Identifying Sensitivities in Flood Frequency Analyses using a Stochastic Hydrologic Modeling System

Abstract. This study assesses sources of variance in stochastic hydrologic modelling to support flood frequency analyses. The major components of the modelling chain, including model structure, model parameter estimation, initial conditions, and precipitation inputs were examined across return periods from 2 to 100,000 years at two watersheds representing different hydro-climates across the western United States. Ten hydrologic model structures were configured, calibrated and run within the Framework for Understanding Structural Errors (FUSE) modular modelling framework for each of the two watersheds. Model parameters and initial conditions were derived from long-term calibrated simulations using a 100-member historical meteorology ensemble. A stochastic event-based hydrologic modelling workflow was developed using the calibrated models; millions of flood event simulations were performed at each basin. The analysis of variance method was then used to quantify the relative contributions of model structure, model parameters, initial conditions, and precipitation inputs to flood magnitudes for different return periods. The attribution of the variance of flood frequencies to each component of a stochastic hydrological modelling framework, including several hydrological model structures, is a novel contribution to the flood modelling literature. Results demonstrate that different components of the modelling chain have different sensitivities for different return periods. Precipitation inputs contribute most to the variance of rare events, while initial conditions are most influential for the more frequent events. However, the hydrological model structure and structure-parameter interactions together play an equally important role in specific cases, depending on the basin characteristics and type of flood metric of interest. This study highlights the importance of critically assessing model underpinnings, understanding flood generation processes, and selecting appropriate hydrological models that are consistent with our understanding of flood generation processes.

Download Full-text

Identifying sensitivities in flood frequency analyses using a stochastic hydrologic modeling system

Hydrology and Earth System Sciences ◽

10.5194/hess-25-5603-2021 ◽

2021 ◽

Vol 25 (10) ◽

pp. 5603-5621

Author(s):

Andrew J. Newman ◽

Amanda G. Stone ◽

Manabendra Saharia ◽

Kathleen D. Holman ◽

Nans Addor ◽

...

Keyword(s):

Hydrologic Modeling ◽

Initial Conditions ◽

Hydrologic Model ◽

Flood Frequency ◽

Model Parameters ◽

Model Structure ◽

Structure Model ◽

Return Periods ◽

Modeling Framework ◽

Flood Generation

Abstract. This study employs a stochastic hydrologic modeling framework to evaluate the sensitivity of flood frequency analyses to different components of the hydrologic modeling chain. The major components of the stochastic hydrologic modeling chain, including model structure, model parameter estimation, initial conditions, and precipitation inputs were examined across return periods from 2 to 100 000 years at two watersheds representing different hydroclimates across the western USA. A total of 10 hydrologic model structures were configured, calibrated, and run within the Framework for Understanding Structural Errors (FUSE) modular modeling framework for each of the two watersheds. Model parameters and initial conditions were derived from long-term calibrated simulations using a 100 member historical meteorology ensemble. A stochastic event-based hydrologic modeling workflow was developed using the calibrated models in which millions of flood event simulations were performed for each basin. The analysis of variance method was then used to quantify the relative contributions of model structure, model parameters, initial conditions, and precipitation inputs to flood magnitudes for different return periods. Results demonstrate that different components of the modeling chain have different sensitivities for different return periods. Precipitation inputs contribute most to the variance of rare floods, while initial conditions are most influential for more frequent events. However, the hydrological model structure and structure–parameter interactions together play an equally important role in specific cases, depending on the basin characteristics and type of flood metric of interest. This study highlights the importance of critically assessing model underpinnings, understanding flood generation processes, and selecting appropriate hydrological models that are consistent with our understanding of flood generation processes.

Download Full-text

Contrasting phenotypes emerging from stable rules: A model based on self-regulated control loops captures the dynamics of shoot extension in contrasting maize phenotypes

Annals of Botany ◽

10.1093/aob/mcz168 ◽

2019 ◽

Vol 126 (4) ◽

pp. 615-633 ◽

Cited By ~ 1

Author(s):

T Vidal ◽

B Andrieu

Keyword(s):

Initial Conditions ◽

Growth Conditions ◽

Internal Variables ◽

Model Parameters ◽

Crop Models ◽

Control Loops ◽

Modelling Framework ◽

Appearance Rate ◽

Leaf Appearance Rate ◽

Leaf Appearance

Abstract Background and Aims The dynamics of plant architecture is a central aspect of plant and crop models. Most models assume that whole shoot development is orchestrated by the leaf appearance rate, which follows a thermal time schedule. However, leaf appearance actually results from leaf extension and taking it as an input hampers our ability to understand shoot construction. The objective of the present study was to assess a modelling framework for grasses, in which the emergence of leaves and other organs is explicitly calculated as a result of their extension. Methods The approach builds on a previous model, which uses a set of rules co-ordinating the timing of development within and between phytomers. We first assessed rule validity for four experimental datasets, including different cultivars, planting densities and environments, and accordingly revised the equations driving the extension of the upper leaves and of internodes. We then fitted model parameters for each dataset and evaluated the ability to simulate the measured phenotypes across time. Finally, we carried out a sensitivity analysis to identify the parameters that had the greatest impact and to investigate model behaviour. Key Results The modified version of the model simulated correctly the contrasting maize phenotypes. Co-ordination rules accounted for the observations in all studied cultivars. Factors with major impact on model output included extension rates, the time of tassel initiation and initial conditions. A large diversity of phenotypes could be simulated. Conclusions This work provides direct experimental evidence for co-ordination rules and illustrates the capacity of the model to represent contrasting phenotypes. These rules play an important role in patterning shoot architecture and some of them need to be assessed further, considering contrasting growth conditions. To make the model more predictive, several parameters could be considered in the future as internal variables driven by plant status.

Download Full-text

Towards Conditional Parameter Estimation for Automatic Model Structure Identification: Using Mixed-Integer Calibration for Model Development

10.5194/egusphere-egu2020-13562 ◽

2020 ◽

Author(s):

Diana Spieler ◽

Juliane Mai ◽

Bryan Tolson ◽

James Craig ◽

Niels Schütze

Keyword(s):

Parameter Estimation ◽

Mixed Integer ◽

Structure Identification ◽

Model Parameters ◽

Suitable Model ◽

Model Structure ◽

Hydrologic Modelling ◽

Vast Number ◽

Calibration Process ◽

Decision Variables

A recently introduced framework for Automatic Model Structure Identification (AMSI) allows to simultaneously optimize model structure choices (integer decision variables) and parameter values (continuous decision variables) in hydrologic modelling. By combining the mixed-integer optimization algorithm DDS and the flexible hydrologic modelling framework RAVEN, AMSI is able to test a vast number of model structure and parameter combinations in order to identify the most suitable model structure for representing the rainfall runoff behavior of a catchment. The model structure and all potentially active model parameters are calibrated simultaneously. This causes a certain degree of inefficiency during the calibration process, as variables might be perturbed that are not currently relevant for the tested model structure. In order to avoid this, we propose an adaption of the current DDS algorithm allowing for conditional parameter estimation. Parameters will only be perturbed during the calibration process if they are relevant for the model structure that is currently tested. The conditional parameter estimation setup will be compared to the standard DDS algorithm for multiple AMSI test cases. We will show if and how conditional parameter estimation increases the efficiency of AMSI.

Download Full-text

Incremental model breakdown to assess the multi-hypotheses problem

10.5194/hess-2017-691 ◽

2017 ◽

Author(s):

Florian U. Jehn ◽

Lutz Breuer ◽

Tobias Houska ◽

Konrad Bestian ◽

Philipp Kraft

Keyword(s):

Model Performance ◽

Complex Model ◽

Model Parameters ◽

Model Structure ◽

Objective Functions ◽

Final Model ◽

Incremental Model ◽

Model Efficiency ◽

Improved Model ◽

Model Structures

Abstract. The ambiguous representation of hydrological processes have led to the formulation of the multiple hypotheses approach in hydrological modelling, which requires new ways of model construction. However, most recent studies focus only on the comparison of predefined model structures or building a model step-by-step. This study tackles the problem the other way around: We start with one complex model structure, which includes all processes deemed to be important for the catchment. Next, we create 13 additional simplified models, where some of the processes from the starting structure are disabled. The performance of those models is evaluated using three objective functions (logarithmic Nash-Sutcliffe, percentage bias and the ratio between root mean square error to the standard deviation of the measured data). Through this incremental breakdown, we identify the most important processes and detect the restraining ones. This procedure allows constructing a more streamlined, subsequent 15th model with improved model performance, less uncertainty and higher model efficiency. We benchmark the original Model 1 with the final Model 15 and find that the incremental model breakdown leads to a structure with good model performance, fewer but more relevant processes and less model parameters.

Download Full-text

Review of Electric Load Modeling Research

Advanced Materials Research ◽

10.4028/www.scientific.net/amr.811.627 ◽

2013 ◽

Vol 811 ◽

pp. 627-630 ◽

Cited By ~ 1

Author(s):

Xue Song Zhou ◽

Huan Liang ◽

You Jie Ma

Keyword(s):

Voltage Stability ◽

Transient Stability ◽

Load Flow ◽

Model Parameters ◽

Model Structure ◽

Electric Load ◽

Structure Model ◽

Load Modeling ◽

Load Model ◽

Model Research

The effect of load model on the analyses of load flow, transient stability, small disturbance stability and voltage stability is analyzed. The importance of the load modeling research is emphasized. The development of component-based method and measurement-based method is reviewed. The advances on the load model research including the select ion of load model structure, model parameters identification, load model with the voltage stability analysis and the sensitivity of load model to the transient stability is summarized.

Download Full-text

Parametrization consequences of constraining soil organic matter models by total carbon and radiocarbon using long-term field data

10.5194/bg-2015-630 ◽

2016 ◽

Cited By ~ 1

Author(s):

L. Menichetti ◽

T. Kätterer ◽

J. Leifeld

Keyword(s):

Data Streams ◽

Parameter Uncertainty ◽

Total Carbon ◽

Model Parameters ◽

Model Structure ◽

Relative Importance ◽

Organic C ◽

Control Feedback ◽

The Impact ◽

Model Structures

Abstract. Soil organic carbon (SOC) dynamics result from different interacting processes and controls on spatial scales from sub-aggregate to pedon to the whole ecosystem. These complex dynamics are translated into models as abundant degrees of freedom. This high number of not directly measurable variables and, on the other hand, very limited data at disposal result in equifinality and parameter uncertainty. Carbon radioisotope measurements are a proxy for SOC age both at annual to decadal (bomb peak based) and centennial to millennial time scales (radio decay based), and thus can be used in addition to total organic C for constraining SOC models. By considering this additional information, uncertainties in model structure and parameters may be reduced. To test this hypothesis we studied SOC dynamics and their defining kinetic parameters in the ZOFE experiment, a >60-years old controlled cropland experiment in Switzerland, by utilising SOC and SO14C time-series. To represent different processes we applied five model structures, all stemming from a simple mother model (ICBM): I) two decomposing pools, II) an inert pool added, III) three decomposing pools, IV) two decomposing pools with a substrate control feedback on decomposition, V) as IV but with also an inert pool. These structures were extended to explicitly represent total SOC and 14C pools. The use of different model structures allowed us to explore model structural uncertainty and the impact of 14C on kinet ic parameters. We considered parameter uncertainty by calibrating in a formal Bayesian framework. By varying the relative importance of total SOC and SO14C data in the calibration, we could quantify the effect of the information from these two data streams on estimated model parameters. The weighing of the two data streams was crucial for determining model outcomes, and we suggest including it in future modelling efforts whenever SO14C data are available. The measurements and all model structures indicated a dramatic decline in SOC in the ZOFE experiment after an initial land use change in 1949 from grass- to cropland, followed by a constant but smaller decline. According to all structures, the three treatments (control, mineral fertilizer, farmyard manure) we considered were still far from equilibrium. The estimates of mean residence time (MRT) of the C pools defined by our models were sensitive to the consideration of the SO14C data stream. Model structure had a smaller effect on estimated MRT, which ranged between 5.91 and 4.22 years and 78.93 and 98.85 years for young and old pool, respectively, for structures without substrate interactions. The simplest model structure performed the best according to information criteria, validating the idea that we still lack data for mechanistic SOC models. Although we could not exclude any of the considered processes possibly involved in SOC decomposition, it was not possible to discriminate their relative importance.

Download Full-text

Parametrization consequences of constraining soil organic matter models by total carbon and radiocarbon using long-term field data

Biogeosciences ◽

10.5194/bg-13-3003-2016 ◽

2016 ◽

Vol 13 (10) ◽

pp. 3003-3019 ◽

Cited By ~ 10

Author(s):

Lorenzo Menichetti ◽

Thomas Kätterer ◽

Jens Leifeld

Keyword(s):

Kinetic Parameters ◽

Data Streams ◽

Parameter Uncertainty ◽

Total Carbon ◽

Model Parameters ◽

Model Structure ◽

Relative Importance ◽

Control Feedback ◽

The Impact ◽

Model Structures

Abstract. Soil organic carbon (SOC) dynamics result from different interacting processes and controls on spatial scales from sub-aggregate to pedon to the whole ecosystem. These complex dynamics are translated into models as abundant degrees of freedom. This high number of not directly measurable variables and, on the other hand, very limited data at disposal result in equifinality and parameter uncertainty. Carbon radioisotope measurements are a proxy for SOC age both at annual to decadal (bomb peak based) and centennial to millennial timescales (radio decay based), and thus can be used in addition to total organic C for constraining SOC models. By considering this additional information, uncertainties in model structure and parameters may be reduced. To test this hypothesis we studied SOC dynamics and their defining kinetic parameters in the Zürich Organic Fertilization Experiment (ZOFE) experiment, a > 60-year-old controlled cropland experiment in Switzerland, by utilizing SOC and SO14C time series. To represent different processes we applied five model structures, all stemming from a simple mother model (Introductory Carbon Balance Model – ICBM): (I) two decomposing pools, (II) an inert pool added, (III) three decomposing pools, (IV) two decomposing pools with a substrate control feedback on decomposition, (V) as IV but with also an inert pool. These structures were extended to explicitly represent total SOC and 14C pools. The use of different model structures allowed us to explore model structural uncertainty and the impact of 14C on kinetic parameters. We considered parameter uncertainty by calibrating in a formal Bayesian framework. By varying the relative importance of total SOC and SO14C data in the calibration, we could quantify the effect of the information from these two data streams on estimated model parameters. The weighing of the two data streams was crucial for determining model outcomes, and we suggest including it in future modeling efforts whenever SO14C data are available. The measurements and all model structures indicated a dramatic decline in SOC in the ZOFE experiment after an initial land use change in 1949 from grass- to cropland, followed by a constant but smaller decline. According to all structures, the three treatments (control, mineral fertilizer, farmyard manure) we considered were still far from equilibrium. The estimates of mean residence time (MRT) of the C pools defined by our models were sensitive to the consideration of the SO14C data stream. Model structure had a smaller effect on estimated MRT, which ranged between 5.9 ± 0.1 and 4.2 ± 0.1 years and 78.9 ± 0.1 and 98.9 ± 0.1 years for young and old pools, respectively, for structures without substrate interactions. The simplest model structure performed the best according to information criteria, validating the idea that we still lack data for mechanistic SOC models. Although we could not exclude any of the considered processes possibly involved in SOC decomposition, it was not possible to discriminate their relative importance.

Download Full-text

How good does Automatic Model Structure Identification work? A Benchmark Study with 6915 Model Structures.

10.5194/egusphere-egu21-12232 ◽

2021 ◽

Author(s):

Diana Spieler ◽

Niels Schütze

Keyword(s):

Optimization Algorithms ◽

Calibration Procedure ◽

Mixed Integer ◽

Structure Identification ◽

Model Parameters ◽

Model Structure ◽

Integer Optimization ◽

Mixed Integer Optimization ◽

Standard Calibration ◽

Model Structures

Recent investigations have shown it is possible to simultaneously calibrate model structures and model parameters to identify appropriate models for a given task (Spieler et al., 2020). However, this is computationally challenging, as different model structures may use a different number of parameters. While some parameters may be shared between model structures, others might be relevant for only a few structures, which theoretically requires the calibration of conditionally active parameters. Additionally, shared model parameters might cause different effects in different model structures, causing their optimal values to differ across structures. In this study, we tested how two current &#8220;of the shelf&#8221; mixed-integer optimization algorithms perform when having to handle these peculiarities during the automatic model structure identification (AMSI) process recently introduced by Spieler et al. (2020).To validate the current performance of the AMSI approach, we conduct a benchmark experiment with a model space consisting of 6912 different model structures. &#160;First, all model structures are independently calibrated and validated for three hydro-climatically differing catchments using the CMA-ES algorithm and KGE as the objective function. This is referred to as standard calibration procedure. We identify the best performing model structure(s) based on validation performance and analyze the range of performance as well as the number of structures performing in a similar range. Secondly, we run AMSI on all three catchments to automatically identify the most feasible model structure based on the KGE performance. Two different mixed-integer optimization algorithms are used &#8211; namely DDS and CMA-ES. Afterwards, we compare the results to the best performing models of the standard calibration of all 6912 model structures.Within this experimental setup, we analyze if the best performing model structure(s) AMSI identifies are identical to the best performing structures of the standard calibration and if there are differences in performance when using different optimization algorithms for AMSI. We also validate if AMSI can identify the best performing model structures for a catchment at a fraction of the computational cost than the standard calibration procedure requires by using &#8220;off the shelf&#8221; mixed-integer optimization algorithms.&#160;&#160;&#160;Spieler, D., Mai, J., Craig, J. R., Tolson, B. A., & Sch&#252;tze, N. (2020). Automatic Model Structure Identification for Conceptual Hydrologic Models. Water Resources Research, 56(9). https://doi.org/10.1029/2019WR027009

Download Full-text

Comparing classical performance measures with signature indices derived from flow duration curves to assess model structures as tools for catchment classification

Hydrology Research ◽

10.2166/nh.2015.221 ◽

2015 ◽

Vol 47 (1) ◽

pp. 1-14 ◽

Cited By ~ 6

Author(s):

Rita Ley ◽

Hugo Hellebrand ◽

Markus C. Casper ◽

Fabrizio Fenicia

Keyword(s):

Performance Measures ◽

Model Comparison ◽

Hydrological Model ◽

Performance Measure ◽

Model Structure ◽

Flow Duration ◽

Modelling Framework ◽

Flow Duration Curves ◽

Model Structures ◽

Meso Scale

The ability of a hydrological model to reproduce observed streamflow can be represented by a large variety of performance measures. Although these metrics may suit different purposes, it is unclear which of them is most appropriate for a given application. Our objective is to investigate various performance measures to assess model structures as tools for catchment classification. For this purpose, 12 model structures are generated using the SUPERFLEX modelling framework, which are then applied to 53 meso-scale basins in the Rhineland-Palatinate (Germany). Statistical and hydrological performance measures are compared with signature indices derived from the flow duration curve and combined into a new performance measure, the standardized signature index sum (SIS). The performance measures are evaluated in their ability to distinguish the relative merits of various model alternatives. In many cases, classical and hydrological performance measures assign similar values to different hydrographs. These measures, therefore, are not well suited for model comparison. The proposed SIS is more effective in revealing differences between model results. It allows for a more distinctive identification of a best performing model for individual basins. A best performing model structure obtained through the SIS can be used as basin classifier.

Download Full-text

Estimation of Gear Backlash: Theory and Simulation

Journal of Dynamic Systems Measurement and Control ◽

10.1115/1.2801324 ◽

1998 ◽

Vol 120 (1) ◽

pp. 74-82 ◽

Cited By ~ 13

Author(s):

Jeffrey L. Stein ◽

Churn-Hway Wang

Keyword(s):

Computer Simulation ◽

Quality Control ◽

Open Loop ◽

Measurement Noise ◽

Model Parameters ◽

Model Structure ◽

Estimation Technique ◽

Structure Model ◽

Product Quality Control ◽

Gear Backlash

Machine and product condition monitoring is important to product quality control, especially for unmanned manufacturing. This paper proposes a technique for the estimation of clearance in mechanical systems under dynamic conditions with specific application to the estimation of backlash in gear systems of servomechanisms. The technique is based on a momentum transfer analysis that shows that the change in the speed (defined as bounce) of the primary gear due to impact with the secondary gear is related to the magnitude of the backlash. An algorithm is presented to estimate the bounce in real-time. The algorithm estimates the bounce by computing the standard bounce which is defined as the standard deviation of the demodulated envelope of the primary gear speed. The standard bounce is shown to be a good measure of the bounce when the system is excited sinusoidally. The algorithm’s accuracy and sensitivity are verified through computer simulation of an open-loop DC servomechanism. An approximately linear relationship between the standard bounce and the backlash magnitude is observed. This holds for backlash values exceeding recommended tolerances by ±100 percent. The algorithm is also shown to be insensitive to changes in the simulation model structure, model parameters as well as system and measurement noise. The estimation technique is accurate, computationally simple, and requires no additional sensors if the servosystem to be monitored already has a conventional tachometer.

Download Full-text