How Do You Find Out Whether a Sensor Was Replaced?
From the earlier reference collection. Read the current software blog.
Station metadata, instrument identifiers, and maintenance records can reveal whether a sensor changed and how to interpret the series.
Check the station metadata
Start with the station’s published metadata. Look for a sensor or instrument identifier, model description, installation date, and any recorded period of operation. A station name can remain constant while the hardware changes, so the station label alone does not establish that every historical reading came from the same instrument.
If metadata is versioned, compare the current record with earlier versions. Note when an identifier appears or disappears and whether the source explains the transition. Avoid inferring a replacement solely from a change in values; environmental conditions and processing changes can also cause shifts.
Look for maintenance history
Change logs, service notes, and maintenance records may document replacement, repair, recalibration, relocation, or other work. Search by the station identifier and relevant dates. A record that names the removed and installed instruments is more conclusive than a generic maintenance entry, but even brief notes can help identify a period requiring closer review.
If records are not published, the data provider may have a station page or downloadable metadata that supplies instrument history. State what records were available and what they establish. Do not say that a sensor was replaced merely because a log is silent; lack of documentation leaves the question unresolved.
Inspect the data cautiously
A discontinuity in the series may be a clue. Look for an abrupt level shift, changed precision, a new sampling interval, or a gap around a suspected maintenance date. These patterns can suggest a change in instrument or processing, but they are not proof of replacement. A real change in conditions or a revised import can produce similar effects.
Compare the timeline with recorded station and processing changes. If an instrument identifier changes on a known date, mark that boundary in charts and preserve the earlier and later records. If there is only a suspected boundary, label it as such and explain the evidence rather than presenting it as confirmed.
Preserve continuity and provenance
When a replacement is documented, retain both the station identity and the instrument identity in the data model. Record the date range for each instrument and any changes to location, depth, calibration, or reporting method. This lets readers distinguish a continuous station series from a single unchanged device series.
A replacement does not automatically make data incomparable, but it may affect interpretation. Readers can check whether a transition creates a measurement discontinuity and account for it in analysis. Clear metadata and change history help answer the practical question: which sensor produced each value, and what changed around that time?