diameter (m)
CryoSat-2	SIRAL (radar)	2010-present	369	∼300	∼1650 × 300 (SARIn) ∼1650 (LRM)
ICESat	GLAS (laser)	2003–2009	91	∼170	∼70
ICESat-2	ATLAS (laser)	2018-present	91	∼0.7	∼17

Figure 2. Flowchart of deriving the time series of lake water level by synthesizing multiple satellite altimetry data.

Figure 3. Data difference between ICESat and ICESat-2 in relation to location latitude (a) and lake area (b).

Figure 4. Pairing results for CryoSat-2 and ICESat-2 in Lake Ontario.

Figure 5. Relationship between CryoSat-2 systematic deviation and (a) location latitude; (b) lake area; and (c) the number of pairs.

Figure 6. Comparison of lake water levels derived from satellite altimetry data and in situ measurements for Ontario Lake during 2003–2020. Water level time series (a); ICESat (b), ICESat-2 (c), CryoSat-2 (d), ICESat & ICESat-2 & CryoSat-2 (e).

Figure 7. Box plot representing R² and RMSE of satellite and validation data (gauge station and Hydroweb data).

Figure 8. Box plot representing R² and RMSE of synthesized data (using CryoSat-2 SARIn and LRM modes, respectively) and validation data.

Figure 9. R² and RMSE of synthesized and validation data versus the number of CryoSat-2 and ICESat-2 pairs.

Figure 10. Scatter plot of lake level change rates between the validation and synthesized data.

Figure 11. R² (a) and RMSE (b) of the joint satellite and validation data are obtained by three different correction methods: Method-1 (Month paired), Method-2 (Fitting), and Method-3 (No pairing).

Figure 12. Systematic deviations of CryoSat-2 relative to ICESat-2 calculated for different matching times.

Figure 13. Systematic bias of CryoSat-2 data for different matching time windows.

Figure 14. Coverage of global lakes by different satellites.

Figure 15. Relationship between the number of pairs of CryoSat-2 and ICESat-2 and the lake area.

Table A1. Information for study lakes.

ID	Lake_name	Lat (°)	Lon (°)	Area (km^2)	Validation type	N	Bias (cm)	R^2	RMSE (cm)
1	Caspian Sea	41.71	50.74	377001.91	1	19	−12.13	0.98	6.31
2	Great Slave	61.78	−113.74	26734.29	2	13	−5.44	0.82	12.12
3	Winnipeg	52.53	−97.81	23923.04	2	14	−12.58	0.85	12.75
4	Superior	47.53	−87.76	81843.92	2	15	−14.47	0.85	11.36
5	Michigan	44.01	−86.76	57726.84	2	15	−10.00	0.98	17.88
6	Ontario	43.66	−77.76	19347.36	2	16	−11.64	0.86	11.05
7	Huron	44.95	−82.26	59399.30	2	16	−11.48	0.97	14.83
8	Erie	42.15	−81.25	25767.79	2	15	−8.13	0.94	7.91
9	Ladoga	60.83	31.47	17444.01	1	20	−6.16	0.92	15.61
10	Balkhash	46.29	75.63	16717.89	1	20	−15.17	0.79	8.18
11	Kara-Bogaz-Gol	41.33	53.64	18666.80	1	20	−10.41	0.98	9.33
12	Victoria	−1.10	32.91	67166.22	1	21	3.72	0.99	8.49
13	Amadjuak	64.91	−71.21	2994.94	1	10	−9.72	0.67	101.63
14	Kasba	60.31	−102.19	1330.11	1	7	−8.73	0.72	5.56
15	Athabasca	59.19	−109.32	7528.73	2	13	−10.66	0.95	9.30
16	Reindeer	57.31	−102.36	5435.48	2	15	−51.77	0.59	15.91
17	La Ronge	55.13	−104.93	1296.47	2	6	−73.08	0.38	11.58
18	Lesser Slave	55.44	−115.39	1186.51	2	11	−0.67	0.95	11.52
19	Moose	53.94	−100.11	1311.82	2	5	−23.46	0.46	59.98
20	Manitoba	50.90	−98.65	4751.05	2	13	−15.25	0.72	31.69
21	Nipigon	49.82	−88.52	4505.95	1	9	−17.07	0.55	29.52
22	Saint-Jean	48.59	−72.05	1112.79	2	5	−4.90	0.97	36.87
23	Saint Clair	42.45	−82.67	1161.17	2	8	−8.32	0.93	38.88
24	Nicaragua	11.54	−85.40	7833.32	1	18	−5.04	0.98	21.09
25	Chiquita	−30.54	−62.62	6132.89	1	20	−13.20	0.95	21.17
26	Argentino	−50.22	−72.63	1319.94	1	21	12.83	0.96	8.46
27	Peipsi	58.55	27.55	3489.00	1	11	−15.63	0.89	21.31
28	Small Aral Sea	46.44	60.65	2964.43	1	20	2.77	0.97	18.62
29	Issyk Kul	42.44	77.27	6195.93	1	20	−2.12	0.79	11.69
30	Sarygamysh	41.94	57.42	3772.03	1	19	−11.07	0.99	12.03
31	Aydar	40.86	66.92	2742.41	1	17	−10.10	0.89	14.34
32	Hungtze	33.32	118.60	1374.36	1	5	−61.20	0.43	29.35
33	Taihu	31.20	120.19	2329.14	2	15	−13.14	0.83	13.53
34	Turkana	3.55	36.18	7473.43	1	20	−4.16	0.99	28.69
35	Kyoga	1.54	32.90	2788.19	1	20	10.70	0.99	9.90
36	Rukwa	−7.97	32.19	5894.78	1	21	4.29	0.97	26.53
37	Mweru	−9.00	28.73	5042.56	1	18	4.19	0.98	11.77
38	Bangweulu	−11.18	29.81	1763.53	1	9	−11.91	0.95	19.90
39	Aylmer	64.13	−108.52	800.94	1	7	−29.69	0.47	16.45
40	Biche	54.87	−112.08	238.98	2	6	−2.24	0.83	34.64
41	Cormorant	54.22	−100.80	338.73	2	6	−78.35	0.57	20.03
42	St. Martin	51.66	−98.44	331.34	2	3	−11.08	0.65	8.22
43	Dauphin	51.26	−99.77	515.73	2	5	−2.99	0.87	33.05
44	Nipissing	46.26	−79.78	878.30	2	5	5.35	0.74	14.48
45	Yellowstone Lake	44.44	−110.37	340.47	1	3	1.87	0.81	10.71
46	Har	48.10	93.21	574.52	1	7	−11.35	0.85	11.05
47	Ulungur	47.26	87.29	854.89	1	9	−9.09	0.96	24.47
48	Bosten	41.97	87.04	961.84	1	16	0.68	0.98	9.52

Validation type 1 means that the validation data is Hydroweb and 2 means that the validation data is in situ data.

N represents the number of months in which CryoSat-2 and ICESat-2 both have data.

Bias represents the systematic deviation of CryoSat-2 relative to ICESat-2.

R² indicates the correlation between the combined satellite data and the validation data.

RMSE indicates the Root Mean Square Error between the synthesized satellite data and the validation data.

Figure A1. Water level time series obtained from synthesized data and validation data for lakes 1 to 48.

Figure A1. Continued

Figure A2. CryoSat-2 and ICESat-2 pairs data for lakes 1 to 48.

Figure A2. Continued