Why Can Two Nearby Temperature Sensors Disagree?
From the earlier reference collection. Read the current software blog.
Nearby sensors may sample different water, operate under different conditions, or report measurements taken at different times.
Nearby does not mean equivalent
Two temperature sensors can be close on a map and still measure different conditions. One may sit in a sheltered cove while the other is exposed to a current. One may be near a warm inflow, a sunlit bank, or a structure that changes local circulation. The first step in interpreting a difference is to compare the stations’ descriptions, not to assume that one instrument is wrong.
Consider a pair of probes on opposite sides of a pier. A tide or wind may push water past one probe while the other sits in a quiet pocket. Their distance could be short, but the water reaching each sensor is not necessarily the same.
Compare depth and exposure
Depth is especially important when water is layered. A shallow probe can respond to surface heating sooner than a deeper probe. Differences in shade, enclosure, mounting, or distance from shore can also affect local conditions or how quickly the instrument responds. Check the stated depth and placement details before treating values as direct duplicates.
A good comparison records the station coordinates, depth, and basic site description side by side. If those details are absent, the uncertainty is about what each reading represents. Avoid filling that gap with a confident explanation based only on proximity.
Check calibration and timing
Sensors can differ because of calibration, instrument condition, or maintenance history. Look for calibration dates, methods, stated uncertainty, and documented repairs or replacements. A small difference may fall within the instruments’ stated uncertainty; a larger one deserves investigation, but it still does not establish which reading is correct by itself.
Timing can create an apparent disagreement too. One display may show a new observation while another shows a value from earlier in the hour. Compare observation timestamps, not just the time the dashboard refreshed. If one station samples every minute and another every fifteen minutes, a rapid change can produce different reported values even when both are accurate for their respective observation times.
Make a fair comparison
A practical comparison begins by aligning units and observation times, then reviewing depth, exposure, calibration, and summary method. For example, convert both readings to the same unit, choose observations made at similar times, and confirm whether each value is an instantaneous sample or an interval average. Keep the original values and metadata so the comparison can be repeated.
If the difference remains, describe it as an observed difference and identify plausible causes as possibilities. Do not silently average the readings or label one faulty without evidence. More context—such as maintenance records or an independent nearby measurement—may help resolve the question. The result should make clear what is known, what has been checked, and what remains uncertain.