Occurrence forms and environmental characteristics of phosphorus in water column and sediment of urban waterbodies replenished by reclaimed water.

To illustrate the evolution process, environmental feature and phytoremediation effect of phosphorus (P) in waterbodies with long-term replenishment by reclaimed water (RW), an urban waterbody using RW as the sole supply was selected as a case study. The concentration and distribution of soluble reactive phosphate (SRP), dissolved organic P (DOP) and particulate P (PP) in water column, as well as organic P (OP), inorganic P (IP), exchangeable P (Ex-P), BD-P (redox-sensitive P), NaOH-P (P bound to Fe and Al oxyhydroxides) and HCl-P (P bound to Ca) in sediment were investigated. Results showed that the seasonal average concentrations of total phosphorus (TPw) in water column ranged from 0.048 to 0.130 mg∙L-1, with the highest in summer and the lowest in winter. P in water column was predominantly present in dissolved state, with the similar proportions of SRP and DOP. SRP decreased apparently in midstream, where the phytoremediation was applied extensively. PP content obviously increased in non-phytoremediation area in downstream, due to visitor activity and sediment resuspension. Total phosphorus (TPs) in sediments was between 352.9 and 1331.3 mg∙kg-1, with average 365.7 mg∙kg-1 of IP and 382.8 mg∙kg-1 of OP. Among IP, HCl-P had the highest proportion, followed by BD-P, NaOH-P and Ex-P. OP was significantly higher in phytoremediation areas than that in non-phytoremediation areas. Coverage of aquatic plants was positively correlated with TP, OP, BAP, while it was negatively correlated with BD-P. Hydrophyte stabilized and conserved active P in sediment and prevented release of active P. Moreover, hydrophyte increased NaOH-P and OP contents in sediment through regulating the abundance of phosphorus-solubilizing bacteria (PSB), such as Lentzea and Rhizobium. Four sources were identified through two multivariate statistical models. RW and runoff were the dominant sources of P accounting for 52.09 %, which mainly contributed to P in sediment especially IP.

Soil Heavy Metal Content Prediction Based on a Deep Belief Network and Random Forest Model.

In order to extract useful information from X-ray fluorescence (XRF) spectra and establish a high-accuracy prediction model of soil heavy metal contents, a hybrid model combining a deep belief network (DBN) with a tree-based model was proposed. The DBN was first introduced into feature extraction of XRF spectral data, which can obtain deep layer features of spectra. Owing to the strong regression ability of the tree-based model, it can offset the deficiency of DBN in prediction ability so it was used for predicting heavy metal contents based on the extracted features. In order to further improve the performance of the model, the parameters of model can be optimized according to the prediction error, which was completed by sparrow search algorithm and the gird search. The hybrid model was applied to predict the contents of As and Pb based on spectral data of overlapping peaks. It can be obtained that R2 of As and Pb reached 0.9884 and 0.9358, the mean square error of As and Pb are as low as 0.0011 and 0.0058, which outperform other commonly used models. That proved the combination of DBN and tree-based model can obtain more accurate prediction results.