Search

We assessed 6122 environmental modelling papers published since 2005 to determine whether the number of citations each paper had received by September 2014 could be predicted with no knowledge of the paper's quality. A random forest was applied, using a range of easily quantified or classified variables as predictors. The 511 papers published in two key journals in 2008 were further analysed to consider additional variables. Papers with no differential equations received more citations. The topic of the paper, number of authors and publication venue were also significant. Ten other factors, some of which have been found significant in other studies, were also considered, but most added little to the predictive power of the models. Collectively, all factors predicted 16–29% of the variation in citation counts, with the remaining variance (the majority) presumably attributable to important subjective factors such as paper quality, clarity and timeliness. [ABSTRACT FROM AUTHOR]

Sensitivity analysis (SA) has been used to evaluate the behavior and quality of environmental models by estimating the contributions of potential uncertainty sources to quantities of interest (QoI) in the model output. Although there is an increasing literature on applying SA in environmental modeling, a pragmatic and specific framework for spatially distributed environmental models (SD-EMs) is lacking and remains a challenge. This article reviews the SA literature for the purposes of providing a step-by-step pragmatic framework to guide SA, with an emphasis on addressing potential uncertainty sources related to spatial datasets and the consequent impact on model predictive uncertainty in SD-EMs. The framework includes: identifying potential uncertainty sources; selecting appropriate SA methods and QoI in prediction according to SA purposes and SD-EM properties; propagating perturbations of the selected potential uncertainty sources by considering the spatial structure; and verifying the SA measures based on post-processing. The proposed framework was applied to a SWAT (Soil and Water Assessment Tool) application to demonstrate the sensitivities of the selected QoI to spatial inputs, including both raster and vector datasets - for example, DEM and meteorological information - and SWAT (sub)model parameters. The framework should benefit SA users not only in environmental modeling areas but in other modeling domains such as those embraced by geographical information system communities. • A pragmatic framework of sensitivity analyses is provided for spatially distributed environmental models. • The framework prescribes sequential steps in which important considerations are highlighted. • The framework benefits users of sensitivity analyses in environmental modeling and GIS communities. [ABSTRACT FROM AUTHOR]

Hydrodynamic models with rain-on-the-grid capabilities are usually computationally expensive for automatic parameter estimation. In this paper, we present a global optimization-based algorithm to calibrate a fully distributed hydrologic-hydrodynamic and water quality model (HydroPol2D) using observed data (i.e., discharge, or pollutant concentration) as input. The algorithm finds near-optimal set of parameters to explain observed gauged data. This framework, although applied in a poorly-gauged urban catchment, is adapted for catchments with more detailed observations. The results of the automatic calibration indicate NSE = 0.99 for the V-Tilted catchment, RMSE = 830 mg L-1 for salt concentration pollutograph in a wooden-plane (i.e., 8.3% of the event mean concentration), and NSE = 0.89 in a urban real-world catchment. This paper also explores the issue of equifinality (i.e., multiple parameters giving the same calibration performance) in model calibration indicating the performance variation of calibrating only with an outlet gauge or with multiple gauges within the catchment. [Display omitted] • An automatic calibration algorithm for distributed flood and water quality modeling is developed. • It uses HydroPol2D model and calibrate water quantity and quality parameters globally. • Data from observed gauges such as discharges, depths, and concentration is used for calibration. • Poorly placed gauges and low runoff events can increase equifinality during calibration. [ABSTRACT FROM AUTHOR]

This paper presents improved explicit local time-stepping (LTS) schemes of both first and second order accuracy for storm surge modeling. The two-dimensional shallow water equations are numerically solved on unstructured triangular meshes using finite volume method with Roe's approximate Riemann solver. The LTS algorithms are designed based on explicit Euler and strong stability preserving Runge–Kutta time integration methods. A single-layer interface prediction–correction scheme is adopted to combine coarse and fine time discretization, further enhancing the stability of the LTS schemes, particularly at higher LTS levels and during long time simulations. An ideal numerical test validates the efficiency of the improved LTS models, revealing their capability to improve computational speed while preserving conservation properties and reducing accuracy loss as LTS levels increase. We further apply the LTS models to cross-scale simulations of storm surges in the Northwest Pacific. Results show that compared to the global time-stepping (GTS) models, the LTS models significantly boost computational speed by up to 37%, all while delivering equally reliable computational outcomes. With expanding high-resolution coastal data and the need for high-resolution modeling, the improved LTS models show great potential for cross-scale storm surge modeling. • This paper develops improved explicit first and second-order local time-stepping (LTS) schemes for two-dimensional shallow water equations. • The improved LTS schemes enhance computational efficiency while ensuring stability at higher LTS levels and during extended simulations, surpassing original LTS methods. • The LTS models achieve up to 37% efficiency gain in Northwest Pacific surge simulations, revealing their broad application potential. [ABSTRACT FROM AUTHOR]

This paper presents a comprehensive Virtual Reality (VR) based framework for visualizing numerical simulations of geophysical flows in a realistic and immersive manner. The framework allows integrating output data from various mesh-based Eulerian numerical models into a VR environment, enabling users to interact with and explore the terrain and geophysical flows through the VR experience. Three case studies, including a snow avalanche, quick clay landslide, and flash flood in Norway, demonstrate its versatility. The VR environment offers intuitive menus and user interactions, allowing users to read flow depth and velocity values, facilitating a direct link between numerical data and their visual representation. This framework can reshape geophysical flow hazard identification and disaster management by integrating physics-based numerical modeling results into VR Environments, thus enhancing knowledge dissemination among experts, the general public, non-expert stakeholders, and policymakers. The paper also highlights challenges and opportunities identified during the integration, guiding future developments. • A framework that integrates numerical simulations of geophysical flows into VR. • The framework is versatile across different geophysical flow types and numerical models. • Real time visualization of numerical simulation results is still a challenge. • The framework has a potential for disaster preparedness and emergency training. [ABSTRACT FROM AUTHOR]

With the development of geographic simulation methods in recent decades, a great deal of resources have accumulated to support their implementation. These resources can be divided into model resources for analyzing or predicting geographic phenomena or processes, data resources for representing the characteristics of real or simulated environment, and computing resources for supporting simulation tasks. These resources are characterized by geospatial distribution and are difficult to discover and reuse. OpenGMS has carried out a series of fundamental research to sharing and collaborating distributed resources on the web. On this basis, this paper presents the concept of the OpenGMS open portal. The portal adopts FAIR principles and supports resources sharing and reuse to facilitate collaboration and exchange between resource contributors and users. This paper takes applications of the portal in resource sharing and reuse case and online training courses as examples to illustrate how the portal can bridge contributors and users of resources. • The portal adopts FAIR principles and is integrated into OpenGMS platform. • The functional design for resource contributors and users is presented. • The portal attracts many users and amasses lots of geographic simulation resources. • The portal assists in model sharing and reuse among researchers in cyberspace. [ABSTRACT FROM AUTHOR]

Abstract: Treatment of industrial wastewaters is currently confronting important challenges concerning both cost management of treatment plants and fulfillment of tightening environmental regulations. Online monitoring of wastewater treatment is critical, because changes in the performance of treatment can lead to various problems such as decreased efficiency of purification, decreased energy efficiency, or ineffective use of chemicals. Moreover, changes in the operation of a treatment process can inflict changes that have unforeseen consequences, including an increased amount of harmful effluents, and therefore it is essential for a monitoring system to be able to adapt to various process conditions. It seems, however, that the monitoring systems used currently by the industry are lacking this functionality and are therefore only partially able to meet the needs of modern industry. In addition, there is typically a large amount of measurement data available in the industry, for which advanced data processing and computational tools are needed for monitoring, analysis, and control. For this reason, it would be useful to have a monitoring system which could be able to handle a large amount of measurement data and present the essential information on the state and evolution of the process in a simple, user-friendly and flexible manner. In this paper, we introduce an adaptive multivariable approach based on self-organizing maps (SOM) which can be utilized for advanced monitoring of industrial processes. The system developed can provide a new kind of tool for illustrating the condition and evolution of an industrial wastewater treatment process. The operation of the system is demonstrated using process measurements from an activated sludge treatment plant, which is a part of a pulp and paper plant. [Copyright &y& Elsevier]

In view of upcoming more stringent air quality limits and the ambition to align with the WHO guidelines, nitrogen dioxide (NO 2) pollution from traffic and other sources will remain a problem in the EU. To assess the impact of traffic measures and emission reductions in other sectors on NO 2 -concentrations, an EU-wide high-resolution NO 2 source apportionment web-application was developed. The application allows users to define scenarios in a user-friendly way and quickly visualize the NO 2 -concentrations at measurement stations and in cities. The user can configure a new Euro 7/VII emission standard and additionally define urban access regulations scenarios in cities. To capture the spatial scales of NO 2 pollution, the SHERPA source-receptor model was used in combination with the QUARK kernel dispersion model. The first model considers long-distance impacts, the latter considers the strong concentration gradients close to roads. This paper focuses on the methodology, a follow-up paper describes the web-application. • We present a method for fast sectoral and spatial NO 2 source apportionment. • The road transport sector is considered in high detail and at 100-m resolution. • The impact of a new vehicle emission standard for NO x can be simulated. • The effect of urban access regulations can be simulated in 948 European cities. • NO 2 concentration effects are visualized at measurement stations and over cities. [ABSTRACT FROM AUTHOR]

Geographic modeling is considered as an effective way to solve geographic problems. Geographic conceptual modeling is the first step of geographic modeling, in which modelers can fully exchange modeling ideas, thereby achieving a common understanding of geographic system and guiding subsequent geographic modeling process. However, due to the diverse research backgrounds of modelers, it is challenging for them to exchange modeling ideas with one another. This paper provides a visual cognitive communication method for modelers by constructing a conceptual graph gallery, and then improves the efficiency of geographic conceptual modeling. A multidimensional description method that describes modeling cognition from multiple perspectives, the conceptual graph gallery that includes concept items and conceptual graphs, and the conceptual graph gallery-based cognitive communication methods are designed in this paper. Finally, a case study involving the construction of a hydrological conceptual graph gallery is designed to illustrate the feasibility of conceptual graph gallery-based cognitive communication. • Cognitive communication is a key component in geographic conceptual modeling. • Geographical concepts can be described from different perspectives. • Concept graphs can describe geographic concepts visually. • Conceptual graph gallery can support cognitive communication. [ABSTRACT FROM AUTHOR]

This paper describes the DAnish Lagrangian Model (DALM), which is a new high-resolution air pollution model based on the concept of Lagrangian particles. The new model is developed from the simpler Gaussian plume-in-grid Urban Background Model (UBM) and is to be integrated with the DEHM/UBM/AirGIS modeling system, developed at Aarhus University. In the first part of the paper, the theoretical foundation of the model is presented, and the implementation of a large set of physical and numerical parameterizations is discussed. The second part describes the validation of DALM against measurements, applying different combinations of the implemented parameterizations. This validation demonstrates that DALM can accurately reproduce spatiotemporal patterns in the measured data and that its performance is more sensitive to parameterizations of vertical compared to horizontal transport. Conclusively, the combination of parameterizations yielding the best model performance is determined based on a ranking system, and future improvements to DALM are outlined. [Display omitted] • A new comprehensive local-scale Lagrangian air pollution model is in development. • The new model is validated against measurements from Danish monitoring stations. • Some sets of planetary boundary layer parameterizations show a better performance. [ABSTRACT FROM AUTHOR]

Groundwater level (GWL) prediction is important for ecological protection and resource utilization; it helps in formulating policies for artificial groundwater recharge, modifying the number of extraction wells, etc., and can support sustainable human development as well as inform water resource management decisions. However, climate change, anthropogenic impacts, and the complex coupling between surface water and groundwater increase the difficulty of predicting groundwater levels. The model proposed in this paper combines external data as well as multiple models. The method leverages long and short-term memory (LSTM) and convolutional neural network (CNN) models, combined with secondary modal decomposition and slime mould algorithm (SMA), together with an adaptive weight module (AWM). The study applies this method to predict GWL for three different hydrological conditions in China, specifically for the Jinan Baotu Spring, Heihu Spring, and Zhongtianshe watershed of Taihu Lake. A comparison of metrics such as mean absolute error and Nash efficiency coefficient for single and hybrid models shows that the model in this paper is more advantageous than the single model and other hybrid models. The interpretability of the model is enhanced by SHAP values that demonstrate the degree of contribution of the input variables. This paper uses SHAP analyses to identify the key drivers affecting groundwater levels. These factors must be detected in order to develop groundwater resource protection measures. [Display omitted] • Multivariate fusion data including hydrology and meteorology are used as model input. • A secondary modal decomposition module for historical groundwater level data was utilized. • The neural network hyperparameters are optimized using the slime mould algorithm. • Aggregate subsets of prediction with adaptive modules rather than linear summation. • The interpretable SHAP model measures the degree of influence of external variables. [ABSTRACT FROM AUTHOR]

Reliable probabilistic hydrological prediction requires appropriate handling of residual errors, which can pose considerable complexity. This paper proposes a nonparametric residual error (NRE) model that effectively captures the statistical characteristics of raw residuals. The NRE model employs a local linear estimator with a robust bandwidth selector to estimate the regression and conditional volatility functions of raw residuals. Additionally, the AR(1) model and location-mixture Gaussian distribution are used to estimate the temporal correlation structure and innovation distribution. Through two case studies in South-East China, this research demonstrates the superiority of the NRE model over the benchmark Box-Cox transformation approach in terms of prediction reliability, precision, and bias correction capabilities. Simulation experiments further reveal that the NRE model can effectively fit the residual regression function, conditional volatility function, and innovation distribution under varying scenarios. The proposed residual error model is anticipated to promote the adoption of probabilistic predictions in hydrological modeling applications. • Residual Error Handling: This paper tackles complex residual errors in hydrological modeling. • Accuracy Improvement: The NRE model boosts accuracy with a local linear estimator. • Predictive Superiority: The NRE model excels in predicting reliability, precision, and bias correction. • Data-Driven Flexibility: The NRE model reduces reliance on assumptions for hydrological modelers. [ABSTRACT FROM AUTHOR]

The ability of Bayesian networks (BNs) to model complex systems and uncertainties makes it a perfect tool for the research on subfields related to climate change. In fact, in the past 30 years, BNs have been widely used in this field, with 1502 articles in total. Quantitatively understanding influential researchers, institutions, mainstream topics and research trends will help us quickly go deeper into this field. Thus, a scientometric method was conducted. In this paper, we identified the influential authors, journals, countries, institutions, topics and disciplines, key articles and research trends by collaboration network analysis, keyword co-occurrence network and document co-citation network analysis. As a result, we found that environmental sciences technology and water resources were the most popular research subfields, followed by energy fuels and meteorological atmospheric sciences. While as time goes on, research focuses have gradually shifted. Public environmental occupational health will become one of the most popular research subfields in the future. • We review papers of Bayesian networks in the subfields of climate change. • Influential journals, institutions, landmark papers, citation bursts are identified. • Recent topics focus on environmental sciences technology. • Public environmental occupational health will become new hot issues. [ABSTRACT FROM AUTHOR]

Computational models for understanding and predicting fire in wildland and managed lands are increasing in impact. Data characterizing the fuels and environment is needed to continue improvement in the fidelity and reliability of fire outcomes. This paper addresses a gap in the characterization and population of mid-story fuels, which are not easily observable either through traditional survey, where data collection is time consuming, or with remote sensing, where the mid-story is typically obscured by forest canopy. We present a methodology to address populating a mid-story using a generative model for fuel placement that captures key concepts of spatial density and heterogeneity that varies by regional or local environmental conditions. The advantage of using a parameterized generative model is the ability to calibrate (or 'tune') the generated fuels based on comparison to limited observation datasets or with expert guidance, and we show how this generative model can balance information from these sources to capture the essential characteristics of the wildland fuels environment. In this paper we emphasize the connection of terrestrial LiDAR (TLS) as the observations used to calibrate of the generative model, as TLS is a promising method for supporting forest fuels assessment. Code for the methods in this paper is available. • A spatial model generates representative mid-story fuels for wildland fire simulation. • Model parameters are learned to match observed spatial heterogeneity and density. • Terrestrial LiDAR plot observations are used for characterization. • Practical aspects of calibrating stochastic models to limited data are presented. • Spatial covariates inform fuel placement for an application dataset. [ABSTRACT FROM AUTHOR]

This paper presents a dynamical neural network framework to understand how catchment systems respond to daily rainfall-runoff processes over time. We developed an interactive Python-based deep neural network (DNN) package called FlowDyn (presented through a JS-based web platform) to simulate and forecast daily streamflow data for >180 gauging stations across the globe. Several DNN models, including long short-term memory (LSTM), Gated Recurrent Unit (GRU), and a hybrid network of convolutional neural network and LSTM (ConvLSTM), as well as an auto encoder (AE) model were developed and integrated into the FlowDyn pipeline to analyze and forecast sequential daily streamflow values that are embedded within a web-based application for demonstration and visualization. Inputs were gathered from different web services, including the catchment attributes and meteorology for large-sample studies (CAMELS), the national climatic data center (NCDC), and the global runoff data center (GRDC). DNN configurations were trained and tested with an average accuracy rating of 0.83 across 183 river basins globally. FlowDyn simulation and performance demonstrated that different DNN models were able to learn both regionally consistent and location-specific hydrological behaviors. Through the findings of this paper, we advocate the merit of applying FlowDyn package in the field of daily rainfall-runoff prediction at both local and global scales. • A dynamic framework was developed to predict daily streamflow records of >180 gauging stations across the globe. • The FlowDyn pipeline offers functionalities, such as employing multiple DNNs, various performance metrics, and visualization. • FlowDyn has the ability to retrain and provide more accurate results in case of extending the datasets. [ABSTRACT FROM AUTHOR]

Abstract: The Environmental Effects Monitoring (EEM) Statistical Assessment Tool (SAT) Decision Support System (DSS) has been developed to provide a user-friendly data analysis, display and decision support tool for Canada''s federal environmental effects monitoring program for the pulp and paper and mining industries. The target users include industries, consultants, regional EEM coordinators, National EEM Office and scientists involved in EEM-related research. The tool allows the assessment of the effects of effluent from industrial or other sources on fish and benthic populations. Effect endpoints, which are used as indicators of potentially important effluent effects, are measured at effluent-exposed sites and are compared statistically to measures at reference sites, in order to determine if changes have occurred and the magnitude of the changes. The main driver of the EEM-SAT DSS is its rule-based expert system. The results are used in assessing the adequacy of existing regulations for protecting aquatic environments. [Copyright &y& Elsevier]

The most common approach to mitigation of lake eutrophication is reduction of phosphorus emissions, in particular by changing farm management. This reduction can be combined with landscaping retention structures upstream of the lake, the analyses of which the paper is based on. The management of these structures currently focuses on maximising the quantity of phosphorus trapped, regardless of lake dynamics. This paper adapts a dynamical model of lake phosphorus to examine the effects of these phosphorus retention structures. We highlight two effects: first, a structure that traps some of the phosphorus load before it reaches the lake reduces the amount of phosphorus in lake water. Second, some retention structures slow down lake phosphorus dynamics in a way that may perversely slow lake restoration. We propose a cleaning strategy that maximises the chances of restoring a lake to an oligotrophic condition. We demonstrate our model with a real-world case study. • We consider the mitigation of lake eutrophication though different types of retention structures. • We analysed three type of retention effects of a structure on the phosphorus dynamics in the lake. • The efficiency of this structure depends on a trade-off between its delayed effect and sedimentation processes. • Minimising phosphorus input is unexpectedly not always the best solution. [ABSTRACT FROM AUTHOR]

Research activity in the field of air pollution forecasting using artificial neural networks (ANNs) has increased dramatically in recent years. However, the development of ANN models entails levels of uncertainty given the black-box nature of ANNs. In this paper, a protocol by Maier et al. (2010) for ANN model development is presented and applied to assess journal papers dealing with air pollution forecasting using ANN models. The majority of the reviewed works are aimed at the long-term forecasting of outdoor PM 10 , PM 2.5 , and oxides of nitrogen, and ozone. The vast majority of the identified works utilised meteorological and source emissions predictors almost exclusively. Furthermore, ad-hoc approaches are found to be predominantly used for determining optimal model predictors, appropriate data subsets and the optimal model structure. Multilayer perceptron and ensemble-type models are predominantly implemented. Overall, the findings highlight the need for developing systematic protocols for developing powerful ANN models. • Research activity in ambient air pollution forecasting with ANNs continues to grow. • Forecasting of outdoor PM10, PM2.5, nitrogen oxides and ozone levels was widely done. • Feedforward and hybrid ANN model types were predominantly used. • Most of the identified model building steps were done in an ad-hoc manner. [ABSTRACT FROM AUTHOR]

Maximizing crop production with minimal resources such as water and energy is the primary focus of sustainable agriculture. Subsurface water retention technology (SWRT) is a stable approach that preserves water in sandy soils using water saving membranes. An optimal use of SWRT depends on its shape, location and other factors. In order to predict crop yield for different irrigation schedule, we require at least two computational processes: (i) a crop growth modeling process and (ii) a water and nutrient permeation process through soil to the root system. Validation of software parameters to suit properties of specific field becomes increasingly hard since they involve a coordination with field data and coordination between two software. In this paper, we propose a computationally fast approach that utilizes HYDRUS-2D software for water and nutrient flow simulation and DSSAT crop simulation software with an evolutionary multi-objective optimization (EMO) procedure in a coordinated manner to minimize water utilization and maximize crop yield prediction. Our proposed method consists of training one-dimensional crop model (DSSAT) on data generated by two dimensional model calibrates and validates (HYDRUS-2D), that accounts for water accumulation in the SWRT membranes. Then we used DSSAT model to find the best irrigation schedules for maximizing crop yield with the highest plant water use efficiency (Tambussi et al., 2007; Blum, 2009) using for the EMO methodology. The optimization procedure minimizes water usage with the help of rainfall water and increases corn yield prediction as much as six times compare to a non-optimized and random irrigation schedule without any SWRT membrane. Our framework also demonstrates an integration of latest computing software and hardware technologies synergistically to facilitate better crop production with minimal water requirement. • Precision irrigation using subsurface water retention technology (SWRT) is optimized • Water and nutrient mobility are simulated using HYDRUS-2D software • HYDRUS-2D's computational complexity is alleviated using a calibration procedure of DSSAT software which is fast • A multi-objective optimization method is employed to obtain optimal irrigation practices for minimum water usage and maximize crop growth. • This paper depicts how recent computational intelligence methods can be utilized to integrate two irrigation-based simulation software with weather and soil characteristics to obtain two important goals of agriculture practices. [ABSTRACT FROM AUTHOR]

This paper presents a pattern language for developing Object-Oriented Bayesian Networks (OOBN), as a member of the component-based probabilistic models family, to tackle complex problems. The proposed pattern language integrates knowledge from various domains, such as modeling, software engineering, and Bayesian networks, to provide a comprehensive solution for developing OOBNs. The paper also provides a validation framework to evaluate the pattern language. As a practical application for the OOBN pattern language, a case study of using it to develop an OOBN is presented. The model in the case study aims to represent the complex interconnections among the Sustainable Development Goals (SDG), long-term sustainability and resilience. The results of the case study validate the effectiveness of the pattern language and highlight its potential for future applications. The proposed OOBN pattern language provides a systematic approach to the development of OOBN, reducing the complexity and increasing the efficiency of their modeling process • New pattern language for Object-Oriented Bayesian Networks. • The pattern language Integrates multidisciplinary knowledge to handle large networks. • Validation framework for language evaluation. • Case study models SDG, sustainability, and nation resilience. [ABSTRACT FROM AUTHOR]

Marine plastic pollution has in recent decades become ubiquitous, posing threats to flora, fauna, and potentially human health. Proper monitoring and registration of litter occurrences are, therefore, of paramount importance to support better-informed decision-making. In this paper, we introduce the PlastOPol marine litter monitoring system. PlastOPol integrates external data sources on beached litter with data collected through citizen science initiatives based on the use of a mobile application (App). The App relies on state-of-the-art machine-learning approaches for litter detection and registration. The system also supports a human-in-the-loop strategy based on which improved versions of litter detection models are created over time thanks to annotations by citizen scientists. Finally, the system includes a geographic visualization tool to support the analysis of litter distribution data by decision-makers. This system has the potential to create a direct path between citizens, researchers, and decision-makers on the issue of marine litter. Finally, the paper presents compelling usage scenarios of the proposed monitoring system and discusses the evaluation of the App through a user study. The user study suggests that the PlastOPol system is an effective and valuable tool to monitor and communicate marine litter. [Display omitted] • This paper introduces PlastOPol, a new system for marine litter monitoring based on citizen science. • This paper introduces a data model for litter occurrence registration and data sharing. • The model evolves based on the annotation of citizen scientists made through a mobile application. • This paper introduces a map-based visualization that supports the analysis of litter distribution. • This paper presents and discusses a user study to assess the potential of PlastOPol to support citizen science tasks. [ABSTRACT FROM AUTHOR]

This paper presents the PAVICS-Hydro platform, the result of long-term strategic considerations and compromises between functionality and ease-of-use, which is developed to provide users with direct access to a comprehensive hydroclimatic modeling and analysis framework over North America. This includes a Python hydrological modelling package, efficient access to meteorological data, a large array of post-processed hydroclimatic data, numerical weather forecasts and bias-corrected CMIP6 climate projections at the scale of 5797 catchments, thus providing the basic components necessary for hydrological modeling and forecasting, as well as climate change impact studies. This paper presents the principles on which the PAVICS-Hydro platform was developed while explaining the choices behind its key features, with a specific focus on lessons learned for creators of the next generations' scientific research platforms. A case study using PAVICS-Hydro is given to illustrate how this platform can be used to support different needs in the field of water resources. • We present the PAVICS-Hydro platform for hydrological modelling and research. • The platform can be used for simulation, forecasting and climate change studies. • It helps automate repetitive tasks such as model building and data processing. • Access to multiple large-scale hydrometeorological and climate datasets is provided. • The platform is built on open-source software and free to use. [ABSTRACT FROM AUTHOR]

There has been some discussion about agent-based modelling (ABM) and theory, particularly how ABM might facilitate theory building. However, there is confusion about the different ways they could relate and some scepticism as to whether theory is needed if one has an ABM. This paper distinguishes some of the different ways that the term "theory" is used in ABM papers in three important ABM journals: Environmental Modelling & Software , Computers, Environment and Urban Systems and the Journal of Artificial Societies and Social Simulation. Apart from the simple-minded identification of theory with mathematics, we distinguish nine different ways that theory and ABM relate. This analysis is situated with respect to some of the expectations and philosophical background behind the idea of "theory". The paper concludes with some ways in which theory and ABM could work better together, some possible ways forward and suggests that a more cautious approach to generalisation might be more appropriate. • A brief review of some of the expectations and philosophical background concerning the role of "theory". • An analysis of nine different roles of "theory" can play with respect to Agent-Based Modelling (ABM). • Pointing out that some of the hopes for theory-building using ABM have yet to be realised so more development is needed. • Suggesting that rhetoric about theory generation should adopt a less ambitious approach to generalisation. • Some possible ways forward. [ABSTRACT FROM AUTHOR]

River water surface slopes (WSS) can improve the performance of rating curves for discharge estimates and provides information about river hydraulic processes and phenomena. Utilizing the beam pattern of ICESat-2, the WSS can be estimated by the slope of a linear regression of the water surface elevation and the distance along the river. This paper presents a robust processing scheme for calculating WSS along rivers in the SWOT River Database. We demonstrate the processing scheme for the Amur River basin using four years of data. This yields 7306 WSS estimates across 1693 reaches approximately 10 km in length. The slopes were estimated between 0.60 cm/km and 651.04 cm/km with a median standard error of 0.47 cm/km. The method works for slopes larger than 0.5 cm/km and river widths larger than 50 m. The software presented in this paper is available as the R package ICE2WSS on Github (https://github.com/lindchr/ICE2WSS). • Utilizing the unique beam pattern of ICESat-2 to estimate river water surface slopes • Open-source R package for robust estimation with customizable parameters • R package enhances river water surface slope availability and reduces data latency • Enables identification of areas with temporal variability of the water surface slope [ABSTRACT FROM AUTHOR]

Currently, the deep geological repository approach for spent nuclear fuel is regarded as the most dependable and secure method for permanently disposing of this kind of waste. Among its key safety components is an engineered barrier made from compacted bentonite, which isolates the encapsulated waste from the surrounding host rock. As a result, understanding how bentonites react to varying compositions of groundwater is crucial. This is where numerical modelling becomes essential. It is generally approved by the scientific community to idealise bentonite as a material structured under a double porosity system composed of the macro and microstructure. In this context, this paper illustrates the capabilities of a double-porosity reactive transport model for bentonites fully implemented in the multiphysics COMSOL platform. For this purpose, different experimental tests were simulated based on the evaluation of diffusive ion transport, mineral dissolution and cation exchange processes in MX-80 bentonite, obtaining very satisfactory results. • A new reactive transport model for bentonites was implemented in COMSOL Multiphysics. • The model includes a geochemical system composed of 14 chemical species and gypsum. • The model considers diffusive-dispersive-advective transport in double-porosity media. • The model validation was run by 3 tests set in the Task Force on Engineered Barrier. [ABSTRACT FROM AUTHOR]

Machine learning (ML) models are increasingly popular in environmental and hydrologic modeling, but they typically contain uncertainties resulting from noisy data (erroneous or outlier data). This paper presents a novel probabilistic approach that combines ML and Markov Chain Monte Carlo simulation to (1) detect and underweight likely noisy data, (2) develop an approach capable of detecting noisy data during model deployment, and (3) interpret the reasons why a data point is deemed noisy to help heuristically distinguish between outliers and erroneous data. The new algorithm recognizes that there is no unique way to split the training data into noisy and clean data, and thus produces an ensemble of plausible splits. The algorithm successfully detected noisy data in synthetic benchmark problems with varying complexity and a real-world public supply water withdrawal dataset. The algorithm is generic and flexible, making it suitable for application across a broad range of hydrologic and environmental disciplines. • The study presents a new probabilistic method to identify and reduce the impact of noisy data in machine learning datasets. • The approach generates a supervised noise detection model to identify noisy data during both model development and deployment. • The supervised noise detection model is interpreted to identify factors causing data to appear as noisy. • Interpretation of the supervised noise detection model is used to heuristically distinguish between erroneous and outlier data. [ABSTRACT FROM AUTHOR]

Each submodel in a hydro-morphodynamic model has its own local calibration parameters, leading to a high degree of uncertainty in their application. This paper proposes a global parameterization framework of hydro-morphodynamic models, which involves the development and implementation of submodels that share some common calibration parameters. The proposed model reduces the total number of adjustable parameters while helping to better understand the physics of the problem. As a case study, a holistic 1D vertical numerical simulation of the Ems estuary has been established. This simulation is proficient in qualitatively reproducing observed profiles of vertical velocity, concentration, and velocity shear. Using the proposed global parameterization, the model is calibrated using only measured rheological data from the Ems estuary, with these parameters universally applied to all submodels, eliminating the need for separate calibration for other submodels. The simulation demonstrates a commendable agreement with measurements while concurrently reducing the number of calibration parameters. • Introduces a global framework sharing calibration parameters among submodels. • Streamlines adjustable parameters and enhances understanding of underlying physics. • Tested in the Ems estuary case study using a 1D vertical numerical simulation. • Demonstrates strong measurement agreement and reduces calibration parameters. [ABSTRACT FROM AUTHOR]

The Full of Economic-Environment Linkages and Integration dX/dt (FeliX) model is a System Dynamics-based Integrated Assessment Model (IAM), explicitly incorporating human behaviors and their dynamic interactions among global systems. This paper presents FeliX 2.0, describing the detailed framework and key interactions among nine integrated modules. FeliX 2.0 refined its original version in population dynamics, food and land use systems, and socioeconomic settings for poverty analysis. Robust calibration is applied to key variables against their historical data since 1950. Future projections of multiple variables up to 2100 demonstrate coherences between FeliX 2.0 and the IAMs used in IPCC assessments. Both outputs (the robust calibration results and future projections) underscore the efficacy of FeliX 2.0 in capturing complex interdependencies within global systems. FeliX 2.0 stands as an informative tool and offers insights into interactions within the human-Earth system and the analysis of complex economic-environmental-social challenges in short- and long-term future. • A SD-based IAM model—FeliX 2.0—integrating human behavior for the human-Earth system simulation. • FeliX 2.0 refines FeliX 1.0 in population dynamics, food and land use, and poverty modeling. • Coherent projections up to 2100 for key variables in human-Earth systems by FeliX 2.0 • FeliX 2.0 stands as a pioneering tool for analyzing economic, environmental and social challenges. [ABSTRACT FROM AUTHOR]

Mixed enterprise farming systems that integrate more than one production system are important in agricultural production world-wide. Understanding and improving them can be made easier by modelling them with software tools. Modelling mixed enterprise farming systems can be a complex task as the interaction between the enterprises will introduce many dependencies. There are many software tools available that can model single enterprise systems, while there are few with the ability to model the biophysical systems in mixed farming. AusFarm has been designed and used to model mixed enterprise farming systems, integrating livestock, pasture, and crop models in one software tool and allowing flexible management of the whole farm. This paper demonstrates some key techniques that have been used for building and simulating mixed enterprise Australian farm systems in AusFarm. Examples of how to structure a cropping system and a livestock system are given. Key livestock and crop management tasks are implemented using flexible management rules. • Component based modelling in AusFarm can represent many types of mixed farming systems. • AusFarm's flexible management allows modelling of many types of farming practices. • Multi-dimensional experiments can be constructed using AusFarm for examining comparable scenarios. • AusFarm is designed to encapsulate a workflow for building, testing and analysing farming systems. [ABSTRACT FROM AUTHOR]

Wastewater utilities face competing priorities as they work to protect human health and water quality, and to maintain infrastructure in their communities. Budgetary constraints can be especially pronounced among small to medium-sized utilities. Utilities are increasingly turning to so-called intelligent water approaches as a cost-effective alternative to upgrading aging infrastructure. Intelligent water encompasses automated control and real-time decision support technologies and can be applied at scale to large and small utilities alike accommodating differences in needs, capabilities, and funds. Intelligent water upgrades can be designed to optimize existing conveyance, storage, and treatment during storms to help mitigate flooding and combined sewer overflows. The most promising real-time control algorithms coordinate control of upstream and downstream assets and are designed using urban hydrologic and hydraulic modeling software. The capabilities of legacy software, however, can sometimes inhibit the creation of sophisticated control algorithms. In this paper, we present PySWMM — an open-source Python wrapper developed for the EPA Storm Water Management Model (SWMM). PySWMM enables runtime interactions with the SWMM computational engine to flexibly read, modify system parameters, and control digital infrastructure during a simulation. Crucially, it allows modelers to easily combine SWMM with the rich set of scientific computing, big data, and machine learning modules found in the Python ecosystem. We highlight two real-world intelligent water case studies utilizing PySWMM in the cities of Cincinnati and Columbus, Ohio where it has helped to eliminate tens of millions of gallons of combined sewer overflows annually. • PySWMM is an open-source Python wrapper for EPA SWMM. • Embedding SWMM into Python's scientific computing ecosystem expands its capabilities. • Two utility CSO management applications leveraging PySWMM for real-time control and decision support are described. • Application results demonstrate how intelligent control can help reduce CSOs by tens of millions of gallons annually. [ABSTRACT FROM AUTHOR]

Fuel moisture content (FMC) is a critical parameter in fire and plume behaviors, showing diurnal and spatial variations influenced by local meteorological conditions, soil characteristics, and fuel properties. In low-intensity fires, small-scale FMC variations intensify, leading to an amplification of their effects on fire physics. In an effort to capture these variations, this paper presents the development of a physics-based model that couples a thermodynamic-based FMC prediction model for dead fuels with the Fire Dynamics Simulator of the National Institute of Standards and Technology. The model accuracy is validated against several existing experimental data, showing improvements over the baseline model which uses the kinetic-based Arrhenius drying approach. A case study of flame propagation in a small fuel bed is also presented, indicating the improved performance of the new model and its novel capabilities in capturing complex processes of fuel drying and moisture flux exchanges between the fuel and ambient atmosphere. • Fuel moisture is an important factor in shaping fire behavior and plume dynamics. • A detailed fuel moisture model is integrated into a fire-atmosphere interaction model. • The model is physics-based, based on dead-fuel energy-water balance analysis. • The model is integrated into the Fire Dynamics Simulator (FDS) of NIST. • It improves fire behavior prediction, capturing the complex fuel drying dynamics. [ABSTRACT FROM AUTHOR]

The increase in air pollution has posed numerous new challenges for human society, making the exploration of an effective method for predicting air pollutant concentrations highly significant. The current research faces several primary challenges: the neglect of non-Euclidean characteristics of site distribution on data and the strong spatiotemporal dependencies in the dispersion process of pollutants. To address these issues, this paper constructs a spatiotemporal hybrid prediction model – the MGAtt-LSTM method – for predicting PM 2.5 concentrations, which employs the dynamic multi-graph attention module (MGAtt) to tackle spatial dependencies and Long Short-Term Memory networks (LSTM) to address temporal dependencies. Additionally, extensive experiments are conducted by using historical air pollutant monitoring data and meteorological data from the Beijing-Tianjin-Hebei region. The results demonstrate that the proposed MGAtt-LSTM model achieved superior performance in concentration prediction compared to existing benchmark models. • The neural network consists of multi-graph attention network and LSTM network. • Air pollution data in North China and meteorological data are used to forecast. • The uneven distribution of pollutant sites is considered to predict PM 2.5 concentrations precisely. • The use of multi-graph attention networks addresses the issue of traditional GCN methods relying on fixed graph structures. [ABSTRACT FROM AUTHOR]

Given climate change, trend detection is gaining increasing attention in the context of multivariate frequency analysis. In this paper, we propose new statistical tests for multivariate trend detection. The first one, a multivariate overall trend (MOT) test, is designed to detect trend in all components of the multivariate distribution (margins and dependence structure) whereas the second test is a multivariate dependence trend (MDT) test focusing on detecting trend in the dependence structure. A simulation study is used to evaluate the performance of the proposed tests. Results show that the proposed MOT test performs well when trend is present in margins, in the dependence structure and/or in both. Likewise, results of the proposed MDT test indicate a higher power when the trend is in the dependence structure. Moreover, an application to a real-world dataset is provided. Performing the proposed tests with the univariate tests provides a complete overview of trend detection. • Two multivariate trend tests for multivariate hydrological series are proposed. • New multivariate overall trend (MOT) test dealing with trend in all the components of the whole multivariate distribution. • New multivariate dependence trend (MDT) test focuses on trend in the dependence structure. • Vast simulation study is considered to evaluate the performance of the tests. • The developed tests show high performance, with increasing power observed as the trend slope and sample size increase. [ABSTRACT FROM AUTHOR]

When performing modeling routines with iterative simulations, a research gap exists in developing the parallelizing frameworks that are adaptable to the increasing complexities of the hydrologic models. This study addresses this gap by proposing a parallel simulation stack for SWAT (PASS4SWAT) that leverages the power of cloud-native infrastructure. Cloud-native applications are designed to be scalable, agile, and resilient, making them ideal for computationally intensive tasks such as hydrologic modeling. To achieve this goal, PASS4SWAT was created by integrating a message broker, a container runtime and orchestration tool, and certain components from SWAT-CUP. We evaluated PASS4SWAT using the Jinjiang watershed model and a synthetic model with high input/output demands to demonstrate its effectiveness. The results show that PASS4SWAT can consistently replicate the results obtained with SWAT-CUP and significantly reduce the runtime, achieving 91.5% and 93.2% reductions in single-node and multinode Kubernetes clusters, respectively. Therefore, in this paper, we conclude that PASS4SWAT can effectively address the high computational demands of the SWAT model by scaling out parallel tasks on a cluster, and can adapt flexibly to diverse environments, including single-node clusters, multinode clusters, and potential cloud platforms. [Display omitted] • PASS4SWAT significantly reduces the runtime needed for calibration and uncertainty analysis. • PASS4SWAT can adapt to diverse environments, including single-node and multinode clusters. • PASS4SWAT can autonomously manage subtasks and obviate the need for explicit failover and load balancing. • PASS4SWAT has great potential for application in other models and can be linked with other calibration tools. [ABSTRACT FROM AUTHOR]

The exponential growth in data generated by satellites, radars, sensors, and analysis and reanalysis from model outputs for the hydrological domain requires efficient real-time data management and distribution mechanisms. This paper introduces HydroRTC, a web-based data transfer and communication library designed to accelerate large-scale data sharing and analysis. Leveraging next-generation web technologies like WebSockets, WebRTC and Node.js, the library enables seamless peer-to-peer sharing, smart data transmission, and large dataset streaming. Three primary scenarios are presented as use cases, demonstrating the potential of HydroRTC as server-to-peer with intelligent data scheduling and large data streaming, peer-to-peer data sharing, and peer-to-server for data exchange. HydroRTC offers a promising solution for collaborative infrastructures in the hydrological and environmental domain, allowing real-time and high-throughput data sharing and transfer for enhancing research efficiency and collaboration capabilities. • HydroRTC accelerates large-scale data sharing with next-gen web technologies. • Three primary scenarios: server-to-peer, peer-to-peer, peer-to-server data exchange. • Promising solution for collaborative infrastructures in hydrological and environmental domains. • Exponential growth in data necessitates efficient real-time management mechanisms. • Leverages WebSockets, WebRTC, Node.js for seamless peer-to-peer sharing. [ABSTRACT FROM AUTHOR]

Although traditional physical models have been used to analyse groundwater systems, the emergence of novel machine learning models can improve the accuracy of the predictions. Deep learning has been prominent in environmental and climate change problems. In this paper, we present a framework for utilising deep learning models to predict groundwater levels based on nearby streamflow and rainfall data. We address the missing data problem using a Bayesian linear regression model within the deep learning framework. Our deep learning framework utilises models such as long-short term memory (LSTM) networks and convolutional neural networks (CNN) for multi-step ahead time series prediction. We examine the fluctuations in groundwater levels at various boreholes located near Middle Creek in New South Wales, Australia. We use the National Collaborative Research Infrastructure Strategy (NCRIS) groundwater database and utilise Bayesian linear regression to impute missing data. We investigate the accuracy of the selected models for individual and regional basins and univariate and multivariate strategies. Our results show that the LSTM-based regional model with multivariate strategy using rainfall data provided the best accuracy. • We predict groundwater levels based on stream flow and rainfall data. • Our framework uses Bayesian data imputation and deep learning for multi-step prediction. • We examine the groundwater fluctuations at various boreholes located in Australia. • Results show a fine-tuned regional model with rainfall data yielded the best accuracy. • We find that LSTM models outperformed CNNs using the multivariate approach. [ABSTRACT FROM AUTHOR]

Cellular automata (CA) based models are practical tools to simulate the spatiotemporal landscape evolution induced by the land use/-cover change (LUCC). Existing models are struggling to comprehensively handle the spatiotemporal driving relationships amid the nonlinear LUCC process. Besides, the landscape patterns are not considered in most models, making them struggled to support the development strategies. Aiming at overcoming these obstacles, a novel land use/-cover change model concerning spatiotemporal dependency and properties related to landscape evolution (STAPLE) is proposed in this paper. A potential generating module establishing the nonlinear spatiotemporal driving relationship and a spatial allocating module employing a landscape-based CA are integrated for realistic LUCC simulations. As a case study, the proposed model is applied in Zhengzhou, China to assess its performance. It is indicated that the STAPLE model achieves a higher simulation accuracy, and the landscape properties are effectively manipulated. It provides a reproducible tool for policy-makers to explore a low-ecological-risk landscape under different future scenarios and achieve sustainable developments. • Propose an LUCC model concerning spatiotemporal dependency and landscape evolution. • Control the spatial evolution of landscape using a novel CA algorithm. • Promote accuracy by employing ST-CNN to assimilate the latent spatiotemporal dependency. • Provide a reproducible tool for policy-makers to achieve sustainable development. [ABSTRACT FROM AUTHOR]

Multi-source heterogeneous, multi-modal, and multi-type open scientific data (e.g., thematic sharing sites, metadata, journal articles, etc.) on earth surface systems (EES) provide important data sources for knowledge mining, discovery, and accurate recommendations, and also pose increasing challenges, resulting in the need to develop appropriate tools to address these challenges and support decision-making. This paper constructs an interoperable software system to store, visualize, and publish open science data of ESS. Utilizing an open scientific data catalogue repository encompassing EES information as foundational input and employing an integrated modeling methodology, this system endeavors to synthesize heterogeneous surface data of diverse linguistic, sourced, and typological origins. The objective is to facilitate multidimensional data retrieval and precise data auto-recommendation, thereby fostering the dissemination of scientific data and facilitating value-added services within EES domain. The tool may be used by stakeholders including researchers, data analysts, policymakers and national authorities to support decision-making on questions ranging from locating the location of open data related to the topic, to discovering high-quality data, selecting the data with the better overall evaluation. Along with a description of the system/platform design process, its structure, and the constituent models, key results are presented relating to the user interface, and several application examples. Software systems can help modelers to use the best features of a single software tool to answer open scientific data-related questions that seek to discovery, use, comparison or synthesis within or across topics of ESS. • Developed a novel tool for constructing a scientific data association network. • Presented an interoperable software system. • Proposed a 'time-space-topic' model for building associative networks. [ABSTRACT FROM AUTHOR]

Extreme events have the potential to significantly impact transportation infrastructure performance. For example, in the case of bridges, climate change impacts the river discharge, hence scouring patterns, which in turn, affects the bridge foundation stability. Therefore, extreme events (river flow) forecasting is mandatory in bridge reliability analysis. This paper approaches this river flow forecasting problem by developing a Markov-Switching Autoregressive model coupled with a conditional hidden seasonal Markov component. In addition, the proposed model is also combined with the deep machine learning neural networks method to forecast river flow from a dataset or from simulations. The proposed method is illustrated by using realistic data: historic river flow values of the Thames River. The results indicate that the proposed model well represented the extreme events within the dataset. In terms of river flow forecasting, the results indicate that the forecasts improve when the training period changes from 20 years to 40 years. • We proposed a Markov-Switching Autoregressive model with a conditional hidden seasonal Markov component. • The proposed model accurately captures and represents extreme river flow events. • Deep machine learning neural networks were used to forecast river flow. • River flow forecasts improve when the training period increases. [ABSTRACT FROM AUTHOR]

To enable organizations to achieve carbon neutrality through agile capabilities, this paper establishes an integrative framework of carbon-zero agility consisting of three dimensions: search scope agility, search locus agility, and search pace agility. However, applying common agile methodologies like Scrum, Kanban, and Lean to cultivate these capabilities inevitably introduces feedback ambiguity, which can paralyze decision-making and increase errors due to inherent human cognitive limitations. To address this, tailored carbon-zero feedback algorithms are proposed to complement human judgment in agile workflows. Specifically, prescriptive analytics, federated learning, and probabilistic programming are injected into Scrum, Kanban, and Lean respectively to restore clarity amidst ambiguity. The framework is grounded in cases from the textile industry to demonstrate applicability in practical settings. By targeting the roots of distortions with human-algorithm collaborations, it provides an actionable roadmap to implement carbon-zero agility. • Establishes carbon-zero agility framework: search scope, locus, and pace agility. • Identifies behavioral challenges, and feedback ambiguity in agile for carbon-zero. • Proposes tailored carbon-zero feedback algorithms to enhance agility. • Connects agile processes with broader organizational sustainability capabilities. [ABSTRACT FROM AUTHOR]

Forest fires cause every year damages to biodiversity, atmosphere, and economy activities. Forest fire simulation have improved significantly, but input data describing fire scenarios are subject to high levels of uncertainty. In this work the two-stage prediction scheme is used to adjust unknown parameters. This scheme relies on an input data calibration phase, which is carried over following a genetic algorithm strategy. The calibrated inputs are then pipelined into the actual prediction phase. This two-stage prediction scheme is leveraged by the cloud computing paradigm, which enables high level of parallelism on demand, elasticity, scalability and low-cost. In this paper, all the models designed to properly allocate cloud resources to the two-stage scheme in a performance-efficient and cost-effective way are described. This Cloud-based Urgent Computing (CuCo) architecture has been tested using, as study case, an extreme wildland fire that took place in California in 2018 (Camp Fire). • Data-driven calibration to deal with uncertainty in forest fire spread prediction. • Cloud-based urgent computing implementation of a two-stage prediction model. • Use of utility function to deal with the cost-performance trade-off. • Validation against a deadly and destructive wildfire with promising results. [ABSTRACT FROM AUTHOR]

Many environmental variables, in particular, related to air or water quality, are measured in a limited number of points and often for a limited time span. This forbids the development of accurate models for interesting locations with missing or insufficient data and poses the question of whether a model developed for another measurement site can be reliably applied. Such a question is particularly critical when the model is entirely data-driven, such as a neural network. In this context, the paper proposes a procedure to evaluate the expected performance of an existing neural network model applied to a new unmonitored station. This transferability assessment is exemplified by the problem of forecasting ozone concentrations in different environmental settings around the Alpine Arc. Long Short-Term Memory (LSTM) neural network models are applied for predicting hourly concentrations in 20 stations of different types (urban, rural, and mountain). The analysis of the results allows us to determine the expected performance of such models in new cases and reduce the transferability uncertainty when the existing models can be partitioned into clusters. The LSTM models demonstrate the possibility of high accuracy in ozone forecasting at all sites. Given the significant impacts of this gas on human health and the environment, this can contribute to better decision-making and mitigation strategies for air pollution control. • We investigate the possibility of training a model in a site and reuse it elsewhere. • We introduce the concept of transferability matrix and transferability indicators. • The transfer learning reliability is enhanced by clustering the sites. • Ozone concentration forecasting for 20 locations in the Alpine region is considered. [ABSTRACT FROM AUTHOR]