This directory contains updated estimates of drogue off dates for drifters in the time period October 1992-November 2010, sorted by severity (amount of drogue-off time misdiagnosed as drogue-on prior to this reevaluation). For more information, see: Lumpkin, R., S. Grodsky, M.-H. Rio, L. Centurioni, J. Carton and D. Lee, 2013: Removing spurious low-frequency variability in drifter velocities. J. Atmos. Oceanic Technol., 30 (2), 353-360, doi:10.1175/JTECH-D-12-00139.1. Drifters more recent than the ones in this study are being evaluated and included in standard Global Drifter Program updates. The purpose of these files is to offer automatic and manual reevaluation of the historical data, i.e., only the drifters that were present in the study period of Lumpkin et al. (2013). Note that Lumpkin et al. (2013) used CCMP winds; for more recent data, we have derived a wind model (following that study) using 6h NCEP/NCAR reanalysis winds. This has allowed us to evaluate the historical drifters for drogue loss after November 2010 (results included in these files) and drogue loss of more recent drifters (results not included here, but reflected in our standard metadata files). ASCII FILE: Each file is named YYYY_MM_DD_drogueoff.dat according to its creation date. Older versions are kept in the folder "archive". The first column gives the ID of the drifter. The second column is the original drogue off date in the drifter Data Assembly Center metadata file (the date before Lumpkin et al, 2012). "9990-00-00" indicates that the drogue was still attached as of the end of the study. Subsequent updates of the DAC metadata may have included a drogue off date. "1979-1-1" indicates drogue status uncertain from beginning of record (no drifters were deployed before Feb. 1979 in the GDP data). The third column is the automatically detected drogue off date generated by the methodology described in Lumpkin et al. (2012). If this date is earlier that the drogue off date in the second column, we recommend using this date until the data have been examined manually (fourth column). The fourth column is the result of a manual reevaluation of the data, concurrently using submergence/tether strain, anomalous downwind motion, and transmission frequency variation. This reevaluation began after the first version of the update (using the automatic procedure) was created, and was completed on 14 February 2013. Special dates are used in the following cases: "9999-00-00": not yet reassessed manually. "9998-00-00": reassessed but drogue loss unclear. "9990-00-00": drogue still on as of 19 Nov 2010, the most recent CCMP wind data at the time of this study, or still on at death (whichever is first). Any other value should be considered the best possible estimate of drogue-off date and will be included in a future update of the DAC metadata. In all cases, a drogue off date of "1979-01-01" indicates that the drogue status was uncertain from the beginning. MATLAB FILE: The data in the ASCII file (see above) are also included in a Matlab file. In this file, dates are in Julian drifter days since 31 December 1978. For example, Doff_orig(1)=8268= 20 August 2001. 99999 = not yet reassessed manually, 99998 = reassessed but drogue loss unclear, 99990 = drogue still on at end of study, or still on as of drifter death (whichever was first). 1 = drogue status uncertain. The Matlab file also includes the latitude and longitude of each drifter at the time of the automatic and manual drogue off determination, and a vector "Cat" which indicates the status of the manual reevaluation compared to the original value in the directory file metadata: Cat = 1 : Drogue loss confirmed at time given in directory file. Cat = 1.5 : directory file confirmed by manual reassessment (Cat.1 or 2) Cat = 2 : Drogue on until end in directory file, confirmed by manual reevaluation. Cat = 3: drogue loss misdiagnosed in directory file. Metadata will need to be changed. Cat = 7 : drogue loss unclear from manual reevaluation. Will need more consideration.