Andreas Scheidegger

Predation influences the structure of biofilm developed on ultrafiltration membranes

This study investigates the impact of predation by eukaryotes on the development of specific biofilm structures in gravity-driven dead-end ultrafiltration systems. Filtration systems were operated under ultra-low pressure conditions (65 mbar) without the control of biofilm formation. Three different levels of predation were evaluated: (1) inhibition of eukaryotic organisms, (2) addition of cultured protozoa (Tetrahymena pyriformis), and (3) no modification of microbial community as a control. The system performance was evaluated based on permeate flux and structures of the biofilm. It was found that predation had a significant influence on both the total amount and also the structure of the biofilm. An open and heterogeneous structure developed in systems with predation whereas a flat, compact, and thick structure that homogeneously covered the membrane surface developed in absence of predation. Permeate flux was correlated with the structure of the biofilm with increased fluxes for smaller membrane coverage. Permeate fluxes in the presence or absence of the predators was 10 and 5 L m(-2) h(-1), respectively. It was concluded that eukaryotic predation is a key factor influencing the performance of gravity-driven ultrafiltration systems.

Strategic rehabilitation planning of piped water networks using multi-criteria decision analysis

To overcome the difficulties of strategic asset management of water distribution networks, a pipe failure and a rehabilitation model are combined to predict the long-term performance of rehabilitation strategies. Bayesian parameter estimation is performed to calibrate the failure and replacement model based on a prior distribution inferred from three large water utilities in Switzerland. Multi-criteria decision analysis (MCDA) and scenario planning build the framework for evaluating 18 strategic rehabilitation alternatives under future uncertainty. Outcomes for three fundamental objectives (low costs, high reliability, and high intergenerational equity) are assessed. Exploitation of stochastic dominance concepts helps to identify twelve non-dominated alternatives and local sensitivity analysis of stakeholder preferences is used to rank them under four scenarios. Strategies with annual replacement of 1.5-2% of the network perform reasonably well under all scenarios. In contrast, the commonly used reactive replacement is not recommendable unless cost is the only relevant objective. Exemplified for a small Swiss water utility, this approach can readily be adapted to support strategic asset management for any utility size and based on objectives and preferences that matter to the respective decision makers.

Network condition simulator for benchmarking sewer deterioration models.

An accurate description of aging and deterioration of urban drainage systems is necessary for optimal investment and rehabilitation planning. Due to a general lack of suitable datasets, network condition models are rarely validated, and if so with varying levels of success. We therefore propose a novel network condition simulator (NetCoS) that produces a synthetic population of sewer sections with a given condition-class distribution. NetCoS can be used to benchmark deterioration models and guide utilities in the selection of appropriate models and data management strategies. The underlying probabilistic model considers three main processes: a) deterioration, b) replacement policy, and c) expansions of the sewer network. The deterioration model features a semi-Markov chain that uses transition probabilities based on user-defined survival functions. The replacement policy is approximated with a condition-class dependent probability of replacing a sewer pipe. The model then simulates the course of the sewer sections from the installation of the first line to the present, adding new pipes based on the defined replacement and expansion program. We demonstrate the usefulness of NetCoS in two examples where we quantify the influence of incomplete data and inspection frequency on the parameter estimation of a cohort survival model and a Markov deterioration model. Our results show that typical available sewer inventory data with discarded historical data overestimate the average life expectancy by up to 200 years. Although NetCoS cannot prove the validity of a particular deterioration model, it is useful to reveal its possible limitations and shortcomings and quantifies the effects of missing or uncertain data. Future developments should include additional processes, for example to investigate the long-term effect of pipe rehabilitation measures, such as inliners.

Combining expert knowledge and local data for improved service life modeling of water supply networks

The presented approach aims to overcome the scarce data problem in service life modeling of water networks by combining subjective expert knowledge and local replacement data. A procedure to elicit imprecise quantile estimates of survival functions from experts, considering common cognitive biases, was developed and applied. The individual expert priors of the parameters of the service life distribution are obtained by regression over the stated distribution quantiles and aggregated into a single prior distribution. Furthermore, a likelihood function for the commonly encountered censored and truncated pipe replacement data is formulated. The suitability of the suggested Bayesian approach based on elicitation data from eight experts and real network data is demonstrated. Robust parameter estimates could be derived in data situations where frequentist maximum likelihood estimation is unsatisfactory, and to show how the consideration of imprecision and in-between-variance of experts improves posterior inference.

pH-Dependent Biotransformation of Ionizable Organic Micropollutants in Activated Sludge

Removal of micropollutants (MPs) during activated sludge treatment can mainly be attributed to biotransformation and sorption to sludge flocs, whereby the latter process is known to be of minor importance for polar organic micropollutants. In this work, we investigated the influence of pH on the biotransformation of MPs with cationic-neutral speciation in an activated sludge microbial community. We performed batch biotransformation, sorption control, and abiotic control experiments for 15 MPs with cationic-neutral speciation, one control MP with neutral-anionic speciation, and two neutral MPs at pHs 6, 7, and 8. Biotransformation rate constants corrected for sorption and abiotic processes were estimated from measured concentration time series with Bayesian inference. We found that biotransformation is pH-dependent and correlates qualitatively with the neutral fraction of the ionizable MPs. However, a simple speciation model based on the assumption that only the neutral species is efficiently taken up and biotransformed by the cells tends to overpredict the effect of speciation. Therefore, additional mechanisms such as uptake of the ionic species and other more complex attenutation mechanisms are discussed. Finally, we observed that the sorption coefficients derived from our control experiments were small and showed no notable pH-dependence. From this we conclude that pH-dependent removal of polar, ionizable organic MPs in activated sludge systems is less likely an effect of pH-dependent sorption but rather of pH-dependent biotransformation. The latter has the potential to cause marked differences in the removal of polar, ionizable MPs at different operational pHs during activated sludge treatment.

Sewer deterioration modeling with condition data lacking historical records.

Accurate predictions of future conditions of sewer systems are needed for efficient rehabilitation planning. For this purpose, a range of sewer deterioration models has been proposed which can be improved by calibration with observed sewer condition data. However, if datasets lack historical records, calibration requires a combination of deterioration and sewer rehabilitation models, as the current state of the sewer network reflects the combined effect of both processes. Otherwise, physical sewer lifespans are overestimated as pipes in poor condition that were rehabilitated are no longer represented in the dataset. We therefore propose the combination of a sewer deterioration model with a simple rehabilitation model which can be calibrated with datasets lacking historical information. We use Bayesian inference for parameter estimation due to the limited information content of the data and limited identifiability of the model parameters. A sensitivity analysis gives an insight into the models robustness against the uncertainty of the prior. The analysis reveals that the model results are principally sensitive to the means of the priors of specific model parameters, which should therefore be elicited with care. The importance sampling technique applied for the sensitivity analysis permitted efficient implementation for regional sensitivity analysis with reasonable computational outlay. Application of the combined model with both simulated and real data shows that it effectively compensates for the bias induced by a lack of historical data. Thus, the novel approach makes it possible to calibrate sewer pipe deterioration models even when historical condition records are lacking. Since at least some prior knowledge of the model parameters is available, the strength of Bayesian inference is particularly evident in the case of small datasets.

Statistical failure models for water distribution pipes - A review from a unified perspective.

This review describes and compares statistical failure models for water distribution pipes in a systematic way and from a unified perspective. The way the comparison is structured provides the information needed by scientists and practitioners to choose a suitable failure model for their specific needs. The models are presented in a novel framework consisting of: 1) Clarification of model assumptions. The models originally formulated in different mathematical forms are all presented as failure rate. This enables to see differences and similarities across the models. Furthermore, we present a new conceptual failure rate that an optimal model would represent and to which the failure rate of each model can be compared. 2) Specification of the detailed data assumptions required for unbiased model calibration covering the structure and completeness of the data. 3) Presentation of the different types of probabilistic predictions available for each model. Nine different models and their variations or further developments are presented in this review. For every model an overview of its applications published in scientific journals and the available software implementations is provided. The unified view provides guidance to model selection. Furthermore, the model comparison presented herein enables to identify areas where further research is needed.

A simulated annealing approach to approximate Bayes computations

Approximate Bayes computations (ABC) are used for parameter inference when the likelihood function of the model is expensive to evaluate but relatively cheap to sample from. In particle ABC, an ensemble of particles in the product space of model outputs and parameters is propagated in such a way that its output marginal approaches a delta function at the data and its parameter marginal approaches the posterior distribution. Inspired by Simulated Annealing, we present a new class of particle algorithms for ABC, based on a sequence of Metropolis kernels, associated with a decreasing sequence of tolerances w.r.t. the data. Unlike other algorithms, our class of algorithms is not based on importance sampling. Hence, it does not suffer from a loss of effective sample size due to re-sampling. We prove convergence under a condition on the speed at which the tolerance is decreased. Furthermore, we present a scheme that adapts the tolerance and the jump distribution in parameter space according to some mean-fields of the ensemble, which preserves the statistical independence of the particles, in the limit of infinite sample size. This adaptive scheme aims at converging as close as possible to the correct result with as few system updates as possible via minimizing the entropy production of the process. The performance of this new class of algorithms is compared against two other recent algorithms on two toy examples as well as on a real-world example from genetics.

Assessing Wastewater Micropollutant Loads with Approximate Bayesian Computations

Wastewater production, like many other engineered and environmental processes, is inherent stochastic in nature and requires the use of complex stochastic models, for example, to predict realistic patterns of down-the-drain chemicals or pharmaceuticals and personal care products. Up until now, a formal method of statistical inference has been lacking for many of those models, where explicit likelihood functions were intractable. In this Article, we investigate Approximate Bayesian Computation (ABC) methods to infer important parameters of stochastic environmental models. ABC methods have been recently suggested to perform model-based inference in a Bayesian setting when model likelihoods are analytically or computationally intractable and have not been applied to environmental systems analysis or water quality modeling before. In a case study, we investigate the performance of three different algorithms to infer the number of wastewater pulses contained in three high-resolution data series of benzotriazole and total nitrogen loads in sewers. We find that all algorithms perform well and that the uncertainty in the inferred number of corresponding wastewater pulses varies between 6% and 28%. In our case, the results are more sensitive to substance characteristics than to catchment properties. Although the application of ABC methods requires careful tuning and attention to detail, they have a great general potential to update stochastic model parameters with monitoring data and improve their predictive capabilities.

Assessing the performance of sewer rehabilitation on the reduction of infiltration and inflow.

Inflow and Infiltration (I/I) into sewer systems is generally unwanted, because, among other things, it decreases the performance of wastewater treatment plants and increases combined sewage overflows. As sewer rehabilitation to reduce I/I is very expensive, water managers not only need methods to accurately measure I/I, but also they need sound approaches to assess the actual performance of implemented rehabilitation measures. However, such performance assessment is rarely performed. On the one hand, it is challenging to adequately take into account the variability of influential factors, such as hydro-meteorological conditions. On the other hand, it is currently not clear how experimental data can indeed support robust evidence for reduced I/I. In this paper, we therefore statistically assess the performance of rehabilitation measures to reduce I/I. This is possible by using observations in a suitable reference catchment as a control group and assessing the significance of the observed effect by regression analysis, which is well established in other disciplines. We successfully demonstrate the usefulness of the approach in a case study, where rehabilitation reduced groundwater infiltration by 23.9%. A reduction of stormwater inflow of 35.7%, however, was not statistically significant. Investigations into the experimental design of monitoring campaigns confirmed that the variability of the data as well as the number of observations collected before the rehabilitation impact the detection limit of the effect. This implies that it is difficult to improve the data quality after the rehabilitation has been implemented. Therefore, future practical applications should consider a careful experimental design. Further developments could employ more sophisticated monitoring methods, such as stable environmental isotopes, to directly observe the individual infiltration components. In addition, water managers should develop strategies to effectively communicate statistically not significant I/I reduction ratios to decision makers.