How to Translate Easily | How to Read Shipboard Thermosalinograph Salinity, Intake Temperature, Flow and Bottle Calibration

When a ship measures seawater through a pipe, interpretation begins upstream

A shipboard thermosalinograph reads conductivity and temperature in a seawater flow-through system, then derives practical salinity. Translating the resulting data correctly requires attention to intake depth, timing, pump flow, laboratory warming, sensor drift and discrete bottle calibration. A plotted salinity point represents a chain of measurements and processing, not simply ‘the sea at the ship’s GPS coordinate’.

This guide turns TSG information into a set of reproducible decisions. It explains exactly what was sensed, why a time lag matters, how a pump or bubble can imitate a plume, how to compute bottle residuals, and how salinity language changes under different conventions. All numerical exercises are hypothetical worked examples unless specifically marked as quoted methodology.

NOAA AOML’s original system descriptions and the University of Hawai‘i cruise reports form the specialist method sources for the explanation. A reader should use the configuration of their own vessel rather than assume that intake depth, sampling frequency or pumping rate remains the same across ships.

The route from seawater to a number on a chart

How to translate thermosalinograph data easily begins at a ship’s hull intake, not at the salinity graph. Water is pumped through pipes to conductivity and temperature sensors, and the resulting readings are combined with navigation and quality control. Practical salinity is derived from conductivity and temperature; intake water temperature may require a different sensor from the internal cell thermistor. Water transport produces time lag, flow problems can create false peaks, and bottle samples provide an independent calibration check. Treat this as a measured system with physical plumbing, not a single number automatically attached to the ship’s current GPS position.

First-principles measurement and timing

Suppose a pipe system contains 10 litres of water between intake and sensor and pumps 2 litres per minute at steady state. The nominal transport delay is 10/2 = 5 minutes. If a ship travels 4 metres per second, it moves 4 × 300 = 1200 metres during that time. An uncorrected TSG salinity observation might therefore be plotted over one kilometre from the original intake location. Real systems require verified flow and effective volume rather than these hypothetical values. The same reasoning explains why internal temperature may differ from ambient water temperature and why abrupt salinity fronts can appear displaced or smoothed.

A complete calibration example

Suppose three properly aligned discrete bottle salinity measurements give 34.96, 35.01 and 35.04, while the corresponding TSG readings are 34.93, 34.98 and 35.01. Under the residual convention TSG minus bottle, differences are −0.03, −0.03 and −0.03 practical-salinity units, giving mean bias −0.03. An additive +0.03 correction could remove this constant illustrative offset. But only if the sampling point, response time, cell temperature, bottle analysis and stability were independently checked. Three pairs cannot establish that the offset remained constant through an entire long voyage, especially across maintenance and flow interruptions.

What the original source documents establish

NOAA AOML describes the TSG as a flow-through system measuring conductivity and temperature and deriving salinity. NOAA’s quality-control descriptions include time, geographic position, plausible ranges, spikes, gradients, flat-line checks and climatology; delayed procedures compare against additional ocean observations and samples. The University of Hawai‘i HOT and WHOTS cruise manuals document actual shipboard configurations, bottle samples and processing. These are examples of methods, not claims that every ship uses identical plumbing or reference sampling intervals. Always consult the vessel-specific metadata before quoting intake depth, transit time or manufacturer accuracy.

A reader-friendly reporting standard

A well-translated TSG result says the variable, the depth of intake, the measurement and sampling interval, the calibration status, the number and distribution of bottle comparisons, and whether a value is provisional or delayed-mode checked. It separates a true salinity gradient from plumbing or sensor artifacts and never erases unusual measurements merely because they differ from climatology. In a cross-language report preserve Practical Salinity terminology and units, distinguish it from mass-based Absolute Salinity, and retain corrected positions, uncertainty and provenance.

Thirty complete diagnosis and interpretation lessons

1. A thermosalinograph observes water brought into a ship

A thermosalinograph, or TSG, is commonly a flow-through conductivity and temperature measurement system. Seawater enters a hull intake, travels through plumbing to the sensors, and is logged with ship navigation. The measurement corresponds to sampled intake water, not a magical direct observation of the whole sea surface. Separate the ocean property being investigated from the behaviour of the measurement apparatus. This distinction matters most when a result is surprising, because plausible environmental values can also be produced by instrument faults.

A vessel sailing through a narrow front may record a salinity change after the ship has already moved beyond the point where that water entered the intake. The delay follows the water pathway and pump flow, not the GPS timestamp alone. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Translate the device as a sample-flow system. A report that treats the sensor laboratory position as the original sampling position can place the anomaly on the wrong part of the ship track. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

2. Salinity is derived from conductivity and temperature

Most routine ship TSGs measure electrical conductivity and temperature; practical salinity is calculated from those quantities using the applicable calibration and salinity algorithms. Salinity is not usually a direct reading from a separate magic salt sensor. To check the calculation, express time in seconds or minutes and distance in consistent SI units. State the water pathway or residual convention before interpreting a number, because reversing the order can reverse the apparent conclusion.

A conductivity change could reflect a true salt change, a temperature effect, bubbles, contamination, or sensor drift. Only after the instrument variables and quality checks are consistent should the resulting salinity be interpreted as a seawater property. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Preserve the measurement chain: conductivity plus sensor temperature and appropriate pressure assumptions lead to computed salinity. Do not reduce the chain to ‘the ship weighed its salt’. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

3. Intake temperature and internal temperature have different jobs

A temperature sensor near the intake can better represent ambient water temperature than a thermistor in a shipboard laboratory after water has passed along warm piping. The conductivity cell still needs a matched water temperature for salinity calculation. A reliable chart needs a reversible explanation: another reader should be able to locate the raw sensor stream, reproduce the correction and see why any interval was flagged. The line’s neat appearance does not establish its quality.

If intake temperature is 24.0 °C and the laboratory cell reads 24.7 °C, the shipboard warming does not mean the ocean warmed by 0.7 °C between those positions. The two sensors observe the same inflowing water at different points. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Use intake temperature for the intended near-surface seawater-temperature product if the processing convention prescribes it, and use the cell-associated temperature correctly for conductivity-to-salinity conversion. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

4. A sensor is not the intake depth

The hull intake may sit several metres below the surface depending on vessel design, draft and operating state. TSG seawater-temperature and salinity data represent that intake layer rather than a satellite’s skin temperature or every depth. Before reporting a freshwater or warming mechanism, test the competing explanations of plumbing, calibration, location alignment and sampling depth. Unsupported attribution is harder to repair than a candidly qualified measurement.

A ship drawing seawater at roughly 3 m can differ from a satellite estimate of the very near surface during rain or strong daytime heating. The readings may both be valid even when their values differ. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Write the intake depth or documented depth range. Do not call a three-metre salinity record a direct top-centimetre measurement unless the application explicitly justifies the approximation. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

5. Ship motion and sampling time need reconciliation

Navigation systems often log GPS location independently of the conductivity, intake and internal-temperature sensors. A time stamp can represent acquisition by a sensor after water travelled through pipes, while the navigation point reflects the ship’s current position. Review whether the example uses a directly observed quantity, a conductivity-based calculation, a statistical quality flag or an independent bottle reference. These categories should remain distinct when the result is rewritten for a general audience.

If transit from hull to sensor takes 45 seconds and the ship travels at 5 m/s, the sampling location could be displaced by about 225 m along track if no delay correction is applied. The numbers are an illustrative multiplication, not a ship’s reported specifications. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Measure or estimate the lag from plumbing volume and flow, then align to intake time. Record the correction rather than quietly shifting positions on a map. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

6. Pipe volume and pump rate determine transit delay

For a steady incompressible flow, transit time is approximately water volume between intake and sensor divided by volumetric flow. Both quantities matter; pipe length alone cannot establish the delay without diameter and actual flow. A careful translation retains units, time frame, intake configuration and uncertainty. It is better to state that a result is unresolved than to remove a critical qualifier merely to make the prose sound smooth.

Suppose the effective plumbing volume is 12 L and flow is 2 L/min. Transit delay is about six minutes. At vessel speed 4 m/s, that represents about 1.44 km of travel, enough to smear narrow coastal features. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Document actual volume and flow where available. A nominal pump specification is not proof of steady flow through every branch, especially during valves changes or filter maintenance. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

7. Water temperature can change in plumbing

Shipboard plumbing may exchange heat with machinery spaces, sunlight or air-conditioned compartments. A thermosalinograph’s internal thermistor reads the water at the instrument, not necessarily its ambient condition at intake. Separate the ocean property being investigated from the behaviour of the measurement apparatus. This distinction matters most when a result is surprising, because plausible environmental values can also be produced by instrument faults.

If an intake thermistor shows 22.5 °C and an internal cell reports 23.0 °C, thermal exchange along the line is plausible. Selecting one number without naming the sensor location changes the target variable. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Compare paired sensors, time lag and stable sections. Correct temperature products using a documented sensor hierarchy instead of assuming all ship temperatures are interchangeable. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

8. A stopped pump invalidates the water-source assumption

A TSG can continue logging while the pump flow slows, stops or changes source. The cell then contains water that may warm, cool, evaporate or be influenced by maintenance, no longer fresh seawater from the hull. To check the calculation, express time in seconds or minutes and distance in consistent SI units. State the water pathway or residual convention before interpreting a number, because reversing the order can reverse the apparent conclusion.

A sudden apparent temperature rise during a period of near-zero flow should prompt an operational check rather than a headline about an abrupt ocean heatwave. The ship may simply be sampling stagnant water. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Keep pump status and flow flags in the quality record. Discard or flag intervals when the intended intake path did not operate normally. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

9. Air bubbles can mimic low conductivity

Air occupies volume and changes electrical conduction through a cell. Entrained bubbles, pump cavitation or inadequate debubbling can produce suspicious conductivity and derived-salinity excursions. A reliable chart needs a reversible explanation: another reader should be able to locate the raw sensor stream, reproduce the correction and see why any interval was flagged. The line’s neat appearance does not establish its quality.

A sharp salinity dip that coincides with erratic conductivity while intake temperature and bottle samples remain stable is consistent with a flow-cell disturbance. It is not sufficient to label a freshwater plume. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Check bubble traps, flow, raw conductivity and any diagnostic flags. Only describe freshwater influence when independent evidence supports it. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

10. Biofouling can cause progressive drift

Material accumulating on a conductivity cell can alter its response over days or weeks. The resulting salinity bias may drift smoothly and look like a broad environmental trend on a long voyage. Before reporting a freshwater or warming mechanism, test the competing explanations of plumbing, calibration, location alignment and sampling depth. Unsupported attribution is harder to repair than a candidly qualified measurement.

A sensor slowly reading 0.03 salinity units lower than nearby bottle samples over several days suggests calibration or fouling concerns. That trend is not automatically a continuous decline in the ocean’s salinity. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Track residuals against discrete reference samples and maintenance logs. A change in the instrument is a competing explanation that must be tested before interpreting a geophysical pattern. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

11. Bottle samples provide an independent anchor

Discrete seawater samples collected from a closely matched sampling point can be analysed with a calibrated laboratory salinometer. They provide reference values against which continuous TSG estimates can be compared. Review whether the example uses a directly observed quantity, a conductivity-based calculation, a statistical quality flag or an independent bottle reference. These categories should remain distinct when the result is rewritten for a general audience.

Suppose the TSG reads 34.91 while a paired analysed sample gives 34.95. The TSG-minus-bottle residual is −0.04. Correcting that difference may be warranted if repeated pairs support a systematic calibration bias. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Verify time alignment and sampling point, and inspect bottle handling before calculating a correction. One discrepant bottle is not a trustworthy universal offset. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

12. Pairing times matters before calibration

The water captured in a bottle must correspond to the water passing through the thermosalinograph. Sampling-port geometry and transport delay can make a nominally simultaneous bottle and sensor timestamp refer to different seawater parcels. A careful translation retains units, time frame, intake configuration and uncertainty. It is better to state that a result is unresolved than to remove a critical qualifier merely to make the prose sound smooth.

A fast-moving ship can cross a narrow front in a minute. Comparing a bottle collected after the front with a TSG average before the front manufactures an apparent calibration bias that is really a timing mismatch. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Align the water flow path, log times and location. Near sharp fronts, use carefully defined averaging windows rather than convenient whole-hour values. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

13. Calculating mean salinity bias

For paired estimates, define residual as TSG minus bottle salinity, then compute its mean over accepted comparison pairs. Its sign tells you whether the continuous system tends high or low under that convention. Separate the ocean property being investigated from the behaviour of the measurement apparatus. This distinction matters most when a result is surprising, because plausible environmental values can also be produced by instrument faults.

For three residuals of −0.02, −0.03 and −0.01, their mean is −0.02 salinity units. An additive +0.02 adjustment would remove that mean offset in this toy example, but only if the instrument behaves stably. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Keep the sign convention printed beside any correction and avoid applying a cruise-wide offset when evidence indicates drift or multiple operating regimes. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

14. Scatter and mean bias tell different stories

The average residual can be near zero even when individual TSG–bottle differences are large. Root-mean-square error, or RMSE, describes typical squared deviation and should not be substituted for the signed mean. To check the calculation, express time in seconds or minutes and distance in consistent SI units. State the water pathway or residual convention before interpreting a number, because reversing the order can reverse the apparent conclusion.

For residuals +0.10 and −0.10, mean bias is zero but RMSE is 0.10. The instrument cannot be called perfectly matched simply because opposing errors cancel. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Report a bias and a scatter statistic with sample size and sampling distribution. Discuss whether short-lived fronts, lag or bottle precision contribute to scatter. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

15. A calibration line is not always constant

A TSG may need a time-dependent or conductivity-dependent calibration rather than a single offset, but complex models can overfit sparse bottle samples. Calibration should be physically motivated and checked independently. A reliable chart needs a reversible explanation: another reader should be able to locate the raw sensor stream, reproduce the correction and see why any interval was flagged. The line’s neat appearance does not establish its quality.

If early-cruise bottles suggest +0.01 correction and late-cruise bottles suggest +0.06, a constant +0.035 may leave systematic residuals at both ends. A justified drift model may fit better but needs sufficient comparisons. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Show the correction’s time dependence and acceptance tests. Never infer a sophisticated calibration curve solely to make every plotted point match. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

16. Outlier bottles require investigation, not automatic deletion

A discrete bottle can be contaminated, mislabeled, mistimed or compromised in laboratory analysis. A surprising residual is a reason to inspect chain of custody before deciding it belongs outside calibration. Before reporting a freshwater or warming mechanism, test the competing explanations of plumbing, calibration, location alignment and sampling depth. Unsupported attribution is harder to repair than a candidly qualified measurement.

If one bottle differs by 0.6 units while adjacent pairs differ by about 0.02, check sample handling and independent repeat analyses. The outlier may be faulty, but it could also represent a real sharp front. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Apply a documented rejection rule and retain excluded points in an audit. A correction designed to remove inconvenient readings can conceal real ocean variability. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

17. Practical salinity is not a mass percentage

Practical Salinity Scale values are derived from conductivity ratio and are conventionally treated as dimensionless. Casual labels such as PSU are widespread, but practical salinity is not simply grams of dissolved salts per kilogram. Review whether the example uses a directly observed quantity, a conductivity-based calculation, a statistical quality flag or an independent bottle reference. These categories should remain distinct when the result is rewritten for a general audience.

A reported practical salinity of 35 cannot be translated as 35 percent salt. The corresponding mass fraction is on the order of tens of grams per kilogram, not thirty-five grams per hundred grams. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

For rigorous reporting specify Practical Salinity and its convention, and distinguish it from Absolute Salinity in g/kg when TEOS-10 conversions are actually performed. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

18. Absolute salinity needs additional information

TEOS-10 Absolute Salinity is related to but not universally numerically identical to Practical Salinity. Conversions account for the reference composition of seawater and regional composition anomalies as appropriate. A careful translation retains units, time frame, intake configuration and uncertainty. It is better to state that a result is unresolved than to remove a critical qualifier merely to make the prose sound smooth.

A TSG practical-salinity value of 35 must not be relabeled 35 g/kg as though no conversion or convention existed. For many open-ocean applications numerical differences may be small but technically meaningful. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Identify whether the source reports SP, reference salinity or SA, and specify units accordingly. A translation should not silently replace a dimensionless conductivity-derived measure with mass per mass. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

19. Flow interruption can create spikes at restart

When a pump restarts, a TSG may flush stale water and dislodged bubbles before fresh seawater reaches the cells. The first new readings can be transient and do not necessarily reflect the ocean. Separate the ocean property being investigated from the behaviour of the measurement apparatus. This distinction matters most when a result is surprising, because plausible environmental values can also be produced by instrument faults.

Suppose conductivity rises in two seconds as flow returns, followed by a minute of unstable data. Mapping that point to the current ship position can create a false narrow high-salinity patch. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Use startup flags, flow-stability checks and a reasonable flushing interval. Document which readings were omitted and how flow restoration was confirmed. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

20. A ship can sample its own disturbed water

Intake location, ship speed, turning, thrusters, discharge outlets and bubble entrainment can affect the sampled water. A TSG does not operate independently of vessel plumbing and hydrodynamics. To check the calculation, express time in seconds or minutes and distance in consistent SI units. State the water pathway or residual convention before interpreting a number, because reversing the order can reverse the apparent conclusion.

An apparent low-salinity anomaly during a manoeuvre near a discharge outlet should trigger review of intake configuration, ship operations and surrounding water measurements. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Distinguish a water-mass feature from the sampling platform’s influence. Cross-check repeated transects rather than treating every short spike as environmental truth. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

21. Crossing a front creates a response-time problem

A TSG’s flow-through system and sensor response may smooth a rapid salinity step. The observed front width along track can be broader than the true water-mass transition. A reliable chart needs a reversible explanation: another reader should be able to locate the raw sensor stream, reproduce the correction and see why any interval was flagged. The line’s neat appearance does not establish its quality.

If water salinity changes over 200 m while the intake-plus-cell system responds over an additional 300 m of ship travel, the plotted edge can appear far wider. This is an illustrative conceptual distinction, not a measured transfer function. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Account for transit lag and response time before deriving frontal gradient from underway data. A smooth trace can reflect instrument physics as well as mixing. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

22. Location coordinates must match the corrected time

A ship’s GPS position usually updates frequently, while the TSG sample is linked to intake time through transport and electronics delays. A navigation correction alters the data’s geographic interpretation. Before reporting a freshwater or warming mechanism, test the competing explanations of plumbing, calibration, location alignment and sampling depth. Unsupported attribution is harder to repair than a candidly qualified measurement.

At 10 knots the ship travels about 5.14 m/s. A 120-second timing mismatch corresponds to roughly 617 m along the track, which is considerable for a narrow coastal feature. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Preserve the original timestamp and a documented adjusted sampling time. Do not change coordinates without showing the physical basis and direction of adjustment. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

23. Salinity and temperature need different uncertainty budgets

Intake temperature, internal cell temperature, conductivity calibration, flow and bottle reference precision contribute in different ways. It is unsafe to attach a single generic instrument accuracy number to every derived quantity. Review whether the example uses a directly observed quantity, a conductivity-based calculation, a statistical quality flag or an independent bottle reference. These categories should remain distinct when the result is rewritten for a general audience.

If temperature is stable but conductivity calibration is drifting, the salinity series may be biased while intake-temperature measurements remain valuable. Conversely, an intake thermometer fault need not invalidate an independently calibrated cell temperature. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Report variable-specific flags and uncertainties. A sensor package can be partially usable even if one data stream fails. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

24. Real-time QC and delayed-mode calibration answer different needs

NOAA TSG procedures describe real-time checks and additional delayed-time comparisons, including independent observations and bottle calibration. Real-time availability is not identical to final expert-reconciled quality. A careful translation retains units, time frame, intake configuration and uncertainty. It is better to state that a result is unresolved than to remove a critical qualifier merely to make the prose sound smooth.

A preliminary transect may show a freshwater dip that delayed processing later flags as a pump artifact. The earlier data were useful for rapid monitoring but should not be presented retrospectively as definitive. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Cite the data stream’s processing level and update status. A corrected later product does not mean the raw instrument observations were fabricated. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

25. Gross-range checks cannot detect every error

Testing whether values fall inside broadly plausible temperature and salinity ranges catches impossibilities and severe sensor failures. It does not identify every small bias, timing error or physically plausible but false spike. Separate the ocean property being investigated from the behaviour of the measurement apparatus. This distinction matters most when a result is surprising, because plausible environmental values can also be produced by instrument faults.

A TSG reporting 34.7 salinity when the true value is 34.9 may pass global plausibility checks yet still be scientifically inaccurate for calibration-grade applications. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Use regional tests, gradients, flat-line tests, bottle comparisons and sensor diagnostics in addition to simple upper and lower bounds. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

26. A constant-value series can be suspicious

A perfectly flat temperature or conductivity trace may indicate stalled flow, a frozen logger, an electronic fault or a genuinely stable water mass. The flat line is an anomaly to diagnose, not an automatic verdict. To check the calculation, express time in seconds or minutes and distance in consistent SI units. State the water pathway or residual convention before interpreting a number, because reversing the order can reverse the apparent conclusion.

If conductivity remains identical to many decimal places for an hour while the ship crosses very different waters, check raw counts and instrument status. But do not flag a stable ocean region solely because it looks boring. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Compare independent temperature, salinity, flow and navigation information. State the evidence for a flat-line flag in the final data description. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

27. Regional climatology is a diagnostic, not a substitute

Comparing underway observations with regional climate data can expose gross problems, but real freshwater plumes and rare conditions may legitimately depart from historical typical ranges. A reliable chart needs a reversible explanation: another reader should be able to locate the raw sensor stream, reproduce the correction and see why any interval was flagged. The line’s neat appearance does not establish its quality.

A strong river plume can yield salinity outside an offshore climatology without any sensor defect. Automatically replacing it with the climatological mean would erase potentially important science. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Use climatology to trigger a check rather than overwrite observations. Preserve verified extremes with their uncertainty and local context. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

28. Satellite salinity and ship salinity can observe different depths

L-band satellite sea-surface salinity senses a very shallow electromagnetic sampling layer, whereas a ship intake may be metres below. Rain-induced stratification can make simultaneous values differ. Before reporting a freshwater or warming mechanism, test the competing explanations of plumbing, calibration, location alignment and sampling depth. Unsupported attribution is harder to repair than a candidly qualified measurement.

After rainfall, a fresh lens in the upper centimetres can register lower salinity than a 4 m ship intake. The difference may be physically real and need not indicate that the satellite or TSG is wrong. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Compare depth, footprint, rainfall history and time before declaring a calibration discrepancy. Translate both measurements with their spatial and vertical supports. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

29. Bottle sampling cadence limits drift detection

Discrete bottle checks anchor a continuous record only at their sampling times. If too few bottles are collected or they cluster in one regime, short periods of instrument drift can remain unseen. Review whether the example uses a directly observed quantity, a conductivity-based calculation, a statistical quality flag or an independent bottle reference. These categories should remain distinct when the result is rewritten for a general audience.

Taking three bottles on one morning cannot reliably establish that a ten-day voyage stayed calibrated. A regularly spaced sequence and targeted sampling after maintenance can reveal changing residuals. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Report number of samples, their time distribution and which operating states they cover. Precision claims should reflect actual reference coverage. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

30. A transparent data record includes repairs and exclusions

A useful final underway dataset includes provenance: raw variable, correction algorithm, bottle records, rejected intervals, navigation alignment, quality flags and version. Without provenance the smooth final series is difficult to audit. A careful translation retains units, time frame, intake configuration and uncertainty. It is better to state that a result is unresolved than to remove a critical qualifier merely to make the prose sound smooth.

If a front disappears after a flow-status filter, the report should say that the old feature was rejected because the sampling system was not drawing stable seawater. The revised interpretation then becomes traceable. These values are constructed teaching cases, not a report from a named ship. Check the dimension and whether the measurement belongs to seawater at intake, water inside a cell, or a laboratory reference sample.

Translate corrections as part of the science, not as inconvenient internal housekeeping. Readers need to know which observations were used and which were rejected. As a transfer check, imagine the flow rate, time lag, bottle result or sampling depth changed. Explain which part of the interpretation would still hold and which would need recalculation. Making that boundary explicit is the core of trustworthy TSG reporting.

Independent problems and checked answers

A 6 L water path runs at 1.5 L/min. Find nominal delay.

The delay is 6 / 1.5 = 4 minutes, assuming steady flow and that 6 L is the effective volume between intake and measuring cell. The useful habit is to name the measured variable, the conversion, its unit and the assumption on which the calculation depends. In an actual cruise report, cross-check the answer against logs and calibration records before accepting a strong environmental explanation.

At 6 m/s, how far does a ship travel during that four-minute delay?

Four minutes is 240 seconds. Distance is 6 × 240 = 1440 m or 1.44 km. This is an illustrative location offset, not a universal lag correction. The useful habit is to name the measured variable, the conversion, its unit and the assumption on which the calculation depends. In an actual cruise report, cross-check the answer against logs and calibration records before accepting a strong environmental explanation.

TSG values 35.02, 35.04 and 35.06 match bottles 35.00, 35.02 and 35.04. Find bias.

Each TSG-minus-bottle residual is +0.02. Mean bias is +0.02 practical-salinity units, suggesting a potential −0.02 additive offset if calibration stability is independently supported. The useful habit is to name the measured variable, the conversion, its unit and the assumption on which the calculation depends. In an actual cruise report, cross-check the answer against logs and calibration records before accepting a strong environmental explanation.

Residuals are +0.10 and −0.10. What does zero mean bias conceal?

Mean bias is zero, but RMSE is 0.10. Agreement must be assessed with scatter as well as signed mean; opposing errors can cancel. The useful habit is to name the measured variable, the conversion, its unit and the assumption on which the calculation depends. In an actual cruise report, cross-check the answer against logs and calibration records before accepting a strong environmental explanation.

Which temperature should represent seawater at the hull intake?

Use the properly calibrated intake-associated thermometer according to the ship’s processing plan, not automatically the warmer internal conductivity-cell temperature. The useful habit is to name the measured variable, the conversion, its unit and the assumption on which the calculation depends. In an actual cruise report, cross-check the answer against logs and calibration records before accepting a strong environmental explanation.

Why can a real front shift along track in the raw data?

Transport delay and sensor response place the measured signal later than the intake water’s original sampling time, moving it relative to the ship’s current navigation position. The useful habit is to name the measured variable, the conversion, its unit and the assumption on which the calculation depends. In an actual cruise report, cross-check the answer against logs and calibration records before accepting a strong environmental explanation.

What does salinity of 35 mean in Practical Salinity?

It is a conductivity-derived practical-salinity value, conventionally dimensionless, not thirty-five percent salt by mass. Converting to Absolute Salinity needs the appropriate framework. The useful habit is to name the measured variable, the conversion, its unit and the assumption on which the calculation depends. In an actual cruise report, cross-check the answer against logs and calibration records before accepting a strong environmental explanation.

A salinity dip coincides with low pump flow. Do you call it a plume?

No. First test a flow, bubble or stale-water artifact, and compare independent observations before giving it an oceanographic cause. The useful habit is to name the measured variable, the conversion, its unit and the assumption on which the calculation depends. In an actual cruise report, cross-check the answer against logs and calibration records before accepting a strong environmental explanation.

A satellite and ship salinity disagree just after heavy rain. Which extra information matters?

Compare satellite sampling depth, ship intake depth, rain timing, spatial footprint and ship calibration. A real vertical freshwater lens can explain part of the difference. The useful habit is to name the measured variable, the conversion, its unit and the assumption on which the calculation depends. In an actual cruise report, cross-check the answer against logs and calibration records before accepting a strong environmental explanation.

Why should a bottle outlier not be deleted without explanation?

The discrepancy could indicate sampling error, mistiming, a real front or instrument bias. Document the investigation and any rejection rule so calibration remains reproducible. The useful habit is to name the measured variable, the conversion, its unit and the assumption on which the calculation depends. In an actual cruise report, cross-check the answer against logs and calibration records before accepting a strong environmental explanation.

Frequently asked questions

Is a thermosalinograph the same as a CTD cast?

No. A shipboard TSG usually samples water continuously from a hull intake along the cruise track, while a CTD profiler measures variables through a vertical water column. They can be cross-checked but have different spatial and temporal support. Readers should verify vessel-specific data definitions and processing status. The generic explanation describes a method, not a guarantee that every ship has identical equipment or operating conditions.

Can the sensor measure salinity without temperature?

Conductivity depends on temperature, and practical salinity calculation requires suitable temperature information. Misalignment or thermal drift can bias the derived salinity even when raw conductivity looks plausible. Readers should verify vessel-specific data definitions and processing status. The generic explanation describes a method, not a guarantee that every ship has identical equipment or operating conditions.

Why are discrete bottles still needed with modern sensors?

They provide independent checks against conductivity drift and contamination. The usefulness of bottles depends on collection, lab analysis, timing and sufficient coverage. Readers should verify vessel-specific data definitions and processing status. The generic explanation describes a method, not a guarantee that every ship has identical equipment or operating conditions.

What if the salinity seems too low near a river mouth?

It may reflect true fresh water or a sensor or intake problem. Check operational status and a regional physical context before rejecting the result or accepting a freshwater-plume interpretation. Readers should verify vessel-specific data definitions and processing status. The generic explanation describes a method, not a guarantee that every ship has identical equipment or operating conditions.

Does one calibration offset fix the entire voyage?

Not necessarily. Bias can evolve with time, biofouling, flow, maintenance or instrument state. Apply an offset only with supporting comparison samples and checks. Readers should verify vessel-specific data definitions and processing status. The generic explanation describes a method, not a guarantee that every ship has identical equipment or operating conditions.

Is practical salinity in PSU identical to g/kg?

No. Practical Salinity is based on a conductivity ratio and conventionally dimensionless. Absolute Salinity in g/kg is a different thermodynamic quantity calculated under TEOS-10 conventions. Readers should verify vessel-specific data definitions and processing status. The generic explanation describes a method, not a guarantee that every ship has identical equipment or operating conditions.

What is the minimum metadata for a useful TSG chart?

Identify the ship or platform, intake depth, sensor chain, sampling interval, water-flow and lag treatment, geographic coordinates, QC status and calibration evidence. Otherwise a clean line can hide important ambiguities. Readers should verify vessel-specific data definitions and processing status. The generic explanation describes a method, not a guarantee that every ship has identical equipment or operating conditions.

Original methodology, field and calibration sources

NOAA AOML thermosalinograph introduction and measurement chain.

NOAA AOML thermosalinograph real-time and delayed-mode quality-control procedures.

University of Hawai‘i HOT programme thermosalinograph and bottle sampling manual.

WHOI–University of Hawai‘i WHOTS cruise shipboard TSG calibration and processing.

University of Hamburg thermosalinograph observations for satellite salinity calibration.