The Ocean Archive System searches our original datasets as they were submitted to us, not individual points or profiles. If you want to search and retrieve ocean profiles in a common format, or objectively analyzed fields, your better option may be to use one of our project applications. See: Access Data
OAS accession Detail for 0259491, meta_version: 1. Current meta_version is: 3
| << previous | |revision: 1| | next >> |
| accessions_id: | 0259491 | archive |
|---|---|
| Title: | Real time multivariate Madden Julian Oscillation projection coefficients for Global Ensemble Forecast System v12 Reanalysis (2000-01-01 to 2020-03-31) and Reforecasts (weekly initializations between 2000-01-05 and 2019-12-25) (NCEI Accession 0259491) |
| Abstract: | Overview: This text provides details about the Real time multivariate MJO (RMM) indices available for the GEFSv12 reanalysis and reforecast datasets. Methodology: The MJO indices are calculated using the methodology outlined in Wheeler and Hendon (2004). Reanalysis RMMs: For the GEFSv12 reanalysis, anomalies of outgoing longwave radiation (OLR), 200 hPa zonal winds (U200), and 850 hPa zonal winds (U850) are calculated by removing the mean and climatological seasonal cycle (first three harmonics) of each of these fields from their full January 1st 2000 – December 31st 2019 timeseries. Next, for one of the aforementioned fields of anomalies on a given day in January 1st 2000 – December 31st 2019 timeseries, the mean of the most recent 120 days of anomaly data for that field is subtracted to remove the influence of low frequency variability (e.g., ENSO). In order to remove the means of the most recent 120 days for the early 2000 period, year 1999 data is needed, therefore we use anomalies derived from Climate Forecast System Reanalysis (Saha et al. 2010). CFSR data is also used to extend the MJO indices beyond 2020 for comparison with reforecasts initialized in December of 2019. Next, the anomalous fields are latitudinally averaged between 15S and 15N using square root of cosine weighting. Then each of these latitudinally averaged sets of anomalies is divided by the normalization factor provided by WH04: 15.1 W/m2 for OLR; 4.81 m/s for U200 1.81 m/s for U850. The anomalous fields are then projected onto the GEFSv12 reanalysis EOFs. Finally, the projection coefficients are standardized by their respective standard deviations (8.6 and 8.4). Reanalysis data is provided for January 1, 2000 through March 31, 2020. Reforecast ensemble mean RMMs: For the GEFSv12 reforecasts, each of the 11 ensemble members for a given reforecast initialization are first averaged together. The MJO indices are then calculated using the adaptations of WH04 methodology applied to forecast datasets detailed by Gottschalck et al. (2010). Different from step (i) of G2010, here, anomalies of OLR, U200, and U850 are calculated relative to lead-time dependent reforecast climatologies as described in Pegion et al. (2019), their Appendix B, which helps to remove biases in the anomalies originating from the natural drift of the reforecast towards a model’s own internal mean state (Becker et al. 2014). While removing the means of the most recent 120 days of anomaly data for a given field in step (ii) from G2010, a combination of GEFSv12 reanalysis and GEFSv12 reforecast data is used. For instance, if we seek to remove the mean of the most recent 120 days of U200 anomaly data from lead 15 of the January 2nd, 2019 reforecast, noting that lead 15 will correspond to January 17th, 2019, the 120 days of anomalies will be made up of 105 days of GEFSv12 reanalysis data derived from the period preceding January 2nd and 15 days of reforecast data between January 2nd 2019 and January 17th 2019. Following this step, all steps are the same as in the reanalysis: latitudinal averaging in the tropics, dividing the fields by their normalization factors, projecting onto the GEFSv12 reanalysis EOFs, and standardizing. Reforecast individual member RMMs: Rather than averaging together the 11 ensemble members for a given reforecast initialization as in the ensemble mean calculations, each member can be preserved and from it, the corresponding lead time dependent climatology for the relevant field is removed. This allows us to calculate anomalies for each member of each reforecast and subsequently calculate RMMs exactly the same way as described in the preceding section. Additional notes: The reforecasts are 36 leads long in which "lead 0" coincides with the day of initialization and "lead 1" coincides with the day following the initialization date. So for the first reforecast available, January 5th, 2000, this day will coincide with "lead 0" and January 6th will coincide with "lead 1." Both the reforecast and reanalysis are derived from 6-hourly data, which where averaged together to produce daily means. Relevant publications: Guan, H., Zhu, Y., Sinsky, E., Fu, B., Li, W., Zhou, X., ... & Kumar, A. (2022). GEFSv12 reforecast dataset for supporting subseasonal and hydrometeorological applications. Monthly Weather Review, 150(3), 647-665.////Hamill, T. M., Whitaker, J. S., Shlyaeva, A., Bates, G., Fredrick, S., Pegion, P., ... & Woollen, J. (2022). The Reanalysis for the Global Ensemble Forecast System, Version 12. Monthly Weather Review, 150(1), 59-79. |
| Date received: | 20220818 |
| Start date: | 20000101 |
| End date: | 20200331 |
| Seanames: | |
| West boundary: | -180 |
| East boundary: | 180 |
| North boundary: | 15 |
| South boundary: | -15 |
| Observation types: | model output |
| Instrument types: | |
| Datatypes: | |
| Submitter: | |
| Submitting institution: | |
| Collecting institutions: | |
| Contributing projects: | |
| Platforms: | |
| Number of observations: | |
| Supplementary information: | Submission Package ID: 358KB4 |
| Availability date: | |
| Metadata version: | 1 |
| Keydate: | 2022-09-13 11:30:38+00 |
| Editdate: | 2022-09-13 11:31:28+00 |