Model Output Uncertainty Research Articles

Can the young water fraction reduce predictive uncertainty in water transit time estimations?

Transit time distributions (TTDs) of streamflow are informative descriptors of catchment hydrological functioning and solute transport mechanisms. Conventional methods for estimating TTDs generally require model calibration against extensive tracer data time series, which are often limited to well-studied experimental catchments. We challenge this limitation and propose an alternative approach that uses the young water fraction (Fywobs), an increasingly used water age metric which represents the proportion of streamflow with a transit time younger than 2-3 months, and that can be robustly estimated with sparsely measured tracer data. To this end, we conducted a proof of concept study by modelling TTDs using StorAge Selection (SAS) functions with oxygen isotopes (δ18O) measurements for 23 diverse catchments in Germany. In a Monte-Carlo approach, we computed the (averaged) marginal TTDs of a prior parameter distribution and derived a model-based Fyw (Fywsim). We compared Fywsim with Fywobs, obtained from δ18O measurements, and constrained the prior SAS parameters distribution. Subsequently, we derived a posterior distribution of parameters and resulting model simulations. Our findings showed that using Fywobs to constrain the model effectively reduced parameter equifinality and simulation uncertainty. However, the value of Fywobs on reducing model uncertainty varied across sites, with larger values (Fywobs≥0.10) leading to simulations with a narrower uncertainty band and higher model efficiency, whilst smaller values (Fywobs≤0.05) had limited influence on reducing model output uncertainty. We discussed the potential and limitations of combining SAS functions with Fywobs, and considered broader implications of this approach for enhancing our understanding of catchment functioning and water quality status.

Efficient uncertainty quantification in a spatially multiscale model of pulmonary arterial and venous hemodynamics.

Pulmonary hypertension (PH) is a debilitating disease that alters the structure and function of both the proximal and distal pulmonary vasculature. This alters pressure-flow relationships in the pulmonary arterial and venous trees, though there is a critical knowledge gap in the relationships between proximal and distal hemodynamics in disease. Multiscale computational models enable simulations in both the proximal and distal vasculature. However, model inputs and measured data are inherently uncertain, requiring a full analysis of the sensitivity and uncertainty of the model. Thus, this study quantifies model sensitivity and output uncertainty in a spatially multiscale, pulse-wave propagation model of pulmonary hemodynamics. The model includes fifteen proximal arteries and twelve proximal veins, connected by a two-sided, structured tree model of the distal vasculature. We use polynomial chaos expansions to expedite sensitivity and uncertainty quantification analyses and provide results for both the proximal and distal vasculature. We quantify uncertainty in blood pressure, blood flow rate, wave intensity, wall shear stress, and cyclic stretch. The latter two are important stimuli for endothelial cell mechanotransduction. We conclude that, while nearly all the parameters in our system have some influence on model predictions, the parameters describing the density of the microvascular beds have the largest effects on all simulated quantities in both the proximal and distal arterial and venous circulations.

Uncertainty Quantification of a Machine Learning Subgrid‐Scale Parameterization for Atmospheric Gravity Waves

AbstractSubgrid‐scale processes, such as atmospheric gravity waves (GWs), play a pivotal role in shaping the Earth's climate but cannot be explicitly resolved in climate models due to limitations on resolution. Instead, subgrid‐scale parameterizations are used to capture their effects. Recently, machine learning (ML) has emerged as a promising approach to learn parameterizations. In this study, we explore uncertainties associated with a ML parameterization for atmospheric GWs. Focusing on the uncertainties in the training process (parametric uncertainty), we use an ensemble of neural networks to emulate an existing GW parameterization. We estimate both offline uncertainties in raw NN output and online uncertainties in climate model output, after the neural networks are coupled. We find that online parametric uncertainty contributes a significant source of uncertainty in climate model output that must be considered when introducing NN parameterizations. This uncertainty quantification provides valuable insights into the reliability and robustness of ML‐based GW parameterizations, thus advancing our understanding of their potential applications in climate modeling.

Maize yield prediction and condition monitoring at the sub-county scale in Kenya: synthesis of remote sensing information and crop modeling

Agricultural production assessments are crucial for formulating strategies for closing yield gaps and enhancing production efficiencies. While in situ crop yield measurements can provide valuable and accurate information, such approaches are costly and lack scalability for large-scale assessments. Therefore, crop modeling and remote sensing (RS) technologies are essential for assessing crop conditions and predicting yields at larger scales. In this study, we combined RS and a crop growth model to assess phenology, evapotranspiration (ET), and yield dynamics at grid and sub-county scales in Kenya. We synthesized RS information from the Food and Agriculture Organization (FAO) Water Productivity Open-access portal (WaPOR) to retrieve sowing date information for driving the model simulations. The findings showed that grid-scale management information and progressive crop growth could be accurately derived, reducing the model output uncertainties. Performance assessment of the modeled phenology yielded satisfactory accuracies at the sub-county scale during two representative seasons. The agreement between the simulated ET and yield was improved with the combined RS-crop model approach relative to the crop model only, demonstrating the value of additional large-scale RS information. The proposed approach supports crop yield estimation in data-scarce environments and provides valuable insights for agricultural resource management enabling countermeasures, especially when shortages are perceived in advance, thus enhancing agricultural production.

Using Median Point in Keeling Plot to Reduce the Uncertainty of the Isotopic Composition of Evapotranspiration

Abstract The isotopic composition of evapotranspiration δET is a crucial parameter in isotope-based evapotranspiration (ET) partitioning and moisture recycling studies. The Keeling plot method is the most prevalent method to calculate δET, though it contains large extrapolated uncertainties from the least squares regression. Traditional Keeling regression uses the mean point of individual measurements. Here, a modified Keeling plot framework was proposed using the median point of individual measurements. We tested the δET uncertainty using the mean point [σET (mean)] and median point [σET (median)]. Multiple resolutions of input and output data from six independent sites were used to test the performance of the two methods. The σET (mean) would be greater than σET (median) when the mean value of inverse vapor concentration () is greater than the median value of inverse vapor concentration []. When applying the filter of r2 > 0.8, around 70% of σET (mean) was greater than σET (median). This phenomenon might be due to the normality of the vapor concentration Cυ producing the asymmetric distribution of 1/Cυ. The median method could perform significantly better than the mean method when inputting high-resolution measurements (e.g., 1 Hz) and when the water vapor concentration Cυ is relatively low. Compared to the mean method, applying the median method could on average reduce 6.88% of ET partitioning uncertainties and could on average reduce 9.00% of moisture recycling uncertainties. This study provided a new insight of the Keeling plot method and emphasized handling model output uncertainty from multiple perspectives instead of only from input parameters.

Multi-objective calibration and uncertainty analysis for the event-based modelling of flash floods

ABSTRACT This study investigates the best approach to calibrate an event-based conceptual Hydrologiska Byråns Vattenbalansavdelning (HBV) model, comparing different trials of single-objective, single-event multi-objective (SEMO), and multi-event-multi-objective (MEMO) model calibrations using root mean squared error (RMSE), Nash-Sutcliffe model efficiency coefficient (NSE), and Bias as objective functions. Model performance was validated for several peak events via 90% confidence interval (CI)-based output uncertainty quantification of relative error of discharges. Multi-objective optimization yielded more accurate and robust solutions compared to single-objective calibrations. Ensembles of Pareto solutions from the multi-objective calibrations better characterized the flood peaks within the uncertainty intervals. MEMO calibration exhibited lower uncertainties and better prediction of peak events versus SEMO calibration. Moreover, the MEMO_6D (six-dimensional) approach outperformed the SEMO_3D and MEMO_3D in capturing the larger peak events. This study suggests that the MEMO_6D is the best approach for predicting large flood events with lower model output uncertainties when the calibration is performed with a better combination of peak events.

Sensitivity of long-term productivity estimations in mixed forests to uncertain parameters related to fine roots

Forest growth models are increasingly being used in forestry and ecology research as predictive tools to help developing practical guidelines and to improve understanding of the drivers of forest ecosystem functioning. Models are usually calibrated using parameters directly obtained or estimated from empirical field observation, and hence are subject to uncertainty. Thus, output accuracy depends on input parameters precision and on how influential is each parameter on model behaviour. Hence, it is important to analyse parameter-related uncertainty and its effects on model outputs. This can be done by performing sensitivity analyses, which allow to explore the influence of one or several calibration parameters on model outputs. As studies on tree root parameters are particularly scarce, the aim of the present work was to evaluate the influence of parameters related to fine roots on estimations of long-term forest growth patterns in pure and mixed forests, using FORECAST (a hybrid forest growth model) as a virtual lab. The fine root parameters assessed were biomass, turnover rate, and nitrogen content. The analysis was performed by simulating monospecific stands of two contrasting species (Pinus sylvestris L. and Fagus sylvatica L.), and mixed stands formed by both species. In all cases, FORECAST showed good capability to contain uncertainty propagation during the first and middle stages of stand development (<40 years). After that moment, model output uncertainty steadily increased, but it reached different maximum uncertainty levels depending on stand type. Simulations of the less nutrient demanding P. sylvestris manifested very little sensitivity when growing in monospecific stands. However, F. sylvatica monospecific stands showed intermediate sensitivity, but significant species interactions occurred in mixed stands that determined the biggest impact detected of uncertainty related to fine root parameters over model outputs. All things considered, FORECAST displayed an interesting capability to capture some of the interspecific interactions that are key in mixed forests functioning. Our results suggest an acceptable model performance under uncertain parameterization but also caution against expecting accurate quantitative estimations of forest growth, especially when considering long-term scenarios in complex mixed stands.

OP46 The Decision Uncertainty Toolkit: Risk Measures And Visual Outputs To Support Health Technology Decision-making During Public Health Crises

IntroductionDuring public health crises such as the COVID-19 pandemic, decision-makers have relied on infectious disease models to predict and estimate the impact of various health technologies. The difficulties associated with capturing and representing uncertainty using infectious disease models leads to a high risk of making decisions that are misaligned to policy objectives. Even when uncertainty is adequately captured in the analysis, the tools for communicating the risks and harms of making wrong decisions have proved inadequate, which can lead to the suboptimal adoption of critical health technologies including vaccines and antivirals. We aim to adapt and extend health economic methods for the characterization, estimation, and communication of uncertainty to infectious disease modeling.MethodsEconomic and infectious disease models share many features, including the comparison of policy alternatives on outcomes important to decision-makers (such as hospital census, total infections), but each takes a different approach to analysis of uncertainty. We extend best practices from health economics to infectious disease modeling and develop a suite of tools and visualization techniques which represent parameter uncertainty and the risk these unknowns present to decision-makers.ResultsIn consultation with decision-makers and infectious disease modeling experts we developed the ‘Decision Uncertainty Toolkit’ of model outputs and visuals. Visual tools for uncertainty are developed to: (i) accurately capture uncertainty in key infectious disease model outputs, and (ii) support intuitive and direct interpretation by infectious disease modelers and decision-makers. We also developed quantitative measures for the downside risk of policy alternatives, specified to capture both the probability and magnitude of losses relative to policy targets for a range of infectious disease model outputs. Together, these outputs can support decision-making by quantifying outcome uncertainty and the risks associated with policy alternatives.ConclusionsWe developed the toolkit visuals and risk measures alongside infectious disease modelers and decision makers. The toolkit is designed to improve decision-maker understanding of decision risk in order to improve outcomes during future public health crises.

Comparative Sensitivity Analysis of Hydrology and Relative Corn Yield under Different Subsurface Drainage Design Using DRAINMOD

DRAINMOD is a process-based hydrologic model used to analyze the effectiveness of various drainage systems and management strategies. In this study, a sensitivity analysis of DRAINMOD hydrologic parameters for two different field settings located at Champaign, Illinois, was performed to determine the most sensitive parameters that affect the subsurface flow and relative productivity of corn. Latin-Hypercube One-Factor-at-a-Time (LH-OAT) was used to determine the sensitivity index of 17 parameters for six objective functions for daily flow, water balance, and relative yield for the productivity of corn. The results indicated that flow and yield were highly sensitive to drainage design parameters such as drainage depth and spacing. Winter flow and the water balance were sensitive to soil thermal conductivity parameters; however, they had no impact on the relative corn yield. The significant difference in sensitivity of the two fields was observed in the hydraulic conductivity of soil layers due to varying thicknesses for different soil types. This study highlights the need for more careful calibration of these sensitive parameters to reduce equifinality and model output uncertainty and appropriate drainage design for optimizing crop productivity and drainage outflow.

Investigating the effects of APSIM model configuration on model outputs across different environments

IntroductionSoil type plays a major role in nutrient dynamics and soil water which impacts crop growth and yield. The influence of soil characteristics on crop growth is usually evaluated through field experimentation (in the short term) and through crop-soil modelling (in the long-term). However, there has been limited research which has looked at the effect of model structural uncertainty of model outputs in different soil types.MethodsTo analyze the impact of soil inputs on model structural uncertainty, we developed eight model structures (a combination of two crop models, two soil water models and two irrigation models) within the Agricultural Production Systems sIMulator (APSIM) across three soil types (Ferralsols, Alisols and Chernozems). By decomposing the mean proportion of variance and simulated values of the model outputs (yield, irrigation, drainage, nitrogen leaching and partial gross margin) we identified the influence of soil type on the magnitude of model structural uncertainty.ResultsFor all soil types, crop model was the most significant source of structural uncertainty, contributing &amp;gt;60% to variability for most modelled variables, except irrigation demand which was dominated by the choice of irrigation model applied. Relative to first order interactions, there were minimal (&amp;lt;12%) contributions to uncertainty from the second order interactions (i.e., inter-model components). We found that a higher mean proportion of variance does not necessarily imply a high magnitude of uncertainty in actual values. Despite the significant impact of the choice of crop model on yield and PGM variance (contributing over 90%), the small standard deviations in simulated yield (ranging from 0.2 to 1 t ha-1) and PGM (ranging from 50.6 to 374.4 USD ha-1) compared to the mean values (yield: 14.6 t ha-1, PGM: 4901 USD ha-1) indicate relatively low actual uncertainty in the values. Similarly, the choice of irrigation model had a contribution of over 45% to variance, but the relatively small standard deviations ranging from 11 to 33.3 mm compared to the overall mean irrigation of 500 mm suggest low actual uncertainty in the values. In contrast, for the environmental variables- drainage and nitrogen leaching, the choice of crop model had contributions of more than 60% and 70% respectively, yet the relatively large standard deviations ranging from 7.1 to 30.6 mm and 0.6 to 7.7 kg ha-1 respectively, compared to the overall mean values of drainage (44.4 mm) and nitrogen leaching (3.2 kg ha-1), indicate significant actual uncertainty.DiscussionWe identified the need to include not only fractional variance of model uncertainty, but also magnitude of the contribution in measured units (e.g. t ha-1, mm, kg ha-1, USD ha-1) for crop model uncertainty assessments to provide more useful agronomic or policy decision-making information. The findings of this study highlight the sensitivity of agricultural models to the impacts of moisture availability, suggesting that it is important to give more attention to structural uncertainty when modelling dry/wet conditions depending on the output analyzed.

The Impact of Meteorological Forcing Uncertainty on Hydrological Modeling: A Global Analysis of Cryosphere Basins

AbstractMeteorological forcing is a major source of uncertainty in hydrological modeling. The recent development of probabilistic large‐domain meteorological data sets enables convenient uncertainty characterization, which however is rarely explored in large‐domain research. This study analyzes how uncertainties in meteorological forcing data affect hydrological modeling in 289 representative cryosphere basins by forcing the Structure for Unifying Multiple Modeling Alternatives (SUMMA) and mizuRoute models with precipitation and air temperature ensembles from the Ensemble Meteorological Data set for Planet Earth (EM‐Earth). EM‐Earth probabilistic estimates are used in ensemble simulation for uncertainty analysis. The results reveal the magnitude, spatial distribution, and scale effect of uncertainties in meteorological, snow, runoff, soil water, and energy variables. There are three main findings. (a) The uncertainties in precipitation and temperature lead to substantial uncertainties in hydrological model outputs, some of which exceed 100% of the magnitude of the output variables themselves. (b) The uncertainties of different variables show distinct scale effects caused by spatial averaging or temporal averaging. (c) Precipitation uncertainties have the dominant impact for most basins and variables, while air temperature uncertainties are also nonnegligible, sometimes contributing more to modeling uncertainties than precipitation uncertainties. We find that three snow‐related variables (snow water equivalent, snowfall amount, and snowfall fraction) can be used to estimate the impact of air temperature uncertainties for different model output variables. In summary, this study provides insight into the impact of probabilistic data sets on hydrological modeling and quantifies the uncertainties in cryosphere basin modeling that stem from the meteorological forcing data.

Global Sensitivity Analysis of a Dam Breaching Model: To Which Extent Is Parameter Sensitivity Case‐Dependent?

AbstractFailure of dams and dikes often leads to devastating consequences in protected areas. Numerical models are instrumental tools to assess flood risk and guide emergency management, but numerous uncertainties affect model outcomes. Identifying uncertain model input parameters that induce high uncertainties in model outputs is essential. This paper focuses on two model outputs: the maximum breach discharge and the time to reach this peak. Using our implementation of a simplified physically based dam breaching model developed by Wu (2016), a global sensitivity analysis was conducted based on Sobol indices of total order. Unlike in most previous studies, many different configurations (twenty‐seven), both at laboratory and field scales, were considered. For each of them, input variables were ranked according to the significance of the contribution of their uncertainty to the output variability, and the dependency between reference configurations and sensitivity analysis results was highlighted. Depending on the considered case study, input parameter uncertainties with the largest impact on output variability were different, and so was the magnitude of output uncertainties. We demonstrate that sensitivity analysis results obtained for a specific configuration cannot be transferred as it to other configurations. Finally, sensitivity and uncertainty quantification results were combined in a decision tree to determine which input parameter uncertainty is the most critical in each configuration and what standard deviation in the output variables is expected. The global sensitivity analysis procedure presented here may apply to a wide variety of models in environmental sciences.

Uncertainty quantification for computational modelling of laser powder bed fusion

Additive manufacturing (AM) may have many advantages over traditional casting and wrought methods, but our understanding of the various processes is still limited. Computational models are useful to study and isolate underlying physics and improve our understanding of the AM process-microstructure-property relations. However, these models necessarily rely on simplifications and parameters of uncertain value. These assumptions reduce the overall reliability of the predictive capabilities of these models, so it is important to estimate the uncertainty in model output. In doing so, we quantify the effect of model limitations and identify potential areas of improvement, a procedure made possible by uncertainty quantification (UQ). Here we highlight recent work which coupled and propagated statistical and systematic uncertainties from a melt pool transport model based in OpenFOAM, through a grain scale cellular automaton code. We demonstrate how a UQ framework can identify model parameters which most significantly impact the reliability of model predictions through both models and thus provide insight for future improvements in the models and suggest measurements to reduce output uncertainty.

Bayesian Calibration of Computational Knee Models to Estimate Subject-Specific Ligament Properties, Tibiofemoral Kinematics, and Anterior Cruciate Ligament Force With Uncertainty Quantification.

High-grade knee laxity is associated with early anterior cruciate ligament (ACL) graft failure, poor function, and compromised clinical outcome. Yet, the specific ligaments and ligament properties driving knee laxity remain poorly understood. We described a Bayesian calibration methodology for predicting unknown ligament properties in a computational knee model. Then, we applied the method to estimate unknown ligament properties with uncertainty bounds using tibiofemoral kinematics and ACL force measurements from two cadaver knees that spanned a range of laxities; these knees were tested using a robotic manipulator. The unknown ligament properties were from the Bayesian set of plausible ligament properties, as specified by their posterior distribution. Finally, we developed a calibrated predictor of tibiofemoral kinematics and ACL force with their own uncertainty bounds. The calibrated predictor was developed by first collecting the posterior draws of the kinematics and ACL force that are induced by the posterior draws of the ligament properties and model parameters. Bayesian calibration identified unique ligament slack lengths for the two knee models and produced ACL force and kinematic predictions that were closer to the corresponding in vitro measurement than those from a standard optimization technique. This Bayesian framework quantifies uncertainty in both ligament properties and model outputs; an important step towards developing subject-specific computational models to improve treatment for ACL injury.

Addressing complexities in MPM modeling of calibration chamber cone penetrometer tests in homogenous and highly interlayered soils

Cone penetrometer tests (CPTs) are used to characterize soil for a variety of geotechnical engineering applications, including earthquake-induced liquefaction triggering assessment. Numerical modeling of CPTs is frequently used to better understand soil behavior, soil-penetrometer interaction, and engineering estimates made from CPT data. However, calibrating and validating numerical CPT simulations with experimental calibration chamber (CC) data can be challenging. Specifically, uncertainties in the interpretation of laboratory strength and compression data compound with uncertainties in the CC testing and the assumptions made when developing the numerical model. This article provides a comprehensive review of uncertainties in the calibration and validation of CPT numerical simulations performed in homogenous sand, homogenous clay, and layered sand-clay soil profiles, comparing numerical results with well-documented experimental calibration chamber tests performed at Deltares. In particular, the Material Point Method (MPM) is used to perform the numerical analyses. A framework is presented to assess how uncertainty in the numerical model output is attributed to each input parameter. It is demonstrated that uncertainties can be explored numerically. Finally, recommendations for future experimental and numerical studies of CPTs are provided.

Source apportionment of polychlorinated biphenyls (PCBs) using different receptor models: A case study on sediment from the Portland Harbor Superfund Site (PHSS), Oregon, USA

Multivariate modelling techniques are used by a wide variety of investigations in environmental chemistry. It is surprisingly rare for studies to show a detailed understanding of uncertainties created by modelling or how uncertainties in chemical analysis impact model outputs. It is common to use untrained multivariate models for receptor modelling. These models produce a slightly different output each time they are run. The fact that a single model can provide different results is rarely acknowledged. In this manuscript, we attempt to address this by investigating differences that can be generated using four different receptor models (NMF, ALS, PMF & PVA) to perform source apportionment of polychlorinated biphenyls (PCBs) in surface sediments from Portland Harbor. Results showed that models generally had a strong agreement and identified the same main signatures that represented commercial PCB mixtures, however, subtle differences were identified by; different models, same models but with a different number of end members (EM), and the same model with the same number of end members. As well as identifying different Aroclor-like signatures, the relative proportion of these sources also varied. Depending on which method is selected it may have a significant impact on conclusions of a scientific report or litigation case and ultimately, allocation on who is responsible for paying for remediation. Therefore, care must be taken to understand these uncertainties to select a method that produces consistent results with end members that can be chemically explained. We also investigated a novel approach to use our multivariate models to identify inadvertent sources of PCBs. By using a residual plot produced from one of our models (NMF) we were able to suggest the presence of approximately 30 different potentially inadvertently produced PCBs which account for 6.6 % of the total PCBs in Portland Harbor sediments.

Demand Forecasting Approaches for New Contraceptive Technologies: A Landscape Review and Recommendations for Alignment.

Market size estimations and demand forecasts use a variety of methodological approaches to inform decision-making around new (and lesser-used) contraceptive methods. For contraceptive products already available at scale in a market, historical procurement and consumption data can help to inform these forecasts. However, little published guidance is available on appropriate approaches to estimating contraceptive demand in the absence of historical data. This landscape review aimed to describe the variety of approaches for modeling demand for new contraceptive methods, highlight opportunities for alignment around forecasting practices, and make recommendations to support more accurate forecasting and sound decision-making based on forecasts. We used the published scientific and gray literature to inform the development of a semistructured guide for key informant interviews. We conducted interviews with 29 experts representing a spectrum of interests in market size estimation and demand forecasting for new contraceptive methods (e.g., ministries of health, donors, manufacturers, technical assistance providers, and demand forecasting specialists). We coded notes from the interviews using thematic content analysis. The purposes of market size estimation and demand forecasting for new contraceptive methods vary widely, as do associated model inputs and outputs. Key informants revealed a need for more standardized language around market size estimation and demand forecasting and highlighted key recommendations: select models that are fit-for-purpose, clearly articulate assumptions and uncertainty in model outputs, consider a broad range of contraceptive options in a forecast to capture the complete contraceptive supply environment, and perform a reality check of results and refresh assumptions. We recommend following a simple decision pathway to ensure that forecasts are fit-for-purpose, with appropriate inputs, outputs, and assumptions clearly articulated. Common pitfalls around overestimating demand should be avoided. Incorporating best practices into forecasting exercises will ensure that models are useful for the stakeholders.

Computationally-feasible uncertainty quantification in model-based landslide risk assessment

Introduction: Increasing complexity and capacity of computational physics-based landslide run-out modelling yielded highly efficient model-based decision support tools, e.g. landslide susceptibility or run-out maps, or geohazard risk assessments. A reliable, robust and reproducible development of such tools requires a thorough quantification of uncertainties, which are present in every step of computational workflow from input data, such as topography or release zone, to modelling framework used, e.g. numerical error.Methodology: Well-established methods from reliability analysis such as Point Estimate Method (PEM) or Monte Carlo Simulations (MCS) can be used to investigate the uncertainty of model outputs. While PEM requires less computational resources, it does not capture all the details of the uncertain output. MCS tackles this problem, but creates a computational bottleneck. A comparative study is presented herein by conducting multiple forward simulations of landslide run-out for a synthetic and a real-world test case, which are used to construct Gaussian process emulators as a surrogate model to facilitate high-throughput tasks.Results: It was demonstrated that PEM and MCS provide similar expectancies, while the variance and skewness differ, in terms of post-processed scalar outputs, such as impact area or a point-wise flow height. Spatial distribution of the flow height was clearly affected by the choice of method used in uncertainty quantification.Discussion: If only expectancies are to be assessed then one can work with computationally-cheap PEM, yet MCS has to be used when higher order moments are needed. In that case physics-based machine learning techniques, such as Gaussian process emulation, provide strategies to tackle the computational bottleneck. It can be further suggested that computational-feasibility of MCS used in landslide risk assessment can be significantly improved by using surrogate modelling. It should also be noted that the gain in compute time by using Gaussian process emulation critically depends on the computational effort needed to produce the training dataset for emulation by conducting simulations.

UncertainSCI: Uncertainty quantification for computational models in biomedicine and bioengineering

Background:Computational biomedical simulations frequently contain parameters that model physical features, material coefficients, and physiological effects, whose values are typically assumed known a priori. Understanding the effect of variability in those assumed values is currently a topic of great interest. A general-purpose software tool that quantifies how variation in these parameters affects model outputs is not broadly available in biomedicine. For this reason, we developed the ‘UncertainSCI’ uncertainty quantification software suite to facilitate analysis of uncertainty due to parametric variability. Methods:We developed and distributed a new open-source Python-based software tool, UncertainSCI, which employs advanced parameter sampling techniques to build polynomial chaos (PC) emulators that can be used to predict model outputs for general parameter values. Uncertainty of model outputs is studied by modeling parameters as random variables, and model output statistics and sensitivities are then easily computed from the emulator. Our approaches utilize modern, near-optimal techniques for sampling and PC construction based on weighted Fekete points constructed by subsampling from a suitably randomized candidate set. Results:Concentrating on two test cases—modeling bioelectric potentials in the heart and electric stimulation in the brain—we illustrate the use of UncertainSCI to estimate variability, statistics, and sensitivities associated with multiple parameters in these models. Conclusion:UncertainSCI is a powerful yet lightweight tool enabling sophisticated probing of parametric variability and uncertainty in biomedical simulations. Its non-intrusive pipeline allows users to leverage existing software libraries and suites to accurately ascertain parametric uncertainty in a variety of applications.

Uncertainty Analysis of HYDRUS-1D Model to Simulate Soil Salinity Dynamics under Saline Irrigation Water Conditions Using Markov Chain Monte Carlo Algorithm

Utilizing degraded quality waters such as saline water as irrigation water with proper management methods such as leaching application is a potential answer to water scarcity in agricultural systems. Leaching application requires understanding the relationship between the amount of irrigation water and its quality with the dynamic of salts in the soil. The HYDRUS-1D model can simulate the dynamic of soil salinity under saline water irrigation conditions. However, these simulations are subject to uncertainty. A study was conducted to assess the uncertainty of the HYDRUS-1D model parameters and outputs to simulate the dynamic of salts under saline water irrigation conditions using the Markov Chain Monte Carlo (MCMC) based Metropolis-Hastings algorithm in the R-Studio environment. Results indicated a low level of uncertainty in parameters related to the advection term (water movement simulation) and water stress reduction function for root water uptake in the solute transport process. However, a higher level of uncertainty was detected for dispersivity and diffusivity parameters, possibly because of the study’s scale or some error in initial or boundary conditions. The model output (predictive) uncertainty showed a high uncertainty in dry periods compared to wet periods (under irrigation or rainfall). The uncertainty in model parameters was the primary source of total uncertainty in model predictions. The implementation of the Metropolis-Hastings algorithm for the HYDRUS-1D was able to conveniently estimate the residual water content (θr) value for the water simulation processes. The model’s performance in simulating soil water content and soil water electrical conductivity (ECsw) was good when tested with the 50% quantile of the posterior distribution of the parameters. Uncertainty assessment in this study revealed the effectiveness of the Metropolis-Hastings algorithm in exploring uncertainty aspects of the HYDRUS-1D model for reproducing soil salinity dynamics under saline water irrigation at a field scale.