Cold chain record-keeping for packhouses: what to log and when

The three moments that matter

Receival: pulp temperature (not air temperature) on arrival, per lot, with the time. This is the record that separates your problem from the grower's or the carrier's.

Storage: which cold room the lot went into and when. Room-level logging usually comes from the room itself; what your lot records need is the link, this lot, this room, these dates, so a room excursion can be mapped to affected lots.

Dispatch: pulp temperature at loading, per lot or pallet, with time and truck. After the truck door closes, this is your last defensible data point.

A pulp thermometer reading photographed against the carton label is close to indisputable, and takes five seconds on a phone. When a buyer claims fruit arrived at 12 degrees, a timestamped photo showing 4 degrees at loading changes the conversation immediately.

Pulp, air, and surface temperature are not interchangeable

Air temperature inside a cold room or a reefer container tells you what the refrigeration unit is doing, not what the product actually experienced. A room can read 2 degrees Celsius on its display while a pallet loaded four hours earlier still holds field heat at its core. Surface temperature, read with an infrared gun off the outside of a carton, is faster to take but reflects whatever the carton surface has been exposed to in the last few minutes, a forklift ride across a warm dock, direct sun through a loading bay door, and can be several degrees off the product inside.

Pulp temperature, taken with a probe thermometer pushed into the flesh of the fruit or the core of a carton, is the only reading that reflects what the product itself has actually reached. To take it correctly: select a unit from the middle of the pallet, not a corner carton that has had the most airflow, open the carton, and insert the probe at least 2 to 3 centimetres into the flesh of a piece near the centre of the pack, away from any ice or gel packs that would give a false low reading. Hold it until the display stabilises, usually 10 to 20 seconds for a fast-response probe, before recording. For product that cannot be probed without damage, such as leafy greens or delicate berries, use a probe inserted between units at the pack centre instead, which approximates pulp closely enough for record purposes.

Choosing a logger: single-use, reusable, and real-time

Single-use loggers are disposable PDF or paper-strip devices, roughly 8 to 15 US dollars each, that record a temperature curve for one shipment and are discarded, or filed, at the far end. They are the right choice for a one-off shipment to a new buyer who wants proof, or for lanes where the receiving party will not return equipment.

Reusable loggers, USB or Bluetooth devices in the 30 to 80 US dollar range, download their recorded curve to a laptop or phone and get redeployed for the next trip. They cost more per unit but far less per trip over a season, and they are the standard choice for a packhouse running the same lanes repeatedly, provided someone owns getting them back.

Real-time loggers use cellular or satellite connectivity to report temperature, and often GPS position, continuously during transit, typically 150 to 400 US dollars per unit or on a subscription model. They earn their cost on long ocean legs or high-value cargo, where an alert that a reefer has drifted out of range at day 6 of an 18-day voyage lets someone intervene before day 18, rather than finding out on arrival.

The trade-off is simple: single-use buys you a record after the fact, reusable buys you the same record more cheaply if you run the lane often, and real-time buys you the chance to act before the claim exists.

How many loggers, and where

For a 40-foot reefer container, use a minimum of three loggers: one near the nose, close to the unit's own return-air sensor, one at the mid-point of the load, and one at the door end, which is the position most exposed to ambient heat during loading and unloading and typically shows the widest swings. A single logger at the nose alone will read close to the setpoint even when the door-end pallets are running several degrees warmer, because that is exactly where the refrigeration unit is working hardest to compensate.

For a pallet moving within a cold room or a short domestic haul, one logger per pallet is standard, taped to a carton at the pallet's core, roughly the geometric centre between two layers, rather than left loose on top where it reads ambient air instead of product temperature.

A fixed cold room needs its own continuous logger recording room air temperature at a set interval, commonly every 5 to 15 minutes, independent of any per-lot logger. The room logger tells you when an excursion happened; the lot record tells you which lots were in the room at that time. Both are needed, because a room reading alone cannot tell you whether lot 4471 was affected or had already shipped out that morning.

What a defensible record contains, and how long to keep it

A temperature reading without supporting detail is an assertion, not evidence. A record that will hold up needs five fields every time: the lot id, the time to the minute, the instrument id (the serial number or asset tag of the specific thermometer or logger used), that instrument's most recent calibration date, and who took the reading. Drop the instrument id and a buyer's claims team can argue you cannot prove which device, or how reliable it was. Drop the time and they can argue the reading was taken before or after the moment in dispute.

Retention periods are set by the buyer contract far more often than by a food safety regulator. Most contracts we see specify 2 years; some importers, particularly those selling on to retail chains with their own claims windows, ask for 3. The one regulatory floor worth knowing is the US FDA's Food Traceability Rule (FSMA 204), which requires 2 years of retention for records tied to Critical Tracking Events on covered high-risk commodities, but that rule governs traceability records, not temperature specifically, and most packhouses end up holding temperature records to whichever is longer: the contract term or that floor.

Calibration, audits, and the handover points that decide who pays

An uncalibrated instrument loses an argument even when the reading itself was accurate, because the other side only has to raise reasonable doubt about the number, not disprove it. Check a probe thermometer against an ice bath, which should read 0 degrees Celsius, at least monthly, and send it or its logger equivalent for a certified, traceable calibration annually. If a check shows more than 0.5 degrees Celsius of drift, pull the instrument from use and recalibrate before its next reading is used in a record. Keep the calibration certificate on file and reference its date in every reading that instrument produces.

When an auditor or a buyer's claims team investigates a temperature dispute, they typically ask for three things in this order: the temperature log for the specific lot number in question, the calibration certificate for the instrument that produced it, and a chain of custody showing who handled the product and who took each reading. They will usually cross-check your figures against the carrier's own reefer data logger or a competing set of readings taken by the receiving party, so a record that is internally consistent, same instrument id, same person, same format, across the whole chain is worth more than one perfect number in isolation.

Responsibility for temperature transfers at defined handover points, and the reading that protects you is the one taken right at that transfer, not before or after it. At grower to packhouse, a receival pulp reading protects the packhouse from being blamed for field heat it inherited. At packhouse to cold store, a reading logging which room and when protects the packhouse if the room later has an excursion. At cold store to carrier, the dispatch pulp reading at loading protects the packhouse from anything that happens on the truck, since it is the last point before the carrier takes physical control. At carrier to importer, a reading taken at discharge, ideally by both parties or witnessed, protects whichever side is not at fault for anything that happened in transit. Miss the reading at any one of these points and that leg of the journey becomes unaccountable, which in practice means whoever holds the weaker paperwork absorbs the claim.

Attach records to the lot, not the day

A folder of daily temperature sheets answers "what happened on Tuesday"; a dispute asks "what happened to THIS lot". Records attached to the lot follow it through packing, palletising and dispatch, so the evidence is already assembled when a claim arrives. Fernable's QC checks attach readings and photos directly to the lot at receival and dispatch, which is exactly this pattern.