As one of the earth's key ecosystems, rivers have been intensively studied and modelled through the application of machine learning (ML). With the amount of large data available, these computer ...algorithms are ever increasing in numerous fields, although there is ongoing scepticism and scholars still question the actual impact and deliverables of algorithms.
This study aims to provide a systematic review of the state‐of‐the‐art ML‐based techniques, trends, opportunities and challenges in river research by applying text mining and automated content analysis.
Unsupervised and supervised learning have dominated river research while neural networks and deep learning have also gradually gained popularity. Matrix factorisation and linear models have been the most popular ML algorithms, with around 1300 and 800 publications on these topics in 2020 respectively. In contrast, river researchers have had few applications in multiclass and multilabel algorithm, associate rule and Naïve Bayes.
The current article proposes an end‐to‐end workflow of ML applications in river research in order to tackle major ML challenges, including four steps: (1) data collection and preparation; (2) model evaluation and selection; (3) model application; and (4) feedback loops. Within this workflow, river modellers have to balance numerous trade‐offs related to model traits, such as complexity, accuracy, interpretability, bias, data privacy and accessibility and spatial and temporal scales. Any choices made when balancing the trade‐offs can lead to different model outcomes affecting the final applications. Hence, it is necessary to carefully consider and specify modelling goals, understand the data collected and maintain feedback loops in order to continuously improve model performance and eventually reach the research objectives. Moreover, it remains crucial to address the users' needs and demands that often entail additional elements, such as computational cost, development time and the quantity, quality and compatibility of data. Furthermore, river researchers should account for new technologies and regulations in data collection and protection that are transforming the development and applications of ML, most notably data warehouse and information management with multiple‐cycles that are becoming a cornerstone of the integration of ML in decision‐making in river and ecosystem management.
Despite covering a small portion of the earth’s surface, lakes and reservoirs offer enormous benefits to human society, environmental well-being, and economic welfare. Previous studies have provided ...insights into specific subjects, yet integrated perspectives on the development of the two waterbodies are missing. To this end, we conducted a bibliometric analysis as a systematic data gathering to perform a large-scale overview and assess global trends of their scientific publications. Moreover, a second goal is to differentiate their research hotspots and current challenges given the different nature of their origin and functionality. 147,811 publications from 1955 to 2019 were retrieved from the database of the Science Citation Index Expanded, and then, divided into four research lines, (1) design and operation; (2) environment and ecology; (3) sanitation and human health; (4) socioeconomics. Bibliometric indicators showed that the number of publications sustained a rapid growth, from 100 during the 1950s to around 7800 publications per year during the past few years. The United States and EU 28 have long been world leaders in lake and reservoir research yet China has tremendously boosted its publications within the past 20 years, advancing this nation to the new world leader in both categories in 2019. Taking a closer look at research hotspots, design and operation have been the main topics for reservoir research while environment and ecology topics are the hotspots in lakes-related studies. This reflected the intensive human interventions in reservoirs, whose major purposes are to supply hydropower energy, irrigation, water storage, and aquaculture. Conversely, the impacts of eutrophication, heavy metals, and climate change have become more severe with the increase of species extinction and biodiversity loss, leading to urgent needs for lake restoration. Both freshwater bodies show comparable attention on their roles in socioeconomics while much higher concerns about sanitation and human health have been paid in reservoirs compared to its counterpart. Clear obtained distinctions in the hotspots and challenges of lake and reservoir research can contribute to better decision support systems of the two waterbodies.
Display omitted
•Introduce Learned Binary Masks (LBM) to interpret an RNN’s ICU mortality predictions.•Attribute individual RNN predictions to their input features using LBM & KernelSHAP.•Aggregate ...attributions from KernelSHAP and LBM to interpret the RNN at various scales.•Introduce a patient data representation that facilitates use of LBM and KernelSHAP.
Deep learning has demonstrated success in many applications; however, their use in healthcare has been limited due to the lack of transparency into how they generate predictions. Algorithms such as Recurrent Neural Networks (RNNs) when applied to Electronic Medical Records (EMR) introduce additional barriers to transparency because of the sequential processing of the RNN and the multi-modal nature of EMR data. This work seeks to improve transparency by: 1) introducing Learned Binary Masks (LBM) as a method for identifying which EMR variables contributed to an RNN model’s risk of mortality (ROM) predictions for critically ill children; and 2) applying KernelSHAP for the same purpose. Given an individual patient, LBM and KernelSHAP both generate an attribution matrix that shows the contribution of each input feature to the RNN’s sequence of predictions for that patient. Attribution matrices can be aggregated in many ways to facilitate different levels of analysis of the RNN model and its predictions. Presented are three methods of aggregations and analyses: 1) over volatile time periods within individual patient predictions, 2) over populations of ICU patients sharing specific diagnoses, and 3) across the general population of critically ill children.
Though aquaculture plays an important role in providing foods and healthy diets, there are concerns regarding the environmental sustainability of prevailing practices. This study examines the trends ...and changes in fisheries originating from aquaculture production in Thailand and provides insights into such production’s environmental impacts and sustainability. Together with an extensive literature review, we investigated a time series of Thai aquaculture production data from 1995 to 2015. Overall, Thai aquaculture production has significantly increased during the last few decades and significantly contributed to socio-economic development. Estimates of total aquaculture production in Thailand have gradually grown from around 0.6 to 0.9 million tons over the last twenty years. Farmed shrimp is the main animal aquatic product, accounting for an estimated 40% of total yields of aquaculture production, closely followed by fish (38%) and mollusk (22%). Estimates over the past decades indicate that around 199470 ha of land is used for aquaculture farming. Out of the total area, 61% is used for freshwater farms, and 39% is used for coastal farms. However, this industry has contributed to environmental degradation, such as habitat destruction, water pollution, and ecological effects. Effective management strategies are urgently needed to minimize the environmental impacts of aquaculture and to ensure it maximally contributes to planetary health. Innovative and practical solutions that rely on diverse technology inputs and smart market-based management approaches that are designed for environmentally friendly aquaculture farming can be the basis for viable long-term solutions for the future.
A better design instruction for waste stabilization ponds is needed due to their growing application for wastewater purification, increasingly strict environmental regulations, and the fact that most ...of previous design manuals are outdated. To critically review model-based designs of typical pond treatment systems, this paper analyzed more than 150 articles, books, and reports from 1956 to 2016. The models developed in these publications ranged from simple rules and equations to more complex first-order and mechanistic models. From a case study on all four approaches, it appeared that rules of thumb is no longer a proper tool for pond designs due to its low design specification and very high output variability and uncertainty. On the other hand, at the beginning phase of design process or in case of low pressure over land and moderate water quality required, regression equations can be useful to form an idea for pond dimensions. More importantly, mechanistic models proved their capacity of generating more precise and comprehensive designs but still need to overcome their lack of calibration and validation, and overparameterization. In another case study, an essential but often overlooked role of uncertainty analysis in pond designs was investigated via a comparison between deterministic and uncertainty-based approaches. Unlike applying a safety factor representing all uncertainty sources, probabilistic designs quantify the uncertainty of model outputs by including prior uncertainty of inputs and parameters, which generates more scientifically reliable outcomes for decision makers. Based on these findings, we advise engineers and designers to shift from the conventional approaches to more innovative and economic tools which are suitable for dealing with large variations of natural biological systems.
Display omitted
•More than 150 publications related to pond designs from 1956 to 2016 were analyzed.•It is difficult for data-driven models to be extrapolated to future designs.•It is advised to apply process-driven models for more precise pond designs.•Deterministic designs frequently produced over-sized systems.•For better cost effective designs, we advised to apply probabilistic approaches.
This special issue consists of fourteen selected articles, that cover a wide spectrum of Ecosystem Services (ES) of lakes and reservoirs, including: (1) water purification ...
Emerging global threats, such as biological invasions, climate change, land use intensification, and water depletion, endanger the sustainable future of lakes and reservoirs. To deal with these ...threats, a multidimensional view on the protection and exploitation of lakes and reservoirs is needed. The holistic approach needs to contain not just the development of economy and society but also take into account the negative impacts of this growth on the environment, from that, the balance between the three dimensions can be sustained to reach a sustainable future. As such, this paper provides a comprehensive review on future opportunities and challenges for the sustainable development of lakes and reservoirs via a critical analysis on their contribution to individual and subsets of the Sustainable Development Goals (SDGs). Currently, lakes and reservoirs are key freshwater resources. They play crucial roles in human societies for drinking water provision, food production (via fisheries, aquaculture, and the irrigation of agricultural lands), recreation, energy provision (via hydropower dams), wastewater treatment, and flood and drought control. Because of the (mostly) recent intensive exploitations, many lakes and reservoirs are severely deteriorated. In recent years, physical (habitat) degradation has become very important while eutrophication remains the main issue for many lakes and ponds worldwide. Besides constant threats from anthropogenic activities, such as urbanization, industry, aquaculture, and watercourse alterations, climate change and emerging contaminants, such as microplastics and antimicrobial resistance, can generate a global problem for the sustainability of lakes and reservoirs. In relation to the SDGs, the actions for achieving the sustainability of lakes and reservoirs have positive links with the SDGs related to environmental dimensions (Goals 6, 13, 14, and 15) as they are mutually reinforcing each other. On the other hand, these actions have direct potential conflicts with the SDGs related to social and economic dimensions (Goals 1, 2, 3 and 8). From these interlinkages, we propose 22 indicators that can be used by decision makers for monitoring and assessing the sustainable development of lakes and reservoirs.
Video monitoring is a rapidly evolving tool in aquatic ecological research because of its non-destructive ability to assess fish assemblages. Nevertheless, methodological considerations of video ...monitoring techniques are often overlooked, especially in more complex sampling designs, causing inefficient data collection, processing, and interpretation. In this study, we discuss how video transect sampling designs could be assessed and how the inter-observer variability, design errors and sampling variability should be quantified and accounted for. The study took place in the coastal areas of the Galapagos archipelago and consisted of a hierarchical repeated-observations sampling design with multiple observers. Although observer bias was negligible for the assessment of fish assemblage structure, diversity and counts of individual species, sampling variability caused by simple counting/detection errors, observer effects and instantaneous fish displacement was often important. Especially for the counts of individual species, sampling variability most often exceeded the variability of the transects and sites. An extensive part of the variability in the fish assemblage structure was explained by the different transects (13%), suggesting that a sufficiently high number of transects is required to account for the within-location variability. Longer transect lengths allowed a better representation of the fish assemblages as sampling variability decreased by 33% if transect length was increased from 10 to 50 meters. However, to increase precision, including more repeats was typically more efficient than using longer transect lengths. The results confirm the suitability of the technique to study reef fish assemblages, but also highlight the importance of a sound methodological assessment since different biological responses and sampling designs are associated with different levels of sampling variability, precision and ecological relevance. Therefore, besides the direct usefulness of the results, the procedures to establish them may be just as valuable for researchers aiming to optimize their own sampling technique and design.
Imitating natural lakes, pond treatment systems inherit a high complexity with interconnected web of biochemical reactions and complex hydraulic processes. As such, its simulation requires a large ...and integrated model, which has been a challenge for pond engineers. In this study, we develop an all-encompassing model to gain a quantitative and comprehensive understanding of the hydraulic, physicochemical and microbiological conversion processes in the most common pond, a facultative pond. Moreover, to deal with an evitable issue of large mechanistic models as overparameterization leading to poor identifiability, a systematic parameter estimation was implemented. The application of sensitivity analysis reveals the most influential parameters on pond performance. Particularly, physical parameters, such as vertical eddy diffusivity, water temperature, and maximum growth rate of heterotrophs induce the most changes of organic matters while microbial assimilation and ammonia volatilization appear to be main processes for nutrient removal. In contrast, the efficiency of phosphate precipitation and nutrient biological removal via polyphosphate accumulating organisms and denitrifying bacteria is limited. Identifiability problems are addressed mainly by the characterization of light dependence of algal growth, interaction between water temperature and its coefficient, and the growth of autotrophic bacteria while based on the determinant measures, the most important parameter subsets affecting model outputs are related to physical processes and algal activity. After the establishment of the influential and identifiable parameter subset, an automatic calibration with the data collected from Ucubamba pond system (Ecuador) demonstrates the effect of high-altitude climatic conditions on pond behaviors. An aerobic prevailing condition is observed as a result of high light intensity causing accelerated algal activities, hence, leading to the limitation of hydrolysis, anaerobic processes, and the growth of anoxic heterotrophs for denitrification. Furthermore, the output of uncertainty analysis indicates that a large avoidable uncertainty as a result of vast complexity of the applied model can be reduced greatly via a systematic approach for parameter estimation.
Display omitted
•An all-encompassing mechanistic model of a facultative pond is developed.•A sensitivity analysis reveals the most influential parameters on pond behavior.•Poor identifiability is caused mainly by the depiction of algal light dependence.•Model calibration exposes the distinctive characteristics of high-altitude ponds.•Avoidable uncertainty can be reduced via a systematic parameter estimation.