University of Washington
Japan/East Sea Profiling Drifters

---

Table of Contents

1. Introduction
2. Automatic data processing & web-page generation
3. Reliability statistics
4. The Sea-Bird CTD module
5. Deployment locations for the Japan/East Sea profiling drifters
6. Argos Data Telemetry
7. GTS publication of float profiles
8. Contact Information

Introduction

This project is designed to examine the formation and circulation of water masses in the Japan/East Sea, using profiling drifters. Components for these instruments are purchased from Webb Research, Inc., of Falmouth, MA, USA, and final assembly, calibration, and ballasting operations are carried out at the University of Washington in Seattle. In all, 36 drifters have been deployed as part of this project. Deployment of 32 floats were carried out during a summer, 1999, expedition of the research vessel "Professor Khromov", based at the Far East Regional Hydrometeorological Research Institute in Vladivostok. These deployments were conducted exclusively by Russian scientists during this expedition.

Our goal here is to provide nearly real-time information on profiling drifters that will be launched in the region. Trajectories and profiles appear here as soon the data are received from the ARGOS system, normally within a day of data collection. Using this web page it is possible to view individual trajectories as well as temperature and salinity profiles along those trajectories, and to see a spaghetti plot of all trajectories. In addition, by clicking a mouse in any figure, it is possible to download a postscript file of the figure. As time goes on, new features will be added.

A similar project, with about 70 PALACE floats, has been underway in the N. Atlantic Ocean since the summer of 1997. The techniques, scientific results, and general information for the Atlantic project are analogous to the results that are expected from the Japan/East Sea program.

Automatic data processing, data distribution, & web-pages generation

Reliability Statistics

The APEX Float: The APEX is a newer model of profiling float that were first subjected to trials in the subtropical North Atlantic. Seven Atlantic VOS deployments of APEX instruments equipped with Sea-Bird CTD modules were conducted between September, 1998 and February, 1999. These trials were designed to test the APEX in preparation for this experiment in the Japan/East Sea. The APEX is also manufactured by WRC.

The Sea-Bird CTD module

Most of the US floats in the JES were deployed during July and August of 1999. The basic APEX engine appears to be performing well. However, there have been problems with the CTD measurements collected by the SeaBird CTD sensor package on some of the instruments. Approximately one-half of the the floats deployed showed an initial salinity anomaly in the range of 0.02 to 0.47 practical salinity units (PSU). This determination was made by comparing the salinity values measured at the bottom of each profile (a depth of about 800 m) to the mean salinity at the same depth in the JES. The mean salinity profile was determined by taking the average of 21 CTD stations collected by the PI on this project in the JES in 1995 using a Neil Brown Mark 3 System CTD package. These CTD data have been reported in a published paper (Riser et al., Journal of Oceanography, V55, pp. 133-156, 1999). Based on these data, the standard deviation of the salinity below 500 m is approximately 0.002 PSU and so the salinity at a depth of 800 m is constant to within 0.004 PSU or so. Thus, the "correct" value of salinity at a depth of 800 m in the JES is known quite accurately. Sea-Bird's SBE-41 specifications indicate a stated accuracy of 0.005 PSU. By considering both natural variability and instrument uncertainty then it is possible to say that floats with a salinity anomaly greater than approximately 0.01 PSU are likely operating outside the manufacturer's specifications. Conversely, floats that have a salinity anomaly less than approximately 0.006 PSU are likely operating within the manufacturer's stated specifications. The band from 0.006 PSU to 0.01 PSU is a statistical grey area in terms of judging the correctness of the salinity measurements.

According to a detailed explanation by SeaBird, the likely cause of the salinity errors is contamination of the conductivity cell by a foreign substance, at some time after calibration at the factory and before executing the first profile. The most likely candidate is contamination by TBTO, the substance used as the biocide to eliminate biological fouling inside the conductivity cell. Since this biocide is slightly water soluble, the contamination is slowly washing out the conductivity cell. According to SeaBird, contamination effects the conductivity cell geometry such that a simple offset is introduced into the whole measured salinity profile. For most affected floats, the size of the salinity offset can be seen to steadily decrease (with each profile) in a quasi-asymptotic fashion. The number of floats with a salinity offset greater than 0.010 PSU has decreased from approximately 24 to 14 after the first 8 profiles, or a time period of 62 days. Many of the remaining outliers continue to show improvement and we anticipate that the number of floats showing significant salinity offsets after 20 profiles will be perhaps a half dozen or so. The plot below shows the most recent (updated daily) salinity offsets for floats affected by TBTO contamination.

To the extent that the salinity error is a simple offset of the whole salinity profile then we have exploited the very stable and uniform hydrography of the deep water (ie., near 800 m) to correct the offset. We estimate that the corrected salinities reported here have an uncertainty of approximately 0.006 PSU. Some uncertainty is due to finite instrument precision and flushing out some of the TBTO contamination during the profile but most of the uncertainty stems from variations in the deep-water hydrography.

In the Profile Summary for each float, the salinity offset "delta S" (ie., the measured value of S at the bottom depth minus the mean value) is given. In the Float Profiles section, the corrected salinity profiles are plotted in blue while the measured salinity profiles are shown in grey. In many profiles the offset is so small that the grey curve is plotted directly beneath the blue curve and cannot be seen.

Argos Data Telemetry

• Float Id: A unique 3 digit identifier assigned to each float.

• Number of Profiles: The number of profiles the float has executed so far.

• Incomplete Profiles: The number of profiles that have an incomplete Argos message set. This condition results in a missing segment of data in the profile plots. The accompanying percentage gives the ratio of the Incomplete Profiles to the Number of Profiles.

• Satellite Passes: The average number of times that a satellite passed overhead while the float was on the surface.

• Messages Expected: This is the average number of messages that should have been received when satellites were overhead. If two or more satellites are overhead simultaneously then each transmission was multiply-counted. The number of messages expected for a given satellite pass is determined from the Argos data stream by counting the number of messages received and adding the number of messages missing. Due to averaging and rounding effects, these entries in the table below may not add up exactly.

• Messages Received: This is the average number of messages that were actually received by satellites (without regard to the CRC check). The accompanying percentage gives the average ratio of the messages received to the messages expected.

• Messages Missing: This is the average number of messages that were transmitted while a given satellite was overhead but that were not received by the satellite. These are detected in the Argos data stream by gaps in time and message number. The accompanying percentage gives the average ratio of the messages missing to the messages expected.

• Good Messages: This is the average number of messages received that passed the CRC check. The accompanying percentage gives the average ratio of the good messages to the messages received.

• Bad Messages: This is the average number of messages received that failed the CRC check. The accompanying percentage gives the average ratio of the bad messages to the messages received.

• Telemetric Efficiency: This measures the average rate of successful transmission when satellites are overhead. It is the average ratio of the number of good messages to the number of messages expected.

Publication of Profiles to the Global Telecommunications System (GTS)

Quality control of profile data is managed automatically by a QC module that is integrated into the GTS publication software. Each profile must pass a number of quality control criteria in order to be published via GTS. If even a single observation in the profile violates any of the criteria then the profile is not published. For the Japan/East Sea, the following criteria are used to assess the quality of profiles.

1. A profile that is older than 30 days will be rejected as required for GTS publication.

2. A profile that does not have a valid location fix will be rejected as required for GTS publication. This might potentially happen if ARGOS reception is poor.

3. A profile with a gap between adjacent observations that exceeds 300 decibars will be rejected. A profile with the shallowest observation deeper than 300 decibars will also be rejected. This is to maintain reasonable accuracy in the pressure to depth conversion calculation as required for GTS publication. This might happen if ARGOS reception is poor.

4. A profile with an observation where the pressure is reported to be less than zero or greater than 3000 decibars will be rejected as indicative of a pressure sensor problem or malfunction.

5. A profile with deepest observation less than 50 decibars is rejected. For example, this might happen if the float drifts into shallow water or becomes entangled in mats of surface-drifting seaweed.

6. A profile with pressures that are not monotonic will be rejected. This is intended as a general criterion that a properly collected profile should satisfy.

7. A profile with fewer than 5 observations will be rejected. This is intended as a general criterion that a properly collected profile should satisfy.

8. A profile containing an apparent static instability will be rejected. Specifically, if the instability between two observations exceeds 0.004 Kg m^-3 then the profile will be rejected. This test is applied only to observations deeper than 500 decibars. Shallower observations are not tested for static instability.

9. A profile that contains a salinity observation that deviates from the mean by more than 3 standard deviations plus the instrument error will be rejected (see the blue envelope in the plot below). The mean and standard deviation are computed as described in the section describing the Sea-Bird CTD module. The plot below shows the mean salinity profile (bold blue). This test is applied only to observations deeper than 500 decibars. The salinities posted to GTS are offset to match the mean profile at the deepest observation.

10. A profile that contains a temperature observation that deviates from the mean by more than 3 standard deviations plus the instrument error will be rejected (see the red envelope in the plot below). The mean and standard deviation are computed from the ensemble of Japan/East Sea profiles. The plot below shows the mean temperature profile (bold red). This test is applied only to observations deeper than 500 decibars.
    

Visitor Number: