Parameters sensitivity analysis for a~crop growth model applied to winter wheat in the Huanghuaihai Plain in China

Abstract. Parameters sensitivity analysis is a crucial step in effective model calibration. It quantitatively apportions the variation of model output to different sources of variation, and identifies how "sensitive" a model is to changes in the values of model parameters. Through calibration of parameters that are sensitive to model outputs, parameter estimation becomes more efficient. Due to uncertainties associated with yield estimates in a regional assessment, field-based models that perform well at field scale are not accurate enough to model at regional scale. Conducting parameters sensitivity analysis at the regional scale and analyzing the differences of parameter sensitivity between stations would make model calibration and validation in different sub-regions more efficient. Further, it would benefit the model applied to the regional scale. Through simulating 2000 × 22 samples for 10 stations in the Huanghuaihai Plain, this study discovered that TB (Optimal temperature), HI (Normal harvest index), WA (Potential radiation use efficiency), BN2 (Normal fraction of N in crop biomass at mid-season) and RWPC1 (Fraction of root weight at emergency) are more sensitive than other parameters. Parameters that determine nutrition supplement and LAI development have higher global sensitivity indices than first-order indices. For spatial application, soil diversity is crucial because soil is responsible for crop parameters sensitivity index differences between sites.

Download Full-text

Sensitivity analysis and the challenges posed by multiple approaches: a multifaceted mess

10.5194/egusphere-egu2020-6626 ◽

2020 ◽

Author(s):

Monica Riva ◽

Aronne Dell'Oca ◽

Alberto Guadagnini

Keyword(s):

Sensitivity Analysis ◽

Model Calibration ◽

Global Sensitivity Analysis ◽

System Model ◽

Model Parameters ◽

Model Output ◽

Multiple Model ◽

Total Uncertainty ◽

Global Sensitivity ◽

Sensitivity Indices

Modern models of environmental and industrial systems have reached a relatively high level of complexity. The latter aspect could hamper an unambiguous understanding of the functioning of a model, i.e., how it drives relationships and dependencies among inputs and outputs of interest. Sensitivity Analysis tools can be employed to examine this issue.Global sensitivity analysis (GSA) approaches rest on the evaluation of sensitivity across the entire support within which system model parameters are supposed to vary. In this broad context, it is important to note that the definition of a sensitivity metric must be linked to the nature of the question(s) the GSA is meant to address. These include, for example: (i) which are the most important model parameters with respect to given model output(s)?; (ii) could we set some parameter(s) (thus assisting model calibration) at prescribed value(s) without significantly affecting model results?; (iii) at which space/time locations can one expect the highest sensitivity of model output(s) to model parameters and/or knowledge of which parameter(s) could be most beneficial for model calibration?The variance-based Sobol&#8217; Indices (e.g., Sobol, 2001) represent one of the most widespread GSA metrics, quantifying the average reduction in the variance of a model output stemming from knowledge of the input. Amongst other techniques, Dell&#8217;Oca et al. [2017] proposed a moment-based GSA approach which enables one to quantify the influence of uncertain model parameters on the (statistical) moments of a target model output.Here, we embed in these sensitivity indices the effect of uncertainties both in the system model conceptualization and in the ensuing model(s) parameters. The study is grounded on the observation that physical processes and natural systems within which they take place are complex, rendering target state variables amenable to multiple interpretations and mathematical descriptions. As such, predictions and uncertainty analyses based on a single model formulation can result in statistical bias and possible misrepresentation of the total uncertainty, thus justifying the assessment of multiple model system conceptualizations. We then introduce copula-based sensitivity metrics which allow characterizing the global (with respect to the input) value of the sensitivity and the degree of variability (across the whole range of the input values) of the sensitivity for each value that the prescribed model output can possibly undertake, as driven by a governing model. In this sense, such an approach to sensitivity is global with respect to model input(s) and local with respect to model output, thus enabling one to discriminate the relevance of an input across the entire range of values of the modeling goal of interest. The methodology is demonstrated in the context of flow and reactive transport scenarios.&#160;ReferencesSobol, I. M., 2001. Global sensitivity indices for nonlinear mathematical models and their Monte Carlo estimates. Math. Comput. Sim., 55, 271-280.Dell&#8217;Oca, A., Riva, M., Guadagnini, A., 2017. Moment-based metrics for global sensitivity analysis of hydrological systems. Hydr. Earth Syst. Sci., 21, 6219-6234.

Download Full-text

Comprehensive global sensitivity analysis of a repository model using different types of transformations and metamodeling techniques

10.5194/egusphere-egu21-6131 ◽

2021 ◽

Author(s):

Sabine M. Spiessl ◽

Dirk-A. Becker ◽

Sergei Kucherenko

Keyword(s):

Sensitivity Analysis ◽

Global Sensitivity Analysis ◽

Higher Order ◽

Model Parameters ◽

High Dimensional Model Representation ◽

Model Output ◽

Global Sensitivity ◽

Sparse Polynomial ◽

Sensitivity Indices ◽

Highly Nonlinear

Due to their highly nonlinear, non-monotonic or even discontinuous behavior, sensitivity analysis of final repository models can be a demanding task. Most of the output of repository models is typically distributed over several orders of magnitude and highly skewed. Many values of a probabilistic investigation are very low or even zero. Although this is desirable in view of repository safety it can distort the evidence of sensitivity analysis. For the safety assessment of the system, the highest values of outputs are mainly essential and if those are only a few, their dependence on specific parameters may appear insignificant. By applying a transformation, different model output values are differently weighed, according to their magnitude, in sensitivity analysis. Probabilistic methods of higher-order sensitivity analysis, applied on appropriately transformed model output values, provide a possibility for more robust identification of relevant parameters and their interactions. This type of sensitivity analysis is typically done by decomposing the total unconditional variance of the model output into partial variances corresponding to different terms in the ANOVA decomposition. From this, sensitivity indices of increasing order can be computed. The key indices used most often are the first-order index (SI1) and the total-order index (SIT). SI1 refers to the individual impact of one parameter on the model and SIT represents the total effect of one parameter on the output in interactions with all other parameters. The second-order sensitivity indices (SI2) describe the interactions between two model parameters.In this work global sensitivity analysis has been performed with three different kinds of output transformations (log, shifted and Box-Cox transformation) and two metamodeling approaches, namely the Random-Sampling High Dimensional Model Representation (RS-HDMR) [1] and the Bayesian Sparse PCE (BSPCE) [2] approaches. Both approaches are implemented in the SobolGSA software [3, 4] which was used in this work. We analyzed the time-dependent output with two approaches for sensitivity analysis, i.e., the pointwise and generalized approaches. With the pointwise approach, the output at each time step is analyzed independently. The generalized approach considers averaged output contributions at all previous time steps in the analysis of the current step. Obtained results indicate that robustness can be improved by using appropriate transformations and choice of coefficients for the transformation and the metamodel.[1] M. Zuniga, S. Kucherenko, N. Shah (2013). Metamodelling with independent and dependent inputs. Computer Physics Communications, 184 (6): 1570-1580.[2] Q. Shao, A. Younes, M. Fahs, T.A. Mara (2017). Bayesian sparse polynomial chaos expansion for global sensitivity analysis. Computer Methods in Applied Mechanics and Engineering, 318: 474-496.[3] S. M. Spiessl, S. Kucherenko, D.-A. Becker, O. Zaccheus (2018). Higher-order sensitivity analysis of a final repository model with discontinuous behaviour. Reliability Engineering and System Safety, doi: https://doi.org/10.1016/j.ress.2018.12.004.[4] SobolGSA software (2021). User manual https://www.imperial.ac.uk/process-systems-engineering/research/free-software/sobolgsa-software/.

Download Full-text

Parameter Screening in Statistical Dynamic Computer Model Calibration Using Global Sensitivities

Journal of Mechanical Design ◽

10.1115/1.4006874 ◽

2012 ◽

Vol 134 (8) ◽

Cited By ~ 10

Author(s):

Dorin Drignei ◽

Zissimos P. Mourelatos

Keyword(s):

Sensitivity Analysis ◽

Computer Model ◽

Model Calibration ◽

Simulation Models ◽

Computer Models ◽

Automotive Application ◽

Transient Signals ◽

Sensitivity Indices ◽

Calibration Algorithm ◽

Computationally Intensive

Computer, or simulation, models are ubiquitous in science and engineering. Two research topics in building computer models, generally treated separately, are sensitivity analysis and computer model calibration. In sensitivity analysis, one quantifies the effect of each input factor on outputs, whereas in calibration, one finds the values of input factors that provide the best match to a set of test data. In this article, we show a connection between these two seemingly separate concepts for problems with transient signals. We use global sensitivity analysis for computer models with transient signals to screen out inactive input factors, thus making the calibration algorithm numerically more stable. We show that the computer model does not vary with respect to parameters having zero total sensitivity indices, indicating that such parameters are impossible to calibrate and must be screened out. Because the computer model can be computationally intensive, we construct a fast statistical surrogate of the computer model which is used for both sensitivity analysis and computer model calibration. We illustrate our approach with both a simple example and an automotive application involving a road load data acquisition (RLDA) computer model.

Download Full-text

Sensitivity-guided evaluation of the HBV hydrological model parameterization

Journal of Hydroinformatics ◽

10.2166/hydro.2012.011 ◽

2012 ◽

Vol 15 (3) ◽

pp. 967-990 ◽

Cited By ~ 8

Author(s):

M. B. Zelelew ◽

K. Alfredsen

Keyword(s):

Sensitivity Analysis ◽

Hydrological Model ◽

River Basin Management ◽

Low Flow ◽

Model Parameters ◽

Sensitivity Index ◽

Analysis Method ◽

Flow Conditions ◽

Model Parameterization ◽

Sensitivity Analysis Method

Applying hydrological models for river basin management depends on the availability of the relevant data information to constrain the model residuals. The estimation of reliable parameter values for parameterized models is not guaranteed. Identification of influential model parameters controlling the model response variations either by main or interaction effects is therefore critical for minimizing model parametric dimensions and limiting prediction uncertainty. In this study, the Sobol variance-based sensitivity analysis method was applied to quantify the importance of the HBV conceptual hydrological model parameterization. The analysis was also supplemented by the generalized sensitivity analysis method to assess relative model parameter sensitivities in cases of negative Sobol sensitivity index computations. The study was applied to simulate runoff responses at twelve catchments varying in size. The result showed that varying up to a minimum of four to six influential model parameters for high flow conditions, and up to a minimum of six influential model parameters for low flow conditions can sufficiently capture the catchments' responses characteristics. To the contrary, varying more than nine out of 15 model parameters will not make substantial model performance changes on any of the case studies.

Download Full-text

Sensitivity analysis of the Chaohu Lake eutrophication model with a new index based on the Morris method

Water Science & Technology Water Supply ◽

10.2166/ws.2017.204 ◽

2017 ◽

Vol 18 (4) ◽

pp. 1375-1387 ◽

Cited By ~ 1

Author(s):

Yulin Wang ◽

Zulin Hua ◽

Liang Wang

Keyword(s):

Sensitivity Analysis ◽

Life Stage ◽

Eastern China ◽

Parameter Sensitivity ◽

Chaohu Lake ◽

Sensitivity Index ◽

Lake Eutrophication ◽

Multiple Model ◽

Morris Method ◽

Sensitivity Indices

Abstract Chaohu Lake is a large shallow lake in eastern China, and few eutrophication model studies have been conducted there. We present practical sensitivity indices based on the Morris method to compare the sensitivity of a parameter group on one model output with that of one parameter on multiple model outputs. The new sensitivity indices were employed to measure the parameter sensitivity of the Chaohu Lake eutrophication model. The results of the sensitivity analysis demonstrate that the most sensitive parameters on cyanobacteria biomass, NH4, NO3, and PO4 were BMR, KDN, Nitm, and KRP, and the most sensitive parameter groups were algae-related, nitrogen-related, and phosphorus-related, which all directly participate in their cycles. Furthermore, Nitm, KRP, KDN, KHP, BMR, KTB, KTHDR, and KTCOD were the most important for the Chaohu Lake eutrophication model. The water environment characteristics, such as the cyanobacteria life stage in the simulated period, significantly affected parameter sensitivity. The power-law relationship between the new sensitivity index and the standard deviation of model variables in the Chaohu Lake model were also determined. This finding allows us to estimate the interactions between parameters using their sensitivity index. The results provide a basis for further improvement of the Chaohu Lake eutrophication model.

Download Full-text

Strategic Framework for Parameterization of Cell Culture Models

Processes ◽

10.3390/pr7030174 ◽

2019 ◽

Vol 7 (3) ◽

pp. 174

Author(s):

Pavlos Kotidis ◽

Cleo Kontoravdi

Keyword(s):

Experimental Data ◽

Cell Culture ◽

Sensitivity Analysis ◽

Goodness Of Fit ◽

Sampling Methods ◽

Model Parameters ◽

Sensitivity Index ◽

Cho Cell ◽

Culture Models ◽

Cell Culture Models

Global Sensitivity Analysis (GSA) is a technique that numerically evaluates the significance of model parameters with the aim of reducing the number of parameters that need to be estimated accurately from experimental data. In the work presented herein, we explore different methods and criteria in the sensitivity analysis of a recently developed mathematical model to describe Chinese hamster ovary (CHO) cell metabolism in order to establish a strategic, transferable framework for parameterizing mechanistic cell culture models. For that reason, several types of GSA employing different sampling methods (Sobol’, Pseudo-random and Scrambled-Sobol’), parameter deviations (10%, 30% and 50%) and sensitivity index significance thresholds (0.05, 0.1 and 0.2) were examined. The results were evaluated according to the goodness of fit between the simulation results and experimental data from fed-batch CHO cell cultures. Then, the predictive capability of the model was tested against four different feeding experiments. Parameter value deviation levels proved not to have a significant effect on the results of the sensitivity analysis, while the Sobol’ and Scrambled-Sobol’ sampling methods and a 0.1 significance threshold were found to be the optimum settings. The resulting framework was finally used to calibrate the model for another CHO cell line, resulting in a good overall fit. The results of this work set the basis for the use of a single mechanistic metabolic model that can be easily adapted through the proposed sensitivity analysis method to the behavior of different cell lines and therefore minimize the experimental cost of model development.

Download Full-text

Applicability of a distributed watershed pollution model in a data-poor environment in Santa Catarina State, Brazil

Revista Brasileira de Ciência do Solo ◽

10.1590/s0100-06832008000400035 ◽

2008 ◽

Vol 32 (4) ◽

pp. 1699-1712 ◽

Cited By ~ 2

Author(s):

Ivan Luiz Zilli Bacic ◽

David G. Rossiter ◽

Christiaan Mathias Mannaerts

Keyword(s):

Sensitivity Analysis ◽

Model Calibration ◽

Management Practices ◽

Expert Knowledge ◽

Stream Water ◽

Model Parameters ◽

Calibration Data ◽

Santa Catarina ◽

Environmental Models ◽

Santa Catarina State

Intensification of agricultural production without a sound management and regulations can lead to severe environmental problems, as in Western Santa Catarina State, Brazil, where intensive swine production has caused large accumulations of manure and consequently water pollution. Natural resource scientists are asked by decision-makers for advice on management and regulatory decisions. Distributed environmental models are useful tools, since they can be used to explore consequences of various management practices. However, in many areas of the world, quantitative data for model calibration and validation are lacking. The data-intensive distributed environmental model AgNPS was applied in a data-poor environment, the upper catchment (2,520 ha) of the Ariranhazinho River, near the city of Seara, in Santa Catarina State. Steps included data preparation, cell size selection, sensitivity analysis, model calibration and application to different management scenarios. The model was calibrated based on a best guess for model parameters and on a pragmatic sensitivity analysis. The parameters were adjusted to match model outputs (runoff volume, peak runoff rate and sediment concentration) closely with the sparse observed data. A modelling grid cell resolution of 150 m adduced appropriate and computer-fit results. The rainfall runoff response of the AgNPS model was calibrated using three separate rainfall ranges (< 25, 25-60, > 60 mm). Predicted sediment concentrations were consistently six to ten times higher than observed, probably due to sediment trapping along vegetated channel banks. Predicted N and P concentrations in stream water ranged from just below to well above regulatory norms. Expert knowledge of the area, in addition to experience reported in the literature, was able to compensate in part for limited calibration data. Several scenarios (actual, recommended and excessive manure applications, and point source pollution from swine operations) could be compared by the model, using a relative ranking rather than quantitative predictions.

Download Full-text

Systematic evaluation of biofilm models for engineering practice: components and critical assumptions

Water Science & Technology ◽

10.2166/wst.2011.709 ◽

2011 ◽

Vol 64 (4) ◽

pp. 930-944 ◽

Cited By ~ 30

Author(s):

J. P. Boltz ◽

E. Morgenroth ◽

D. Brockmann ◽

C. Bott ◽

W. J. Gellner ◽

...

Keyword(s):

Sensitivity Analysis ◽

Model Calibration ◽

Biofilm Reactor ◽

Systematic Evaluation ◽

Engineering Practice ◽

Model Parameters ◽

Concentration Gradients ◽

Reactor Model ◽

Biofilm Model ◽

Calibration Protocol

Biofilm models are valuable tools for the design and evaluation of biofilm-based processes despite several uncertainties including the dynamics and rate of biofilm detachment, concentration gradients external to the biofilm surface, and undefined biofilm reactor model calibration protocol. The present investigation serves to (1) systematically evaluate critical biofilm model assumptions and components and (2) conduct a sensitivity analysis with the aim of identifying parameter subsets for biofilm reactor model calibration. AQUASIM was used to describe submerged-completely mixed combined carbon oxidation and nitrification IFAS and MBBR systems, and tertiary nitrification and denitrification MBBRs. The influence of uncertainties in model parameters on relevant model outputs was determined for simulated scenarios by means of a local sensitivity analysis. To obtain reasonable simulation results for partially penetrated biofilms that accumulated a substantial thickness in the modelled biofilm reactor (e.g. 1,000 μm), an appropriate biofilm discretization was applied to properly model soluble substrate concentration gradients and, consistent with the assumed mechanism for describing biofilm biomass distribution, biofilm biomass spatial variability. The MTBL thickness had a significant impact on model results for each of the modelled reactor configurations. Further research is needed to develop a mathematical description (empirical or otherwise) of the MTBL thickness that is relevant to modern biofilm reactors. No simple recommendations for a generally applicable calibration protocol are provided, but sensitivity analysis has been proven to be a powerful tool for the identification of highly sensitive parameter subsets for biofilm (reactor) model calibration.

Download Full-text

Hierarchical Sensitivity Analysis for Large-scale Process-based Hydrological Modeling with Application in an Amazonian Watershed

10.5194/hess-2020-87 ◽

2020 ◽

Author(s):

Haifan Liu ◽

Heng Dai ◽

Jie Niu ◽

Bill X. Hu ◽

Han Qiu ◽

...

Keyword(s):

Sensitivity Analysis ◽

Vadose Zone ◽

Model Calibration ◽

Large Scale ◽

Global Sensitivity Analysis ◽

Hydrologic Model ◽

Global Sensitivity ◽

Uncertainty Sources ◽

Sensitivity Indices ◽

Zone Parameters

Abstract. Sensitivity analysis is an effective tool for identifying important uncertainty sources and improving model calibration and predictions, especially for integrated systems with heterogeneous parameter inputs and complex process coevolution. In this work, an advanced hierarchical global sensitivity analysis framework, which integrates the concept of variance-based global sensitivity analysis with a hierarchical uncertainty framework, was implemented to quantitatively analyse several uncertainties associated with a three-dimensional, process-based hydrologic model (PAWS). The uncertainty sources considered include model parameters (three vadose zone parameters, two groundwater parameters, and one overland flow parameter), model structure (different thicknesses to represent unconfined and confined aquifer layers) and climate scenarios. We apply the approach to an ~ 9,000 km2 Amazon catchment modeled at 1 km resolution to provide a demonstration of multiple uncertainty source quantification using a large-scale process-based hydrologic model. The sensitivity indices are assessed based on two important hydrologic outputs: evapotranspiration (ET) and groundwater contribution to streamflow (QG). It was found that, in general, parameters, especially the vadose zone parameters, are the most important uncertainty contributors for all sensitivity indices. In addition, the influence of climate scenarios on ET predictions is also nonignorable. Furthermore, the thickness of the aquifers along the river grid cells is important for both ET and QG. These results can assist in model calibration and provide modelers with a better understanding of the general sources of uncertainty in predictions associated with complex hydrological systems in Amazonia. We demonstrated a pilot example of comprehensive global sensitivity analysis of large-scale, complex hydrological and environmental models in this study. The hierarchical sensitivity analysis methodology used is mathematically rigorous and capable of being implemented in a variety of large-scale hydrological models with various sources of uncertainty.

Download Full-text

Sensitivity Analysis of Biome-BGCMuSo for Gross and Net Primary Productivity of Typical Forests in China

10.21203/rs.3.rs-757618/v1 ◽

2021 ◽

Author(s):

Hongge Ren ◽

Li Zhang ◽

Min Yan ◽

Xin Tian ◽

Xingbo Zheng

Keyword(s):

Sensitivity Analysis ◽

Primary Productivity ◽

Model Calibration ◽

Field Experiments ◽

Net Primary Productivity ◽

Phase 1 ◽

Light Extinction ◽

Sensitivity Index ◽

Sensitivity Testing ◽

Climate Gradient

Abstract Background: Process-based models are widely used to simulate forest productivity, but complex parameterization and calibration challenge the application and development of these models. Sensitivity analysis of numerous parameters is an essential step in model calibration and carbon flux simulation. However, parameters are not dependent on each other, and the results of sensitivity analysis usually vary due to different forest types and regions. Hence, global and representative sensitivity analysis would provide reliable information for simple calibration. Methods: To determine the contributions of input parameters to gross primary productivity (GPP) and net primary productivity (NPP), regression analysis and extended Fourier amplitude sensitivity testing (EFAST) were conducted for Biome-BGCMuSo to calculate the sensitivity index of the parameters at four observation sites under climate gradient from ChinaFLUX. Results: Generally, GPP and NPP were highly sensitive to C:N leaf (C:N of leaves), W int (canopy water interception coefficient), k (canopy light extinction coefficient), FLNR (fraction of leaf N in Rubisco), MR pern (maintenance respiration in kg C/day per kg of tissue N), VPD f (vapor pressure deficit complete conductance reduction), and SLA1 (canopy average specific leaf area in phenological phase 1) at all observation sites. Various sensitive parameters occurred at four observation sites within different climate zones. GPP and NPP were particularly sensitive to FLNR, SLA1 and W int , and C:N leaf in temperate, alpine and subtropical zones, respectively. Conclusions: The results indicated that sensitivity parameters of China's forest ecosystems change with climate gradient. We found that parameter calibration should be performed according to plant functional type (PFT), and more attention needs to be paid to the differences in climate and environment. These findings contribute to determining the target parameters in field experiments and model calibration.

Download Full-text