How to choose a sampling interval for mapping

A sampling interval is the frequency at which a data logger records a temperature measurement. It determines how much of what actually happened in your environment ends up in the data record.

A mapping studyis only as reliable as the data it captures. Calibration gets most of the attention, but a factor of equal importance often goes unchallenged in protocol reviews: The sampling interval.

As we know, a data logger does not record temperature continuously, but "wakes up" at a set interval, takes a snapshot, saves the value, and goes back to sleep. The software then connects those snapshots into a line graph, and everything between two data points is interpolation, not measurement.

In other words, the sampling interval is simply how much time passes between those snapshots.

The frequency at which a data logger records a measurement determines how much of the thermal picture you actually see. Set the interval too wide, and the final report may present a stable environment that never existed. Set it too narrow, and you are drowning in data points without gaining useful resolution.

For a validation engineer, the question is straightforward: what happened to the temperature between those two dots? If the interval is short enough, the answer is "probably nothing significant." If it is not, you may have gaps in your data that no amount of post-hoc analysis can fill.

This matters for three reasons that show up in audits and batch decisions.

Missed excursions

A compressor failure that lasts eight minutes will appear in data logged every minute. It may be completely invisible in data logged every fifteen minutes if the system recovers between samples. The mapping report would show a stable environment, even though a real excursion occurred and went unrecorded.

In GxP environments, undocumented excursions are not harmless. If an inspector reviews your mapping data alongside HVAC event logs and finds a cooling failure that does not appear in the temperature record, the gap itself becomes the finding.

MKT calculation accuracy

Mean kinetic temperature calculations depend on the granularity of the input data. USP <1079.2> defines specific evaluation windows for excursion assessment — 30 consecutive days for controlled room temperature, 24 consecutive hours for controlled cold temperature. If the underlying data has 15-minute gaps, short but significant temperature spikes will be averaged into the calculation rather than captured as discrete events. The result is an MKT value that understates actual thermal exposure.

Audit defensibility

When you present mapping results during an inspection, the sampling interval is part of the story. Inspectors expect the interval to be justified by a documented rationale tied to the thermal characteristics of the environment. "We used 5 minutes because that is what the logger defaults to" is not a defensible answer. "We used 5 minutes because the area's recovery time from door-opening events is 8–12 minutes, and this interval captures at least two data points per event" is.

Also read: GDP temperature mapping: protocol template and guidelines

A common rule of thumb is that GxP guidelines suggest intervals between 1 and 15 minutes. That is roughly accurate but worth unpacking.

WHO TRS 961 Annex 9 Supplement 8 specifies that data loggers should have a programmable sampling period with intervals ranging from one second to beyond one hour. This describes the capability requirement for loggers, not a recommended interval for studies. The WHO does not prescribe a single interval for all applications.

The ISPE Good Practice Guide for Controlled Temperature Chambers elaborates on recording intervals and provides rationales that connect interval selection to the thermal behavior of the mapped space. Dickson and other industry sources cite 1–15 minutes as a typical range for warehouse mapping, with shorter intervals for equipment that changes temperature more rapidly, such as freezers.

Eupry's own WHO mapping guidelines page notes that five-minute intervals will often be sufficient, but emphasizes that this should be based on your risk assessment, not applied as a blanket default.

The common thread across all of these sources is that the interval should be justified by the rate of thermal change in the environment, not chosen by convenience or logger defaults.

ow do you choose the right sampling interval for mapping your environment?

Shorter is not always better. A 10-second interval on a 72-hour warehouse study generates over 25,000 data points per sensor. Multiply that across 30 or more loggers, and you have a dataset that is expensive to store, slow to analyze, and adds no value if the warehouse temperature changes meaningfully only over minutes, not seconds.

The decision depends on three factors.

Rate of change. How fast does the environment respond to disturbances? A shipping dock with frequent door openings and direct exposure to outside air changes temperature rapidly, sometimes within a few minutes. A large cold room with high thermal mass and stable HVAC may take an hour to shift by a single degree. The interval should be short enough to capture at least two data points during the fastest expected temperature change.

Also read: Thermal mass and thermal lag: why air alarms do not tell the full story

Risk assessment. What is the consequence if a short excursion goes unrecorded? For a warehouse storing controlled room temperature products with broad stability budgets, a missed two-minute fluctuation is unlikely to matter. For a freezer holding biologics at -70°C (-94°F), the same gap could mask a critical event.

Practical constraints. Battery life, memory capacity, and data processing time all scale with logging frequency. If your quality system cannot efficiently handle the data volume, the extra resolution creates a new problem without solving the original one. The right interval balances fidelity with operational feasibility.

A reasonable starting framework for most pharmaceutical mapping studies:

• Freezers and ULT equipment: 10 seconds to 3 minutes, depending on door frequency and recovery characteristics
• Refrigerators and cold rooms: 1 to 5 minutes
• Ambient warehouses: 5 to 15 minutes, with shorter intervals near docks and doors

These ranges align with what we recommends for mapping equipment selection and should always be documented in the mapping protocol with a rationale tied to the specific environment.

Also see: How to create a temperature mapping protocol in pharma