Satellite estimation of dissolved organic carbon in eutrophic Lake Taihu, China

Dissolved organic carbon (DOC) in lakes serves as a substrate for heterotrophic bacterial growth, regulator of the global cycle, and light absorption agent. DOC eutrophic is greatly influenced by phytoplankton phenology terrigenous input rivers. Therefore, it necessary significant to dynamically monitor concentration, storage, riverine exchange flux DOC. By using in-situ measurements from 2004 until 2018 ( N = 2019), machine learning algorithm, namely, multilayer back-propagation neural network (MBPNN) model, was developed this work improve remote sensing estimation concentrations Lake Taihu. The model yielded mean error 15.14% testing dataset. monthly concentration significantly increased 2003 192, p < 0.01). High DOCs were observed lake bays with high chlorophyll (Chl-a) levels, growth explained more than 50% variations. Then, given evenly mixed water column on annual scales, we further estimated storage fluxes (input output) due rivers during 2008–2018. Although net (50.56 ± 32.22 × 10 3 t C) approximately 5.2 times average (9.73 1.23 C), controlled variations, which indicated that much transformed into other forms after entering This study verified feasibility (surface flux) lakes. • Machine improved variations Taihu growth. Net ~5.2