Article
Engineering, Marine
Xiang Xing, Bainian Liu, Weimin Zhang, Jianping Wu, Xiaoqun Cao, Qunbo Huang
Summary: An adaptive scheme for Schur product covariance localization is proposed in this paper, which helps to significantly reduce spurious correlations and provide accurate covariances by adaptively obtaining the localization radius through a certain criterion of correlations with the background ensembles.
JOURNAL OF MARINE SCIENCE AND ENGINEERING
(2021)
Article
Meteorology & Atmospheric Sciences
Anthony T. Weaver, Marcin Chrust, Benjamin Menetrier, Andrea Piacentini
Summary: Modeling and cycling the background-error covariance matrix is an active area of research in data assimilation, especially when using filters to model background-error correlations. Updating the normalization factors on each assimilation cycle can be costly, but methods like randomization can provide accurate estimates. By approximating the normalization matrix as a separable product of two matrices, one for the horizontal and one for the vertical components, accurate estimates can be obtained with a large random sample.
QUARTERLY JOURNAL OF THE ROYAL METEOROLOGICAL SOCIETY
(2021)
Article
Meteorology & Atmospheric Sciences
Shujun Zhu, Bin Wang, Lin Zhang, Juanjuan Liu, Yongzhu Liu, Jiandong Gong, Shiming Xu, Yong Wang, Wenyu Huang, Li Liu, Yujun He, Xiangjun Wu, Bin Zhao, Fajing Chen
Summary: This study developed an ensemble four-dimensional variational (En4DVar) hybrid data assimilation system and evaluated its performance in terms of analysis quality and forecast skill. The results showed that the En4DVar system has the ability to improve the accuracy of forecasts, mainly due to the flow-dependent ensemble covariance provided by 4DEnVar.
JOURNAL OF ADVANCES IN MODELING EARTH SYSTEMS
(2022)
Article
Meteorology & Atmospheric Sciences
Loik Berre, Etienne Arbogast
Summary: The global data assimilation system at Meteo-France is based on a 4D-Var formulation with wavelet-based background-error covariances. Further research is conducted in collaboration with ECMWF in the framework of OOPS, resulting in a new 4D-hybrid formulation that combines the attractive features of 4D-Var and 4DEnVar. This new approach improves covariance handling and shows competitive forecast quality in experiments.
QUARTERLY JOURNAL OF THE ROYAL METEOROLOGICAL SOCIETY
(2023)
Article
Environmental Sciences
Jiandong Qiao, Chen Zhou, Yi Liu, Jiaqi Zhao, Zhengyu Zhao
Summary: Based on GNSS data from CMONOC and the IRI model, a fast three-dimensional electron density nowcasting model for China and its adjacent regions was developed. The covariance localization technique was introduced to reduce the memory requirement and improve the calculation speed of the model. Experimental results showed that the assimilation algorithm effectively improved the accuracy of ionospheric electron density nowcast.
Article
Environmental Sciences
Shujun Zhu, Bin Wang, Lin Zhang, Juanjuan Liu, Yongzhu Liu, Jiandong Gong, Shiming Xu, Yong Wang, Wenyu Huang, Li Liu, Yujun He, Xiangjun Wu, Bin Zhao, Fajing Chen
Summary: Many ensemble-based data assimilation methods use observation space localization to mitigate sampling errors. This study proposes a weighted average hypsometry to determine the vertical coordinates of radiance observations and successfully integrates it with an ensemble four-dimensional variational DA system. Experimental results show significant improvements in the analysis quality and forecast skills of the system, especially in the Southern Hemisphere.
Article
Meteorology & Atmospheric Sciences
Eun-Gyeong Yang, Hyun Mee Kim
Summary: This study investigated the performance of three data assimilation methods based on the WRF model over East Asia. The hybrid E3DVAR method outperformed 3DVAR and EnKF for both January and July seasons. Adjusting background error covariance can improve forecast accuracy, and each method has different strengths in different seasons.
ATMOSPHERIC RESEARCH
(2021)
Article
Meteorology & Atmospheric Sciences
Chris Snyder, Gregory J. Hakim
Summary: This article introduces a linear transformation in which state variables and observations have uncorrelated errors in the transformed space, with a diagonal gain matrix in the update step. The transformation includes canonical observation operators (COOs) in the update step, which rank pairs of transformed observations and state variables based on their influence.
JOURNAL OF ADVANCES IN MODELING EARTH SYSTEMS
(2022)
Article
Meteorology & Atmospheric Sciences
Haohao Sun, Lili Lei, Zhengyu Liu, Liang Ning, Zhe-Min Tan
Summary: An analog offline ensemble Kalman filter (AOEnKF) is proposed, which constructs ensemble priors from a control climate simulation for each assimilation time based on an analog criterion using proxy observations. AOEnKF generates smaller posterior errors and requires much less computational cost compared to the online cycling EnKF (CEnKF). It has the advantages of having a more accurate prior ensemble mean and flow-dependent background error covariances compared to the commonly applied offline EnKF (OEnKF).
JOURNAL OF ADVANCES IN MODELING EARTH SYSTEMS
(2022)
Article
Computer Science, Interdisciplinary Applications
Habib Toye, Peng Zhan, Furrukh Sana, Sivareddy Sanikommu, Naila Raboudi, Ibrahim Hoteit
Summary: EnOI is a variant of EnKF that uses a static ensemble to reduce computational cost. A new adaptive EnOI approach is proposed to better represent the time-varying circulation of the Red Sea by adaptively selecting ensemble members at each assimilation cycle.
JOURNAL OF COMPUTATIONAL SCIENCE
(2021)
Article
Computer Science, Interdisciplinary Applications
Habib Toye, Peng Zhan, Furrukh Sana, Sivareddy Sanikommu, Naila Raboudi, Ibrahim Hoteit
Summary: EnOI is a variant of EnKF that reduces computational cost by using a static ensemble to parameterize the background covariance matrix. An adaptive EnOI approach is proposed to better represent the time-varying circulation of the Red Sea, with different schemes for selecting ensemble members tested. Results from assimilating real remote sensing data into a high resolution circulation model of the Red Sea are presented and discussed, demonstrating the relevance of the proposed schemes.
JOURNAL OF COMPUTATIONAL SCIENCE
(2021)
Article
Mathematics, Applied
T. Malou, J. Monnier
Summary: The estimation of the background error covariance operator in data assimilation (DA) is a classical and open topic. This study proposes a method to derive covariance operators from the underlying equations and models them using Green's kernels. The physically-derived operators show better accuracy and faster convergence compared to empirical operators.
Article
Meteorology & Atmospheric Sciences
Shujun Zhu, Bin Wang, Lin Zhang, Juanjuan Liu, Yongzhu Liu, Jiandong Gong, Shiming Xu, Yong Wang, Wenyu Huang, Li Liu, Yujun He, Xiangju Wu
Summary: A new 4DEnVar data assimilation system is developed based on GRAPES-GFS, which improves the performance of four-dimensional variational data assimilation algorithm through dimension-reduced projection technique and ensemble-sample-based localization method. Experimental results show that the new system has better forecast performance and higher forecast skills.
JOURNAL OF ADVANCES IN MODELING EARTH SYSTEMS
(2022)
Article
Engineering, Petroleum
Ricardo Vasconcellos Soares, Xiaodong Luo, Geir Evensen, Tuhin Bhakto
Summary: This study demonstrates the practical advantages of a new local analysis scheme in a 4D seismic history-matching problem. Compared to the Kalman gain localization scheme, the proposed local analysis scheme has improved capacity in handling big models and data sets, leading to faster convergence to the same level of data mismatch values.
Article
Computer Science, Interdisciplinary Applications
Thiago M. D. Silva, Sinesio Pesco, Abelardo Barreto, Mustafa Onur
Summary: ES-MDA is a powerful tool for history matching problems that can provide good data matching and model parameter estimates. This study introduces a new method for generating data covariance inflation factors for ES-MDA, which computes the correct number of data assimilations to achieve better model parameter match and data match.
COMPUTERS & GEOSCIENCES
(2021)
Article
Meteorology & Atmospheric Sciences
Mark Buehner, Dominik Jacques
MONTHLY WEATHER REVIEW
(2020)
Article
Meteorology & Atmospheric Sciences
Xiangming Sun, Xiang-Yu Huang, Chris Gordon, Marion Mittermaier, Rebecca Beckett, Wee Kiong Cheong, Dale Barker, Rachel North, Allison Semple
WEATHER AND FORECASTING
(2020)
Article
Meteorology & Atmospheric Sciences
Maziar Bani Shahabadi, Mark Buehner, Josep Aparicio, Louis Garand
MONTHLY WEATHER REVIEW
(2020)
Review
Meteorology & Atmospheric Sciences
Anurag Dipankar, Stuart Webster, Xiangming Sun, Claudio Sanchez, Rachel North, Kalli Furtado, Jonathan Wilkinson, Adrian Lock, Simon Vosper, Xiang-Yu Huang, Dale Barker
QUARTERLY JOURNAL OF THE ROYAL METEOROLOGICAL SOCIETY
(2020)
Article
Geosciences, Multidisciplinary
Bruce Ingleby, Brett Candy, John Eyre, Thomas Haiden, Christopher Hill, Lars Isaksen, Daryl Kleist, Fiona Smith, Peter Steinle, Stewart Taylor, Warren Tennant, Christopher Tingwell
Summary: Aircraft reports are crucial for numerical weather prediction, but the COVID-19 pandemic caused a loss of aircraft data with no significant degradation in forecast skill. Forecast skill is highly variable and depends on daily noise as well as the mean state of the atmosphere. Using a data denial experiment is the best way to assess the impact of aircraft data on weather forecasts.
GEOPHYSICAL RESEARCH LETTERS
(2021)
Article
Geochemistry & Geophysics
Alexander S. Komarov, Mark Buehner
Summary: A new technique for automated retrieval of ice concentration from RADARSAT-2 images was introduced, showing high accuracy across various spatial scales. The method effectively estimates ice concentration in different spatial scales and produces better agreement with Canadian Ice Service image analysis data.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2021)
Article
Meteorology & Atmospheric Sciences
K. Andrew Peterson, Gregory C. Smith, Jean-Francois Lemieux, Francois Roy, Mark Buehner, Alain Caya, Pieter L. Houtekamer, Hai Lin, Ryan Muncaster, Xingxiu Deng, Frederic Dupont, Normand Gagnon, Yukie Hata, Yosvany Martinez, Juan Sebastian Fontecilla, Dorina Surcel-Colan
Summary: The Global Ensemble Prediction System (GEPS) of Environment and Climate Change Canada has been upgraded to a coupled atmosphere, ocean, and sea-ice version, and shows improved sea-ice prediction compared to persistence and a deterministic Global Deterministic Prediction System (GDPS). The ensemble system offers enhanced benefits over a single deterministic forecast during the minimum and maximum extent periods and the early freeze-up period, although further improvement of the spread/error relationship is needed.
QUARTERLY JOURNAL OF THE ROYAL METEOROLOGICAL SOCIETY
(2022)
Article
Meteorology & Atmospheric Sciences
Craig H. Bishop, Nathan W. Eizenberg
Summary: This paper introduces an implicit ensemble TLM (IETLM) to predict the difference between perturbed and unperturbed nonlinear forecasts. The accuracy of the IETLM is confirmed in the linear regime and a diagonally robust (DR) IETLM is developed for ensemble perturbations in the nonlinear regime. The performance of the DR IETLM is found to match that of the traditional TLM over a wide range of non-linearity.
QUARTERLY JOURNAL OF THE ROYAL METEOROLOGICAL SOCIETY
(2022)
Article
Meteorology & Atmospheric Sciences
G. W. Inverarity, W. J. Tennant, L. Anton, N. E. Bowler, A. M. Clayton, M. Jardak, A. C. Lorenc, F. Rawlins, S. A. Thompson, M. S. Thurlow, D. N. Walters, M. A. Wlasak
Summary: The MOGREPS-G system underwent several enhancements and upgrades since September 2008, including the application of hybrid four-dimensional ensemble variational data assimilation (En-4DEnVar) and improvements in inflation and ensemble spread. These changes have significantly improved ensemble forecasts, but initially had a more neutral impact on deterministic forecasts. A subsequent operational upgrade in December 2020 further improved the deterministic forecast's hybrid data assimilation by introducing shifting in addition to lagging. The adoption of hybrid 4DEnVar in MOGREPS-G also reduces maintenance overheads and enables the assimilation of various observation types.
QUARTERLY JOURNAL OF THE ROYAL METEOROLOGICAL SOCIETY
(2023)
Article
Meteorology & Atmospheric Sciences
Marco Milan, Adam Clayton, Andrew Lorenc, Bruce Macpherson, Robert Tubbs, Gareth Dow
Summary: This study improved the representation of large-scale dynamics in the Met Office limited-area model (LAM) data assimilation system by utilizing the better estimation of these scales from the host model. The method called large-scale blending (LSB) constrained the LAM evolution by a host model with a better representation of the larger scales while preserving the smaller scales predicted by the LAM. The introduction of LSB improved the LAM forecast for various variables.
QUARTERLY JOURNAL OF THE ROYAL METEOROLOGICAL SOCIETY
(2023)
Article
Astronomy & Astrophysics
Daniel Hodyss, Douglas R. Allen, Daniel Tyndall, Peter Caffrey, Sarah E. McDonald
Summary: Data assimilation (DA) is the process of merging information from prediction models with observations to estimate the state of a physical system. In the context of ionospheric physics-based models, this study focuses on understanding how a DA algorithm responds to estimating an external parameter driving the model's interpretation of solar ionizing radiation. The results demonstrate the impact of solar forcing and recombination on the estimation process in a linear and Gaussian framework.
JOURNAL OF SPACE WEATHER AND SPACE CLIMATE
(2023)
Article
Astronomy & Astrophysics
Douglas R. Allen, Daniel Hodyss, Victoriya V. Forsythe, Sarah E. McDonald
Summary: Comparisons of TEC variability among two SAMI3 model runs and JPL/GIM show that there is a non-zero large-scale base correlation in all three datasets for the year 2014 with high solar activity. The SAMI3 runs generally exhibit higher correlations than JPL/GIM, and the correlation values strongly correlate with monthly F10.7 standard deviations.
JOURNAL OF SPACE WEATHER AND SPACE CLIMATE
(2023)
Article
Geochemistry & Geophysics
Alexander S. Komarov, Sergey A. Komarov, Mark Buehner
Summary: New techniques for automated retrieval of ocean surface wind speed from the RADARSAT Constellation mission have been developed and tested, showing lower root-mean square errors compared to existing models that require input wind direction.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Meteorology & Atmospheric Sciences
Greg Roff, Ilia Bermous, Gary Dietachmayer, Joan Fernon, Jim Fraser, Wenming Lu, Susan Rennie, Peter Steinle, Yi Xiao
Summary: The APS2 ACCESS-C2, an upgraded weather prediction system developed by the Australian Bureau of Meteorology, shows significant improvements in forecast accuracy and resolution, making it the first operational convection-permitting model in Australia.
JOURNAL OF SOUTHERN HEMISPHERE EARTH SYSTEMS SCIENCE
(2022)
Article
Geochemistry & Geophysics
Nazanin Asadi, K. Andrea Scott, Alexander S. Komarov, Mark Buehner, David A. Clausi
Summary: This article explores the classification of SAR sea ice imagery into ice, water, or unknown using a multilayer perceptron neural network. Uncertainties in the MLP models, including epistemic uncertainty and aleatoric uncertainty, are also considered. The inclusion of uncertainties in the MLP models slightly reduces accuracies but also reduces misclassification rates, which is important for data assimilation applications.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2021)
