Spatial variation of soil phosphorus in the water level fluctuation zone of the Three Gorges Reservoir: Coupling effects of elevation and artificial restoration.

The water level fluctuation zone (WLFZ) is a distinctive and important component of the reservoir ecosystem. Due to periodic inundation, the fraction, spatial distribution, and chemical reactivity of soil phosphorus (P) within the WLFZ can potentially impact the loading of P into reservoir waters. However, a detailed study of this subject is lacking. In this study, the soil P in the WLFZ of the Three Gorges Reservoir, China, was examined using a combination of chemical sequential extraction, 31P NMR, and adsorption experiments. The results of chemical sequential extraction showed that HCl-Pi constituted the largest P pool among all P forms, with a mean concentration of 338 mg/kg. The content of HCl-Pi decreased significantly toward the dam, while the content of Res-P decreased in the opposite direction. The highest contents of most P forms and total P were observed at an elevation of 160 m. 31P NMR measurements showed that NaOH-EDTA Pi detectable in WLFZ soils at 145 m, 160 m, and 175 m elevation consisted mainly of orthophosphate and pyrophosphate, while NaOH-EDTA Po contained phosphate monoesters and phosphate diesters, accounting for 1.4 % to 46.2 % of NaOH-EDTA TP. Adsorption experiments showed that soil P in the WLFZ was a potential P source for reservoir waters, with chemisorption being the dominant mechanism of P sequestration. The adsorption equilibrium concentration of WLFZ soil was lower at higher elevations (>170 m) compared to lower elevations (<150 m), exhibiting a decrease in the average maximum adsorption from 271 mg/kg to 192 mg/kg. Statistical analysis suggested that Ca and Fe content, particle size, elevation, and artificial restoration were key factors affecting the fraction and content of soil P in the WLFZ. Our findings contribute to an improved understanding of the behavior of soil P in the WLFZ of large reservoirs and its potential contribution to the reservoir waters.

A novel Encoder-Decoder model based on read-first LSTM for air pollutant prediction.

Accurate air pollutant prediction allows effective environment management to reduce the impact of pollution and prevent pollution incidents. Existing studies of air pollutant prediction are mostly interdisciplinary involving environmental science and computer science where the problem is formulated as time series prediction. A prevalent recent approach to time series prediction is the Encoder-Decoder model, which is based on recurrent neural networks (RNN) such as long short-term memory (LSTM), and great potential has been demonstrated. An LSTM network relies on various gate units, but in most existing studies the correlation between gate units is ignored. This correlation is important for establishing the relationship of the random variables in a time series as the stronger is this correlation, the stronger is the relationship between the random variables. In this paper we propose an improved LSTM, named Read-first LSTM or RLSTM for short, which is a more powerful temporal feature extractor than RNN, LSTM and Gated Recurrent Unit (GRU). RLSTM has some useful properties: (1) enables better store and remember capabilities in longer time series and (2) overcomes the problem of dependency between gate units. Since RLSTM is good at long term feature extraction, it is expected to perform well in time series prediction. Therefore, we use RLSTM as the Encoder and LSTM as the Decoder to build an Encoder-Decoder model (EDSModel) for pollutant prediction in this paper. Our experimental results show, for 1 to 24 h prediction, the proposed prediction model performed well with a root mean square error of 30.218. The effectiveness and superiority of RLSTM and the prediction model have been demonstrated.