A practical, step-by-step SOP for Indonesian tuna processors and exporters to pass EU mercury checks in 2026: lot definition, sampling, composite vs individual testing, method/LOQ choices, measurement uncertainty, and COA essentials.
As an Indonesia-Seafood Team that ships tuna to demanding EU buyers, we’ve learned the hard way that mercury control isn’t about one perfect test. It’s about a simple, repeatable system that keeps every lot onside. Here’s exactly how we build and run that system in 2026.
Hook: How we kept EU tuna lots compliant with one simple SOP
We’ve had 0 detentions for mercury in the last 12 months across yellowfin saku, steaks and cubes. Not because our fish are magically low. Because we define lots clearly, sample smart, and make the lab do the heavy lifting on uncertainty. This is the same playbook we hand to partners packing Yellowfin Saku (Sushi Grade), Yellowfin Steak, Bigeye Loin and Skipjack Cube for the EU.
The 3 pillars of fast, reliable EU compliance
- Clear lot definition. Don’t mix risk. Keep species, size grade, FAO area and processing date aligned. If you merge grades, your variability explodes.
- Fit-for-purpose testing. Pick a 17025-accredited lab, lock LOQ, and demand reported uncertainty. Then your borderline calls won’t turn into port surprises.
- Document for port. Your COA must read like an official-control report. When authorities see the right fields, they move faster.
Quick refresher for 2026: Under Regulation (EU) 2023/915, the maximum level for total mercury in tuna (Thunnus spp.) is 1.0 mg/kg wet weight. That’s the number your process must be built around.
Week 1–2: Map risk and lock the sampling plan
Start with the product lines you’ll ship to the EU in the next quarter. We score each line on species and size:
- Skipjack. Lowest risk, small-bodied. 6–12 fish composites often test far below 1.0 mg/kg.
- Yellowfin. Medium risk. Bigger, older fish push higher.
- Bigeye. Highest risk. Large specimens can exceed 1.0 mg/kg even from clean grounds.
Define a lot practically. We use: same species, same size grade, same FAO area (71 or 57 for most Indonesian tuna), same production date window (≤48 hours), same processing line. Smaller, tighter lots are easier to hold and release if issues arise.
How many units per lot should you sample? For industry self-checks mirroring EU official practice, we recommend:
- Default: 10 primary units per lot up to 20 tons.
- High-risk lots (bigeye, yellowfin >40 kg, mixed vessels): 15–20 units.
- Very small lots (<1 ton): 5 units.
Composite vs individual units. A composite sample is accepted in EU official controls and is efficient for release testing. We do a two-tier approach:
- Tier 1 composite: equal-mass portions from each primary unit, homogenized to one lab sample.
- Triggers for Tier 2: if composite is 0.85–1.05 mg/kg, test 3–5 individual units (or sub-composites by size grade) to localize risk and decide disposition.
Sampling details that avoid rework:
- Take edible dorsal white muscle. Avoid skin, bloodline, and glaze. Hg ML applies to wet-weight muscle meat.
- From each unit, cut 20–30 g. Pool to 200–500 g composite. Homogenize to a paste. Split into A/B retains.
- Keep everything at −18 °C. Use clean plastic or stainless tools. Label units and time-stamp custody.
Takeaway: define lots narrowly, sample 10–20 units, run a composite first, and keep a retain. That one change will solve 80% of mercury headaches.
Week 3–6: Choose methods, set LOQs, and pilot your COA
Which method is best for mercury in tuna? We’ve had excellent results with two routings:
- ICP-MS after microwave digestion. Best for multi-element panels. LOQ typically 0.001–0.005 mg/kg.
- Direct mercury analyzer (thermal decomposition + CV-AAS/CV-AFS, e.g., DMA-80). Fast, robust, no wet digestion. LOQ 0.002–0.01 mg/kg.
What LOQ is acceptable? For a 1.0 mg/kg limit, an LOQ ≤0.01 mg/kg is fit-for-purpose. We aim for ≤0.005 mg/kg to keep uncertainty tight.
ISO/IEC 17025 accreditation. Does the lab need to be in the EU? No. Use an ISO/IEC 17025-accredited lab in Indonesia whose scope specifically covers mercury in fish by your chosen method. KAN accreditation under ILAC MRA is recognized by EU importers. Ask for the latest scope and certificate.
Measurement uncertainty. This is where borderline lots live or die. Insist the lab reports the result with expanded uncertainty (k ≈ 2, 95% confidence). Typical U at 1.0 mg/kg for good labs is 8–20%.
- EU decision logic in practice: a lot is deemed non-compliant when result − U > 1.0 mg/kg. If result is slightly above 1.0 but result − U ≤ 1.0, authorities generally won’t condemn on that basis alone.
- Example: 1.04 ± 0.12 mg/kg. Since 1.04 − 0.12 = 0.92 mg/kg, you likely pass. At 1.18 ± 0.08 mg/kg, 1.18 − 0.08 = 1.10 mg/kg, that’s a fail.
Pilot a COA that satisfies EU checks. Your COA should include, at minimum:
- Product and lot ID, species and scientific name, form (loin, saku, steak), FAO area.
- Sampling plan: number of primary units, composite vs individual testing, dates, sample mass, homogenization procedure.
- Method and lab: ICP-MS or DMA/CV-AAS, ISO/IEC 17025 accreditation number and scope.
- Performance: LOQ, measurement uncertainty (k and confidence), CRM/QA controls used.
- Result: mg/kg wet weight, unit basis, analyst/signatory, report date.
We template this once, then run it for all EU-bound tuna lines like Yellowfin Saku (Sushi Grade) and Yellowfin Steak. If you’d like us to sanity-check your COA format or sampling plan, feel free to Contact us on whatsapp.
Week 7–12: Scale, train, and tighten decision rules
Run three consecutive pilot lots. Track the spread. You’ll quickly see patterns by species, size and ground. Then lock decision thresholds that your QA team can follow without debate.
Our practical thresholds:
- Composite <0.70 mg/kg. Release.
- 0.70–0.85 mg/kg. Release, but increase next lot’s sampling to 15 units.
- 0.85–1.00 mg/kg. Hold. Test 3–5 individuals or sub-composites by size grade to localize risk and consider regrading.
-
1.00 mg/kg. Hold. Do not blend to dilute for EU placement. Regrade or divert per your internal policy.
Turnaround and cost in Indonesia. What we see today:
- Lead times: 3–7 working days standard. 24–48 hours for rush, subject to lab capacity.
- Cost: roughly USD 40–100 per sample for Hg only, more for multi-element panels. Rush adds 30–100%.
Chain-of-custody and logistics that auditors love:
- CoC form linking lot, unit IDs, sampler, time, condition, and seal numbers.
- Frozen sample courier at −18 °C with temp loggers for long hauls.
- Photo of sampling and homogenization setup attached to the report packet.
At this point, QA should own the SOP. Sales shouldn’t be chasing labs. And port questions become routine.
The 5 biggest mistakes we still see (and how to avoid them)
- Vague lots. Mixing size grades or multiple catch areas. Fix by tightening lot definitions. Your variance drops.
- Under-sampling bigeye. Five units won’t cut it. Go 15–20 or don’t ship.
- No reported uncertainty. If the COA has a single number with no U, you’re one inspection away from a headache.
- LOQ too high. A 0.05 mg/kg LOQ is technically fine but inflates U. Ask for ≤0.01 mg/kg.
- Sampling only the lean tail. We’ve seen teams “cherry cut” out of habit. Take consistent dorsal muscle portions across the loin and homogenize properly.
Quick answers to the questions we get most
What is the EU mercury limit for tuna in 2026?
1.0 mg/kg total mercury (wet weight) for tuna under Regulation (EU) 2023/915. That’s unchanged for 2026 as of this writing.
How many units should I sample from a tuna lot?
Ten primary units per lot is a solid baseline. Go to 15–20 for bigeye, large yellowfin, or mixed-risk lots.
Is a composite sample acceptable under EU rules?
Yes. Official controls often test composites. For internal release, use a composite first, then test individuals if you’re near the limit.
Which method is best, ICP-MS or CV-AAS/DMA?
Both work if the lab is competent and accredited. ICP-MS is great for multi-element panels; DMA/CV-AAS is fast and rugged. Choose the one that meets your LOQ and uncertainty targets.
Do borderline results pass with uncertainty?
Authorities typically deem a lot non-compliant only when result − U exceeds 1.0 mg/kg. So uncertainty can be the difference between pass and fail.
What must appear on a COA?
Lot/species IDs, FAO area, sampling plan, method, lab 17025 details, LOQ, measurement uncertainty, result in mg/kg wet weight, dates, and signatory.
Does the lab need ISO/IEC 17025 accreditation in Indonesia or the EU?
Indonesia is fine if the accreditation body (KAN) is an ILAC MRA signatory and the lab’s scope covers mercury in fish by your method.
EU mercury limit tuna 2026 compliance checklist
- Lot defined by species, size grade, FAO area, and date window.
- 10–20 primary units sampled per lot. Composite homogenized, A/B retains kept.
- ISO/IEC 17025 lab selected. ICP-MS or DMA/CV-AAS with LOQ ≤0.01 mg/kg.
- COA includes LOQ and measurement uncertainty (k ≈ 2, 95% CI).
- Decision rules documented for 0.85–1.05 mg/kg results.
- CoC, photos, and temp logs filed with each lot.
If you’re building EU programs around premium tuna lines like Bigeye Loin or Yellowfin Steak, this SOP will keep you out of trouble and speed up port clearance. Want a second set of eyes on your sampling or COA format? Call us. Or if you’re exploring broader assortments beyond tuna, you can also View our products and we’ll advise which items are best suited for strict EU programs.
The reality is simple. When your lot definition, sampling, and COA are tight, mercury stops being a gamble and becomes a checkbox. That’s where we like to operate.