Search

Water parameter estimation based on remote sensing is one of the common water quality evaluation methods. However, it is difficult to describe the relationship between the reflectance and the concentration of non-optically active substances due to their weak optical characteristics, and machine learning has become a viable solution for this problem. Therefore, based on machine learning methods, this study estimated four non-optically active water quality parameters including the permanganate index (CODMn), dissolved oxygen (DO), total nitrogen (TN), and total phosphorus (TP). Specifically, four machine learning models including Support Vector Machine Regression (SVR), Random Forest (RF), Extreme Gradient Boosting (XGBoost), and K-Nearest Neighbor (KNN) were constructed for each parameter and their performances were assessed. The results showed that the optimal models of CODMn, DO, TN, and TP were RF (R2 = 0.52), SVR (R2 = 0.36), XGBoost (R2 = 0.45), and RF (R2 = 0.39), respectively. The seasonal 10 m water quality over the Zhejiang Province was measured using these optimal models based on Sentinel-2 images, and the spatiotemporal distribution was analyzed. The results indicated that the annual mean values of CODMn, DO, TN, and TP in 2022 were 2.3 mg/L, 6.6 mg/L, 1.85 mg/L, and 0.063 mg/L, respectively, and the water quality in the western Zhejiang region was better than that in the northeastern Zhejiang region. The seasonal variations in water quality and possible causes were further discussed with some regions as examples. It was found that DO would decrease and CODMn would increase in summer due to the higher temperature and other factors. The results of this study helped understand the water quality in Zhejiang Province and can also be applied to the integrated management of the water environment. The models constructed in this study can also provide references for related research. [ABSTRACT FROM AUTHOR]

Contaminated sites have a negative impact on human health and the ecological environment, which can potentially lead to major pollution incidents, and frequently receive complaints and reports from the public. Therefore, monitoring public perceptions of contaminated sites is crucial for risk management. However, methods based on traditional questionnaire surveys are limited in terms of time, cost, and target audience size. The purpose of this study was to establish a method using social media to monitor public perceptions of contaminated sites in the Yangtze River Delta urban agglomeration. Thus, 6802 public opinions were collected from social media (mainly microblogs in China) and topic modeling and pollution and spatial mining techniques were employed. The topic modeling results indicated that public perceptions of contaminated sites mainly focused on the construction of prevention and control systems, law enforcement work, developments and challenges applicable to the coal industry, environmental public interest litigation (pertaining to pollution), pollution inspection and rectification, and green development and ecological governance, with intensities of 7.9%, 7.8%, 7.1%, 6.1%, 5.9%, and 4.8%, respectively. Three communities resulting from public opinions in the study area included "environment and pollution," "enterprises," and "environmental protection," with proportions of 37.68%, 34.78%, and 27.54%, respectively. The field investigation results indicated that approximately 90% of the tweets in three typical cities (i.e., Taizhou City in Zhejiang Province and Wuxi and Changzhou Cities in Jiangsu Province) involved key industrial enterprises or contaminated sites located within 1 km of the surrounding areas. The emotional analysis indicated that >3401 tweets dealt with a pollution probability (i.e., the possibility of potentially contaminated sites mentioned in social media becoming contaminated sites) exceeding 0.90 for the period 2011–2021. This finding suggests that the pollution probability for the sites involved in these tweets was high. Our study provides methodological references for monitoring public perceptions of contaminated sites for large-scale, long-term, and high-resolution studies. • Social media can be used for monitoring public perceptions of contaminated sites. • Public perception mainly focused on the construction of prevention and control systems. • >3401 tweets dealt with a pollution probability exceeding 0.90 from 2011 to 2021. • Proposed suggestions for managing public opinion of contaminated sites. [ABSTRACT FROM AUTHOR]

The spatial distribution of farmland productivity is important for reducing the loss of productive farmland and developing sustainable agro-ecosystems in China. This study calculated theoretical, realizable and practical productivity for farmland, taking Zhejiang province as a case. It analyzed their statistical characteristics and spatial distribution, explored the factors influencing productivity and classified the agricultural production zones based on the three indicators for spatial clustering using a fuzzy cluster algorithm. The results indicated there existed a significant difference among the three types of productivity, the theoretical productivity exceeded the realizable productivity by 17.69 % and the realizable productivity exceeded the practical productivity by 20.31 %; their spatial distribution patters were similar, with the highest productivity in the north-central plain region and the lowest in the southwestern mountainous region. Although the land in the study area has been cultivated intensively, the potential to increase production is still enormous, which was up to 593 × 10 kg and 556 × 10 for total theoretical and realizable potential, respectively. The theoretical productivity was affected more by the farmland quality, the realizable productivity more by terrain and socioeconomic factors. The whole province was classified spatially into four agricultural production zones based on the calculated three indicators with three types of productivity data: modern high-efficiency agriculture development zone, characteristic agriculture development zone, ecological agriculture development zone and ocean organic agriculture development zone. Each zone is best used for different purposes and requires different protection and management strategies. Such spatial zoning is essential for stabilizing agricultural production, ensuring food security and promoting the sustainable use of farmland. [ABSTRACT FROM AUTHOR]

Agricultural pollution poses a considerable challenge to grain security and human health, especially in economically developed areas. Mineral exploitation, chemical enterprise operation, pesticide and fertilizer application, sewage discharge, and vehicle emissions are the pollution sources of agricultural land. Identifying and assessing potential agricultural pollution (PAP) is, therefore, the most urgent task to achieve grain security and the human health. Large-scale (e.g., regional or national) PAP assessment can be very expensive, which could also generate a certain amount of information that usually discourages evaluation by decision-makers. To identify areas for regional priority investigation, here we proposed an assessment framework for PAP in economically developed areas. The framework consisted of PAP assessment, vulnerability assessment, hazard assessment, and socio-economic assessment. Then, we conducted a case study by using the proposed framework in one of China's economically developed areas, Zhejiang Province. The results showed that PAP, especially soil heavy metal pollution, soil acidification, and surface water pollution involved almost the entire study area. High-vulnerability high-hazard areas were mainly associated with high socio-economic development or high grain yield. These areas had negatively affected grain security and increased carcinogenic risk, potentially contributing to the formation of cancer villages. Based on the results, we proposed measures for environmental risk managers to alleviate the impact of PAP on grain security and human health in economically developed areas. Image 1 • A new assessment framework for potential agricultural pollution (PAP) was proposed. • Exposure to PAP impacted grain security and human health. • Hazard of mines and chemical companies in economically developed areas were highlighted. • The distribution of centers of grain production shifted toward PAP areas in Zhejiang. The proposed assessment framework can effectively assess priority investigation areas of PAP in China's economically developed areas. [ABSTRACT FROM AUTHOR]

Wise soil management requires detailed soil information, but conventional soil class mapping in a rather coarse spatial resolution cannot meet the demand for precision agriculture. With the advantages of non-destructiveness, rapid cost-efficiency, and labor savings, the spectroscopic technique has proved its high potential for success in soil classification. Previous studies mainly focused on predicting soil classes using a single sensor. In this study, we attempted to compare the predictive ability of visible near infrared (vis-NIR) spectra, mid-infrared (MIR) spectra, and their fused spectra for soil classification. A total of 146 soil profiles were collected from Zhejiang, China, and the soil properties and spectra were measured by their genetic horizons. Along with easy-to-measure auxiliary soil information (soil organic matter, soil texture, color and pH), four spectral data, including vis-NIR, MIR, their simple combination (vis-NIR-MIR), and outer product analysis (OPA) fused spectra, were used for soil classification using a multiple objectives mixed support vector machine model. The independent validation results showed that the classification model using MIR (accuracy of 64.5%) was slightly better than that using vis-NIR (accuracy of 64.2%). The predictive model built on vis-NIR-MIR did not improve the classification accuracy, having the lowest accuracy of 61.1%, which likely resulted from an over-fitting problem. The model based on OPA fused spectra performed best with an accuracy of 68.4%. Our results prove the potential of fusing vis-NIR and MIR using OPA for improving prediction ability for soil classification. [ABSTRACT FROM AUTHOR]

• Laboratory vis–NIR and MIR spectra were compared and fused to predict SOC. • A parallel-input convolutional neural network data fusion framework was presented. • Proposed method outperformed commonly-used data fusion strategies for SOC. • The combinational use of vis–NIR and MIR could improve the prediction for SOC. Visible-to-near-infrared (vis–NIR) and mid-infrared (MIR) spectroscopy have been widely utilized for the quantitative estimation of soil organic carbon (SOC). The fusion of vis–NIR and MIR data can be hypothesized to provide accurate and reliable prediction for SOC because spectral data within a specific range of each individual sensor may lack important absorptive features associated with SOC. In this study, six data fusion strategies, principally direct concatenation-partial least squares regression (DC-PLSR), outer product analysis-PLSR (OPA-PLSR), OPA-competitive adaptive reweighted sampling-PLSR (OPA-CARS-PLSR), sequentially orthogonalized-PLSR (SO-PLSR), DC-convolutional neural network (DC-CNN), and parallel input-CNN (PI-CNN), were compared for the spectral estimations of SOC. The data fusion and individual sensor models were developed using soil samples collected from Zhejiang Province, East China, and scanned under laboratory conditions with both vis–NIR and MIR spectrophotometers. The validation results of vis–NIR (validation coefficient of determination [ R 2] = 0.63–0.73) were generally better than those of MIR (validation R 2 = 0.45–0.59). For data fusion, the best validation accuracy was achieved by the PI-CNN (validation R 2 = 0.84), followed in descending order by DC-CNN (validation R 2 = 0.78), SO-PLSR (validation R 2 = 0.73), OPA-CARS-PLSR (validation R 2 = 0.69), OPA-PLSR (validation R 2 = 0.66), and DC-PLSR (validation R 2 = 0.64). The better performance of PI-CNN over DC-CNN demonstrates the necessity of using different sizes of convolutional kernels before feeding into the fully connected layers in the CNN network for fusing vis–NIR and MIR spectral data. The deep-learning fusion method based on PI-CNN can be considered an efficient tool for integrating data from multiple sensors for estimating soil properties in the field of soil spectral modeling. [ABSTRACT FROM AUTHOR]

Understanding terrestrial ecosystem dynamics requires a comprehensive examination of vegetation changes. Remote sensing technology has been established as an effective approach to reconstructing vegetation change history, investigating change properties, and evaluating the ecological effects. However, current remote sensing techniques are primarily focused on break detection but ignore long-term trend analysis. In this study, we proposed a novel framework based on a change detection algorithm and a trend analysis method that could integrate both short-term disturbance detection and long-term trends to comprehensively assess vegetation change. With this framework, we characterized the vegetation changes in Zhejiang Province from 1990 to 2020 using Landsat and landcover data. Benefiting from combining break detection and long-term trend analysis, the framework showcased its capability of capturing a variety of dynamics and trends of vegetation. The results show that the vegetation was browning in the plains while greening in the mountains, and the overall vegetation was gradually greening during the study period. By comparison, detected vegetation disturbances covered 57.71% of the province's land areas (accounting for 66.92% of the vegetated region) which were mainly distributed around the built-up areas, and most disturbances (94%) occurred in forest and cropland. There were two peak timings in the frequency of vegetation disturbances: around 2003 and around 2014, and the proportions of more than twice disturbances in a single location were low. The results illustrate that this framework is promising for the characterization of regional vegetation growth, including long-term trends and short-term features. The proposed framework enlightens a new direction for the continuous monitoring of vegetation dynamics. • A novel framework is proposed to comprehensively assess vegetation changes. • The long-term trends can be supplemented with short-term disturbance features. • The framework worked well in our case evaluation. • The framework enlightens a new direction for the continuous monitoring of vegetation dynamics. [ABSTRACT FROM AUTHOR]

• IMERG-Early, GSMaP-NRT and FY4A QPE were evaluated for overall performance and five typhoons during 2018–2019. • IMERG-Early and GSMaP-NRT performed much better than FY4A QPE. • Satellite-based products tended to underestimate heavy precipitation but overestimate light precipitation. • Satellite-based products showed unstable performance and were more likely to present low estimations during typhoons. • Satellite-based products were difficult to detect heavy precipitation centers during typhoons in some parts of southwestern Zhejiang with complex topography. Effective and accurate estimation of heavy precipitation from typhoons is critical for the early warning of hydrological extremes and disaster mitigation. Satellite-based precipitation products are widely used for various purposes, including supporting climate, hydrology, and severe weather event monitoring and research activities. In this study, the performance of Integrated Multi-SatellitE Retrievals for Global Precipitation Measurement (IMERG), Global Satellite Mapping of Precipitation (GSMaP), and Chinese Fengyun-4A quantitative precipitation estimates (FY4A QPE) were compared to daily observations from 2018 to 2019. Further comparison of these products at an hourly scale was implemented for five typhoons that made landfall in the coastal areas of southeastern China. The results demonstrate that IMERG-Early and GSMaP-NRT performed much better than FY4A QPE, in both overall assessment and specific typhoons. All the satellite-based products showed the best correlations with the gauge observations for heavy precipitation events, and their correlations became worse with lower precipitation intensity. They tended to underestimate heavy precipitation but overestimate light precipitation. Furthermore, their random errors increased with increasing precipitation, according to an analysis of unbiased root mean squared error (ubRMSE). Based upon analyses of the probability of detection (POD) and critical success index (CSI), satellite-based products performed better during light precipitation events than moderate and heavy precipitation events. During typhoons, IMERG-Early and GSMaP-NRT had better accuracy in detecting typhoon precipitation with higher POD and CSI values. FY4A QPE showed instability in different typhoons and a low correlation with gauge observations. In addition, satellite-based products produced precipitation estimates more reliably in northern and eastern Zhejiang, but were difficult to detect heavy precipitation centers in some parts of southwestern Zhejiang with complex topography. These findings not only provide valuable references for the development of satellite-based retrieval algorithms to enhance the ability to monitor typhoon heavy precipitation but also support diverse hydrological models and applications. [ABSTRACT FROM AUTHOR]

Modelling the mass balance and forecasting the temporal variations of cadmium (Cd) in farmland soil play a critical role in the development of mitigation strategies for Cd pollution. In this study, a novel framework integrating the mass balance model with model-independent parameter estimation, geostatistics, and bagging algorithms were integrated to simulate the long-term changes in the Cd content of farmland soil in Zhejiang Province, China. The predicted Cd content in farmland soil in 2013 was compared to observed data (R value = 0.568 and root-mean-square error = 0.177 mg kg−1), demonstrating the feasibility of our model. The prediction results for 2050 indicated that the average concentration of Cd in farmland soil from Zhejiang Province will increase to 0.30 mg kg−1 if the current trend continues, and that 37.4% of the farmland soil in the province will be classified as a "security utilisation region", indicating great risk of soil Cd contamination in these areas. Reducing industrial emissions and soil acidification to reduce the Cd pollution risk should receive great attention. This study provides a new perspective for forecasting the temporal trends of Cd accumulation in farmland soil and facilitates improved management and risk prevention of Cd pollution in agricultural soils and products. Image 1 • A novel framework was proposed to model cadmium accumulation in farmland soil. • Sensitivity analysis was used to identify most impact parameters for prediction. • The model had R value of 0.568 and RMSE of 0.177 mg kg−1 validated by data of 2013. • Simulations were employed and averaged to produce the final forecast map. • Cd in farmland soil in Zhejiang Province will increase to 0.30 mg kg−1 in 2050. [ABSTRACT FROM AUTHOR]