Stability of Water Temperature in the conductivity and temperature calibration system and Result of calibration experiments

Asako Inoue*1, Motoki Miyazaki*2, Kenji Izawa*3, Kentaro Ando*3, Yasushi Takatsuki*3, Keisuke Mizuno*3

Abstract

The purpose of this study is to investigate the temperature distribution in the bath of the conductivity and temperature calibration system (CT bath system) equipped for Japanese-Argo program in 2000, and to show the results of calibration of sensors which were actually launched in 2001 in the North Pacific as the Japanese-Argo floats. On the temperature and conductivity sensor calibration, controlling temperature of water in the bath during calibration is the most important issue, so the temperature distribution during calibration was examined. As the results of this experiments, water temperature was well controlled within 1.7mK.
The calibrations before the deployment of floats were also performed for eight sensors. The results of JAMSTEC calibration showed that the difference of temperature calibration from the manufacturer's calibration was within 1mK with the standard deviation of less than 0.5mK, and that of conductivity calibration from the manufacturer's was 0.5mS/m (corresponding to 0.0042psu at 24 degree-C) with the standard deviation of 0.2mS/m (corresponding to 0.0016psu at 24 degree-C).

Keywords : Argo project,Argo profiling float,Calibration bath system,temperature distribution,error range

*1 Department of Marine Science, Marine Works Japan LTD
*2 Frontier Observation System for Global Change
*3 Ocean Observation and Research Department


1. Introduction

2. Calibration bath system for profiling float sensor.


Photo 1 Argo float calibration bath (left) and controller(right).


Photo 2 Argo CT (Conductivity and Temperature) sensors(SBE41) attached to end-cap for connecting to signal cable. Before calibration, sensors have to be removed from Argo float.


Figure 1 Configuration of Argo calibration bath.

3. Validation of calibration bath system.

3.1. Correction of error among temperature sensors used to measure temperature distribution of the bath.


Figure 2 Time change of temperature measured by four SBE3 sensors during one-point calibration near room temperature.
The water temperature in the bath gradually increased towards room temperature,
however differences among each sensor were almost unchanged during calibration.

S/N
SeaBird社での校正日
差の平均値 (器差) [mK]
差の標準偏差 [mK]
ex08
07-Jul-99
-0.9
0.086
ex06
30-Jul-99
-1.8
0.086
2609
15-Sep-00
-1.0
0.087
2522
22-Sep-00
1207
18-Nov-00
Table 1 Serial Numbers of all sensors used for measurement of temperature distribution, and the average and standard deviation of each sensor referred from S/N2522. The date of calibration by manufacture is also shown.

3.2. Method of experiment and location of temperature sensors.


Figure 3 Locations of temperature sensors (SBE3) during the measurement experiment of temperature distribution in radius direction.


Figure 4 Locations of temperature sensors (SBE3) during the experiment of temperature distribution measurement in vertical direction.

3.3. 温度分布測定結果

Figure 5 Temperature distribution relative to reference points. Upper panel shows temperature distribution in radius direction, and lower panel does the same except for in vertical direction.

設定温度
半径方向測定
鉛直方向測定
Ref T[℃]
T Standard Dev [mK]
T Standard Dev <[mK]
30
0.66
0.67
24
0.46
0.62
20
0.73
0.43
15
0.20
0.66
10
0.35
0.46
5
0.53
0.44
1
0.25
0.57
Table 2 The standard deviation of reference sensor at each calibration point during the experiment of temperature distribution measurement.

3.4. Results

4. Calibration results of float sensors by calibration bath system.

4.1.  Calibration results of SBE41 sensor.


Figure 6 Average(◆)and standard deviation (error bars) of temperature differences (residuals) from reference sensor at 7 calibration.


Figure 7 Conductivity ratio (correction factor to reference conductivity sensor) is shown for each calibration date, calculated as the ratio of measured conductivity to true conductivity calculated from sampled salinity measured with Autosal®. The ratio sometimes shows large difference as indicated with a large circle, in most case, such values are caused by a contamination during water sampling.

4.2. calibration results of FSI sensor


Figure 8 Location of FSI sensor for the wall effect experiment of conductivity sensor of "Inductive cell". The wall effects were measured at 9 points horizontally (A1 to C3 in the upper panel), and at 3 levels (middle to bottom in the lower panel).


Figure 9 Results of wall effect experiment for middle layer and bottom layer. The horizontal axis shows the distance from sidewall, and the vertical axis shows the ratio of conductivity reduction.