1.Ho et al. (20 19) used a single near-infrared band to analyze lake blooms, ignoring the influence of suspended sediment changes in shallow lakes on near-infrared signals. Because lakes (especially shallow lakes) are affected by river input and bottom re-suspension, the concentration of suspended sediment in water bodies presents significant temporal and spatial dynamic differences. However, the strong backscattering signal of sediments will lead to the increase of reflectivity of high turbidity water in near infrared band. Therefore, the method adopted by Ho et al. (20 19) may overestimate the lake water bloom process. Based on the data analysis of 7 1 lake studied by Ho et al. (20 19), it is found that ignoring the influence of sediment signals will lead to misjudgment of the water bloom extraction results of at least 58 (82%) lakes (Figure 1).

Figure 1. Schematic diagram of the influence of (20 19) suspended sediment on the extraction results of water blooms in some lakes. The analysis shows that among the 7 1 lakes studied by Ho et al. (20 19), at least 58 (82%) are affected by this factor.

2.Ho et al. (20 19) used Fmask algorithm to distinguish lake boundaries on Landsat remote sensing data. However, Ho et al. (20 19) didn't notice that only using the simple vegetation index threshold to distinguish water bodies in the Fmask algorithm would identify the areas with strong water blooms as land and exclude them from the subsequent analysis, which may lead to the omission of judging the areas with strong water blooms (Figure 2). Data analysis shows that among the 765,438+0 lakes of Ho et al. (2065,438+09), at least 465,438+0 (58%) have missed judgments.

Figure 2. The extraction results of Songkhla Lake (Thailand) and Hongze Lake (China) show that there are serious errors in the information of lake water bloom intensity based on Landsat by Ho et al. (20 19). (a-d) True color composite image and water bloom intensity index of Landsat 5TM data (BNIR, calculated by the method of Ho et al. (20 19)). Obviously, whether it is highly turbid water or aquatic vegetation, the intensity index of water bloom will be large, which will lead to misjudgment of water bloom. (e) The mask of Hongze Lake water body obtained by Fmask will eliminate the strong water bloom area and lead to misjudgment.

3. The signal obtained by satellite includes two parts: surface reflection and atmospheric radiation. The lake water signal accounts for less than 50% of the total satellite signal in some wavelength ranges, and the proportion changes with the turbidity of water. Ho et al. (20 19) did not consider the influence of atmospheric signals, which may lead to errors in the results.

Figure 3. Spectral characteristics of different ground objects on Landsat 5 TM remote sensing images. Among them, the black curve is the airborne reflectivity, the blue curve is the reflectivity after Rayleigh scattering correction, and the gray curve is the atmospheric Rayleigh scattering reflectivity. In the blue band (1 band), the Rayleigh scattering reflection signal of the atmosphere may account for more than 50% of the whole satellite signal.

4. The revisiting period of Landsat data used by 4.Ho et al. (20 19) is long (16 days), and it is difficult to accurately capture the relevant characteristic information of annual changes of lake algal blooms. Therefore, there is great uncertainty in the trend of lake eutrophication in this paper.

5.Ho et al. (20 19) The comprehensive impact analysis of global change on lake eutrophication based on erroneous remote sensing data is not credible, and the research conclusion may mislead readers.

The original author replied to the different opinions of Feng Lian's team. For details, please refer to the paper published in the same issue of Nature (Ho et al. 20 1).

What happened was a special column set up naturally on 20 18. The review articles in the column mainly put forward different views on the papers published in Nature, which aroused readers' extensive concern.

South University of Science and Technology is the only communication unit of the paper, and Feng Lian is the first author and correspondent of the paper. Co-authors include Liu and Zheng Chunmiao, Dean of south university of science and technology of china Institute of Environmental Science and Engineering. This work is supported by the National Natural Science Foundation of China, the Science and Technology Pilot Program of Chinese Academy of Sciences, and the State Key Laboratory of Comprehensive Prevention and Control of Surface Water and Groundwater Pollution in Environmental Protection Watershed.

Links to Feng Lian's papers with different viewpoints.

/articles/s 4 1586-02 1-03254-3

Link to the original author's reply

/articles/s 4 1586-02 1-03255-2