scholarly journals Operational hydrological data assimilation with the recursive ensemble Kalman filter

2013 ◽  
Vol 17 (1) ◽  
pp. 21-38 ◽  
Author(s):  
H. K. McMillan ◽  
E. Ö. Hreinsson ◽  
M. P. Clark ◽  
S. K. Singh ◽  
C. Zammit ◽  
...  
Keyword(s):  
Kalman Filter ◽  
Ensemble Kalman Filter ◽  
Filter Design ◽  
Initial Conditions ◽  
Information Source ◽  
Design Parameters ◽  
Improve Model ◽  
Operational Flow ◽  
Measured Flow ◽  
Forecasting System

Abstract. This paper describes the design and use of a recursive ensemble Kalman filter (REnKF) to assimilate streamflow data in an operational flow forecasting system of seven catchments in New Zealand. The REnKF iteratively updates past and present model states (soil water, aquifer and surface storages), with lags up to the concentration time of the catchment, to improve model initial conditions and hence flow forecasts. We found the REnKF overcame instabilities in the standard EnKF, which were associated with the natural lag time between upstream catchment wetness and flow at the gauging locations. The forecast system performance was correspondingly improved in terms of Nash–Sutcliffe score, persistence index and bounding of the measured flow by the model ensemble. We present descriptions of filter design parameters and explanations and examples of filter behaviour, as an information source for other groups wishing to assimilate discharge observations for operational forecasting.

scholarly journals Operational hydrological data assimilation with the Retrospective Ensemble Kalman Filter: use of observed discharge to update past and present model states for flow forecasts

2012 ◽  
Vol 9 (8) ◽  
pp. 9533-9575 ◽  
Author(s):  
H. K. McMillan ◽  
E. Ö. Hreinsson ◽  
M. P. Clark ◽  
S. K. Singh ◽  
C. Zammit ◽  
...  
Keyword(s):  
Kalman Filter ◽  
Ensemble Kalman Filter ◽  
Present Model ◽  
Filter Design ◽  
Initial Conditions ◽  
Distributed Model ◽  
Design Parameters ◽  
Improve Model ◽  
Measured Flow ◽  
Forecasting System

Abstract. This paper describes the use of the Retrospective Ensemble Kalman Filter (REnKF) to assimilate streamflow data in an operational flow forecasting system of seven catchments in New Zealand. The REnKF updates past and present model states (soil water, aquifer and surface storages), with lags up to the concentration time of the catchment, to improve model initial conditions and hence flow forecasts. To our knowledge, this is the first time the REnKF has been applied in an operational setting, for a distributed model running over large catchments. We found the REnKF overcame instabilities in the standard EnKF which were associated with the natural lag time between upstream catchment wetness and flow at the gauging locations. The forecast system performance was correspondingly improved in terms of Nash Sutcliffe score and bounding of the measured flow by the model ensemble. We present descriptions of filter design parameters and explanations and examples of filter behaviour. The paper provides information and guidance valuable for other groups wishing to apply the REnKF for operational forecasting.

scholarly journals Sensitivity tests for an ensemble Kalman filter for aerosol assimilation

2010 ◽  
Vol 10 (3) ◽  
pp. 5947-5997
Author(s):  
N. A. J. Schutgens ◽  
T. Miyoshi ◽  
T. Takemura ◽  
T. Nakajima
Keyword(s):  
Kalman Filter ◽  
Ensemble Kalman Filter ◽  
Patch Size ◽  
Initial Conditions ◽  
Grid Cell ◽  
Ensemble Size ◽  
Sensitivity Tests ◽  
Local Patch ◽  
The Impact ◽  
Assimilation System

Abstract. We present sensitivity tests for a global aerosol assimilation system utilizing AERONET observations of AOT (aerosol optical thickness) and AAE (aerosol Ångström exponent). The assimilation system employs an ensemble Kalman filter which requires optimization of three numerical parameters: ensemble size nens, local patch size npatch and inflation factor ρ. In addition, experiments are performed to test the impact of various implementations of the system. For instance, we use a different prescription of the emission ensemble or a different combination of observations. The various experiments are compared against one-another and against independent AERONET andMODIS/Aqua observations. The assimilation leads to significant improvements in modelled AOT and AAE fields. Moreover remaining errors are mostly random while they are mostly systematic for an experiment without assimilation. In addition, these results do not depend much on our parameter or design choices. It appears that the value of the local patch size has by far the biggest impact on the assimilation, which has sufficiently converged for an ensemble size of nens=20. Assimilating AOT and AAE is clearly preferential to assimilating AOT at two different wavelengths. In contrast, initial conditions or a description of aerosol beyond two modes (coarse and fine) have only little effect. We also discuss the use of the ensemble spread as an error estimate of the analysed AOT and AAE fields. We show that a very common prescription of the emission ensemble (independent random modification in each grid cell) can have trouble generating sufficient spread in the forecast ensemble.

scholarly journals DATA ASSIMILATION BY ENSEMBLE KALMAN FILTER WITH THE LORENZ EQUATIONS

2016 ◽  
Vol 38 ◽  
pp. 190
Author(s):  
Regis Sperotto de Quadros ◽  
Fabrício Pereira Harter ◽  
Daniela Buske ◽  
Larri Silveira Pereira
Keyword(s):  
Kalman Filter ◽  
Data Assimilation ◽  
Ensemble Kalman Filter ◽  
Weather Forecasting ◽  
Initial Conditions ◽  
Background Field ◽  
Lorenz Equations ◽  
Integration Period ◽  
Numerical Weather Forecasting ◽  
Chaotic Nature

Data Assimilation is a procedure to get the initial condition as accurately as possible, through the statistical combination of collected observations and a background field, usually a short-range forecast. In this research a complete assimilation system for the Lorenz equations based on Ensemble Kalman Filter is presented and examined. The Lorenz model is chosen for its simplicity in structure and the dynamic similarities with primitive equations models, such as modern numerical weather forecasting. Based on results, was concluded that, in this implementation, 10 members is the best setting, because there is an overfitting for ensembles with 50 and 100 members. It was also examined if the EnKF is effective to track the control for 20% and 40% of error in the initial conditions. The results show a disagreement between the “truth” and the estimation, especially in the end of integration period, due the chaotic nature of the system.  It was also concluded that EnKF have to be performed sufficiently frequently in order to produce desirable results.

scholarly journals Comparison of Ensemble Kalman Filter–Based Forecasts to Traditional Ensemble and Deterministic Forecasts for a Case Study of Banded Snow

2012 ◽  
Vol 27 (1) ◽  
pp. 85-105 ◽  
Author(s):  
Astrid Suarez ◽  
Heather Dawn Reeves ◽  
Dustan Wheatley ◽  
Michael Coniglio
Keyword(s):  
Kalman Filter ◽  
Flow Pattern ◽  
Ensemble Kalman Filter ◽  
Initial Conditions ◽  
Closed Surface ◽  
Flow Patterns ◽  
Precipitation Forecast ◽  
Ensemble Mean ◽  
Correct Location

Abstract The ensemble Kalman filter (EnKF) technique is compared to other modeling approaches for a case study of banded snow. The forecasts include a 12- and 3-km grid-spaced deterministic forecast (D12 and D3), a 12-km 30-member ensemble (E12), and a 12-km 30-member ensemble with EnKF-based four-dimensional data assimilation (EKF12). In D12 and D3, flow patterns are not ideal for banded snow, but they have similar precipitation accumulations in the correct location. The increased resolution did not improve the quantitative precipitation forecast. The E12 ensemble mean has a flow pattern favorable for banding and precipitation in the approximate correct location, although the magnitudes and probabilities of relevant features are quite low. Six members produced good forecasts of the flow patterns and the precipitation structure. The EKF12 ensemble mean has an ideal flow pattern for banded snow and the mean produces banded precipitation, but relevant features are about 100 km too far north. The EKF12 has a much lower spread than does E12, a consequence of their different initial conditions. Comparison of the initial ensemble means shows that EKF12 has a closed surface low and a region of high low- to midlevel humidity that are not present in E12. These features act in concert to produce a stronger ensemble-mean cyclonic system with heavier precipitation at the time of banding.

A Probabilistic Collocation-Based Kalman Filter for History Matching

SPE Journal ◽  
2011 ◽  
Vol 16 (02) ◽  
pp. 294-306 ◽  
Author(s):  
Lingzao Zeng ◽  
Haibin Chang ◽  
Dongxiao Zhang
Keyword(s):  
Kalman Filter ◽  
Data Assimilation ◽  
Ensemble Kalman Filter ◽  
History Matching ◽  
Initial Conditions ◽  
Collocation Point ◽  
Computational Effort ◽  
Production Data ◽  
Sampling Errors ◽  
Probabilistic Collocation

Summary The ensemble Kalman filter (EnKF) has been used widely for data assimilation. Because the EnKF is a Monte Carlo-based method, a large ensemble size is required to reduce the sampling errors. In this study, a probabilistic collocation-based Kalman filter (PCKF) is developed to adjust the reservoir parameters to honor the production data. It combines the advantages of the EnKF for dynamic data assimilation and the polynomial chaos expansion (PCE) for efficient uncertainty quantification. In this approach, all the system parameters and states and the production data are approximated by the PCE. The PCE coefficients are solved with the probabilistic collocation method (PCM). Collocation realizations are constructed by choosing collocation point sets in the random space. The simulation for each collocation realization is solved forward in time independently by means of an existing deterministic solver, as in the EnKF method. In the analysis step, the needed covariance is approximated by the PCE coefficients. In this study, a square-root filter is employed to update the PCE coefficients. After the analysis, new collocation realizations are constructed. With the parameter collocation realizations as the inputs and the state collocation realizations as initial conditions, respectively, the simulations are forwarded to the next analysis step. Synthetic 2D water/oil examples are used to demonstrate the applicability of the PCKF in history matching. The results are compared with those from the EnKF on the basis of the same analysis. It is shown that the estimations provided by the PCKF are comparable to those obtained from the EnKF. The biggest improvement of the PCKF comes from the leading PCE approximation, with which the computational burden of the PCKF can be greatly reduced by means of a smaller number of simulation runs, and the PCKF outperforms the EnKF for a similar computational effort. When the correlation ratio is much smaller, the PCKF still provides estimations with a better accuracy for a small computational effort.

scholarly journals Improvement of ozone forecast over Beijing based on ensemble Kalman filter with simultaneous adjustment of initial conditions and emissions

2011 ◽  
Vol 11 (24) ◽  
pp. 12901-12916 ◽  
Author(s):  
X. Tang ◽  
J. Zhu ◽  
Z. F. Wang ◽  
A. Gbaguidi
Keyword(s):  
Kalman Filter ◽  
Air Quality ◽  
Data Assimilation ◽  
Ensemble Kalman Filter ◽  
Initial Conditions ◽  
Urban Site ◽  
Ozone Forecast ◽  
Regional Air Quality ◽  
Simultaneous Adjustment ◽  
Mean Square Errors

Abstract. In order to improve the surface ozone forecast over Beijing and surrounding regions, data assimilation method integrated into a high-resolution regional air quality model and a regional air quality monitoring network are employed. Several advanced data assimilation strategies based on ensemble Kalman filter are designed to adjust O3 initial conditions, NOx initial conditions and emissions, VOCs initial conditions and emissions separately or jointly through assimilating ozone observations. As a result, adjusting precursor initial conditions demonstrates potential improvement of the 1-h ozone forecast almost as great as shown by adjusting precursor emissions. Nevertheless, either adjusting precursor initial conditions or emissions show deficiency in improving the short-term ozone forecast at suburban areas. Adjusting ozone initial values brings significant improvement to the 1-h ozone forecast, and its limitations lie in the difficulty in improving the 1-h forecast at some urban site. A simultaneous adjustment of the above five variables is found to be able to reduce these limitations and display an overall better performance in improving both the 1-h and 24-h ozone forecast over these areas. The root mean square errors of 1-h ozone forecast at urban sites and suburban sites decrease by 51% and 58% respectively compared with those in free run. Through these experiments, we found that assimilating local ozone observations is determinant for ozone forecast over the observational area, while assimilating remote ozone observations could reduce the uncertainty in regional transport ozone.

scholarly journals Application of Ensemble Kalman Filter to Pedestrian Flow

2020 ◽  
Vol 5 ◽  
pp. A101
Author(s):  
Fumiya Togashi ◽  
Takashi Misaka ◽  
Rainald Löhner ◽  
Shigeru Obayashi
Keyword(s):  
Kalman Filter ◽  
Ensemble Kalman Filter ◽  
Degrees Of Freedom ◽  
Initial Conditions ◽  
Computational Simulation ◽  
Atmospheric Condition ◽  
Weather Forecast ◽  
Pedestrian Flow ◽  
Simulation Code ◽  
First Case

We adopted the Ensemble Kalman Filter (EnKF) methodology in our computational simulation code for pedestrian flows. The EnKF, which is a type of data assimilation methodology, has been developed in the field of weather forecast where the atmospheric condition varies hour by hour. The EnKF estimates the parameters or boundary/initial conditions in the numerical model based on the updated measured data. We considered the EnKF a promising tool for the simulation of pedestrian flows, which are notoriously difficult to predict. In this study, two scenarios were conducted to confirm the usefulness of the EnKF. The first case was unidirectional pedestrian flow in straight corridors, and the second case was Mataf scenario at the Kaaba in Mecca. Needless to say, the second scenario was very challenging because of the number of pilgrims and the degrees of freedom. In each scenario, we conducted the numerical simulation using the original parameter set and then applied the EnKF to improve the accuracy of the simulation.

Characterizing and Optimizing Precipitation Forecasts from a Convection-Permitting Ensemble Initialized by a Mesoscale Ensemble Kalman Filter

2014 ◽  
Vol 29 (6) ◽  
pp. 1295-1318 ◽  
Author(s):  
Craig S. Schwartz ◽  
Glen S. Romine ◽  
Kathryn R. Smith ◽  
Morris L. Weisman
Keyword(s):  
Kalman Filter ◽  
Ensemble Kalman Filter ◽  
Initial Conditions ◽  
Characteristic Curve ◽  
Wrf Model ◽  
Weather Research And Forecasting ◽  
Relative Operating Characteristic ◽  
Horizontal Grid ◽  
Grid Length ◽  
Heavy Rainfall Events

Abstract Convection-permitting Weather Research and Forecasting (WRF) Model forecasts with 3-km horizontal grid spacing were produced for a 50-member ensemble over a domain spanning three-quarters of the contiguous United States between 25 May and 25 June 2012. Initial conditions for the 3-km forecasts were provided by a continuously cycling ensemble Kalman filter (EnKF) analysis–forecast system with 15-km horizontal grid length. The 3-km forecasts were evaluated using both probabilistic and deterministic techniques with a focus on hourly precipitation. All 3-km ensemble members overpredicted rainfall and there was insufficient forecast precipitation spread. However, the ensemble demonstrated skill at discriminating between both light and heavy rainfall events, as measured by the area under the relative operating characteristic curve. Subensembles composed of 20–30 members usually demonstrated comparable resolution, reliability, and skill as the full 50-member ensemble. On average, deterministic forecasts initialized from mean EnKF analyses were at least as or more skillful than forecasts initialized from individual ensemble members “closest” to the mean EnKF analyses, and “patched together” forecasts composed of members closest to the ensemble mean during each forecast interval were skillful but came with caveats. The collective results underscore the need to improve convection-permitting ensemble spread and have important implications for optimizing EnKF-initialized forecasts.

scholarly journals Sensitivity tests for an ensemble Kalman filter for aerosol assimilation

2010 ◽  
Vol 10 (14) ◽  
pp. 6583-6600 ◽  
Author(s):  
N. A. J. Schutgens ◽  
T. Miyoshi ◽  
T. Takemura ◽  
T. Nakajima
Keyword(s):  
Kalman Filter ◽  
Ensemble Kalman Filter ◽  
Patch Size ◽  
Initial Conditions ◽  
Grid Cell ◽  
Ensemble Size ◽  
Sensitivity Tests ◽  
Local Patch ◽  
The Impact ◽  
Assimilation System

Abstract. We present sensitivity tests for a global aerosol assimilation system utilizing AERONET observations of AOT (aerosol optical thickness) and AAE (aerosol Ångström exponent). The assimilation system employs an ensemble Kalman filter which requires tuning of three numerical parameters: ensemble size nens, local patch size npatch and inflation factor ρ. In addition, experiments are performed to test the impact of various implementations of the system. For instance, we use a different prescription of the emission ensemble or a different combination of observations. The various experiments are compared against one-another and against independent AERONET and MODIS/Aqua observations. The assimilation leads to significant improvements in modelled AOT and AAE fields. Moreover remaining errors are mostly random while they are mostly systematic for an experiment without assimilation. In addition, these results do not depend much on our parameter or design choices. It appears that the value of the local patch size has by far the biggest impact on the assimilation, which has sufficiently converged for an ensemble size of nens=20. Assimilating AOT and AAE is clearly preferential to assimilating AOT at two different wavelengths. In contrast, initial conditions or a description of aerosol beyond two modes (coarse and fine) have only little effect. We also discuss the use of the ensemble spread as an error estimate of the analysed AOT and AAE fields. We show that a very common prescription of the emission ensemble (independent random modification in each grid cell) can have trouble generating sufficient spread in the forecast ensemble.

Sign in / Sign up

Export Citation Format

Share Document