Evaluation of statistical models for predicting Escherichia coli particle attachment in fluvial systems

Modeling surface water Escherichia coli fate and transport requires partitioning E. coli into particle-attached and unattached fractions. Attachment is often assumed to be a constant fraction or is estimated using simple linear models. The objectives of this study were to: (i) develop statistical models for predicting E. coli attachment and virulence marker presence in fluvial systems, and (ii) relate E. coli attachment to a variety of environmental parameters. Stream water samples (n = 60) were collected at four locations in a rural, mixed-use watershed between June and October 2012, with four storm events (>20 mm rainfall) being captured. The percentage of E. coli attached to particles (>5 μm) and the occurrences of virulence markers were modeled using water quality, particle concentration, particle size distribution, hydrology and land use factors as explanatory variables. Three types of statistical models appropriate for highly collinear, multidimensional data were compared: least angle shrinkage and selection operator (LASSO), classification and regression trees using the general, unbiased, interaction detection and estimation (GUIDE) algorithm, and multivariate adaptive regression splines (MARS). All models showed that E. coli particle attachment and the presence of E. coli virulence markers in the attached and unattached states were influenced by a combination of water quality, hydrology, land-use and particle properties. Model performance statistics indicate that MARS models outperform LASSO and GUIDE models for predicting E. coli particle attachment and virulence marker occurrence. Validating the MARS modeling approach in multiple watersheds may allow for the development of a parameterizing model to be included in watershed simulation models.     

The dynamics of tree cover change in a rural Australian landscape

This study investigated the change in tree cover in a 130 000 ha rural landscape near Bathurst, New South Wales, Australia between 1954 and 1989. Using aerial photographs, tree cover was assessed using five cover classifications (<2% cover, 2.1–25%, 25.1–50%, >50% or exotic pine plantation), with the resulting data analysed using a Geographic Information System. Overall, there was a loss of woodland/forest and an increase in the area of cleared land and Pinus radiata plantations over 35 years. Although some regeneration of woodland/forest cover occurred on marginal farmland, the general trend of loss and fragmentation has serious ramifications for nature conservation and agricultural production. The eastern and western portions of the study area exhibited quite different patterns in tree cover change, highlighting the need for examining landscapes at appropriate spatial and temporal scales.     

A comparative study on machine learning modeling for mass movement susceptibility mapping (a case study of Iran)

Mass movements are among the most dangerous natural hazards in mountainous regions. The present study employs machine learning (ML) models for mass movement susceptibility mapping (MMSM) in Iran based on a comprehensive dataset of 864 mass movements which include debris flow, landslide, and rockfall during the last 42 years (1977–2019) as well as 12 conditional factors. The results of validation stage show that RF (random forest) is the most viable model for mass movement susceptibility maps. In addition, MARS (multivariate adaptive regression splines), MDA (mixture discriminant additive), and BRT (boosted regression trees) models also provide relatively accurate results. Results of the AUC for validation of produced maps were 0.968, 0.845, 0.828, and 0.765 for RF, MARS, MDA, and BRT, respectively. Based on MMSM generated by RF model, 32% of study area is identified to be under high and very high susceptibility classes. Most of the endangered areas for mass movement are in the west and central parts of the Chaharmahal and Bakhtiari Province. In addition, our findings indicate that elevation, slope angle, distance from roads, and distance from faults are critical factors for mass movement. Our results provide a perspective view for decision makers to mitigate natural hazards.

     

A habitat model for Parus major minor using a logistic regression model for the urban area of Osaka, Japan

We developed a habitat model for Great Tits, Parus major minor, in an urban area of Osaka, Japan. Although Great Tits play an important role in the urban food web, there are few in most of Osaka and downtown Tokyo, areas that have low vegetation cover. We derived a habitat model for Great Tits using a logistic regression model with GIS. A bird survey was conducted twice in 85 urban parks using the line census method in the breeding season, May–July 2000. A GIS base map was created from aerial photographs. The following variables were measured for each park from public records and GIS for predicting the habitat: park properties (age, distance to the nearest mountains, distance to the nearest forest of more than 10 ha), the number of other habitats occupied by the tits within 500 m and 1 km and area of tree cover (within it and surrounding parks). Number of nearby habitats and distance from the nearest mountains and forest of more than 10 ha are variables related to the regional populations of the tits. Area of tree cover was measured for five radii (50, 100, 150, 200, 250 m) from park centers, and a logistic regression model was calculated for each of these radii. Variables were selected by a stepwise method. The best fitting model of the five models was selected using Akaike’s Information Criteria. In the bird survey, the tit was recorded in 12 parks. The areas of parks occupied by the tit were 0.56–136.0 ha (mean: 26.0 ha±42.06 S.D., n=12). Variables selected in the best fitting model were area of tree cover in a radius of 250 m from park center and the number of other habitats within 1 km as positive factors. This model shows that area of tree cover within a certain range and number of nearby populations are the keys to the distribution of Great Tits. From this model obtained, 6.0 ha (31%), 4.0 ha (20%), 2.6 ha (13%) and 1.8 ha (9%) of tree area are needed to achieve probability of 0.5 in a radius of 250 m when numbers of other habitats within 1 km was 0–3, respectively. A realistic target figure is 10% of tree cover throughout urban areas to create an ecologically sustainable city.     

Optimising spatial mergers: commercial and regulatory perspectives

This paper explores the potential application of spatial analysis and modelling techniques to the merger problem, and introduces a conceptual spatial modelling framework to improve retail merger and acquisition planning from both a commercial and regulatory perspective. Focusing on the UK retail sector, the paper demonstrates the importance of local scale market ‘dynamics’ in merger and acquisition (M&A) planning and evaluation; reviews the range of existing spatial modelling approaches to retail location problems, and illustrates how these models can be applied to M&A planning.     

Integration of GIS and multicriteria decision analysis for urban aquaculture development in Bangladesh

Site selection is a key factor in any aquaculture operation, affecting both success and sustainability as well as solving land or water use conflicts. This study was conducted to identify suitable sites for carp farming development in urban water bodies (UWBs) of Chittagong, Bangladesh using Geographical Information Systems (GIS) based MultiCriteria Evaluation (MCE) of water, soil and infrastructure database. ASTER imagery and 14 thematic layers were analyzed with ENVI and GIS capabilities, and developed a series of GIS models to identify and prioritize the appropriate UWBs for carp farming. The study identified 487 UWBs occupying 362 ha and revealed 280 ha (77%) is the most suitable, 36 ha (10%) is moderately suitable and 46 ha (13%) is not suitable which was consistent with field verification. The results are encouraging for extension of carp culture and diversify the economic activities of the urban dwellers.     

Exploring,modelling and predicting spatiotemporal variations in house prices

Hedonic price modelling has long been a powerful tool to estimate house prices in the real estate market. Increasingly, traditional global hedonic price models that largely ignore spatial effects are being superseded by models that deal with spatial dependency and spatial heterogeneity. In addition, many novel methods integrating spatial economics, statistics and geographical information science (GIScience) have been developed recently to incorporate temporal effects into hedonic house price modelling. Here, a local spatial modelling technique, geographically weighted regression (GWR), which accounts for spatial heterogeneity in housing utility functions is applied to a 19-year set of house price data in London (1980–1998) in order to explore spatiotemporal variations in the determinants of house prices. Further, based on the local parameter estimates derived from GWR, a new method integrating GWR and time series (TS) forecasting techniques, GWR–TS, is proposed to predict future local parameters and thus future house prices. The results obtained from GWR demonstrate variations in local parameter estimates over both space and time. The forecasted future values of local estimates as well as house prices indicate that the proposed GWR–TS method is a useful addition to hedonic price modelling.

     

Pipe failure modelling for water distribution networks using boosted decision trees

Pipe failure modelling is an important tool for strategic rehabilitation planning of urban water distribution infrastructure. Rehabilitation predictions are mostly based on existing network data and historical failure records, both of varying quality. This paper presents a framework for the extraction and processing of such data to use it for training of decision tree-based machine learning methods. The performance of trained models for predicting pipe failures is evaluated for simple as well as more advanced, ensemble-based, decision tree methods. Bootstrap aggregation and boosting techniques are used to improve the accuracy of the models. The models are trained on 50% of the available data and their performance is evaluated using confusion matrices and receiver operating characteristic curves. While all models show very good performance, the boosted decision tree approach using random undersampling turns out to have the best performance and thus is applied to a real world case study. The applicability of decision tree methods for practical rehabilitation planning is demonstrated for the pipe network of a medium sized city.     

A comprehensive experimental and artificial network investigation of the performance of an ultrafiltration titanium dioxide ceramic membrane: application in produced water treatment

This work is an experimental investigation of the effects of the operating conditions on the performance of a novel titanium‐based ceramic ultrafiltration membrane used to treat field produced water. To design the experiments and optimize the operating conditions, the Taguchi method was used to predict the optimal operating conditions. Under optimal conditions, an almost oil free permeation is obtained (98.93%) along with the removal of more than 99% for the total organic carbon (TOC), a high turbidity removal (99.82%) and a good salinity rejection for a UF membrane. The membrane was capable of treating a high steady flux of 441 L/m² h with an overall flux decay of 28.6%. The Hermia’s cake formation model fitted the flux declining behaviour better than the three other associated models. Finally, four different techniques based on artificial intelligence (AI) methods were used to fit the flux declining behaviour. They seem to outperform the simple Hermiaˊs model for the modelling of oily water filtration.     

Modelling the impacts of payments for biodiversity conservation on regional land-use patterns

We present a land-use allocation model that evaluates the impact of payments for ecosystem services such as biodiversity conservation on land-use patterns. In a non-linear optimisation procedure, land use is allocated at farm level, taking into account risk behaviour, and spatial as well as temporal variability of net revenues of land-use alternatives, using a spatial resolution of 29 m × 29 m. The model is applied to a study area of 30 km × 34 km in western Ecuador, considered a hotspot for biodiversity. In this coffee growing region, agroforestry systems with shade-coffee are important for biodiversity conservation at the landscape level, but under pressure due to low revenues. Through scenario analyses, we analyse the effects of price premiums per kg, payments per ha, and price buffering for coffee, as well as incentives leading to coffee yield increases, and payments per ha of forest remnants. We compare different risk aversion levels of the involved agents. We conclude that payments per ha are more efficient to maintain or increase coffee areas than payments per kg. Price buffering may be the most cost-effective way to support coffee production, especially at high-risk aversion levels. Significant yield increases are necessary to make coffee more attractive compared to the other alternatives. Relatively low payments per ha of forest can significantly increase forested land. Risk behaviour turns out to be an important factor when determining the possible effect of payments for ecosystem services. We conclude that the model is a versatile tool to support planning of payments for conserving ecosystems.     

Pipe failure rate prediction in water distribution networks using multivariate adaptive regression splines and random forest techniques

ABSTRACT

This paper presents the results of a comparison between multivariate adaptive regression splines (MARS) and random forest (RF) techniques in pipe failure prediction in two water distribution networks. In this regard, pipe diameter, pipe length, pipe installation depth, pipe age and average hydraulic pressure are considered as input variables. Results show that the RF outperforms the MARS which is found as an accurate pipe failure rate predictor. The proposed models are further evaluated through dividing the data into three parts of lower, medium and higher pipe failure rate values. According to the equations produced by MARS technique, three variables of pipe diameter, pipe age and average hydraulic pressure are distinguished as the most effective variables in predicting pipe failure rate in the first case study. Four variables of pipe diameter, pipe length, pipe age and average hydraulic pressure are determined as the most effective variables in the second case study.     

Projecting transition probabilities for regular public roads at the ecoregion scale: A Northern Appalachian/Acadian case study

Existing roads have far-reaching effects on biodiversity, and therefore road network expansion is of critical concern to conservation planning. Road density trend analysis is often too coarse and assumes homogeneous landscapes, whereas spatial transition probability analysis captures landscape variability typical of ecoregions. Simple models for projecting road network growth will assist planning agencies and conservation organizations to guide protection efforts. We investigate growth of regular public roads in the State of Maine over a 17-year historical period, and then use the best-selected (AIC) logistic regression model to validate and then project spatial probability of future roads to the Northern Appalachian/Acadian ecoregion. Nearly 2000 km of new roads were constructed in settled landscapes in Maine 1986–2003, influencing 37,000 ha of adjacent habitats. The majority (93.5%) of the new roads performed local functions and were short (<1/3 km in length), characterized as residential roads typical of sprawl. The best-selected logit model [dwelling density (+), elevation (−), distance to urban area (−), distance to existing primary/secondary highway (−)] captured 84% of reserved new road points in Maine, and only 27% of random points at the >0.5 probability level. The projected model forecasts 0.5 million km of new residential public roads in the Northern Appalachian/Acadian ecoregion for the next two decades, suggesting that cumulative effects of residential road network expansion are a serious region-scale biodiversity threat.     

A Novel Multiscale Methodology for Simulating Tunnel Ventilation Flows During Fires

This paper applies a novel and fast modelling approach to simulate tunnel ventilation flows during fires. The complexity and high cost of full CFD models and the inaccuracies of simplistic zone or analytical models are avoided by efficiently combining mono-dimensional (1D) and CFD (3D) modelling techniques. A simple 1D network approach is used to model tunnel regions where the flow is fully developed (far field), and a detailed CFD representation is used where flow conditions require 3D resolution (near field). This multi-scale method has previously been applied to simulate tunnel ventilation systems including jet fans, vertical shafts and portals (Colella et al., Build Environ 44(12): 2357–2367, 2009) and it is applied here to include the effect of fire. Both direct and indirect coupling strategies are investigated and compared for steady state conditions. The methodology has been applied to a modern tunnel of 7 m diameter and 1.2 km in length. Different fire scenarios ranging from 10 MW to 100 MW are investigated with a variable number of operating jet fans. Comparison of cold flow cases with fire cases provides a quantification of the fire throttling effect, which is seen to be large and to reduce the flow by more than 30% for a 100 MW fire. Emphasis has been given to the discussion of the different coupling procedures and the control of the numerical error. Compared to the full CFD solution, the maximum flow field error can be reduced to less than few percents, but providing a reduction of two orders of magnitude in computational time. The much lower computational cost is of great engineering value, especially for parametric and sensitivity studies required in the design or assessment of ventilation and fire safety systems.     

A compartmental model to describe hydraulics in a full-scale waste stabilization pond

The advancement of experimental and computational resources has facilitated the use of computational fluid dynamics (CFD) models as a predictive tool for mixing behaviour in full-scale waste stabilization pond systems. However, in view of combining hydraulic behaviour with a biokinetic process model, the computational load is still too high for practical use. This contribution presents a method that uses a validated CFD model with tracer experiments as a platform for the development of a simpler compartmental model (CM) to describe the hydraulics in a full-scale maturation pond (7 ha) of a waste stabilization ponds complex in Cuenca (Ecuador). 3D CFD models were validated with experimental data from pulse tracer experiments, showing a sufficient agreement. Based on the CFD model results, a number of compartments were selected considering the turbulence characteristics of the flow, the residence time distribution (RTD) curves and the dominant velocity component at different pond locations. The arrangement of compartments based on the introduction of recirculation flow rate between adjacent compartments, which in turn is dependent on the turbulence diffusion coefficient, is illustrated. Simulated RTD’s from a systemic tanks-in-series (TIS) model and the developed CM were compared. The TIS was unable to capture the measured RTD, whereas the CM predicted convincingly the peaks and lags of the tracer experiment using only a minimal fraction of the computational demand of the CFD model. Finally, a biokinetic model was coupled to both approaches demonstrating the impact an insufficient hydraulic model can have on the outcome of a modelling exercise. TIS and CM showed drastic differences in the output loads implying that the CM approach is to be used when modelling the biological performance of the full-scale system.     

Comparison of different uncertainty techniques in urban stormwater quantity and quality modelling

Urban drainage models are important tools used by both practitioners and scientists in the field of stormwater management. These models are often conceptual and usually require calibration using local datasets. The quantification of the uncertainty associated with the models is a must, although it is rarely practiced. The International Working Group on Data and Models, which works under the IWA/IAHR Joint Committee on Urban Drainage, has been working on the development of a framework for defining and assessing uncertainties in the field of urban drainage modelling. A part of that work is the assessment and comparison of different techniques generally used in the uncertainty assessment of the parameters of water models. This paper compares a number of these techniques: the Generalized Likelihood Uncertainty Estimation (GLUE), the Shuffled Complex Evolution Metropolis algorithm (SCEM-UA), an approach based on a multi-objective auto-calibration (a multialgorithm, genetically adaptive multi-objective method, AMALGAM) and a Bayesian approach based on a simplified Markov Chain Monte Carlo method (implemented in the software MICA). To allow a meaningful comparison among the different uncertainty techniques, common criteria have been set for the likelihood formulation, defining the number of simulations, and the measure of uncertainty bounds. Moreover, all the uncertainty techniques were implemented for the same case study, in which the same stormwater quantity and quality model was used alongside the same dataset. The comparison results for a well-posed rainfall/runoff model showed that the four methods provide similar probability distributions of model parameters, and model prediction intervals. For ill-posed water quality model the differences between the results were much wider; and the paper provides the specific advantages and disadvantages of each method. In relation to computational efficiency (i.e. number of iterations required to generate the probability distribution of parameters), it was found that SCEM-UA and AMALGAM produce results quicker than GLUE in terms of required number of simulations. However, GLUE requires the lowest modelling skills and is easy to implement. All non-Bayesian methods have problems with the way they accept behavioural parameter sets, e.g. GLUE, SCEM-UA and AMALGAM have subjective acceptance thresholds, while MICA has usually problem with its hypothesis on normality of residuals. It is concluded that modellers should select the method which is most suitable for the system they are modelling (e.g. complexity of the model’s structure including the number of parameters), their skill/knowledge level, the available information, and the purpose of their study.     

Numerical simulation of dispersion in urban street canyons with avenue-like tree plantings: Comparison between RANS and LES

Previous CFD studies on pollution dispersion problems have largely centred on employing Reynolds-averaged Navier–Stokes (RANS) turbulence closure schemes, which have often been reported to overpredict pollutant concentration levels in comparison to wind tunnel measurement data. In addition, the majority of experimental and numerical investigations have failed to account for the aerodynamic effects of trees, which can occupy a significant proportion of typical urban street canyons. In the present work, the prediction accuracy of pollutant dispersion within urban street canyons of width to height ratio, W/H = 1 lined with avenue-like tree plantings are examined using two steady-state RANS models (the standard k-ε and RSM), and Large Eddy Simulation (LES) to compare their performance against wind tunnel experiments available on the online database CODASC [1]. Two cases of tree crown porosities are investigated, one for a loosely (P_vol = 97.5%) and another for a densely (P_vol = 96%) packed tree crown, corresponding to pressure loss coefficients of λ = 80 m⁻¹ and λ = 200 m⁻¹, respectively. Results of the tree-lined cases are then compared to a tree-free street canyon in order to demonstrate the impact of trees on the flow field and pollutant dispersion, and it is observed that the presence of trees reduces the in-canyon circulation and air exchange, and increases the overall concentration levels. Between the two numerical methods employed, LES performs better than RANS, because it captures the unsteady and intermittent fluctuations of the flow field, and hence, successfully resolves the transient mixing process within the canyons.     

Influence of threshold value in the use of statistical methods for groundwater vulnerability assessment

Statistical techniques can be used in groundwater pollution problems to determine the relationships among observed contamination (impacted wells representing an occurrence of what has to be predicted), environmental factors that may influence it and the potential contamination sources. Determination of a threshold concentration to discriminate between impacted or non impacted wells represents a key issue in the application of these techniques. In this work the effects on groundwater vulnerability assessment by statistical methods due to the use of different threshold values have been evaluated. The study area (Province of Milan, northern Italy) is about 2000 km² and groundwater nitrate concentration is constantly monitored by a net of about 300 wells. Along with different predictor factors three different threshold values of nitrate concentration have been considered to perform the vulnerability assessment of the shallow unconfined aquifer. The likelihood ratio model has been chosen to analyze the spatial distribution of the vulnerable areas. The reliability of the three final vulnerability maps has been tested showing that all maps identify a general positive trend relating mean nitrate concentration in the wells and vulnerability classes the same wells belong to. Then using the kappa coefficient the influence of the different threshold values has been evaluated comparing the spatial distribution of the resulting vulnerability classes in each map. The use of different threshold does not determine different vulnerability assessment if results are analyzed on a broad scale, even if the smaller threshold value gives the poorest performance in terms of reliability. On the contrary, the spatial distribution of a detailed vulnerability assessment is strongly influenced by the selected threshold used to identify the occurrences, suggesting that there is a strong relationship among the number of identified occurrences, the scale of the maps representing the predictor factors and the model efficiency in discriminating different vulnerable areas.     

High-resolution CFD simulations for forced convective heat transfer coefficients at the facade of a low-rise building

High-resolution 3D steady RANS CFD simulations of forced convective heat transfer at the facades of a low-rise cubic (10 × 10 × 10 m³) building are performed to determine convective heat transfer coefficients (CHTC). The focus is on the windward facade. CFD validation is performed based on wind tunnel measurements of velocity and heat transfer for reduced-scale cubic models. The CFD simulations employ a high-resolution grid with, for the 10 m cubic building, cell centres at a minimum distance of 160 μm from the building surface to resolve the entire boundary layer, including the viscous sublayer and the buffer layer, which dominate the convective surface resistance. The results show that: (1) the wind flow around the building results in highly varying CHTC values across the windward facade; (2) standard and non-equilibrium wall functions are not suitable for CHTC calculation, necessitating either low-Reynolds number modelling or specially-adapted wall functions; (3) at every facade position, the CHTC is a power-law function of the mean wind speed; (4) the CHTC distribution at the windward facade is relatively insensitive to wind direction variations in the 0–67.5° angle range; (5) the CHTC shows a stronger spatial correlation with the turbulent kinetic energy than with the mean wind speed across the facade; and (6) the CHTC distribution across the windward facade is quite similar to the distribution of wind-driven rain (WDR), with both parameters reaching high levels near the top edge of the facade. This suggests that also the convective moisture transfer coefficient will be higher at this location and that the facade parts that receive most WDR might also experience a higher drying rate.     

Impact of irrigation water quality and envelope materials around drip line on emitter performance in subsurface drip irrigation

A field experiment was conducted to study the effect of water quality types of fresh water (FW) and treated wastewater (TWW) and envelope materials of coarse sand (CS), fine sand (FS), and control (CO) on emitter performance, dry matter yield (DMY), and water use efficiency (WUE) under subsurface drip irrigation.The main interaction effect of water quality type and envelope material on coefficient of variation, Christiansen uniformity coefficient, and emission uniformity was not significant (P < 0.05), but they have a significant effect on the average emitter discharge (Q_avg), DMY, and WUE. The means of Q_avg for FW with CS, FS, and CO were 7.13, 6.94, 2.65 L/h, and for TWW, they were 6.78, 6.84, and 2.35 L/h, respectively. The DMY under FW with CS, FS, and CO was 3083.87, 1367.95, and 417.45 kg/ha, and under the TWW, it was 2409.5, 1347.4, and 417 kg/ha, respectively.     

A novel flushing strategy for diatom bloom prevention in the lower-middle Hanjiang River

There are many concerns over the environmental consequences of river regulation in China, such as the Three Gorges Project and the South-to-North Water Diversion Project (SNWDP). In this study, however, we attempted to find the positive role of these constructions in solving environmental problems. We explored the possibility for preventing downstream diatom blooms by using the water storage in the Danjiangkou Reservoir. And we developed a flushing strategy accessing the proper flushing time and water quantity to control the diatom growth. First, we set up a Generalized Additive Model (GAM) to analyze the dynamics of the bloom-formation species, Stephanodiscus hantzschii, in response to the environmental variation. The model took into account the time lags between the biovolume and the environmental parameters. The model indicated that, air temperature explained the most variance in biovolume, followed by soluble reactive silicon (SRSi), turbidity, TP, dam release, PAR, pH and total nitrogen (GAM, R² = 0.759). Afterwards, we applied the model to a new predictive dataset, in which values were simulated according to the assumed dam release and air temperature. The GAM predicted fewer releases for flushing by using this dataset than the measured data, implying a prospect of saving water when using this strategy. Finally, we drew a contour map to present the operating procedure of this strategy. Our flushing strategy is to regulate the dam release above a critical value dependent on the air temperatures predicted over the following few days.