Geology topics

Michael N. Fienen

Publications and source records attributed to Michael N. Fienen.

At least 19 recordsLinked to original sources

USGS2026

Inverse modeling for post-wildfire hydrologic process understanding over snow-dominated watersheds

Wildfire-induced changes to vegetation and soil properties can alter snowpack dynamics and hydrologic processes in ways that complicate predictions of post-wildfire water availability. To better understand these effects, we coupled the Parameter ESTimation in C++ (PEST++) software with the Weather Research & Forecasting Hydrologic model (WRF-Hydro) across three burned, snow-dominated watersheds in the western United States. We incorporated three burn-severity-based land cover categories into WRF-Hydro and used an iterative Ensemble Smoother (iES) within PEST++ to estimate model parameters controlling snow accumulation and melt, infiltration, and streamflow generation. This framework supported three complementary objectives: (1) assessing parameter sensitivities under pre- and post-wildfire conditions, (2) comparing pre- and post-wildfire parameter ensembles to identify wildfire-driven shifts in hydrologic processes, and (3) using counterfactual simulations to isolate the hydrologic response to wildfire from interannual climate variability. Our findings indicate that model parameter sensitivity varied by domain and shifted following wildfire. Maximum snow albedo and snow albedo decay were most influential in the higher-elevation East Troublesome domain, while the rain-snow partitioning temperature threshold dominated in the lower-elevation Chippy Creek and Las Conchas domains. Following wildfire, streamflow sensitivity shifted away from snowpack parameters toward soil hydraulic parameters, consistent with wildfire-driven reductions in evapotranspiration and increases in soil moisture. Incorporating wildfire effects led to complex changes in simulated snowpack accumulation and streamflow timing, generally leading to earlier snowmelt freshet initiation and cessation and reduced late season flows. Simulations were more sensitive to interannual climate variations than wildfire land cover changes. These results highlight both the potential and challenges of ensemble-based inverse modeling for understanding post-wildfire hydrologic change in snow-dominated watersheds.

Colorado, Montana, New Mexico↗

USGS2026

Multi-objective optimization of a hydro-economic model in an over-allocated agricultural basin

Groundwater depletion for agricultural irrigation poses significant environmental and economic challenges. This study introduces a proof-of-concept that combines hydro-economic modeling, scenario-based modeling, and multi-objective optimization to manage pumping curtailment in an over-allocated basin in the western United States. Three optimization scenarios were evaluated, each offering different degrees of management flexibility. Results reveal that scenarios with finer spatial resolution achieved greater environmental benefits per unit profit loss. Additionally, strategies allowing fractional reductions in curtailed wells–rather than complete shutdowns based on water rights seniority–substantially improved efficiency, highlighting the value of increased decision-making flexibility. Although scenario testing can aid stakeholder engagement and strategy exploration, multi-objective optimization provides a systematic framework to quantify tradeoffs between competing objectives. This combined approach demonstrates promise for building consensus and supporting the design of sustainable water management strategies that balance agricultural livelihoods with ecosystem preservation.

Oregon↗

USGS2026

Ensemble methods for history matching and uncertainty quantification with a watershed model

History matching of large hydrologic models is challenging due to data sparsity and non-unique process combinations (and associated parameters) that can produce similar model predictions. We develop an ensemble-based history matching (and uncertainty quantification) approach using an iterative ensemble smoother (iES) method for three cutouts of the National Hydrologic Model (NHM) and qualitatively compare the results and performance to the stepwise history matching approach. In the latter approach, subsets of parameters and observations were sequentially calibrated to a diverse range of observations to mitigate non-uniqueness and local minima. In iES, localization simulates the same causal connections between parameters and observations without the need (and computational cost) of sequential history matching steps. iES uses a weighted sum-of-squared-errors objective function which allows differential weighting of multiple data sources. Formal adoption of range observation also pushes results to within ranges of observation values rather than discrete values. Overall, the ensemble approach performs similarly to the stepwise approach. Both approaches performed poorly for the cutout representing a snowmelt-dominated watershed, indicating a structural issue in the process representation of the model. The main advantage of iES is quantification of uncertainty in both the history matching and the predictions of interest.

Journal of the American Water Resources Associatio↗

USGS2026

Teach me how to pycap: A high-capacity well decision support tool using analytical solutions in Python

Regulatory agencies in humid temperate environments rely on timely evaluations of streamflow depletion and drawdown to protect aquatic ecosystems and existing water users. Numerical models offer detailed insights, but their complexity and time demands often preclude their practical use in rapid decision-making. We present pycap-dss, an open-source Python package that implements a suite of analytical solutions for estimating streamflow depletion and drawdown. The tool supports superposition of multiple wells and time-varying pumping, enabling cumulative impact assessments in situations with multiple wells and streams. The software is modular and extensible, allowing users to interchange solutions or add new analytical methods. A YAML-based configuration supports batch processing of multiple wells, and an optional AnalysisProject class facilitates integration with regulatory workflows. Rigorous unit and regression testing ensures computational reliability, and continuous integration supports ongoing development. We demonstrate deterministic examples of drawdown where multiple solutions are readily compared and streamflow depletion with multiple wells in the Central Sands region of Wisconsin. We also show the value of Monte Carlo analyses of streamflow depletion in the same Central Sands example, leveraging computational efficiency to evaluate the uncertainty of individual and cumulative streamflow depletion calculations from over 200 high-capacity wells.

Wisconsin↗

USGS2025

Toward a new framework to evaluate process-based model configurations and quantify data worth prior to calibration

Model criticism, discrimination, and selection methods often rely on calibrated model outputs. Because calibration can be computationally expensive, model criticism can first be undertaken by assessing model outputs obtained from limited prior parameter ensembles. However, such prior-based methods are often heuristic and do not formalize the notion of balancing model consistency with data and model complexity (i.e., model adequacy). We present a new framework to discriminate among candidate models prior to calibration that formalizes prior-to-calibration model adequacy into a metric to implicitly balance prior model output data coverage with model complexity represented by prior output (co)variance. The prior model adequacy metric “Mahalanobis distance deviation” quantifies the deviation of (a) the set of squared Mahalanobis distances of data from a prior model output distribution from (b) the set of squared Mahalanobis distances of data from their own distribution. A new data worth metric “discernment value” is also presented which quantifies the value of data for screening less-adequate models prior to calibration. Discernment value is calculated from the change in variance of a weighted average of prior model outputs from all candidate models due to less-adequate model outputs receiving lower weight. The framework is demonstrated using a one-dimensional groundwater flow model with eight possible configurations. A synthetic data network is used to test the framework. Results show the framework identifies the candidate models most similar to the true model used to create the synthetic data. Discernment values show variation in the value of different data types and locations for screening less-adequate models.

Water Resources Research↗

USGS2025

PEST++IES how many iterations & realizations, finding the point of diminishing returns

PEST++IES (White 2018; White et al. 2020) is widely used in the groundwater modeling community for its ability to perform computationally efficient history matching and uncertainty analysis in a highly parameterized context. One primary advantage of using an iterative ensemble smoother is that the number of model runs required per iteration depends on the number of realizations in an ensemble, not the number of parameters in each realization. However, this raises the question: what is the optimal number of realizations and iterations to use for any one model before the point of diminishing returns? Using a modified version of the Freyberg model (Freyberg 1988; Hunt et al. 2020), different parameter and observation scenarios were evaluated for four iterations and ensembles of 10, 25, 50, 100, 250, 500, 1000, and 2000 realizations. To match observations, PEST++IES altered hydraulic conductivity ( k ), both globally across the model and locally at three different pilot point densities, as well as global recharge (via a single multiplier), global river conductance, and individual well flow rates. Risk-based well capture zone results (Fienen et al. 2022a) and estimated hydraulic conductivity fields from each scenario were quantitatively and qualitatively compared against the “truth” model and its outputs. Across the cases examined, ensemble sizes of 100 to 250 realizations and two PEST++IES iterations were generally sufficient to achieve good results.

Groundwater↗

USGS2025

Karhunen–Loève deep learning method for surrogate modeling and approximate Bayesian parameter estimation

We evaluate the performance of the Karhunen–Loève Deep Neural Network (KL-DNN) framework for surrogate modeling and approximate Bayesian parameter estimation in partial differential equation models. In the surrogate model, the Karhunen–Loève (KL) expansions are used for the dimensionality reduction of the number of unknown parameters and variables, and a deep neural network is employed to relate the reduced space of parameters to that of the state variables. The KL-DNN surrogate model is used to formulate a maximum-a-posteriori-like least-squares problem, which is randomized to draw samples of the posterior distribution of the parameters. We test the proposed framework for a hypothetical unconfined aquifer via comparison with the forward MODFLOW and inverse PEST++ iterative ensemble smoother (IES) solutions as well as the state-of-the-art Fourier neural operator (FNO) and deep operator networks (DeepONets) operator learning surrogate models. Our results show that the KL-DNN surrogate model outperforms FNO and DeepONet for forward predictions. For solving inverse problems, the randomized algorithm provides the same or more accurate Bayesian predictions of the parameters than IES as evidenced by the higher log predictive probability of both the estimated parameter field and the forecast hydraulic head. The posterior mean obtained from the randomized algorithm is closer to the reference parameter field than that obtained with FNO as the maximum a posteriori estimate.

Advances in Water Resources↗

USGS2025

Assessing nonpoint-source uranium pollution in an irrigated stream-aquifer system

Uranium (U) in rocks and soils of arid and semi-arid environments can be mobilized by irrigation and fertilization, posing environmental and health risks. Elevated U, along with selenium (Se) and nitrate (NO 3 ) co-constituents, necessitates careful monitoring and management. We developed a distributed-parameter numerical model to assess U pollution in an irrigated stream-aquifer system, applying it to a 552 km 2 region in Colorado's Lower Arkansas River Valley (LARV) over 14 years. A MODFLOW model, describing groundwater and stream flow, was coupled with an RT3D-OTIS model to portray reactive U transport. Calibration using the PESTPP-iES iterative ensemble smoother (iES) software indicated good agreement with observed U concentrations. The model revealed substantial and variable U levels across the LARV, highlighting potential hotspots and possible contributing factors, such as geological composition of the bedrock and near-surface shale and aquifer sediments derived from them, irrigation practices, and riparian landscape. U levels exceed the chronic standard (85th percentile = 30 μg/L, set by the US Environmental Protection Agency), which is the permissible regulatory threshold, in groundwater across 44 % of the region and along the river by an average factor of 2.9. Simulated average U concentrations in the non-riparian aquifer and river are 124 μg/L and 60 μg/L, respectively, compared with 112 μg/L and 62 μg/L for measured values. The average 85th percentile U concentration is 222 μg/L in the aquifer and 82 μg/L in the river. Average simulated U mass loading to the river is 0.17 kg/day per km, compared to an estimated 0.23 kg/day per km. Findings provide a baseline for comparing future simulated outcomes of alternative best management practices (BMPs) for U pollution mitigation and offer a methodology applicable to other irrigated regions.

Colorado↗

USGS2025

A model uncertainty quantification protocol for evaluating the value of observation data

The history-matching approach to parameter estimation with models enables a powerful offshoot analysis of data worth—using the uncertainty of a model forecast as a metric for the worth of data. Adding observation data will either have no impact on forecast uncertainty or will reduce it. Removing existing data will either have no impact on forecast uncertainty or will increase it. The history-matching framework makes it possible to perform this quantitative analysis leveraging the connections among observations, model parameters, and model forecasts. We show this behavior on a specific groundwater flow model of the Mississippi Alluvial Plain and show where the analysis can be informative for considering the potential design of an observation network based on existing or potential observations.

Scientific Investigations Report↗

USGS2024

Ensemble methods for parameter estimation of WRF-Hydro

The WRF-Hydro hydrological model has been used in many applications in the past with some level of history matching in the majority of these studies. In this study, we use the iterative Ensemble Smoother (iES), a powerful parameter estimation methodology implemented in the open-source PEST++ software. The iES provides an ensemble solution with an uncertainty bound instead of a single best estimate which has been the common approach in the previous WRF-Hydro studies. We discuss the importance of accounting for observation noise which results in a wider spread in the model solution. We investigate the impact of constructing objective functions by differentially weighting the observations to tune the model response toward model outputs appropriate for a specific application. Results confirm the necessity of differentially weighting the observations before calculation of the objective function as the optimization algorithm struggles with calculating parameter updates with uniform weighting. We also show that we achieve better model performance in terms of verification metrics with higher emphasis on the high flow events, when the objective function is tuned toward an application where the extreme events are of importance. We then investigate the impact of estimating more parameters, in particular we estimate a larger number of snow parameters. Results show a large improvement in the model performance. In summary, our study demonstrates the efficacy of employing iES alongside differential weighting of observations, highlighting its potential to enhance hydrological model parameter estimation.

Water Resources Research↗

USGS2024

Applying local and global sensitivity analysis to inform bigheaded carp management

Natural resource managers commonly use population-level models to aid in understanding the status of target populations or the potential implications of management actions. Sensitivity analyses, specifically, local sensitivity analysis (LSA) and global sensitivity analysis (GSA), exist as tools to improve understanding of these models, the importance of specific parameters to model outcomes, and the assumptions inherent to population models. However, applied ecology and closely related fields such as fisheries and wildlife management rarely use LSA or GSA. Here, we apply LSA and GSA to the Spatially Explicit Invasive Carp Population model, a population-level model used to simulate and compare management actions on the invasive silver carp Hypophthalmichthys molitrix population in the Illinois River. The spatial structure within the model arises from the navigation “pools” that exist between navigation lock-and-dam in the river and are also the focus of fisheries management in the river. We completed LSA and GSA for two reasons: (1) to compare the importance of model input parameters on simulated model populations and (2) to demonstrate the use of LSA and GSA for population-level models. In our system, the most important sensitivity values generally were detected in Alton Pool. This pool is the farthest downstream and includes a direct connection with the Mississippi River. These results indicate that the population dynamics in Alton Pool are an important driver of the population dynamics throughout the Illinois River. Management outcomes in this system will, therefore, be strongly dependent on the dynamics of Alton Pool.

Illinois↗

USGS2024

A reproducible manuscript workflow with a Quarto template

Scientists and resource managers increasingly use Markdown-based tools to create reproducible reports and manuscripts. These workflows allow people to use standardized methods that are more reproducible, efficient, and transparent than other standard office tools. We present a Quarto template and demonstrate how this template may be used for a journal, the Journal of Fish and Wildlife Management , in our article. This template may also be readily adapted to other journals that use Microsoft Word-based workflows and for other product types such as annual reports. We also provide a high-level overview of Quarto and other Markdown-based workflows as part of the document. Lastly, we provide examples of some features of the Quarto publishing system that may be helpful for authors when customizing Quarto templates for specific journal formatting requirements and other product types.

Journal of Fish and Wildlife Management↗

USGS2024

Analysis of factors affecting plume remediation in a sole-source aquifer system, southeastern Nassau County, New York

Several plumes of dissolved, chlorinated solvents, including trichloroethylene, have been identified in a sole-source aquifer near the former Northrop Grumman Bethpage Facility and Naval Weapons Industrial Reserve Plant sites in southeastern Nassau County, New York. Past investigations have documented that the groundwater contamination originated from this industrial area and now extends to the south, in the direction of groundwater flow. The intermixed plumes are commonly referred to as the “Navy Grumman groundwater plume.” Detailed groundwater-flow modeling was needed for the New York State Department of Environmental Conservation (NYSDEC) to evaluate design options necessary for the construction, operation, optimization, maintenance, and monitoring of a groundwater extraction and treatment cleanup plan selected in a December 2019 Amended Record of Decision by the NYSDEC to comprehensively address these plumes. Consequently, the NYSDEC began a cooperative study with the U.S. Geological Survey in 2020 to better understand the local hydrogeologic framework using two independent approaches to characterize aquifer heterogeneity and update an existing regional groundwater-flow model to provide transient boundary conditions for new inset groundwater-flow models of the plume area. We developed these detailed inset models for the two independent aquifer characterizations using history-matching techniques coupled with a novel approach to risk-based management optimization of the remedial design. We also used the updated regional model to assess this optimized groundwater extraction and treatment design for potential saltwater intrusion. The ensembles of parameters resulting from history matching provided a platform with which to evaluate capture by water-supply and remedial wells using particle-tracking techniques. Using the ensemble to select a risk stance, we performed multiobjective optimization to identify various configurations of remedial pumping that are consistent with external constraints and that favor potentially competing objectives. Multiple solutions provide tradeoffs that NYSDEC can consider. In general, pumping redistribution may help to prevent further contamination migration downgradient. These and other study results are intended to support decisions for the remedial design focused on the local area encompassing the full extent of the Navy Grumman groundwater plume.

New York↗

USGS2024

Parameter ESTimation with the Gauss–Levenberg–Marquardt algorithm: An intuitive guide

In this paper, we review the derivation of the Gauss–Levenberg–Marquardt (GLM) algorithm and its extension to ensemble parameter estimation. We explore the use of graphical methods to provide insights into how the algorithm works in practice and discuss the implications of both algorithm tuning parameters and objective function construction in performance. Some insights include understanding the control of both parameter trajectory and step size for GLM as a function of tuning parameters. Furthermore, for the iterative Ensemble Smoother (iES), we discuss the importance of noise on observations and show how iES can cope with non-unique outcomes based on objective function construction. These insights are valuable for modelers using PEST, PEST++, or similar parameter estimation tools.

Groundwater↗

USGS2024

Streamflow depletion caused by groundwater pumping: Fundamental research priorities for management-relevant science

Reductions in streamflow caused by groundwater pumping, known as “streamflow depletion,” link the hydrologic process of stream-aquifer interactions to human modifications of the water cycle. Isolating the impacts of groundwater pumping on streamflow is challenging because other climate and human activities concurrently impact streamflow, making it difficult to separate individual drivers of hydrologic change. In addition, there can be lags between when pumping occurs and when streamflow is affected. However, accurate quantification of streamflow depletion is critical to integrated groundwater and surface water management decision making. Here, we highlight research priorities to help advance fundamental hydrologic science and better serve the decision-making process. Key priorities include (a) linking streamflow depletion to decision-relevant outcomes such as ecosystem function and water users to align with partner needs; (b) enhancing partner trust and applicability of streamflow depletion methods through benchmarking and coupled model development; and (c) improving links between streamflow depletion quantification and decision-making processes. Catalyzing research efforts around the common goal of enhancing our streamflow depletion decision-support capabilities will require disciplinary advances within the water science community and a commitment to transdisciplinary collaboration with diverse water-connected disciplines, professions, governments, organizations, and communities.

Water Resource Research↗

USGS2024

Evaluation of an impulse-response emulator for groundwater contaminant transport modeling

There is a significant need to develop decision support tools capable of delivering accurate representations of environmental conditions, such as ground and surface water solute concentrations, in a timely and computationally efficient manner. Such tools can be leveraged to assess a large number of potential management strategies for mitigating non-point source pollutants. Here, we assess the effectiveness of the impulse-response emulation approach to approximate process-based groundwater model estimates of solute transport from MODFLOW and MT3D over a wide range of model inputs and parameters, with the goal of assessing where in parameter space the assumptions underlying this emulation approach are valid. The impulse-response emulator was developed using the sensitivity analysis utilities in the PEST++ software suite and is capable of approximating MODFLOW/MT3D estimates of solute transport over a large portion of the parameter space tested, except in cases where the Courant number is above 0.5. Across all runs tested, the highest percent errors were at the plume fronts. These results suggest that the impulse-response approach may be suitable for emulation of solute transport models for a wide range of cases, except when high-resolution outputs are needed, or when very low concentrations at plume edges are of particular interest.

Groundwater↗

USGS2024

Precipitation uncertainty estimation and rainfall-runoff model calibration using iterative ensemble smoothers

The introduction of iterative ensemble smoothers (IES) for parameter calibration opens avenues for expanding parameter space in surface water hydrologic modeling. Here, we have introduced independent parameters into a model calibration experiment to estimate errors in rainfall forcing data. This approach has the potential to estimate rainfall errors using other hydrological observations and to improve model calibration. Using high-resolution rain gauge data, we estimated “real” rainfall errors across the Turkey River watershed at storm and daily scales. Tests on synthetic and real-world scenarios successfully estimated errors correlated with observed values – even at daily scales. However, a bias remained from model parameter compensation, and identifying errors was challenging for low precipitation and snowfall. Despite synthetic results showing good error correlation, the biases in parameter identification masked potential improvements in hydrological calibration. This study highlights the potential of IES to provide additional information on rainfall errors, even only using streamflow observations.

Advances in Water Resources↗

USGS2024

Comparing single and multiple objective constrained optimization algorithms for tuning a groundwater remediation system

Groundwater flow and particle tracking models are critical tools to simulate the natural system, contaminant fate and transport, and effects of remediation. Constrained optimization uses models to systematically explore the interplay between remedial design and contaminant fate, considering uncertainty. Sequential Linear Programming (SLP) provides a design alternative addressing a single goal (e.g. maximum hydraulic containment, maximum mass removal). Multi-objective algorithms like Nondominated Sorting Genetic Algorithm (NSGA-II) explore the tradeoffs among such objectives and more (e.g. cost, public-supply well contamination). We explore both approaches at a contaminated site in Long Island, New York USA. We compare the algorithms and ramifications on results. NSGA-II explores, at additional computational cost, explicit tradeoffs among multiple objectives, providing additional insights relative to SLP. The NGSA-II algorithm allows for graphical consideration of three objectives. SLP decision variables often settle at predetermined bounds. Bounds assignment thus differs from parameter estimation; bounds must be acceptable rather than safeguards.

New York↗