Lei, SH; Xu, J; Li, YM; Du, CG; Liu, G; Zheng, ZB; Xu, YF; Lyu, H; Mu, M; Miao, S; Zeng, S; Xu, JF; Li, LL (2020). An approach for retrieval of horizontal and vertical distribution of total suspended matter concentration from GOCI data over Lake Hongze. SCIENCE OF THE TOTAL ENVIRONMENT, 700, 134524.

There are a few studies working on the vertical distribution of TSM, however, understanding the underwater profile of TSM is of great benefit to the study of biogeochemical processes in the water column that still require further research. In this study, three data-gathering expeditions were conducted in Lake Hongze (HZL), China, between 2016 and 2018. Based on the in situ optical and biological data, a multivariate linear stepwise regression method was applied for retrieval of the surface horizontal distribution of TSM (TSM0.2) using GOCI (Geostationary Ocean Color Imager) data. Then, the estimation model of vertical structure of underwater TSM was constructed using layer-by-layer recursion. This study drew several crucial findings: (1) the approach proposed in this paper generated very high goodness of fit results, with determination coefficients (R-2) of 0.83 (p < 0.001, N= 54), and with smaller prediction errors (the mean absolute percentage error is determined to be 16.34%, the root mean square error is 9.01 mg(-1), and the mean ratio is 1.00, N = 26). (2) The monthly surface TSM and the column mass of suspended matter (CMSM) are affected by both wind speed and precipitation in HZL. In addition, the hourly variation of surface TSM and CMSM are driven by local wind, most especially in spring and winter. (3) Compared with the non-uniform hypothesis, the CMSM derived by conventional vertical uniformity hypothesis was underestimated by almost 10% in HZL during 2016. This should warrant the attention of lake managers and lake environmental evolution researchers. (C) 2019 Elsevier B.V. All rights reserved.