Japanese Whaling Ship Data (CDMP/MIT digitization) (deck 761; 1946-84) (by ESRL)

1. Introduction

Mierzejewska et al. (1997) described sea surface temperatures (SST) and other data elements observed (once daily at "local noon") from Japanese whaling ships (JWS) traveling to the Southern Ocean region during 1946-84. Up to four ships would leave Japan in October-November at the start of the whaling season, and return in March-April, crossing the warm pool region (10°N-10°S) each way. The article estimated that the JWS dataset could "contribute an additional 5% to 45% of SST observations to COADS in southern high-latitude oceans."

A digital dataset of SST, air temperature, wind (Beaufort force and direction), plus the ship position and date, was prepared at M.I.T (totaling approximately 20K daily ship reports). Unfortunately, sea level pressure (SLP) data from the original Japanese logsheets were not digitized. Observations of present weather (apparently in a Beaufort scale), sea-state, and ice-state also were not digitized, but these appear to be more difficult to use because of non-standard reporting practices and inadequate documentation.

The available M.I.T. digital data (local noon observations of SST, air temperature, wind strength and direction, plus the date and the ship position and name) were provided in 1997 to NOAA/CDC. This directory contains the documentation and data (zip file) that were received from M.I.T. at that time. Photocopies of the original JWS logsheets also were provided in 1997, for digitization of the missing meteorological data elements.

Digitization of the logsheets was completed by NOAA's Climate Database Modernization Program (CDMP) in 2002. Date, position, and SLP plus the other meteorological elements were keyed. This directory contains the keyed data (zip file), as well as documentation describing how the data were keyed. The logsheets also contained some proprietary whaling (fisheries) information, which, we were informed by M.I.T., should not be keyed or publicly released, and thus, while the logsheets were later imaged for preservation by CDMP (around 2008), they are not publicly available.

2. Translation overview

The M.I.T. and CDMP versions of the data were each translated into the LMR format. Having the two versions of the data was advantageous in allowing cross-checks between them (Figure 1), and the logsheets were also used to help resolve some additional problems. Because of the difficulty and uncertainty of translating present weather, sea-state, and ice-state into modern units, those elements were not translated and are available only as supplemental data (as keyed).

Reports from the two versions of the data in LMR format were then selectively merged into the output data (deck 761; 1946-84; 21K reports), which represent a mixture of reports from the two versions, identified by source ID (SID). Generally, the more complete CDMP data (SID=115; 20K reports input; 20K reports output) were used. However, in the event of problems or missing CDMP data, the M.I.T. version of the data (SID=116; 21K reports input; 425 output) was used instead. In conjunction with the merger, some data corrections and substitutions of dates and position between the CDMP and M.I.T. versions were performed. For example, numerous corrections were made to the M.I.T. dates, based on the CDMP data. Conversely, numerous corrections were made to the CDMP positions, based on the M.I.T. data.

Duplicates were removed from the data. This included the removal of a few additional reports introduced from observations taken at midnight, in additional to local noon, according to notes in English on the logsheets. Generally, however, CDMP did not key reports legitimately containing the same date as a previous report, from eastward-bound ships crossing the dateline. These missing reports were taken from the M.I.T. version, and together with a few other reports that ended up missing for various reasons, were inserted at the end of the output file. Otherwise the output file is sorted by ID (i.e., ship name), and within which by time.

The M.I.T. wind directions were supplied as degrees, already converted from other original units (if applicable). Thus we were unable to assign a direction indicator (DI; see Figure 1). In the CDMP data, the logsheets generally contained wind direction symbols such as "SE/S," which we interpreted as "SE by S," and converted to degrees accordingly (sometimes the logsheets instead contained directions in degrees, which were not altered). In contrast, M.I.T. appears to have misinterpreted the "/" in translating the wind data to degrees. However, since CDMP data were generally used this does not appear to be a large problem.

Finally, the LMR data were translated into the LMRF and IMMA formats, for users (the IMMA format includes the supplemental data). Figures 2 and 3 illustrate the temporal and spatial coverage, respectively, of the data output for deck 761.

3. Comparisons with deck 187

A smaller set of "Japanese Whaling Fleet" data (deck 187; 1946-56; 10K reports) was part of the original data mixture included in COADS Release 1 (Air Weather Service and US Weather Bureau, 1961a). Deck 187 appears to be at least partially related to deck 761 in terms of ship names (Tables 1 and 2). However, in contrast to deck 761, deck 187 contains 3- or 6-hourly synoptic observations (00, 03, ..., 21 UTC) "converted to the 1955 WMO Code FM11A before punching in Japan." These were more complete meteorological reports containing additional elements such as dew point temperature, visibility, and cloudiness. Furthermore, deck 187 was limited to observations south of 20°S (Figure 4).

The relationship between decks 761 and 187 has been discussed with the Japan Meteorological Agency (JMA) (Takashi Yoshida, personal communication, 2004). Details about how both these sets of data were observed and handled may have been lost (possibly together with the original forms), but deck 187 appears to have been keyed by JMA's Office of Antarctic Observation. Since 1946, JMA sent meteorological officers on the Antarctic whaling ships to provide meteorological and navigation support.

We speculate that the regular synoptic observations (deck 187) were made by the meteorological officers, whereas noon observations (deck 761) were made on the fishing forms by the ships' officers. Thus the data from the two decks may be independent observations taken from the same ships (and observed at different times, depending on the relationship between the UTC synoptic hours and local noon). Generally, the data appear to overlap well spatially (Figure 5) (except during 1950-51 deck 187 contains an obvious ship track, which is not in deck 761). However, we have not made comparisons of meteorological elements.

Further investigation would be required to determine if any of the synoptic observations, particularly outside the limited time period of deck 187, also appear on the Kobe Collection microfilm (documented in US Weather Bureau and JMA, 1960), or in the decks keyed from that Collection in 1961 (118 covering 1930-53 and 119 covering 1951-61) or more recently (762; pre-1930 data). For example, a ship named "Tonan Maru" is listed in US Weather Bureau and JMA (1960), which might be the same as that listed in Table 2. On the other hand, the original reference manual for deck 118 (Air Weather Service and US Weather Bureau, 1961b) states that observations "from Japanese Antarctic Whaling Ships [among some other selected sources]...were not punched in this deck."

Table 1. Inventory of deck 761 (output) by ID (ship name). Ship name appears in the ID field in ICOADS.

Table 2. Inventory of deck 187 by ship number. Ship number and name, vessel type, and tonnage are from Air Weather Service and US Weather Bureau (1961a). The number of reports and period were tabulated for each ship number from ICOADS Release 2.1. For some ships, no data were output in Release 2.1 (blank entries). Some data were deleted during duplicate elimination processing (11,482 reports were input to dupelim), but the impact of that processing on the ship makeup is not known. Reports may have also been lost in processing predating ICOADS.

Figures

Figure 1. Total number of Japanese Whaling observations, shown separately for each IMMA field that had any data. Blue (red) bars indicate that the translation of the CDMP data agrees (disagrees) with the translation of the M.I.T. data, into IMMA format. Disagreement includes the case of missing versus extant data. The amount of extant data resulting from the CDMP (M.I.T.) translation is indicated by an orange "X" ("+"); if a field is completely missing, the symbol is not shown (i.e., DI and SLP were completely missing in M.I.T.). IMMA documentation explaining the field abbreviations is available in PDF format in this directory.

Figure 2. Output reports per year for deck 761.

Figure 3. Spatial coverage of deck 761.

Figure 4. Spatial coverage of deck 187.

Figures 5. Ship track comparisons for deck 761 (red "X") versus deck 187 (blue "+"), with separate panels for each whaling season 1946-47, 1947-48, 1948-49, 1949-50, 1950-51, 1951-52, 1952-53, 1953-54, 1954-55, and 1955-56.

References

Air Weather Service and US Weather Bureau, 1961a: Reference Manual 187: Japanese Whaling Ship Sfc. Obs. Climatic Center, USAF, Air Weather Service (MATS) and NWRC, Office of Climatology, US Weather Bureau (9 January 1961), 5 pp. [Available from NCDC.]

Air Weather Service and US Weather Bureau, 1961b: Reference Manual 118: Japanese Surface Marine Obs. Climatic Center, USAF, Air Weather Service (MATS) and NWRC, Office of Climatology, US Weather Bureau (18 April 1961), 5 pp. [Available from NCDC.]

Mierzejewska, A.W., Z. Wu, R.E. Newell, and T. Miyashita, 1997: Japanese Whaling Ships' Sea Surface Temperatures 1946-84. Bull. Amer. Meteor. Soc., 78, 443-447.

US Weather Bureau and JMA, 1960: Guide Book of the Japanese Marine Surface Data. US Weather Bureau and Japan Meteorological Agency, 527 pp.

[Documentation and Software][Translation information]