Simple low-cost approaches to semantic segmentation in radiation therapy planning for prostate cancer using deep learning with non-contrast planning CT images

PurposeDeep learning has shown great efficacy for semantic segmentation. However, there are difficulties in the collection, labeling and management of medical imaging data, because of ethical complications and the limited number of imaging studies available at a single facility.This study aimed to find a simple and low-cost method to increase the accuracy of deep learning semantic segmentation for radiation therapy of prostate cancer.MethodsIn total, 556 cases with non-contrast CT images for prostate cancer radiation therapy were examined using a two-dimensional U-Net. Initially, all slices were used for the input data. Then, we removed slices of the cranial portions, which were beyond the margins of the bladder and rectum. Finally, the ground truth labels for the bladder and rectum were added as channels to the input for the prostate training dataset.ResultsThe highest mean dice similarity coefficients (DSCs) for each organ in the test dataset of 56 cases were 0.85 ± 0.05, 0.94 ± 0.04 and 0.85 ± 0.07 for the prostate, bladder and rectum, respectively. Removal of the cranial slices from the original images significantly increased the DSC of the rectum from 0.83 ± 0.09 to 0.85 ± 0.07 (p < 0.05). Adding bladder and rectum information to prostate training without removing the slices significantly increased the DSC of the prostate from 0.79 ± 0.05 to 0.85 ± 0.05 (p < 0.05).ConclusionsThese cost-free approaches may be useful for new applications, which may include updated models and datasets. They may be applicable to other organs at risk (OARs) and clinical targets such as elective nodal irradiation.     

Quantifying Relationships Among Phosphorus, Agriculture, and Lake Depth at an Inter-Regional Scale

To date, studies examining the impact of agriculture on freshwater systems have been spatially confined (that is, single drainage basin or regional level). Across regions, there are considerable differences in a number of factors, including geology, catchment morphometry, and hydrology that affect water quality. Given this heterogeneity, it is unknown whether agricultural activities have a pervasive impact on lake trophic state across large spatial scales. To address this issue, we tested whether the proportion of agricultural land in a catchment (% Agr) could explain a significant portion of the variation in lake water quality at a broad inter-regional scale. As shallow, productive systems have been shown to be particularly susceptible to eutrophication, we further investigated how lake mean depth modulates the relationship between % Agr and lake total phosphorus (TP) concentration. We applied both traditional meta-analytic techniques and more sophisticated linear mixed-effects models to a dataset of 358 temperate lakes that spanned an extensive spatial gradient (5°E to 73°W) to address these issues. With meta-analytical techniques we detected an across-study correlation between TP and % Agr of 0.53 (one-tailed P-value = 0.021). The across-study correlation coefficient between TP and mean depth was substantially lower (r = −0.38; P = 0.057). With linear mixed-effects modeling, we detected among-study variability, which arises from differences in pre-impact (background) lake trophic state and in the relationship between lake mean depth and lake TP. To our knowledge, this is the first quantitative synthesis that defines the influence of agriculture on lake water quality at such a broad spatial scale. Syntheses such as these are required to define the global relationship between agricultural land-use and water quality.     

Spatiotemporal pattern of gross primary productivity and its covariation with climate in China over the last thirty years

The uncertainties of China's gross primary productivity (GPP) estimates by global data‐oriented products and ecosystem models justify a development of high‐resolution data‐oriented GPP dataset over China. We applied a machine learning algorithm developing a new GPP dataset for China with 0.1° spatial resolution and monthly temporal frequency based on eddy flux measurements from 40 sites in China and surrounding countries, most of which have not been explored in previous global GPP datasets. According to our estimates, mean annual GPP over China is 6.62 ± 0.23 PgC/year during 1982–2015 with a clear gradient from southeast to northwest. The trend of GPP estimated by this study (0.020 ± 0.002 PgC/year² from 1982 to 2015) is almost two times of that estimated by the previous global dataset. The GPP increment is widely spread with 60% area showing significant increasing trend (p < .05), except for Inner Mongolia. Most ecosystem models overestimated the GPP magnitudes but underestimated the temporal trend of GPP. The monsoon affected eastern China, in particular the area surrounding Qinling Mountain, seems having larger contribution to interannual variability (IAV) of China's GPP than the semiarid northwestern China and Tibetan Plateau. At country scale, temperature is the dominant climatic driver for IAV of GPP. The area where IAV of GPP dominated by temperature is about 42%, while precipitation and solar radiation dominate 31% and 27% respectively over semiarid area and cold‐wet area. Such spatial pattern was generally consistent with global GPP dataset, except over the Tibetan Plateau and northeastern forests, but not captured by most ecosystem models, highlighting future research needs to improve the modeling of ecosystem response to climate variations.     

Classification of broiler behaviours using triaxial accelerometer and machine learning

Understanding broiler behaviours provides important implications for animal well-being and farm management. The objectives of this study were to classify specific broiler behaviours by analysing data from wearable accelerometers using two machine learning models, K-Nearest Neighbour (KNN) and Support Vector Machine (SVM). Lightweight triaxial accelerometers were used to record accelerations of nine 7-week-old broilers at a sampling frequency of 40 Hz. A total of 261.6-min data were labelled for four behaviours – walking, resting, feeding and drinking. Instantaneous motion features including magnitude area, vector magnitude, movement variation, energy, and entropy were extracted and stored in a dataset which was then segmented by one of the six window lengths (1, 3, 5, 7, 10 and 20 s) with 50% overlap between consecutive windows. The mean, variation, SD, minimum and maximum of each instantaneous motion feature and two-way correlations of acceleration data were calculated within each window, yielding a total of 43 statistic features for training and testing of machine learning models. Performance of the models was evaluated using pure behaviour datasets (single behaviour type per dataset) and continuous behaviour datasets (continuous recording that involved multiple behaviour types per dataset). For pure behaviour datasets, both KNN and SVM models showed high sensitivities in classifying broiler resting (87% and 85%, respectively) and walking (99% and 99%, respectively). The accuracies of SVM were higher than KNN in differentiating feeding (88% and 75%, respectively) and drinking (83% and 62%, respectively) behaviours. Sliding window with 1-s length yielded the best performance for classifying continuous behaviour datasets. The performance of classification model generally improved as more birds were included for training. In conclusion, classification of specific broiler behaviours can be achieved by recording bird triaxial accelerations and analysing acceleration data through machine learning. Performances of different machine learning models differ in classifying specific broiler behaviours.     

Strategies to develop radiomics and machine learning models for lung cancer stage and histology prediction using small data samples

Predictive models based on radiomics and machine-learning (ML) need large and annotated datasets for training, often difficult to collect. We designed an operative pipeline for model training to exploit data already available to the scientific community. The aim of this work was to explore the capability of radiomic features in predicting tumor histology and stage in patients with non-small cell lung cancer (NSCLC).We analyzed the radiotherapy planning thoracic CT scans of a proprietary sample of 47 subjects (L-RT) and integrated this dataset with a publicly available set of 130 patients from the MAASTRO NSCLC collection (Lung1). We implemented intra- and inter-sample cross-validation strategies (CV) for evaluating the ML predictive model performances with not so large datasets.We carried out two classification tasks: histology classification (3 classes) and overall stage classification (two classes: stage I and II). In the first task, the best performance was obtained by a Random Forest classifier, once the analysis has been restricted to stage I and II tumors of the Lung1 and L-RT merged dataset (AUC = 0.72 ± 0.11). For the overall stage classification, the best results were obtained when training on Lung1 and testing of L-RT dataset (AUC = 0.72 ± 0.04 for Random Forest and AUC = 0.84 ± 0.03 for linear-kernel Support Vector Machine).According to the classification task to be accomplished and to the heterogeneity of the available dataset(s), different CV strategies have to be explored and compared to make a robust assessment of the potential of a predictive model based on radiomics and ML.     

Performance evaluation of an integrated constructed wetland used to treat a contaminated aquatic environment

The treatment performance of an integrated constructed wetland (ICW) that was in operation for 3 years was evaluated. Artificial neural network modeling was used to predict contaminant treatment efficiencies based on easily measured field parameters. The estimates for average yearly removals of total phosphorus (TP), total nitrogen (TN), chemical oxygen demand (COD), and total suspended solids (TSS) were 0.81 ± 0.18, 7.17 ± 1.62, 63.80 ± 17.41, and 126.12 ± 48.61 g m^?2 d^?1, respectively. Removal velocities of contaminants were determined from analyses of inlet–outlet datasets. The areal removal rate constants were 0.46, 0.73, 0.44, and 0.82 m d^?1 for TP, TN, COD, and TSS, respectively. The presence of high background concentrations of contaminants (TP: 0.01 mg L^?1, TN: 0.13 mg L^?1, COD: 6.43 mg L^?1, TSS: 14.83 mg L^?1) indicated that the water in the ICW was mesotrophic. Statistical methods (i.e., principal component analysis (PCA), forward selection, and correlation analysis) were used to select optimal input subsets for different contaminants. These data subsets were subsequently used for model development. To find the optimal network architectures, a genetic algorithm was introduced to the learning processes. The models were competent at providing reasonable matches between the measured and the predicted effluent concentrations of TP (R² = 0.9711), TN (R² = 0.8875), COD (R² = 0.9359), and TSS (R² = 0.9164). The results of the models provided information that will be useful for the design and modification of constructed wetlands.     

程海富营养化机理的神经网络模拟及响应情景分析

揭示湖泊的富营养化发生机制、定量了解关键生源要素与藻类爆发的因果关联对有效改善湖泊水质和富营养化状况具有重要的科学与决策意义。本研究以云南省程海为例,建立了基于神经网络的响应模型,对富营养化机理进行了研究,并从富营养化核心驱动因子识别、神经网络模型构建与架构分析以及叶绿素a(Chl a)与TN、TP浓度降低的响应模拟几个方面对面临的科学问题进行探索。模拟结果表明,神经网络模型必须在适当的架构下才能产生科学合理的结果;程海的富营养化机制由一个氮(N)、磷(P)共限制的营养盐-藻类动力结构主导,但在此主导结构下拥有氮型限制的次级结构。基于神经网络模型模拟,推导出一系列基于湖体水质控制的Chl a响应的非线性函数,为程海的富营养化控制提供了快速决策支持。     

Improvement of image quality for pancreatic cancer using deep learning-generated virtual monochromatic images: Comparison with single-energy computed tomography

PurposeTo construct a deep convolutional neural network that generates virtual monochromatic images (VMIs) from single-energy computed tomography (SECT) images for improved pancreatic cancer imaging quality.Materials and methodsFifty patients with pancreatic cancer underwent a dual-energy CT simulation and VMIs at 77 and 60 keV were reconstructed. A 2D deep densely connected convolutional neural network was modeled to learn the relationship between the VMIs at 77 (input) and 60 keV (ground-truth). Subsequently, VMIs were generated for 20 patients from SECT images using the trained deep learning model.ResultsThe contrast-to-noise ratio was significantly improved (p < 0.001) in the generated VMIs (4.1 ± 1.8) compared to the SECT images (2.8 ± 1.1). The mean overall image quality (4.1 ± 0.6) and tumor enhancement (3.6 ± 0.6) in the generated VMIs assessed on a five-point scale were significantly higher (p < 0.001) than that in the SECT images (3.2 ± 0.4 and 2.8 ± 0.4 for overall image quality and tumor enhancement, respectively).ConclusionsThe quality of the SECT image was significantly improved both objectively and subjectively using the proposed deep learning model for pancreatic tumors in radiotherapy.     

Water quality and diversity of the Recent ostracod fauna in lowland springs from Lombardy (northern Italy)

Sedimentary records provide important information for understanding changes in the history of eutrophication in Lake Taihu. In addition, the catchment nutrient model SWAT provides a powerful tool to examine eutrophic changes in a long-term context. Since it is difficult to evaluate impacts of natural eutrophic development and anthropogenic changes in catchment discharge and land use, simulation of past changes provides a mirror on processes and dynamics. Boundaries in the simulations are set to a pre-industrial time to evaluate natural-agricultural nutrient changes in Taihu Basin a 100 years ago. Total nitrogen (TN) and total phosphorus (TP) in the main channel flowing into the lake are simulated in four sub-basins for 200 model years. Results show that modeling can capture basic features of basin nutrient development, where mean TN concentration (0.12 mg l⁻¹) can be compared in broad scale to mean TN concentration (0.17 mg kg⁻¹) from Lake Taihu sedimentary cores dating back about 100 years. Spatial nutrient simulations suggest that the two major nutrient sources are from the southwestern sub-basin (48% TN and 68% TP of the basin total) and the northwestern sub-basin (18% TN and 17% TP). There are differences of +7.3 × 10⁴ kg TN and +2.0 × 10⁵ kg TP between total input and output values, simulating mean annual amounts of nutrient deposited into the lake. TN and TP concentration differences between input and output sub-basins become smaller in the second 100 years than the first 100 years, suggesting a 100 year period to reach a balance of net nutrients. Catchment nutrient modeling provides a basis to evaluate how nutrient production and balance responded to environmental changes over 200 years in Taihu Basin.     

Effects of straw carbon input on carbon dynamics in agricultural soils: a meta‐analysis

Straw return has been widely recommended as an environmentally friendly practice to manage carbon (C) sequestration in agricultural ecosystems. However, the overall trend and magnitude of changes in soil C in response to straw return remain uncertain. In this meta‐analysis, we calculated the response ratios of soil organic C (SOC) concentrations, greenhouse gases (GHGs) emission, nutrient contents and other important soil properties to straw addition in 176 published field studies. Our results indicated that straw return significantly increased SOC concentration by 12.8 ± 0.4% on average, with a 27.4 ± 1.4% to 56.6 ± 1.8% increase in soil active C fraction. CO₂ emission increased in both upland (27.8 ± 2.0%) and paddy systems (51.0 ± 2.0%), while CH₄ emission increased by 110.7 ± 1.2% only in rice paddies. N₂O emission has declined by 15.2 ± 1.1% in paddy soils but increased by 8.3 ± 2.5% in upland soils. Responses of macro‐aggregates and crop yield to straw return showed positively linear with increasing SOC concentration. Straw‐C input rate and clay content significantly affected the response of SOC. A significant positive relationship was found between annual SOC sequestered and duration, suggesting that soil C saturation would occur after 12 years under straw return. Overall, straw return was an effective means to improve SOC accumulation, soil quality, and crop yield. Straw return‐induced improvement of soil nutrient availability may favor crop growth, which can in turn increase ecosystem C input. Meanwhile, the analysis on net global warming potential (GWP) balance suggested that straw return increased C sink in upland soils but increased C source in paddy soils due to enhanced CH₄ emission. Our meta‐analysis suggested that future agro‐ecosystem models and cropland management should differentiate the effects of straw return on ecosystem C budget in upland and paddy soils.     

Optimizing Artificial Neural Network-Based Models to Predict Rice Blast Epidemics in Korea

To predict rice blast, many machine learning methods have been proposed. As the quality and quantity of input data are essential for machine learning techniques, this study develops three artificial neural network (ANN)-based rice blast prediction models by combining two ANN models, the feed-forward neural network (FFNN) and long short-term memory (LSTM), with diverse input datasets, and compares their performance. The Blast_Weather_FFNN model had the highest recall score (66.3%) for rice blast prediction. This model requires two types of input data: blast occurrence data for the last 3 years and weather data (daily maximum temperature, relative humidity, and precipitation) between January and July of the prediction year. This study showed that the performance of an ANN-based disease prediction model was improved by applying suitable machine learning techniques together with the optimization of hyperparameter tuning involving input data. Moreover, we highlight the importance of the systematic collection of long-term disease data.     

Spatio-temporal air quality analysis and PM2.5 prediction over Hyderabad City,India using artificial intelligence techniques

Air pollution is one of the most serious environmental issues faced by humans, and it affects the quality of life in cities. PM_2.5 forecasting models can be used to create strategies for assessing and warning the public about anticipated harmful levels of air pollution. Accurate pollutant concentration measurements and forecasting are critical criteria for assessing air quality and are the foundation for making the right strategic decisions. Data-driven machine learning models for PM_2.5 forecasting have gained attention in the recent past. In this study, PM_2.5 prediction for Hyderabad city was carried out using various machine learning models viz. Multi-Linear Regression (MLR), decision tree (DT), K-Nearest Neighbors (KNN), Random Forest (RF), and XGBoost. A deep learning model, the Long Short-Term Memory (LSTM) model, was also used in this study. The results obtained were finally compared based on error and R² value. The best model was selected based on its maximum R² value and minimal error. The model's performance was further improved using the randomized search CV hyperparameter optimization technique. Spatio-temporal air quality analysis was initially conducted, and it was found that the average winter PM_2.5 concentrations were 68% higher than the concentrations in summer. The analysis revealed that XGBoost regression was the best-performing machine learning model with an R² value of 0.82 and a Mean Absolute Error (MAE) of 7.01 μg/ m³, whereas the LSTM deep learning model performed better than XGBoost regression for PM_2.5 modeling with an R² value of 0.89 and an MAE of 5.78 μg/ m³.     

Estimation of the long-term nutrient budget and thresholds of regime shift for a large shallow lake in China

In this study, we apply an integrated empirical and mechanism approach to estimate a comprehensive long-term (1953–2012) total nitrogen (TN) and total phosphorus (TP) loading budget for the eutrophic Lake Chaohu in China. This budget is subsequently validated, firstly, by comparing with the available measured data in several years, and secondly, by model simulations for long-term nutrient dynamics using both Vollenweider (VW) model and dynamic nonlinear (DyN) model. Results show that the estimated nutrient budget is applicable for further evaluations. Surprisingly, nutrient loading from non-point sources (85% for TN and 77% for TP on average) is higher than expectation, suggesting the importance of nutrient flux from the soil in the basin. In addition, DyN model performs relatively better than VW model, which is attributed to both the additional sediment recycling process and the parameters adjusted by the Bayesian-based Markov Chain Monte Carlo (MCMC) method. DyN model further shows that the TP loading thresholds from the clear to turbid state (631.8 ± 290.16 t y⁻¹) and from the turbid to clear state (546.0 ± 319.80 t y⁻¹) are significantly different (p < 0.01). Nevertheless, the uncertainty ranges of the thresholds are largely overlapped, which is consistent with the results that the eutrophication of Lake Chaohu is more likely to be reversible (74.12%) than hysteretic (25.53%). The ecosystem of Lake Chaohu shifted from the clear to turbid state during late 1970s. For managers, approximately two-thirds of the current TP loading must be reduced for a shift back with substantial improvement in water quality. Because in practice the reduction of loading from a non-point source is very difficult and costly, additional methods beyond nutrient reduction, such as water level regulation, should be considered for the lake restoration.     

Can machines learn the mutation signatures of SARS-CoV-2 and enable viral-genotype guided predictive prognosis?

MotivationContinuous emergence of new variants through appearance/accumulation/disappearance of mutations is a hallmark of many viral diseases. SARS-CoV-2 variants have particularly exerted tremendous pressure on global healthcare system owing to their life threatening and debilitating implications. The sheer plurality of variants and huge scale of genomic data have added to the challenges of tracing the mutations/variants and their relationship to infection severity (if any).ResultsWe explored the suitability of virus-genotype guided machine-learning in infection prognosis and identification of features/mutations-of-interest. Total 199,519 outcome-traced genomes, representing 45,625 nucleotide-mutations, were employed. Among these, post data-cleaning, Low and High severity genomes were classified using an integrated model (employing virus genotype, epitopic-influence and patient-age) with consistently high ROC-AUC (Asia:0.97 ± 0.01, Europe:0.94 ± 0.01, N.America:0.92 ± 0.02, Africa:0.94 ± 0.07, S.America:0.93 ± 03). Although virus-genotype alone could enable high predictivity (0.97 ± 0.01, 0.89 ± 0.02, 0.86 ± 0.04, 0.95 ± 0.06, 0.9 ± 0.04), the performance was not found to be consistent and the models for a few geographies displayed significant improvement in predictivity when the influence of age and/or epitope was incorporated with virus-genotype (Wilcoxon p_BH < 0.05). Neither age or epitopic-influence or clade information could out-perform the integrated features. A sparse model (6 features), developed using patient-age and epitopic-influence of the mutations, performed reasonably well (>0.87 ± 0.03, 0.91 ± 0.01, 0.87 ± 0.03, 0.84 ± 0.08, 0.89 ± 0.05). High-performance models were employed for inferring the important mutations-of-interest using Shapley Additive exPlanations (SHAP). The changes in HLA interactions of the mutated epitopes of reference SARS-CoV-2 were then subsequently probed. Notably, we also describe the significance of a ‘temporal-modeling approach’ to benchmark the models linked with continuously evolving pathogens. We conclude that while machine learning can play a vital role in identifying relevant mutations and factors driving the severity, caution should be exercised in using the genotypic signatures for predictive prognosis.     

Systems Metabolic Engineering Meets Machine Learning: A New Era for Data‐Driven Metabolic Engineering

The recent increase in high‐throughput capacity of ‘omics datasets combined with advances and interest in machine learning (ML) have created great opportunities for systems metabolic engineering. In this regard, data‐driven modeling methods have become increasingly valuable to metabolic strain design. In this review, the nature of ‘omics is discussed and a broad introduction to the ML algorithms combining these datasets into predictive models of metabolism and metabolic rewiring is provided. Next, this review highlights recent work in the literature that utilizes such data‐driven methods to inform various metabolic engineering efforts for different classes of application including product maximization, understanding and profiling phenotypes, de novo metabolic pathway design, and creation of robust system‐scale models for biotechnology. Overall, this review aims to highlight the potential and promise of using ML algorithms with metabolic engineering and systems biology related datasets.     

Efficiency of nutrient management in controlling eutrophication of running waters in the Middle Danube Basin

Nutrient emission dropped significantly during the last two decades in the Danube Basin. To assess the effect of reduced nutrient loads on the trophic status of running waters, this regional study analyzed the relationships between nutrients (P and N) and suspended chlorophyll (Chl) using long-term monitoring data in Hungary. Including the upstream catchments of trans-boundary rivers, the study covered an approximate area of 400,000 km², equivalent to the half of the entire Danube catchment. Decadal median Chl was unrelated to P and N concentrations in the whole data set and weakly related to total P (TP) at natural-moderately polluted (N-MP) sites, which were distinguished from highly polluted (HP) sites by using cutoff values for chloride, chemical oxygen demand and TP. At both the N-MP sites and most of the HP sites, Chl increased with channel length. This indicated that water residence time was a more important determinant of Chl than nutrients. Nutrient concentrations showed a significant downward trend in time at half of our sites. With a nearly equal frequency, a parallel trend might or might not occur in Chl. The apparent efficiency of nutrient management was expressed as the quotient of the slopes of linear trends in Chl and nutrients. At sites within 150 km from source, this efficiency was marginal. In larger rivers, efficiency improved steeply. The highest efficiency was observed in the downstream reach of the Danube (upstream length >1,300 km) where P availability might frequently limit algal growth. The results suggest that eutrophication management in rivers should be based on Chl response functions, rather than universal nutrient criteria. Four Chl response classes were identified based on the observed longitudinal P and Chl gradients.     

The periphyton index of trophic status PIT: a new eutrophication metric based on non-diatomaceous benthic algae in Nordic rivers

Eutrophication is one of the major problems for surface water quality in Norway, particularly in the lowlands near settlements and agricultural areas. Here, we present a new index based on non-diatomaceous benthic algae (Periphyton index of trophic status, PIT) which is developed on a dataset of >500 samples from >350 sites from the Norwegian mainland and can be used to describe trophic status at a river site. PIT indicator values for benthic algae taxa are derived from water total phosphorus concentrations and range from 1.87 for Stigonema hormoides to 68.91 for Tribonema sp. PIT site values range from 3.42 to 44.45 and cover a range from oligotrophic to eutrophic conditions. The relationship between the PIT and the total phosphorus concentration has one major threshold at 10 μg/l TP, with a slow increase below and a steep increase above 10 μg/l. We conclude that benthic algae species composition at nutrient poor sites reacts only slightly to small increases in phosphorus concentration, while it is most sensible to eutrophication in the range between 10 and 30 μg TP/l. For the genus Oedogonium, we found a significant positive correlation between filament width and TP concentration, making Oedogonium an easy to use eutrophication indicator.     

Quantifying species indicator values for trophic diatom indices: a comparison of approaches

This study compares two approaches for constructing diatom-based indices for monitoring river eutrophication. The first approach is based on weighted averaging of species indicator values with the underlying assumption that species have symmetrical unimodal distributions along the nutrient gradient, and their distributions are sufficiently described by a single indicator value per species. The second approach uses multiple indicator values for individual taxa and is based on the possibility that species have complex asymmetrical response curves. Multiple indicator values represent relative probabilities that a species would be found within certain ranges of nutrient concentration. We used 155 benthic diatom samples collected from rivers in the Northern Piedmont ecoregion (Northeastern U.S.A.) to construct two datasets: one used for developing models and indices, and another for testing them. To characterize the shape of species response curves we analyzed changes in the relative abundance of 118 diatom taxa common in this dataset along the total phosphorus (TP) gradient by fitting parametric and non-parametric regression models. We found that only 34 diatoms had symmetrical unimodal response to TP. Among several indices that use a single indicator value for each species, the best was the weighted averaging partial least square (WA-PLS) inference model. The correlation coefficient between observed and inferred TP in the test dataset was 0.67. The best index that employed multiple indicator values for each species had approximately the same predictive power as the WA-PLS based index, but in addition, this index provided a sample-specific measure of uncertainty for the TP estimation.     

Characteristics of a landscape water with high salinity in a coastal city of China and measures for eutrophication control

Eutrophication of landscape waters is drawing public concerns in China but few studies have been conducted on the problem associated with high water salinity as what happens at Sino-Singapore Tianjin Eco-city in Tianjin, a coastal metropolis of northern China. In order to find ways for eutrophication control, a comparative study was conducted on three landscape water bodies, namely Qingjing Lake, Jiyun River and Jiyun River Oxbow, which are under varied conditions of salinity, organic, and nutrients intrusion. The spatial and temporal variations of water quality were revealed by water sampling and analyses, and correlative relationships were obtained between water salinity and other parameters related to eutrophication. By utilizing a trophic level index (TLI), the eutrophication status of the three landscape water bodies in different seasons could further be evaluated. As a result, water temperature, as expected, showed the strongest effect on eutrophication because higher TLI together with higher Chl-a concentrations tended to occur in later spring and summer seasons, while nutrient concentration, especially TP, was also the determinative factor to the eutrophication status. Of the three water bodies, the Jiyun River Oxbow showed a salinity as high as 20 g/L or more in contrast with the other two water bodies with salinity as 4–5 g/L. Although its TP concentration was usually very low (about 0.1 mg/L), it was under a moderate eutrophication status almost in all seasons, indicating that high salinity tends to induce alga growth. Dilution of saline inflow and nutrients reduction could thus be proposed as the main measures for eutrophication control of landscape waters in the study area.