Warehouse Temperature Mapping Study: Protocol and Sensor Placement

The monitoring probe on the pillar at aisle 14 has read between 21 and 23 °C every afternoon for two years, and the quality file says the warehouse is a 15 to 25 °C area. Then a customer's auditor asks for the mapping report, sees that the study ran for three days in April with loggers at chest height, and asks what the top rack level reaches in July. Nobody knows, and 180 pallets sit up there.

A mapping study answers that question with evidence: where the warmest and coldest points are, in which season, at which height, and therefore where products may be slotted and where the permanent sensors belong. Planned with a written protocol, a calculated logger grid and two seasonal runs, it costs less than a single rejected lot and gives the warehouse, quality and customer teams the same map to work from.

• What this guide covers: the protocol, logger selection and calibration, grid spacing and heights, study duration and seasons, data analysis, and the triggers for re-mapping, using WHO's technical supplement on mapping and the USP and Indian regulatory context.
• What it adds: a logger-count formula worked for a high-bay ambient warehouse, a timestamped hot-spot incident, and illustrative study costs in USD and INR.

What a mapping study proves that routine monitoring cannot

Routine monitoring tells you the temperature at a few fixed points; mapping tells you whether those points represent the whole space. Without a map, a compliant probe reading can coexist with product sitting out of range three metres higher or next to a dock door.

WHO's technical supplement on temperature mapping of storage areas, issued in May 2015 as a supplement to Annex 9 of WHO Technical Report Series No. 961, states that new temperature-controlled storage areas must be mapped before commissioning, empty and in a normal loaded condition, and that mapping should define zones not to be used for time- and temperature-sensitive products. It applies to warehouses, cold rooms, freezer rooms, quarantine areas and receiving and loading bays.

Warehouse temperature mapping for a 15 to 25 °C storage area, under section 1.1 of WHO TRS 961 Annex 9 Supplement 8, takes place before commissioning, both empty and loaded, and the approved report held in the site quality system decides which WMS bays may hold labelled products.

Regulators in both markets expect the evidence without prescribing the method. In the US, 21 CFR 211.142(b) requires drug storage under appropriate temperature, humidity and light conditions, and 21 CFR 205.50(c)(2) requires wholesale distributors to use temperature and humidity recording equipment to document proper storage. USP general chapter <1079.4>, Temperature Mapping for the Qualification of Storage Areas, sits in the USP's storage and distribution series.

Writing the protocol before a single logger is switched on

The protocol fixes the questions before the data arrive, so the results cannot be reinterpreted to fit. WHO's supplement says the logger selection, team, site survey, acceptance criteria and logger locations must all be settled before the protocol is approved.

• Scope and condition: state whether the area is mapped empty, as at operational qualification of a new building, or loaded, as after an alteration to an existing store.
• Acceptance criteria: the allowable limits for the products stored, for example +15.0 to +25.0 °C or +2.0 to +8.0 °C; a study may also run without limits to discover what the space can hold.
• Door-opening test: if included, define frequency and duration; WHO's guidance is that temperature should stay within limits except for at most 30 minutes after a door opening.
• Objectives: variations by day of week and time of day, effects of heating and cooling cycles, airflow problems, safe and unsafe storage zones, and the best positions for permanent monitoring sensors.
• Team and signatures: named, trained members with recorded signatures so every sheet traces to a person.

In India, CDSCO's draft Guidelines on Good Distribution Practices for Pharmaceutical Products (document CDSCO/GDP.PP, version 00) states in paragraph 12.5 that storage areas shall be temperature mapped under representative conditions and that monitors should sit where fluctuations are most likely. The draft also asks, in paragraph 12.4, that monitoring records be kept for at least the shelf life of the stored product plus one year.

Choosing loggers: range, interval and calibration evidence

A study is only as good as its weakest logger, so the specification goes into the protocol. Every logger needs a valid three-point calibration traceable to NIST with an error of no more than ±0.5 °C at each point, and all loggers in one study should be the same model, programmed identically.

• Measuring range: wide enough for every extreme you expect; WHO gives -30 °C to +60 °C as an example.
• Recording interval: user-programmable between 1 and 15 minutes, with memory for the whole study at that interval.
• Calibration points: one below the range, one in the middle and one above, completed within the current year; certificates go into the report.
• Synchronised start: the same interval and start time on every unit, with enough delay to fix them all in place, or the readings cannot be compared.
• Software: data storage and analysis software that meets applicable rules; WHO cites 21 CFR Part 11 as an example.

Warehouse temperature mapping loggers under WHO TRS 961 Annex 9 Supplement 8, Step 1, need a NIST-traceable three-point calibration completed within the current year with no more than ±0.5 °C error at each point, a 1 to 15 minute recording interval, and a shared start time so every reading can be time-correlated.

Check memory before buying or hiring. At a 5-minute interval a logger records 288 readings a day, so a 7-day study plus three days of placement and conditioning needs at least 2,880 readings per channel. Add a door-opening sensor if the protocol tests doors; WHO suggests one so that temperature peaks can be aligned with door events instead of argued about afterwards.

Sizing the logger grid for a high-bay racked warehouse

Place loggers on a horizontal grid every 5 to 10 metres, stacked vertically at several heights, and add loggers at known risk points. WHO notes that very large facilities may stretch the spacing to 20 or 30 metres, but positions should always coincide with where product is or will be stored.

Height matters more than floor area in a racked building. WHO's guidance is that where the ceiling is 3.6 metres or lower, loggers go at low, middle and high level, such as floor, 1.2 m and 3.0 m. Above 3.6 metres, use vertical arrays at the bottom, middle and top; for a 6-metre space WHO's example heights are 0.3, 1.8, 3.6 and 5.4 metres. The formula below sizes the grid for an illustrative 7,200 m² ambient DC.

Logger count for one mapping run

  N_grid  = (ceil(L / s) + 1) x (ceil(W / s) + 1)
  N_zone  = N_grid x h
  N_total = sum of N_zone over zones + N_extra

  L, W    = zone length and width (m)
  s       = grid spacing (m), 5 to 10 m; wider only in very large spaces
  h       = logger heights per grid point
  N_extra = dock doors, HVAC outlets, existing monitoring probes

Illustrative ambient DC, 120 m x 60 m, top beam at 9.0 m, +15 to +25 C

  Racked zone      120 x 40 m, s = 10 m  -> 13 x 5 = 65 points
                   heights 0.3, 3.0, 6.0, 9.0 m (h = 4)   -> 260
  Marshalling zone 120 x 20 m, s = 20 m  -> 7 x 2 = 14 points
                   heights 0.3, 1.8 m (h = 2)             -> 28
  Extras           8 dock doors + 6 HVAC outlets
                   + 6 loggers beside fixed probes        -> 20
  N_total                                                 = 308

Memory check at a 5-minute interval:
  readings = (7 study days + 3 setup days) x 24 x 60 / 5 = 2,880

Give every location a unique ID that encodes zone, aisle, bay and height, and record the serial number against it in the logger location table so each trace links to its calibration certificate. Racking layout drives the grid, so the as-built drawing must match reality; our guide to pallet racking types explains why drive-in, push-back and selective racking create different airflow paths.

Seven days, two seasons and the events worth provoking

Run ambient warehouses for at least seven consecutive days, covering five working days and two weekend days, and repeat the study in the warmest and the coldest season if outside weather affects the building. Cold rooms and freezer rooms generally need 24 to 72 hours and usually only one season.

• Weekly cycle: the seven-day minimum captures weekend HVAC set-backs and the Monday morning recovery, which a three-day study misses.
• Seasonal pair: WHO's supplement says at least two studies may be needed where seasons affect the area, one in the warmest and one in the coldest season, compared to separate seasonal effects from local problems.
• Duplicate refrigeration: where a room has two refrigeration units, map with each unit running alone for a similar period, because the temperature pattern can change with the unit in charge.
• Doors and power: run a defined door-opening test, and if required show how long the space stays within limits after a power failure.
• Access log: without automatic door logging, keep a manual log of door openings, personnel movement and power cuts during the study.

For a new building, map empty during operational qualification and then loaded once stock and racking are in place, because full racking changes airflow. The new distribution centre start-up checklist shows where the loaded study fits before the first labelled-product receipt is allowed.

Incident trace: the top-beam hot spot behind a compliant probe

The trace below is illustrative and follows the 308-logger study sized above, run in the hot season. It shows how a report can pass a space that fails, simply by summarising instead of reading each location against the limits.

1. Day 1, Monday 07:30: The team places 308 loggers per protocol MAP-DC2-S1, all programmed for a 5-minute interval and a 10:00 start; location DL-R14-H9 is on the aisle 14 upright at 9.0 m.
2. Day 3, Wednesday 16:20: DL-R14-H9 peaks at 26.6 °C. The fixed BMS probe TP-07 on the aisle 14 pillar at 1.8 m reads 22.4 °C, so no alarm is raised.
3. Day 9, analysis: Eleven loggers at 9.0 m across aisles 12 to 16 exceed 25.0 °C for three to five hours each afternoon. The WMS shows 182 pallets of 15 to 25 °C stock on levels 5 and 6 there.
4. Day 12, first draft: The draft report states "mean 23.1 °C, compliant", because the analyst summarised per-zone averages and never compared each location's maximum with the limit.
5. Controlled version: The reviewer rejects the draft, the WMS blocks levels 5 and 6 in aisles 12 to 16 for 15 to 25 °C classes, TP-07 moves to 9.0 m, destratification fans are adjusted, and those aisles are re-mapped the same season.

Reading the data: hot spots, cold spots and mean temperatures

Compare every logger's minimum and maximum with the acceptance limits first, then use means to identify consistently warm or cool locations. WHO defines hot and cold spots as the highest and lowest readings that remain within the range; readings outside the range are deviations, not hot spots.

WHO's example makes the distinction concrete: in a room specified for +15.0 to +25.0 °C, cold spots might sit between +15.0 and +17.5 °C and hot spots between +23.0 and +25.0 °C. Hot and cold spots identify where routine monitoring sensors should go, and they must be determined per season because they move. Mean temperatures per location expose repetitive warm or cold patterns that single readings hide, but they never replace checking the actual data for out-of-range values.

Hot and cold spots in a warehouse mapping report under WHO TRS 961 Annex 9 Supplement 8, section 2.4.2, are the highest and lowest readings that still fall inside the acceptance range, such as +23.0 to +25.0 °C in a 15 to 25 °C area, and they decide where monitoring sensors are placed.

• Explain the variations: list the causes, such as heating and cooling units, doors, roof heat and building construction, and allow for each logger's stated error.
• Recommend storage zones: place highly sensitive products where fluctuations were smallest and less sensitive items elsewhere.
• Recommend sensor positions: move routine monitoring and control sensors to the locations the data show are most at risk.
• Get an independent review: WHO's supplement asks for the report to be audited and peer-reviewed by a competent independent person, and approved by whoever commissioned a third-party study.

Mapping is a one-off qualification, not live monitoring. Once sensors are placed, the daily question becomes whether their data are arriving at all; our guide to telling a temperature alarm from a sensor-data gap covers that monitoring problem, which this study only sets up.

Illustrative cost of one two-season study in USD and INR

Most of the cost of a mapping campaign is logger hire and analyst time, and both scale with the logger count. The figures below are illustrative for the 308-logger DC above, with one run per season and a ten-day logger hire covering placement, conditioning and the seven-day study.

Set the campaign cost against the stock it protects. In the incident above, 182 pallets sat in the hot zone; at an illustrative $9,000 or ₹6,00,000 per pallet, that is $1.64 million or ₹10.92 crore of product whose storage history could not be defended. Re-mapping a few aisles after a fix costs a fraction of the full run, so budget a partial re-map in the same season as standard.

Which steps the WMS and BMS enforce and where an analyst judges

Turn the map into location rules the WMS can enforce, and keep the judgment calls with qualified people. A storage class allowed into a bay the map has excluded is a configuration error; deciding whether a door-adjacent spike is acceptable is a quality decision.

The same zoning logic belongs in the building design itself. If HVAC and racking are still being specified, the warehouse design process guide shows where temperature zoning enters the layout decisions, which is cheaper than correcting stratification after the racks are up.

Re-mapping triggers and a study readiness checklist

A map stays valid only while the building and its use stay the same. WHO's supplement suggests periodic re-mapping, for example every three years, unless continuous fixed monitoring supports a periodic re-evaluation instead, and a new study after significant changes.

• Layout or loading change: new racking, higher stacking or a different use pattern that affects loading or air circulation.
• Equipment change: replaced or modified heating, cooling or air distribution, including a set-point change.
• Unexplained variability: monitoring data showing variation outside normal operating limits with no known cause.
• Remedial action: a follow-up study to prove that fixes from the last report worked.

For more on warehouse layout and validation, see the logistics and supply chain hub. If you want mapping results turned into WMS location rules and BMS alarm settings, send us an RFP with your floor plan, rack heights and the storage classes you hold.