CORe is an atmospheric reanalysis from the Climate Prediction Center NCEP/NWS/NOAA with collaboration with Environmental Modeling Center NCEP/NWS/NOAA and Physical Sciences Laboratory NOAA. CORe is designed to replace CDAS (real-time extension to NCEP/NCAR Reanalysis). CORe covers a similar period of time (1950 to real time vs 1948 to real time), and was designed for climate monitoring, one of the Climate Prediction Center's major tasks. Unlike most modern reanalyses, CORe does not assimilate satellite data except for Atmospheric Motion Vectors (AMV). This was done to eliminate spurious "climate shifts" caused by the ingest of satellite data. This problem affected the earlier NCEP/NCAR and Climate Forecast System Reanalyses. This design emphasizes improving the climate trends and anomalies over getting the best state for a forecast.

Descriptions of the CORe can be fund in the following presentations and documents

1. A bnief introduction to CORe
2. Improvements that CORe brings to climate monitoring
3. Concepts of the CORe design
4. Configuration of CORe
5. Evaluation data sets

Workshop and Conference Presentations

CORe version 0: GFS global spectral version

Hierarchy of reanalyses, why only AMV for satellite data
A Preliminary Examination of a Conventional EnKF Atmospheric Reanalysis
41th NOAA annual Climate Diagnostics and Prediction Workshop, Oct 3-6, 2016
Ebisuzaki et al
local copy
https://www.weather.gov/media/sti/climate/STIP/41CDPW/41cdpw-WEbisuzaki.pdf

Initial Assessment of the Conventional Observation Reanalysis
41th NOAA annual Climate Diagnostics and Prediction Workshop, Oct 3-6, 2016
Zhang et al
https://www.weather.gov/media/sti/climate/STIP/41CDPW/41cdpw-LZhang.pdf

Results from phase 0 1950-2010
A Conventional Only Atmospheric Reanalysis (CORe)
5th International Conference for Reanalysis, Nov 2017, Ebisuzaki et al
poster

CORe version 1: FV3GFS version

The CORe version 0 had problems. NOAA now required that the FV3 model be used for atmospheric modelling. Other problems were that the global precipitation was too large (GFS spectral model problem) and there were inconsistencies between the streams (human problem). Fortunately moving to the FV3GFS model reduced the bias in the global precipitation, and the human problems were fixed.

Prior to 1979, global snow depth or cover analyses were not available. Usually one uses the model snow in the pre-satellite period. A comparison of the phase 1 reanalysis for 2016 using model and observed snow depth showed that the model snow had more coverage than observed. So a simple parameterization was developed to reduce the bias in the modelled snow cover.

Introduction to phase 1 vs phase 0
Note: presentation used the old names (phase 1, 2) rather than new names (phase 0, 1)
A Conventional Observation ENKF Atmospheric Reanalysis for Climate Monitoring at NCEP
AGU 2018 A13M_2640, Ebisuzaki et al
poster

A Conventional Observation Reanalysis (CORe) for Climate Monitoring
44th NOAA annual Climate Diagnostics and Prediction Workshop, Oct 22-24, 2019
Ebisuzaki et al
local copy
https://www.weather.gov/media/sti/climate/STIP/44CDPW/44cdpw-WEbisuzaki.pdf

Snow in pre-satellite period, satellite-period results
A conventional Observation Reanalysis (CORe) for Climate Monitoring
45th NOAA annual Climate Diagnostics and Prediction Workshop, Oct 20-22, 2020
Ebisuzaki et al (extended abstract)
local copy
https://www.nws.noaa.gov/ost/STIClimateBulletin/45CDPW/45cdpw-WEbisuzaki.pdf

Snow in presatellite period
Difficulties with Snow in the Conventional Observation Reanalysis (CORe)
Workshop on the Future of US Earth System Reanalysis, May 2022, Ebisuzaki et al
pdf

Trends: reanalyses now can get the trends
Trends in a Conventional Observation Reanalysis (CORe) and other Reanalyses
Workshop on the Future of US Earth System Reanalysis, May 2022, Ebisuzaki et al
pdf

Introduction to CORe for Environmental Modeling Center (NCEP)
Differences from CFSR in design and results
Software developments that could be used by EMC
CORe
EMC meeting, August 2022, Ebisuzaki et al
pdf

Improvements of Reanalyses over the last 10 years (2012-2022).
CORe, ERA-5 and JRA-55 agree in large-scale trends, older reanalyses not so much
Trends from Reanalyses: Progress over the Last 10 Years
47th NOAA annual Climate Diagnostics and Prediction Workshop, Oct 25-27, 2022, Ebisuzaki et al
pdf

Quick overview of CORe for ensemble modelers
Best (short) introduction for users?
A Large FV3GFS Ensemble in a Reanalysis System
9th NOAA Ensemble Users Workshop, August 24, 2023, Ebisuzaki et al
pdf

Low frequency variability from various reanalyses
ENSO and non-ENSO low frequency variability from the Modern Reanalysis Data Sets* and Implications!
48th NOAA annual Climate Diagnostics and Prediction Workshop, 26–29 March 2024, M. Chelliah
pdf

Using CORe to force the UFS land model Developing a new long-term and real-time land surface monitoring product
American Meteorology Society, annual meeting, 2024, Xu et al.,
pdf

CPC expert evaluations (CORe version 1)

Before deciding to make CORe operational, many members of the Climate Prediction Center were tasked to evaluate CORe with respect to other reanalyses, observered data sets and how it would affect operational monitoring. These are all the CPC evaluations.

Correcting the CORe downward radiative fluxes to resemble CERES
Evaluating and Correcting CORe Surface Downward Shortwave and Longwave Fluxes using NASA CERES EBAF-Surface Observations, Hailan Wang
pdf

CORe vs Operational MJO indices
Comparison of MJO RMM Indices in CORe and CPC-Ops, Zewdu Segele, Nicholas Novella
pdf

MJO: CORe, R1 (NCEP/NCAR), R2 (NCEP/DOE), CFSv2 and observations (CMORPH), includes spectral analysis
JieShun Zhu
pdf

ENSO comparison
ENSO Comparisons, Michelle L’Heureux
pdf

Stratospheric evaluation
Evaluation of the Stratosphere in the CORe dataset, Laura Ciasto
pdf

Comparison of sea ice in CORe (OSTIA) to CDRv4
Arctic sea ice and variability, Liu Yanyun
pdf CORe uses a prescribed sea ice from OSTIA. CDRv4 is a sea ice concentration from NOAA/NSIDC based on passive microwave satellite data. Both data sets are based on observations.

Evaluation of surface fluxes, and CORe SST (OSTIA) vs OISSTv2.1
Evaluation of Surface fluxes and SST in the CORe data set, Caihong Wen
pdf
CORe SST is prescribed and from OSTIA. OISSTv2.1 is a SST data set from NOAA. Both are based on observations.

Soil moisture
Evaluation of Soil Moisture in the CORe data set, Yun Fan
pdf

Teleconnections
Teleconnections Patterns & Indices in CORe and the other Re-analyses, Mingyue Chen
pdf

2 m temperature validation
Verifying T2m in CORe, Pingping Xie, Saorong Wu
pdf

CORe precipitation vs gauge precipitation
Verifying CORe Precipitation against a Gauge-based Analysis of Daily Precipitation over the Global Land, Pingping Xie, Shaorong Wu
pdf

CORe precipitation vs radar precipitation
Verifying CORe 3-Hourly Precipitation against MRMS Multi-Sensor Radar Observations over CONUS, Pingping Xie, Shaorong Wu
pdf

Outgoing Longwave Radiation (OLR)
A Comparison of the Outgoing Longwave Radiation (OLR) from the CORe and the CPC Blended OLR (CBO), Pingping Xie, Shaorong Wu
pdf

Results from tropical cyclone tracker, #1
CORe TC Analysis and Comparison to CFSR 1950-2020, Lindsey Long
pdf The CORe "analysis" is an mean of the 80 analyses created by an 80 member ensemble. By virtue of being an ensemble mean, the peak winds are reduced depending on the uncertainty of the analysis. The ensemble mean will have a smaller RMS error than any individual member but this plays havoc with the TC tracker. One needs to run the TC tracker on each of the 80 analyses and then average the results. See next result.

Results from tropical cyclone tracker, #2
CORe TC Analysis 2018 Individual Members, Lindsey Long
pdf

Results from storm track algorithm
Storm Track Climatology Evaluation, CORe vs. CFSR (43 years: 1979-2021), Yutong Pan et al
pdf

Low frequency variability from various reanalyses
ENSO and non-ENSO low frequency variability from the Modern Reanalysis Data Sets* and Implications!
48th NOAA annual Climate Diagnostics and Prediction Workshop, 26–29 March 2024, Muthuvel Chelliah
pdf

Short wave, long wave fluxes at top of the atmosphere
CORe TOA Radiation Analyses, Zen-Zhen Hu et al.
pdf
Getting the correct absolute values of the short and long wave fluxes in reanalyses are difficult because both are dependant on the clouds. If the cloud fraction, cloud properties or heights are off, the fluxes will be different than observed. Maybe we will have to wait until we have cloud resolving models to get the absolute values correct. Until then, we can look at the anomalies of the fluxes to get an idea about the low frequency variability.

The figure g on page 10 shows the CORe equatorial downward shortwave flux at the top of the atmosphere. The plot shows a latitudial variablilty of 3.5 W/m&&2 with 24 cycles in July (estimated from July 2020). Since the DSWRFtoa is purely a calculation based on the orbital characteristics of the earth, you would expect the DSSWRFtoa to have a single cycle during any month. Surprising all the reanalyses showed evidence of a 24 longitudinal cycle. I suspect the code that calculates the DSWRFtoa needs to be made more accurate. However, the modelers will have to return the final verdict.

Comments: Wanqui.Wang@noaa.gov cc: Arun.Kumar@noaa.gov Wesley.Ebisuzaki@noaa.gov