Browsing by Subject "Data assimilation"
Now showing 1 - 5 of 5
Results Per Page
Sort Options
Item Comparing observation impact between variational and ensemble data assimilation schemes on short-term, low-level wind forecasting(2012-08) Kashawlic, Erin; Ancell, Brian A.; Schroeder, John L.; Kang, Song-LakA variety of studies have been performed to determine the effectiveness of different data assimilation schemes within numerical weather prediction. For sequential schemes, previous research using mesoscale models at the horizontal grid spacing of tens of kilometers has shown that the ensemble Kalman filter (EnKF) has outperformed a 3DVAR system in producing both analyses and subsequent forecasts. However, deciding if the same holds true for smaller grid spacing has yet to be determined. Also, the type of observations has been shown to make an impact on the resulting assimilation. This study focuses on investigating the relative performance of an EnKF and 3DVAR data assimilation scheme in producing low-level, short-term (0-24 hr) wind forecasts over the western two-thirds of the United States as well as specifically over Texas. This work uses a nested 12km/3km WRF-ARW modeling configuration and compares the Data Assimilation Research Testbed (DART) EnKF and the 3DVAR Gridpoint Statistical Interpolation (GSI) system over both domains for both routine wind forecasting and wind ramp event forecasting. In addition, a data denial experiment is set up with GSI to show the impact of adding a network of profilers and sodars on both domains for routine wind forecasting.Item Data assimilation for parameter estimation in coastal ocean hydrodynamics modeling(2013-12) Mayo, Talea Lashea; Dawson, Clinton N.Coastal ocean models are used for a vast array of applications. These applications include modeling tidal and coastal flows, waves, and extreme events, such as tsunamis and hurricane storm surges. Tidal and coastal flows are the primary application of this work as they play a critical role in many practical research areas such as contaminant transport, navigation through intracoastal waterways, development of coastal structures (e.g. bridges, docks, and breakwaters), commercial fishing, and planning and execution of military operations in marine environments, in addition to recreational aquatic activities. Coastal ocean models are used to determine tidal amplitudes, time intervals between low and high tide, and the extent of the ebb and flow of tidal waters, often at specific locations of interest. However, modeling tidal flows can be quite complex, as factors such as the configuration of the coastline, water depth, ocean floor topography, and hydrographic and meteorological impacts can have significant effects and must all be considered. Water levels and currents in the coastal ocean can be modeled by solv- ing the shallow water equations. The shallow water equations contain many parameters, and the accurate estimation of both tides and storm surge is dependent on the accuracy of their specification. Of particular importance are the parameters used to define the bottom stress in the domain of interest [50]. These parameters are often heterogeneous across the seabed of the domain. Their values cannot be measured directly and relevant data can be expensive and difficult to obtain. The parameter values must often be inferred and the estimates are often inaccurate, or contain a high degree of uncertainty [28]. In addition, as is the case with many numerical models, coastal ocean models have various other sources of uncertainty, including the approximate physics, numerical discretization, and uncertain boundary and initial conditions. Quantifying and reducing these uncertainties is critical to providing more reliable and robust storm surge predictions. It is also important to reduce the resulting error in the forecast of the model state as much as possible. The accuracy of coastal ocean models can be improved using data assimilation methods. In general, statistical data assimilation methods are used to estimate the state of a model given both the original model output and observed data. A major advantage of statistical data assimilation methods is that they can often be implemented non-intrusively, making them relatively straightforward to implement. They also provide estimates of the uncertainty in the predicted model state. Unfortunately, with the exception of the estimation of initial conditions, they do not contribute to the information contained in the model. The model error that results from uncertain parameters is reduced, but information about the parameters in particular remains unknown. Thus, the other commonly used approach to reducing model error is parameter estimation. Historically, model parameters such as the bottom stress terms have been estimated using variational methods. Variational methods formulate a cost functional that penalizes the difference between the modeled and observed state, and then minimize this functional over the unknown parameters. Though variational methods are an effective approach to solving inverse problems, they can be computationally intensive and difficult to code as they generally require the development of an adjoint model. They also are not formulated to estimate parameters in real time, e.g. as a hurricane approaches landfall. The goal of this research is to estimate parameters defining the bottom stress terms using statistical data assimilation methods. In this work, we use a novel approach to estimate the bottom stress terms in the shallow water equations, which we solve numerically using the Advanced Circulation (ADCIRC) model. In this model, a modified form of the 2-D shallow water equations is discretized in space by a continuous Galerkin finite element method, and in time by finite differencing. We use the Manning’s n formulation to represent the bottom stress terms in the model, and estimate various fields of Manning’s n coefficients by assimilating synthetic water elevation data using a square root Kalman filter. We estimate three types of fields defined on both an idealized inlet and a more realistic spatial domain. For the first field, a Manning’s n coefficient is given a constant value over the entire domain. For the second, we let the Manning’s n coefficient take two distinct values, letting one define the bottom stress in the deeper water of the domain and the other define the bottom stress in the shallower region. And finally, because bottom stress terms are generally spatially varying parameters, we consider the third field as a realization of a stochastic process. We represent a realization of the process using a Karhunen-Lo`ve expansion, and then seek to estimate the coefficients of the expansion. We perform several observation system simulation experiments, and find that we are able to accurately estimate the bottom stress terms in most of our test cases. Additionally, we are able to improve forecasts of the model state in every instance. The results of this study show that statistical data assimilation is a promising approach to parameter estimation.Item Development and evaluation of an advanced microwave radiance data assimilation system for estimating snow water storage at the continental scale(2016-05) Kwon, Yonghwan; Yang, Zong-liang; Dickinson, Robert E.; Fu, Rong; Rodell, Matthew; Moser, Robert D.Snow cover modulates the Earth's surface energy and water fluxes, and snowmelt runoff is the principal source of water for humans and ecosystems in many of the middle to high latitudes in the Northern Hemisphere. Understanding spatial and temporal variation in snowpack is crucial for climate studies and water resource management and thus the climate and hydrological research communities have invested in improving large-scale snow estimates. This dissertation aims to develop an advanced snow radiance assimilation (RA) system to improve continental-scale snow water storage estimates. The RA system is comprised of the Community Land Model version 4 (CLM4) (for snow energy and mass balance modeling), radiative transfer models (RTMs) (for brightness temperature estimates), and the Data Assimilation Research Testbed (DART) (for ensemble-based data assimilation). Two snowpack RTMs, the Microwave Emission Model for Layered Snowpacks (MEMLS) and the Dense Media Radiative Transfer--Multi Layers model (DMRT-ML), are used to simulate T[subscript B] of a multi-layered snowpack. Through an error characterization study, this dissertation presents that the correlations between snow water equivalent (SWE) error and brightness temperature (T[subscript B]) error and subsequent RA performance in estimating snow are significantly affected by all physical properties of soil and snow involved in estimating T[subscript B]. Based on the error characterization results, it is hypothesized that the continental-scale RA performance in estimating snow water storage can be improved by simultaneously updating all model physical states and parameters determining T[subscript B] based on a rule, in which prior estimates are updated depending on their correlations with a prior T[subscript B]. The results of a series of RA experiments show that the improved continental-scale snow estimates are obtained by applying the hypothesis. This dissertation also shows that further improvement of the performance of the RA system can be achieved, especially for vegetated areas, by assimilating the best-performing frequency channels (i.e., 18.7 and 23.8 GHz) and by considering the vegetation single scattering albedo to represent the vegetation effect on T[subscript B] at the top of the atmosphere.Item Heterogeneous Reservoir Characterization Utilizing Efficient Geology Preserving Reservoir Parameterization through Higher Order Singular Value Decomposition (HOSVD)(2015-01-21) Afra, SardarPetroleum reservoir parameter inference is a challenging problem to many of the reservoir simulation work flows, especially when it comes to real reservoirs with high degree of complexity and non-linearity, and high dimensionality. In fact, the process of estimating a large number of unknowns in an inverse problem lead to a very costly computational effort. Moreover, it is very important to perform geologically consistent reservoir parameter adjustments as data is being assimilated in the history matching process, i.e., the process of adjusting the parameters of reservoir system in order to match the output of the reservoir model with the previous reservoir production data. As a matter of fact, it is of great interest to approximate reservoir petrophysical properties like permeability and porosity while reparameterizing these parameters through reduced-order models. As we will show, petroleum reservoir models are commonly described by in general complex, nonlinear, and large-scale, i.e., large number of states and unknown parameters. Thus, having a practical approach to reduce the number of reservoir parameters in order to reconstruct the reservoir model with a lower dimensionality is of high interest. Furthermore, de-correlating system parameters in all history matching and reservoir characterization problems keeping the geological description intact is paramount to control the ill-posedness of the system. In the first part of the present work, we will introduce the advantages of a novel parameterization method by means of higher order singular value decomposition analysis (HOSVD). We will show that HOSVD outperforms classical parameterization techniques with respect to computational and implementation cost. It also, provides more reliable and accurate predictions in the petroleum reservoir history matching problem due to its capability to preserve geological features of the reservoir parameter like permeability. The promising power of HOSVD is investigated through several synthetic and real petroleum reservoir benchmarks and all results are compared to that of classic SVD. In addition to the parameterization problem, we also addressed the ability of HOSVD in producing accurate production data comparing to those of original reservoir system. To generate the results of the present work, we employ a commercial reservoir simulator known as ECLIPSE. In the second part of the work, we will address the inverse modeling, i.e., the reservoir history matching problem. We employed the ensemble Kalman filter (EnKF) which is an ensemble-based characterization approach to solve the inverse problem. We also, integrate our new parameterization technique into the EnKF algorithm to study the suitability of HOSVD based parameterization for reducing the dimensionality of parameter space and for estimating geologically consistence permeability distributions. The results of the present work illustrates the characteristics of the proposed parameterization method by several numerical examples in the second part including synthetic and real reservoir benchmarks. Moreover, the HOSVD advantages are discussed by comparing its performance to the classic SVD (PCA) parameterization approach. In the first part of the present work, we will introduce the advantages of a novel parameterization method by means of higher order singular value decomposition analysis (HOSVD). We will show that HOSVD outperforms classical parameterization techniques with respect to computational and implementation cost. It also, provides more reliable and accurate predictions in the petroleum reservoir history matching problem due to its capability to preserve geological features of the reservoir parameter like permeability. The promising power of HOSVD is investigated through several synthetic and real petroleum reservoir benchmarks and all results are compared to that of classic SVD. In addition to the parameterization problem, we also addressed the ability of HOSVD in producing accurate production data comparing to those of original reservoir system. To generate the results of the present work, we employ a commercial reservoir simulator known as ECLIPSE. In the second part of the work, we will address the inverse modeling, i.e., the reservoir history matching problem. We employed the ensemble Kalman filter (EnKF) which is an ensemble-based characterization approach to solve the inverse problem. We also, integrate our new parameterization technique into the EnKF algorithm to study the suitability of HOSVD based parameterization for reducing the dimensionality of parameter space and for estimating geologically consistence permeability distributions. The results of the present work illustrate the characteristics of the proposed parameterization method by several numerical examples in the second part including synthetic and real reservoir benchmarks. Moreover, the HOSVD advantages are discussed by comparing its performance to the classic SVD (PCA) parameterization approach.Item The framework for satellite gravity data assimilation into land surface models(2016-05) Sakumura, Carly Frances; Bettadpur, Srinivas Viswanath, 1963-; Dawson, Clint; Fowler, Wallace; Tapley, Byron; Zong-Liang YangThe Gravity Recovery and Climate Experiment (GRACE) mission has provided an unprecedented global, homogeneous observational dataset of the time variation in terrestrial water storage (TWS) since 2002. This product has seen widespread use in the study of processes in hydrology, oceanography, the cryosphere, and is particularly critical to inform, improve, and validate computational models of the Earth system. Assimilation of the GRACE TWS fields into current land surface models can correct model deficiencies due to errors in the model structure, atmospheric forcing datasets, parameters, etc. However, the assimilation process is complicated by spatial and temporal resolution discrepancies between the model and observational datasets, characterization of the error in each, and requires tuning to the unique characteristics of satellite gravity data. This study establishes a framework for hydrological data assimilation of terrestrial water storage data from GRACE, closes the loop between GRACE product development and its scientific use, and analyzes the assimilated results for use with current GRACE products and future satellite gravity missions. The framework fuses the strengths of the observational and land surface model datasets into an assimilated product representative of the signal strength and large scale structures of the GRACE dataset effectively downscaled to the high resolution land surface dynamics. The data assimilation framework was developed through a comprehensive analysis of the deficiencies and potential improvements of the satellite data products, the assimilation procedures and error characterization, and the assimilation effectiveness over time. This analysis motivated the development of a higher frequency GRACE dataset more representative of the hydrometeorological signal content with reduced temporal aliasing of the TWS signal. Three innovations were implemented in the product development: regularization, sliding windows, and mascon basis functions, to develop a high-fidelity daily gravity field product (RSWM). The signal and error profile of the RSWM product was comprehensively analyzed via an end-to-end simulation analysis of the GRACE mission. The simulation analysis developed an error covariance representative of the magnitude, correlation, and spatial pattern of error in the RSWM dataset available for use in the data assimilation system. The assimilation algorithms and tools were advanced to optimally incorporate the GRACE TWS data and error covariance information. Daily assimilation was performed globally at the one degree gridcell level, significantly reducing spatial and temporal smoothing of the assimilation update from previous basin-scale assimilation of the monthly mean GRACE datasets. Framework elements additionally defined the mechanisms of the assimilation process: (i) the Gaspari-Cohn localization radius to spatially smooth the coarser resolution GRACE data, (ii) the necessary assimilation update rate to balance assimilation performance and computational efficiency, and (iii) open-loop error growth after assimilation has conditioned the system to advise data latency requirements. The GRACE data assimilation framework is versatile and adaptable to other land surface models, different formulations of data from the current GRACE mission, and future satellite gravity datasets.