Safeguard Mechanism Trajectory Check: Knowing Before...

Most covered facilities find out they have breached the Safeguard baseline in October. The NGER report goes in, the numbers get totalled, and someone in the finance team asks why the ACCU line item in the FY close is suddenly seven figures.

That October discovery is nine months late. The reporting year ended 30 June. Every tonne over baseline between 1 July of the prior year and the day the report is finalised is a tonne that could have been avoided, deferred, or offset while credits were still cheap. Instead it becomes a scramble to source ACCUs before the 31 March surrender deadline in a market where every other exceeding facility is buying at the same time.

The reformed Safeguard Mechanism does not forgive this timing. Baselines are declining 4.9% per year for standard facilities and a minimum 1% per year for Trade Exposed Baseline Adjusted (TEBA) facilities. Under the NGER Act, failure to net out excess emissions with ACCUs or Safeguard Mechanism Credits (SMCs) exposes the operator to civil penalty units at $364 per unit (for offences from 1 July 2026), plus daily continuing penalty exposure while the breach persists. The operational cost of finding out late is usually worse than the penalty exposure itself, because sourcing ACCUs in a tight market between February and March is when spot prices spike.

The fix is not more discipline on the October report. It is a live projection against the declining baseline that runs off the same emissions ledger your NGER submission will eventually use. That is what the Safeguard Trajectory Check does.

Why the October discovery costs more than the exceedance itself

Consider a mid-sized cement facility with a fixed-decline baseline of 480,000 tCO2-e in FY2025-26. The 4.9% annual ratchet brings the FY2026-27 baseline to roughly 456,480 tCO2-e. If actual emissions come in at 490,000 tonnes, the facility is 33,520 tonnes over.

At an ACCU spot price of $36.50, that is about $1.22 million in surrender obligation. If the facility had spotted the trajectory in February and secured ACCUs at the pre-March-surrender-window price of around $32, the same position would cost $1.07 million. The delta is $150,000 that was left on the table purely because no one knew the number until October.

That is the small version. The larger cost is the actions that were never taken. A shutdown deferred by two weeks, a fuel switch scheduled for August instead of April, a boiler tune performed in May rather than November. Each of those decisions could have moved the year-end position by hundreds or thousands of tonnes if the trajectory was visible while the year was still in play. By October, the tonnes are locked in.

Related reading: Safeguard Mechanism Compliance: Tracking Baselines covers baseline tracking in more detail.

What the Trajectory Check actually does

The Trajectory Check is a service (safeguardTrajectoryCheck) that runs against every Safeguard-listed facility linked to the organisation. For each facility, it does three things.

First, it projects year-by-year emissions from the live monthly ledger, adjusted by any planned carbon actions (a boiler electrification, a shutdown, a fuel switch) that have start and completion dates in the plan. Actions that are cancelled are excluded. Actions that are partly through their reduction period contribute a fractional benefit for the current year.

Second, it computes the facility's declining baseline for each year in the plan window. The starting point is the facility's baselineEmissions and baselineYear, stored per facility. The decline rate is 4.9% for standard facilities and the TEBA rate (minimum 1%) for trade-exposed facilities, applied as a compound factor: baseline × (1 − rate)^(year − baselineYear).

Third, it compares projected emissions to the declining baseline for each year in the window and classifies the position:

compliant: projected emissions at or below the declining baseline
at_risk: projected emissions above baseline but within 5% of it
non_compliant: projected emissions more than 5% above baseline
unknown: insufficient data to project

The output is a per-facility, per-year row set with projectedEmissions, decliningBaseline, gap, and status. The facility rolls up to an overall status; the plan rolls up to a portfolio status. When status is worse than compliant, someone needs to act.

It runs on demand, per plan, rather than as a monthly export. Run it after documents are processed, a target is updated, or an action is added to a plan, and it reflects the new position.

Each facility's baseline, with its own decline maths

Safeguard baselines are facility-specific and the maths differs by baseline type. Getting this wrong at the storage layer is where a lot of spreadsheet-based tracking systems come apart.

The safeguardFacilities table stores baseline data per facility with the following fields that matter for the trajectory:

baselineEmissions: the reference baseline in tCO2-e
baselineYear: the year the baseline was set
baselineType: either fixed_decline or production_adjusted
declineRate: annual percentage decline (default 4.9%)
isTradeExposed and tebaDeclineRate: TEBA classification and its lower decline rate (minimum 1%)
productionVariables: per-product emissions intensity for production-adjusted facilities
annualProductionData: actual annual production quantities by product
mmpElected, mmpStartYear, mmpEndYear: Multi-Year Monitoring Period fields

For fixed-decline facilities, the math is straightforward: baseline compounds down 4.9% per year off the baseline year. For production-adjusted facilities, the baseline is recalculated each year as the sum across products of production quantity times emissions intensity times the compound decline factor. If production drops sharply, the baseline drops with it. If production rises, the baseline rises to match. This matters for facilities in cyclical industries where output swings 20-30% year on year.

What "trajectory" actually means in the calculation

Trajectory is not a straight-line extrapolation from year-to-date emissions to a year-end estimate. That method is what most spreadsheets do and it is misleading in almost every case.

The Carbonly Trajectory Check pulls three inputs. The current-year ledger gives the actuals-to-date. The forecasting service applies per-facility seasonality, so a plant that runs heating peaks in Q3 does not get its March number annualised as if the whole year looks like March. Planned operational changes registered against the facility (a new furnace commissioning in October, a two-week outage in November, a mine expansion phase starting in January) are folded in with their expected emissions delta.

The math is not clairvoyant. Facility restarts, unplanned outages, and grid factor changes still create surprises. But it moves the discovery window from October to February or March, and that is usually enough to matter.

Which abatement actions still move this year's position

If the trajectory is heading toward a breach, the Carbon Planning module surfaces the marginal abatement cost (MACC) curve for the facility. The MACC ranks abatement actions by cost per tonne, from cheapest to most expensive. Standard usage.

The trajectory-aware wrinkle is that the MACC is filtered for what can still land within the current reporting year. Rooftop solar with an 18-month payback and a June commissioning date contributes only four months of emission benefit against the FY2026-27 baseline. Electrifying two forklifts with an 8-month payback and an August commissioning date captures ten months of benefit. Both make sense on a NPV basis; only one moves the compliance position for the current year.

That distinction matters when a CFO is deciding between a $2 million capex on solar and $1.2 million of ACCU procurement. The trajectory-aware view says: solar makes strategic sense but does not fix this year's breach, so budget it as next-year capex and cover the current-year gap with a smaller intervention plus targeted ACCU purchases. That is a real capex conversation, not a "we need to reduce emissions" slide.

Related reading: Carbon Reduction Planning: Scenario Modelling Actions Before Committing covers how the scenario builder ties into the MACC ranking.

The SMC top-up option

If the operational actions cannot close the gap in time, Safeguard Mechanism Credits are the alternative. The system tracks SMC purchase records against the facility with credit vintage, price paid, and quantity, and applies them at the surrender calculation. The safeguardCalculation service handles the compliance position math: net position is surplus plus ACCUs surrendered plus SMCs banked, and the status classification uses that net figure.

Two limits to be aware of. The 30% ACCU threshold check flags facilities where ACCU surrender exceeds 30% of the baseline, which triggers a written justification requirement to the Clean Energy Regulator covering abatement projects attempted, technical barriers, and cost-benefit analysis. The system generates the justification template with the specific facility and year prefilled; the compliance team completes the substantive content.

Second limit worth stating openly: nothing about SMC procurement is done in the system. Sourcing credits, negotiating price, and settling with the counterparty happens outside Carbonly. What the system does is capture the procurement record (vintage, price, quantity, counterparty) and apply it correctly at surrender time with the audit trail intact.

A dated review on the calendar instead of an October fire alarm

The Trajectory Check surfaces on the Carbon Compliance Calendar as a scheduled review event. Sustainability lead sees "Safeguard trajectory review due 15 March" as a calendar entry rather than as a fire alarm at 6pm on 28 October.

Related reading: Autonomous NGER Submission Workflow covers how the same calendar drives the October NGER submission, and Methane Management Plans Under the Safeguard Mechanism covers the fugitive emissions layer that trajectory-covers for gas-heavy facilities.

What you can show the Regulator about the year's decisions

Writes to the emissions ledger are captured as typed audit events, and each record keeps its source document and emission factor version. Not because auditors ask for it in year one; because they may ask for it in year three when the pattern of decisions is what determines whether the operator has been managing to baseline in good faith or discovering breaches by accident.

When the Clean Energy Regulator or a limited-assurance auditor under ASSA 5010 asks how the reported emissions reconcile to the facility baseline, you can show the ledger record with its source document reference, the emission factor version pinned at calculation time, and the trajectory checks you ran during the year. Under AASB S2 paragraph 29 the same data supports the mandatory climate-related financial disclosures, so the Safeguard trajectory work is not a separate reporting stream. It is the same data with a compliance overlay.

The Evidence Pack export bundles the trajectory reports for the reporting year, the baseline determination, the action-plan projections, and the SMC and ACCU application records. Under NGER audit or Group 2 assurance, that pack lands with the auditor as one artefact.

Related reading: ASRS Assurance Requirements: What Your Auditor Will Actually Ask For and NGER Audit Preparation with the Evidence Pack cover the assurance side of the same evidence trail.

For consulting practices running Safeguard compliance for their industrial clients, the trajectory check is the leverage point. The consultant sets the baseline determination, configures the facility parameters (TEBA status, baseline type, MMP election), and reviews the flagged trajectories monthly. Review time goes to the facilities the check marks at risk or non-compliant, instead of a line-by-line ledger review of every facility every month.

Practical starting point

For a facility over the Safeguard threshold of 100,000 tCO2-e of covered (Scope 1) emissions a year, the sequence is straightforward.

Load the last three years of emissions ledger: utility bills, fuel dockets, refrigerant records, on-site combustion. The AI Document Engine handles the volume in weeks rather than months if the source documents live in email or a folder. Confirm the facility baseline as recorded in the Safeguard Rule determination. Enter the baseline type (fixed decline or production adjusted), the decline rate, and TEBA status. If MMP is elected, enter the start and end years.

Then run the Trajectory Check monthly for a quarter. The first month reveals the projection with wide error bars, because most of the input is model rather than observation. Every subsequent month refines it as more actuals land and the seasonality profile calibrates. By month three, the projection is defensible enough to drive a real conversation with the CFO about where the year is heading.

If the trajectory is flagging at_risk or non_compliant, the MACC filtered for within-year actions is the next step. If the gap cannot be closed operationally, SMC procurement records get loaded and applied.

For a Safeguard-covered facility, that sequence moves the breach discovery point from October to March. Seven months is enough to matter. Two weeks is not.

Carbonly pricing is per-project with a $100/month workspace minimum. To load a Safeguard-covered facility and run the Trajectory Check against your live ledger, email hello@carbonly.ai.