Adsorption and desorption of heavy metals at water sediment interface based on bayesian model

Estuarine areas are not only the main gathering point for human sewage but also the place where one-way and two-way fluids interact, thus forming a complex and changeable geochemical physical field. Here, heavy metals (HMs) are adsorbed and desorbed due to physical, chemical, and biochemical processes. However, the adsorption and desorption behavior of HMs in the aquatic environment is complex, and physicochemical processes occurring in the estuarine sediment-water interface control the direction and boundaries of the system. This study analyzed the migration and distribution of HMs in rivers and lakes, and established a Bayesian network model to quantitatively understand the impact of nutrients and key environmental factors on the adsorption-desorption behavior of HMs in lake and estuaries, as well as the competitive relationship between environmental factors. The influence of environmental factors and the occurrence of HMs are both important model inputs. Our findings indicated that the migration risk of Cd in Qinghai Lake was high. Environmental factors such as Cation exchange capacity (CEC), Organic matter (OM), Soluble fluoride (SFL), and pH play the most important role in the adsorption and desorption of HMs. Our findings also indicated that the exchange and activity of HMs in sediments were much higher than in the overlying water. The organic matter content was the most complex environmental factor affecting HMS adsorption and desorption at the water-sediment interface. Additionally, the mass concentration of dissolved oxygen (DO) has a linear relationship with bioavailable HMs in river and lake sediments, but has no linear relationship with the concentration of water-soluble HMs. Interestingly, there are synergistic effects between environmental factors, which directly or indirectly affect the release of bioavailable HMs. However, it is important to determine whether the effects of different environmental factors on the exchange of bioavailable HMs are negative or positive. Our findings suggested that Bayesian network (BN) signals (positive or negative) could provide insights into the transfer direction of metals in the water-sediment interface.