A5075733431- Meteorological Phenomena and Simulations
- Climate variability and models
- Atmospheric and Environmental Gas Dynamics
- Precipitation Measurement and Analysis
- Tropical and Extratropical Cyclones Research
- Hydrology and Drought Analysis
- Wind and Air Flow Studies
- Oceanographic and Atmospheric Processes
- Plant Water Relations and Carbon Dynamics
- Energy Load and Power Forecasting
- Hydrology and Watershed Management Studies
- Geophysics and Gravity Measurements
- Atmospheric aerosols and clouds
- Fire effects on ecosystems
- Reservoir Engineering and Simulation Methods
- Remote Sensing in Agriculture
- Cryospheric studies and observations
- Solar Radiation and Photovoltaics
- Remote-Sensing Image Classification
- Hydrological Forecasting Using AI
- Soil Geostatistics and Mapping
- Structural Health Monitoring Techniques
- Geochemistry and Geologic Mapping
- Aquatic Ecosystems and Phytoplankton Dynamics
- Regional Economic and Spatial Analysis
Physical Sciences (United States)
2009-2025
NOAA National Weather Service
2005-2025
NOAA Physical Sciences Laboratory
2015-2023
NOAA Earth System Research Laboratory
2011-2020
European Centre for Medium-Range Weather Forecasts
2020
Cooperative Institute for Research in Environmental Sciences
2003-2020
National Oceanic and Atmospheric Administration
2007-2020
University of Colorado Boulder
2004-2020
University of Washington
2020
NOAA National Centers for Environmental Prediction
2019
The ensemble Kalman filter (EnKF) is a data assimilation scheme based on the traditional update equation. An of forecasts are used to estimate background-error covariances needed compute gain. It known that if same observations and gain each member ensemble, will systematically underestimate analysis-error covariances. This cause degradation subsequent analyses may lead divergence. For large ensembles, it this problem can be alleviated by treating as random variables, adding perturbations...
The usefulness of a distance-dependent reduction background error covariance estimates in an ensemble Kalman filter is demonstrated. Covariances are reduced by performing elementwise multiplication the matrix with correlation function local support. This reduces noisiness and results improved estimate, which generates reduced-error model initial conditions. benefits applying can be understood part from examining characteristics simple 2 × matrices generated random sample vectors known...
Rank histograms are a tool for evaluating ensemble forecasts. They useful determining the reliability of forecasts and diagnosing errors in its mean spread. generated by repeatedly tallying rank verification (usually an observation) relative to values from sorted lowest highest. However, uncritical use histogram can lead misinterpretations qualities that ensemble. For example, flat histogram, usually taken as sign reliability, still be unreliable ensembles. Similarly, U-shaped commonly...
Ensemble data assimilation methods assimilate observations using state-space estimation and low-rank representations of forecast analysis error covariances. A key element such is the transformation ensemble into an with appropriate statistics. This may be performed stochastically by treating as random variables, or deterministically requiring that updated perturbations satisfy Kalman filter covariance equation. Deterministic updates are implementations square root filters. The nonuniqueness...
A hybrid ensemble Kalman filter–three-dimensional variational (3DVAR) analysis scheme is demonstrated using a quasigeostrophic model under perfect-model assumptions. Four networks with differing observational densities are tested, including one network data void. The operates by computing set of parallel assimilation cycles, each member the receiving unique perturbed observations. observations generated adding random noise consistent observation error statistics to control Background for...
Abstract Real-data experiments with an ensemble data assimilation system using the NCEP Global Forecast System model were performed and compared Data Assimilation (GDAS). All observations in operational stream assimilated for period 1 January–10 February 2004, except satellite radiances. Because of computational resource limitations, comparison was done at lower resolution (triangular truncation wavenumber 62 28 levels) than GDAS real-time runs 254 64 levels). The outperformed...
Abstract Inflation of ensemble perturbations is employed in Kalman filters to account for unrepresented error sources. The authors propose a multiplicative inflation algorithm that inflates the posterior proportion amount observations reduce spread, resulting more regions dense observations. This justified since variance affected by sampling errors these regions. similar “relaxation prior” proposed Zhang et al., but it relaxes spread back prior instead perturbations. new compared method al....
Motivated by the success of ensemble forecasting at medium range, performance a prototype short-range forecast system is examined. The dataset consists 15 case days from September 1995 through January 1996. There are members ensemble, 10 an 80-km version eta model and five regional spectral model. Initial conditions include various in-house analyses available National Centers for Environmental Prediction as well bred initial interpolated medium-range ensemble. Forecasts 29-km mesoeta were...
Ensemble forecasting is increasingly accepted as a powerful tool to improve early warnings for high-impact weather. Recently, ensembles combining forecasts from different systems have attracted considerable level of interest. The Observing System Research and Predictability Experiment (THORPEX) Interactive Grand Globa l (TIGGE) project, prominent contribution THORPEX, has been initiated enable advanced research demonstration the multimodel ensemble concept pave way toward operational...
A multidecadal ensemble reforecast database is now available that approximately consistent with the operational 0000 UTC cycle of 2012 NOAA Global Ensemble Forecast System (GEFS). The dataset consists an 11-member run once each day from initial conditions. Reforecasts are to +16 days. As GEFS, at T254L42 resolution (approximately 1/2° grid spacing, 42 levels) for week +1 forecasts and T190L42 3/4° spacing) +2 forecasts. were initialized Climate Reanalysis conditions, perturbations generated...
Abstract A hybrid ensemble transform Kalman filter–three-dimensional variational data assimilation (ETKF–3DVAR) system for the Weather Research and Forecasting (WRF) Model is introduced. The based on existing WRF 3DVAR. Unlike 3DVAR, which utilizes a simple, static covariance model to estimate forecast-error statistics, combines covariances with complex, flow-dependent statistics. Ensemble are incorporated by using extended control variable method during minimization. perturbations...
Abstract The International Grand Global Ensemble (TIGGE) was a major component of Observing System Research and Predictability Experiment (THORPEX) research program, whose aim is to accelerate improvements in forecasting high-impact weather. By providing ensemble prediction data from leading operational forecast centers, TIGGE has enhanced collaboration between the meteorological communities enabled studies on wide range topics. paper covers objective evaluation data. For parameters, it...
Abstract A parametric statistical postprocessing method is presented that transforms raw (and frequently biased) ensemble forecasts from the Global Ensemble Forecast System (GEFS) into reliable predictive probability distributions for precipitation accumulations. Exploratory analysis based on 12 years of reforecast data and ⅛° climatology-calibrated analyses shows censored, shifted gamma can well approximate conditional distribution observed accumulations given forecasts. nonhomogeneous...
Abstract When evaluating differences between competing precipitation forecasts, formal hypothesis testing is rarely performed. This may be due to the difficulty in applying common tests given spatial correlation of and non-normality errors. Possible ways around these difficulties are explored here. Two datasets forecasts evaluated, a set two gridded from operational weather prediction models sets probabilistic quantitative model output statistics an ensemble forecasts. For each test, data...
Abstract A general theory is proposed for the statistical correction of weather forecasts based on observed analogs. An estimate sought probability density function (pdf) state, given today’s numerical forecast. Assume that an infinite set reforecasts (hindcasts) and associated observations are available climate stable. it possible to find a past model forecast states nearly identical current state. With dates these forecasts, asymptotically correct probabilistic can be formed from...
The value of the model output statistics (MOS) approach to improving 6–10-day and week 2 probabilistic forecasts surface temperature precipitation is demonstrated. Retrospective 2-week ensemble “reforecasts” were computed using a version NCEP medium-range forecast with physics operational during 1998. An NCEP–NCAR reanalysis initial condition bred modes used initialize 15-member ensemble. Probabilistic generated by logistic regression technique mean (precipitation) or anomaly (temperature)...
Studies using idealized ensemble data assimilation systems have shown that flow-dependent background-error covariances are most beneficial when the observing network is sparse. The computational cost of recently proposed algorithms directly proportional to number observations being assimilated. Therefore, ensemble-based should both be more computationally feasible and provide greatest benefit over current operational schemes in situations Reanalysis before radiosonde era (pre-1931) just such...
Abstract It is common practice to summarize the skill of weather forecasts from an accumulation samples spanning many locations and dates. In calculating these scores, there implicit assumption that climatological frequency event occurrence approximately invariant over all samples. If actually varies among samples, metrics may report a different expected. Many deterministic verification metrics, such as threat are prone mis‐reporting skill, probabilistic forecast Brier score relative...
Abstract Three recently proposed and promising methods for postprocessing ensemble forecasts based on their historical error characteristics (i.e., ensemble-model output statistics methods) are compared using a multidecadal reforecast dataset. Logistic regressions nonhomogeneous Gaussian generally preferred daily temperature, medium-range (6–10 8–14 day) temperature precipitation forecasts. However, the better sharpness of ensemble-dressing sometimes yields best Brier scores even though...
The accuracy of short-range probabilistic forecasts quantitative precipitation (PQPF) from the experimental Eta–Regional Spectral Model ensemble is compared with Nested Grid Model's model output statistics (MOS) over a set 13 case days September 1995 through January 1996. Ensembles adjusted to compensate for deficiencies noted in prior were found be more skillful than MOS all categories except basic probability measurable precipitation. Gamma distributions fit corrected provided an...
A "reforecast" (retrospective forecast) dataset has been developed. This is comprised of a 15-member ensemble run out to 2-week lead. Forecasts have every day from 0000 UTC initial conditions 1979 the present. The model 1998 version National Centers for Environmental Prediction's (NCEP's) Global Forecast System (GFS) at T62 resolution. 15 consist reanalysis and seven pairs bred modes. facilitates number applications that were heretofore impossible. Model errors can be diagnosed past...
The Hydrological Ensemble Prediction Experiment (HEPEX) is an international project to advance technologies for hydrological forecasting. Its goal “to bring the and meteorological communities together demonstrate how produce utilize reliable ensemble forecasts make decisions benefit of public health safety, economy, environment.” HEPEX open group composed primarily researchers, forecasters, water managers, users. welcomes new members. In first workshop, held in spring of2004, participants...
Abstract As a companion to Part I, which discussed the calibration of probabilistic 2-m temperature forecasts using large training datasets, II discusses 12-hourly precipitation amounts. Again, ensemble reforecast datasets from European Centre for Medium-Range Weather Forecasts (ECMWF) and Global Forecast System (GFS) were used testing calibration. North American Regional Reanalysis (NARR) analysis data verification training. Logistic regression was perform calibration, with...
Abstract The hybrid ensemble transform Kalman filter–three-dimensional variational data assimilation (ETKF–3DVAR) system developed for the Weather Research and Forecasting (WRF) Model was further tested with real observations, as a follow-up observation simulation experiment (OSSE) conducted in Part I. A domain encompassing North America considered. Because of limited computational resources large number experiments conducted, forecasts analyses employed relatively coarse grid spacing (200...