To address the challenge of partial missing data in atmospheric CO₂ concentration observations, this study systematically evaluated the performance of 12 mainstream machine learning algorithms for data imputation, utilizing CO₂ measurement data collected at the 30m (L1 layer) and 80m (L2 layer) heights of the Longfengshan Regional Atmospheric Background Station over the period from January 2009 to June 2022. An optimized ensemble machine learning method was proposed, with the distance-weighted K-Nearest Neighbors (KNeighborsDist) algorithm as its core framework. This method reconstructs missing data by leveraging synchronous observational records from the two co-located height layers. Validation analysis of the reconstructed dataset yielded excellent performance (R²≥0.98), which confirms a strong correlation between the reconstructed greenhouse gas data and in-situ measurements, as well as the effectiveness and accuracy of the proposed model in greenhouse gas data reconstruction tasks. The method successfully filled the data gaps in the original dataset of Longfengshan Station, generating a continuous 5minute interval CO₂ dataset with a completeness rate of 100%. Additionally, it well preserved the long-term trends and seasonal, monthly, and diurnal variation characteristics of CO₂. The proposed AI-based data reconstruction algorithm is applicable to other atmospheric background stations and can provide high-quality data support for regional climate change analysis, carbon emission accounting, and inversion research.

Jiangsu Province—one of China's major rice-producing regions—was selected as the research area. A high-precision CH₄ observation system was established at a height of 50m in Taizhou City, central Jiangsu, and the period from 20^th May to 31^st October 2023, covering the local rice-growing season, was used for analysis. Paddy CH₄ flux was retrieved at a 10d scale by coupling an atmospheric transport model with an optimized inversion framework. The main findings are as follows: (1) The "bottom-up" approach based EDGAR prior inventory exhibited neither distinct temporal variability nor emission peaks for paddy CH₄ fluxes, inconsistent with the observed peak emissions during the tillering-booting stage in July~August in Jiangsu, which should be constrained and calibrated by using atmospheric CH₄ observations. (2) Inversions revealed that both Taizhou City and Jiangsu Province experienced peak CH₄ emissions in early to mid-August, followed by a gradual decline. The peak fluxes reached (966.6 ± 104.6) mg/(m²·d) and (1182.9 ± 128.0) mg/(m²·d), respectively, which were (3.27 ± 0.35) times higher than the prior estimates. (3) Over the entire study period, the posterior mean paddy CH₄ flux for Taizhou and Jiangsu were (307.3 ± 40.5) mg CH₄/(m²·d) and (376.7 ± 49.7) mg CH₄/(m²·d), respectively, (1.37 ± 0.18) times the prior values. These results indicate that the prior inventory substantially underestimated both the magnitude and temporal variation of paddy CH₄ emissions.

A thin carbon layer (~1.15μm thick) was fabricated on coarse polyethylene terephthalate (PET) fibers via in-situ modification with sucrose-derived carbon (SDC), which only occupied 10^-3 of the fiber's pore size and significantly improved the dielectric properties. The composite filter material achieved a high filtration efficiency of 98.10% for 0.3μm particles while maintaining low resistance, resulting in a comprehensive quality factor (CQF) significantly higher than that of the origin PET fibers. Furthermore, the carbon layer was strongly bonded to the PET fibers, ensuring excellent long-term stability and suitability for repeated use.

Co-doped ZnO gas-sensitive materials were synthesized via a hydrothermal method, and the influence of varying Co doping concentrations on their gas-sensing properties was systematically investigated. Furthermore, focusing on the emission characteristics of acetone within the multi-component VOCs generated during rubber vulcanization, the feasibility of real-time acetone detection using these materials was explored, aiming to enable cost-effective operation of adsorption systems under compliance conditions. The results indicated that the Co-ZnO-2% sensor achieved a high response value of 1146.8 to 260mg/m³ acetone at 240℃, with rapid response and recovery times of 16s and 8s, respectively. Additionally, it exhibited excellent selectivity, reproducibility, and long-term stability toward acetone. In a simulated VOC adsorption purification experiment, the Co-ZnO-2% sensor was applied, and its response values were demonstrated to effectively indicate both the breakthrough and saturation states of the adsorbent. Under multi-component adsorption conditions, with acetone as a representative indicator, a warning threshold and a concentration prediction model, X_acetone=(Y-16.89)/4.47, were established to estimate VOC concentrations. This approach successfully realized online monitoring of the adsorption process and evaluation of the emission status.

In this study, dewatered sludge pre-granules were used as seed sludge to cultivate aerobic granular sludge (AGS) with high phosphorus removal capabilities over a 120-day operation. Inoculation with pre-granules allowed the system to bypass the initial granulation phase, establishing stable biological phosphorus removal by day 42. During the stable period, the experimental group achieved an average total phosphorus removal efficiency of (94.4±3.7)%, outperforming the control group (85.9±5.1)%. The pre-granulated AGS exhibited a particle size of 1100~1300μm, a sludge volume index of 50~80mL/g, and average removal efficiencies for chemical oxygen demand and total nitrogen of (94.5±2.9)% and (79.3±6.4)%, respectively. Scanning electron microscopy with energy-dispersive X-ray spectroscopy (SEM-EDS) revealed a notably higher iron content in the pre-granulated AGS (3.87% vs. 0.24% in the control), with iron uniformly distributed inside the granules. High-throughput sequencing further indicated enrichment of extracellular polymeric substance (EPS) secretion microorganisms, such as Candidatus_Competibacter (10.9% vs. 5.0% in the control). The results demonstrated that, unlike the primarily biological phosphorus removal mode in the control, the pre-granulated AGS developed a synergistic chemical-biological removal mechanism. Initially, EPS complexed with residual ferric iron from the dewatered sludge pre-granules, enhancing chemical phosphorus adsorption and providing a structural skeleton that improved granule stability. Subsequently, this stable structure facilitated microbial retention, yielding a higher mixed liquor volatile suspended solids concentration (4.3g/L vs.3.1g/L). Although the relative abundance of polyphosphate-accumulating organisms (PAOs) was similar between groups (0.48% vs. 0.47%), the absolute abundance of PAOs was elevated in the pre-granulated system. Moreover, the stable granular structure likely supported higher metabolic activity PAOs, as reflected by an increase in the relative abundance of key functional genes, such as polyphosphate kinase (ppk1, 0.42‰ vs. 0.34‰).

Aiming at the problems of low sewage temperature, poor nitrogen removal efficiency and weak shock resistance of traditional activated sludge process in the alpine regions of southwest China, a pilot-scale experiment of pure moving bed biofilm reactor (MBBR) was conducted. The startup characteristics, nitrogen removal performance, and shock resistance mechanism under long-term ultra-low temperature conditions were systematically investigated. The results showed that the pure MBBR was successfully started up within 25days at a water temperature of (6.14±1.06)℃. During a long-term operation of 245days with an average water temperature of (5.45±1.29)℃, superior effluent quality was achieved with an actual hydraulic retention time (HRT) only 20% of that of the activated sludge process in the same WWTP. After a 1-month shutdown, the treatment capacity was fully restored within only 14days. In the pure-membrane MBBR system, the first-stage nitrification zone realized the removal of inhibitors such as chemical oxygen demand (COD) and preliminary nitrification, the second stage served as the main nitrification zone, and the third stage ensured stable effluent quality. A single-stage process was adopted in the denitrification zone, where the reflux ratio was flexibly adjusted according to the residual nitrate nitrogen in the anoxic zone to enhance nitrogen removal. Regarding nitrogen removal potential, when facing influent quality shock, increasing dissolved oxygen (DO) in the aerobic zone to strengthen mass transfer and oxygen supply could further improve the system's nitrification load by more than 25%, while driving the nitrogen removal capacity to increase by more than 10%. The excellent low-temperature resistance of the pure MBBR was attributed to two aspects: on one hand, the biofilm enhanced the enrichment of nitrogen-removing functional bacteria, especially low-temperature-tolerant strains such as Flavobacterium and Dechloromonas; On the other hand, the biofilm's dense structure and complex microbial community supported thermal insulation. For pure MBBR design, enhanced functional bacteria enrichment and high effective biomass should be concurrently considered, thus conferring higher treatment load.

This study examined a combined potassium permanganate (KMnO₄) and polyaluminium chloride (PAC) process for treating karst waters, with experiments simulating the seasonal variations in manganese concentration. Using algal organic matter (AOM) from Chlorella sp. and Pseudoanabaena sp., the mechanisms for Mn(II) removal, AOM transformation, and DBPs control were investigated. The results indicated that KMnO₄ oxidized Mn(II) to MnO₂, while PAC promoted the co-precipitation of MnO₂ with AOM through adsorption and bridging. TEM analysis confirmed that MnO₂ was encapsulated by PAC flocs and associated with AOM. FTIR and ICP-MS analyses demonstrated enhanced Mn(II) adsorption via coordination with algal surface functional groups. The combined process reduced the formation of carbonaceous DBPs (C-DBPs) and nitrogenous DBPs (N-DBPs) by 16% and 18.5%, respectively. Notably, trichloromethane (TCM) formation was reduced by 51%, underscoring the process's effectiveness against C-DBPs. Compared to KMnO₄ alone, the combined process reduced N-DBPs formation by only 0.6% in the Chlorella sp. system but by 27% in the Pseudoanabaena sp. system. This disparity suggests that DBP formation pathways are differentially regulated by AOM components, which vary with algal species.

The triboelectric nanogenerator (TENG), which was constructed of indium tin oxide (ITO), polytetrafluoroethylene (PTFE), and conductive aluminum tape, was utilized in this study. It was driven by secondary effluent to generate electricity for degrading methyl orange wastewater. Experimental results showed that a maximum output voltage of 23.52V and a current of 2.15μA were achieved by a single TENG unit. When aluminum, iron, copper, and stainless steel were used as electrodes to treat 40mg/L methyl orange wastewater, decolorization rates of 83.21%, 82.88%, 75.27%, and 51.38% were obtained respectively. With aluminum electrodes, decolorization rates exceeding 80% were achieved for methyl orange concentrations<60mg/L, but they declined gradually for methyl orange concentrations>60mg/L, and only 27.85% was reached at 150mg/L. When the number of TENG units was increased from one to six, the decolorization rate was increased from 85.05% to 91.05%. The degradation process was fitted to the first-order kinetic equation: ln(C₀/C_t) = 0.027t-0.0815.

Acidic mine drainage (AMD) containing high concentrations of heavy metal ions such as Fe²⁺ and Mn²⁺ was treated by a freezing method utilizing the purification effect of ice crystals. The process parameters for removing Fe²⁺ and Mn²⁺ from simulated AMD via the freezing method were optimized through single-factor and orthogonal experiments, and the optimal conditions were determined (for Fe²⁺: freezing temperature of -12℃, initial concentration of 0.9mg/L, freezing radius of 4.25cm; for Mn²⁺: freezing temperature of -6℃, initial concentration of 1.2mg/L, freezing radius of 5.00cm). The optimized process was then applied to treat actual AMD. The results showed that the removal rates of Fe²⁺ and Mn²⁺ from the actual AMD reached 60% and 59%, respectively, which confirmed the effectiveness of the technology. Furthermore, barley seeds and seedlings were used as model plants to evaluate the phytotoxicity of the AMD before and after treatment. The results indicated that, compared with the untreated AMD, the treated AMD did not inhibit the germination of barley seeds or the growth of barley seedlings. This study confirmed that the freezing method can not only effectively remove heavy metal pollutants from AMD but also significantly reduce its biotoxicity.

In response to the constrained efficacy of traditional in situ chemical oxidation and enhanced bioremediation in karst groundwater, novel slow-release materials (SRMs) were prepared employing paraffin and concrete as binding agents. Simulated karst column experiments were performed to characterize the release kinetics and assess the removal performance of these SRMs against typical fuel pollutants (BTX). In this study, slow-release materials were prepared using paraffin and concrete as binders. Column experiments simulating a dynamic karst aquatic environment were conducted to investigate the release behavior of these materials and their efficacy in removing typical fuel contaminants (BTX compounds). The results showed that: Paraffin-based slow-release materials exhibited significant media-specific release behaviour, but their initial burst release of persulfate resulted in increased non-target consumption. Concrete-based slow-release materials demonstrated a more stable release profile and achieved a higher nitrate utilization rate in the remediation system. Although ferrous salt activation improved the oxidation efficiency of persulfate, challenges such as short activation duration (<72hours), limited spatial influence, and low ferrous ion utilization (leading to iron hydroxide precipitation) remained. In karst groundwater, the first-order decay rate constant for benzene series concentration under the slow-release dosing method (0.154d^-1) was higher than that under the continuous dosing method (0.044d^-1), indicating superior removal efficiency with slow-release dosing. The pH buffering effect in the limestone medium delayed persulfate decay, while the sandy medium facilitated the enrichment of functional microorganisms, thereby enhancing bioremediation.