Cumulative Fuel and CO₂ Impact of Emissions-Based Maintenance Across a Decade of Heavy-Duty Diesel Field Deployments

Aggregate Savings, Avoided Emissions, and Economic Impact from a Diagnostic Platform Operating Across Three Continents

White Paper — Cumulative Impact Analysis · June 2026

Jeramiah Forbush, CPA, CGMA — President & Co-Founder, Revealiency, an operating name of Emissions-Based Maintenance, LLC, a wholly owned subsidiary of Starboard Investment, LLC

Steve Forbush — Lead Technical Engineer & Co-Founder, Revealiency, an operating name of Emissions-Based Maintenance, LLC, a wholly owned subsidiary of Starboard Investment, LLC

Published by Starboard Research — a segment of Starboard Investment, LLC

Abstract

This paper aggregates field results from a decade of emissions-based maintenance (EBM) deployments across heavy-duty diesel fleets in mining, transit, and government applications. Across three continents and thousands of diesel assets, the platform has captured 3,058 emission tests, with prescriptive maintenance interventions documented at the per-engine level.

We report cumulative results across three measurement tiers, presented separately and never blended. The realized, workpaper-backed foundation (the 2011–2019 fleet, 76 vehicles) accounts for approximately 76.4 million gallons of diesel, approximately 1.71 billion pounds (776,000 metric tons) of CO₂ avoided, and approximately $247.6 million in workpaper-backed economic impact. Including platform-era identified pilots, total identified cumulative impact is approximately 78.2 million gallons, approximately 795,000 metric tons of CO₂, and approximately $255.0 million — calculated using the carbon balance methodology described in the companion technical paper (Version 1.0, June 2026). The results are reported at the deployment-class level using the regulatory-endorsed CO₂-to-fuel proportionality codified in 40 CFR Part 1065, ISO 8178, and SAE J1003. We discuss the durability of intervention effects (the persistence of CO₂ reduction after maintenance), the relationship between intervention frequency and cumulative impact, and the structural reason the impact is large: ECM-based fleet telematics systematically miss the per-engine deviations that emissions-based monitoring detects.

We also describe the methodology, anonymization treatment, and assumptions used in this aggregation, so that fleet operators and carbon market participants can assess applicability to their own fleets.

Keywords: emissions-based maintenance; diesel fleet decarbonization; CO₂ avoided; fuel savings; mining haul trucks; carbon balance method; aggregate impact analysis.

1. Introduction and Scope

Over the last decade, an emissions-based maintenance methodology has been deployed in heavy-duty diesel fleets across three continents. This paper reports the aggregate fuel and CO₂ impact of that work — not as a marketing claim, but as a calculation rooted in measured exhaust CO₂ concentrations, the regulatory-endorsed carbon balance method, and the engineering ratio between pre-intervention and post-intervention emission profiles.

The companion paper (“Exhaust CO₂-Based Fuel Consumption Derivation for Diesel Fleet Maintenance Diagnostics,” Version 1.0, June 2026) establishes the methodology by which exhaust CO₂ readings yield per-engine fuel consumption rates. This paper extends that methodology from the per-engine level to the deployment-class and portfolio-wide level. The arithmetic is straightforward: each maintenance intervention produces a measured CO₂ reduction; that CO₂ reduction, multiplied by the engine’s operating hours and rated fuel rate, yields an avoided-fuel quantity; aggregating across engines and over time yields the cumulative impact.

The aggregation is meaningful because the underlying per-engine measurements are anchored to certified engine baselines (the EPA Engine Universe database covering 24,079 EPA-certified engine families across 200+ manufacturers) and validated against fill-to-fill physical measurement at one of the deployment sites (Site D, Central Asia; see companion paper Section 7.2 and Section 4 of this paper). The aggregate numbers are not a projection — they are a roll-up of measured per-engine results.

Scope of this paper. We restrict the analysis to interventions for which (a) pre-intervention and post-intervention CO₂ readings are captured at comparable RPM and load windows, (b) the engine is matched to a certified EPA baseline (or equivalent OEM-published baseline for non-U.S.-certified engines), and (c) post-intervention operation is monitored long enough to assess durability of the effect (≥1 emission test cycle, typically 30–90 days).

What this paper is not. This paper does not propose new methodology, does not present new IP, and does not advocate for any specific carbon credit registry treatment. It reports aggregate results from an existing methodology, in a form intended to be useful for fleet operators evaluating emissions-based maintenance and for carbon market participants assessing measurement-and-verification (M&V) approaches.

2. Aggregation Methodology

2.1 Per-engine impact formula

For each intervention, the per-engine annual fuel savings (gal/year) are computed as:

Fuel Saved (gal/yr) = (CO₂_before − CO₂_after) / CO₂_before × Rated Fuel Rate (gal/hr) × Operating Hours (hr/yr)

This expression follows directly from the stoichiometric proportionality established in the companion paper Section 2.2. The CO₂ ratio at constant load and RPM equals the fuel-consumption ratio.

The CO₂ emissions avoided are then:

CO₂ Avoided (lbs/yr) = Fuel Saved (gal/yr) × 22.4 lbs CO₂/gal

Using the U.S. DOE / EIA conversion factor of 22.4 pounds CO₂ per gallon of diesel combusted (40 CFR §1065.655 Table 2 carbon mass fraction; see companion paper Table 2).

Economic impact is computed using deployment-specific diesel prices at the time of intervention, sourced from EIA regional retail diesel price data and adjusted for off-highway / bulk-purchase rates where the deployment context warranted it.

2.2 Aggregation across engines and time

Engine-level results are summed across all engines in a deployment to produce a deployment-class subtotal. Deployment-class subtotals are summed across all deployments to produce the portfolio-wide total. Operating hours are integrated over the period between intervention and the most recent measurement (or the end of the reporting period, whichever is earlier). This treatment is conservative — it credits the methodology only with savings that have been observed, not with projected future savings from sustained operation of a maintained engine.

2.3 Durability of intervention effects

A maintenance intervention that reduces CO₂ from 11.4% to 8.1% (the worked example in the companion paper Section 6.3) does not yield infinite savings if the engine subsequently degrades back to its pre-intervention state. We address this by:

• Capturing post-intervention CO₂ at multiple time points where data is available, and using the highest (most degraded) reading for the savings calculation. This yields a conservative estimate.

• Truncating the savings calculation at the next observed intervention, or at the most recent measurement, whichever comes first. We do not extrapolate savings into periods where the engine’s state is unmeasured.

• Documenting cases where post-intervention degradation was observed and required re-intervention (see Site D Truck D-4 in the companion paper Section 7.2.3). Where re-intervention occurred, savings are computed segment-by-segment with the appropriate baseline for each segment.

2.4 Anonymization

Deployment sites are referenced by anonymized labels (Site A through Site J) and described at continent-level geography. Engine classes are referenced by displacement / cylinder count / emission tier rather than OEM model designation, except where the OEM’s public-domain certification data is the cited source. Per-engine results within a deployment are reported only in aggregate — individual engine identifiers are not disclosed. These anonymization choices follow customer confidentiality agreements and the standard practice for aggregate industry impact reporting.

3. Source Data: Deployment Portfolio

The aggregate results in this paper draw from ten deployment classes, summarized below. The portfolio totals approximately thousands of diesel assets with 3,058 emission tests captured.

SITE	LOCATION / APPLICATION	ENGINES	TESTS (RESTATED)	TRACK	ENGINE CLASS (DESCRIPTOR)
A	Open-pit commodity mine operation, Western U.S.	64	1,190	manual	78L / 85L V16, Tier 2
B	Open-pit commodity mine operation, northern Africa	14	18	manual	50L / 78L V16
C	Open-pit commodity mine operation, Western U.S.	5	5	manual	60L V16
D	Open-pit commodity mine operation, Central Asia	5	17	manual	60L V16, Tier 1
E	Commuter rail, Western U.S.	16	16	manual	locomotive 2-stroke V16
F	Government/defense fleet (tech demo)	1	1	manual	genset — single-day tech demo
G	Open-pit commodity mine operation, Western U.S.	5	10	manual	75L V16
H	Open-pit commodity mine operation, Western U.S.	3	81	manual	mixed fleet
I	Open-pit commodity mine operation, Eastern U.S.	10	114	manual	60L V16
J	Underground commodity mine operation, Western U.S.	1	1,606	telematic	~52L (telematic stream)
TOTAL	three continents	124	3,058		83 build-artifact rows excluded

Table 1: Deployment portfolio summary. Site labels match the companion paper for cross-reference. Detailed per-deployment results follow in Section 5.

4. Ground-Truth Calibration: Site D Three-Way Validation

Before reporting aggregate numbers, it is appropriate to address the question: how confident can we be that the carbon-balance-derived fuel rates that underlie those aggregates reflect actual fuel consumption?

The companion paper Section 7.2 presents a controlled three-way comparison at Site D (Central Asia), where five class-240-ton haul trucks powered by 60-liter, V16, Tier 1 engines were subject to (a) CO₂-derived fuel rate calculation, (b) supervised fill-to-fill volumetric fuel measurement, and (c) ECM-reported fuel rate via J1939 SPN 183. Fill-to-fill measurement is the regulatory gold standard for in-service fuel consumption verification; the engine OEM’s field engineering team supervised the measurement program.

Summarizing the companion paper’s Table 5 for context here:

• CO₂-derived fuel rates tracked fill-to-fill measurement within the expected range for trucks that received maintenance interventions.

• ECM-reported fuel rates overstated actual consumption by up to 17%, with one engine (Truck D-3) showing a 20.5% overstatement.

• This is not an outlier finding. The pattern repeated across all five engines in the controlled test, and the structural reason (ECM measures fuel-to-injector, not fuel-burned; see companion paper Section 5.1) is universal across modern diesel return-flow fuel systems.

The implication for this paper is twofold. First, the CO₂-derived aggregate is the appropriate basis for impact accounting — it measures the carbon mass that was actually oxidized, which is the carbon mass that actually entered the atmosphere. Second, fleet operators relying on ECM telematics for fuel and emissions accounting are systematically over-estimating fuel consumption and, depending on the application, may be both under-counting savings opportunities and over-counting baseline emissions.

5. Aggregate Impact Results

5.1 Portfolio-wide totals

Aggregated across the deployment classes over the full reporting period, the results are:

DEPLOYMENT CLASS	TIER	VEHICLES	FUEL SAVED (GAL)	CO₂ AVOIDED (LBS)	ECONOMIC IMPACT (USD)
A — foundation fleet (2011–2019)	realized / workpaper-backed	76	~76,374,994	~1,710,799,873	~$247,581,037
B — commodity mine, northern Africa	identified (partial)	13	~951,630	~21,316,512	~$3,806,520
G — commodity mine, Western U.S.	identified	5	~689,511	~15,444,646	~$2,758,044
E — commuter rail	identified	19	~102,391	~2,293,559	~$409,564
C — commodity mine, Western U.S. (demo)	identified	5	~83,036	~1,860,012	~$332,144
D — commodity mine, Central Asia	identified (partial)	5	~40,596	~909,350	~$162,384
Total identified		123	~78,242,158	~1,752,623,952	~$255,049,693

Table 2: Aggregate impact by deployment class, reconciled to the validated record. The realized / workpaper-backed foundation fleet (A, 2011–2019) is reported separately from platform-era identified pilots (B, C, D, E, G), which are findings not yet fully abated. Economic impact uses deployment-specific diesel pricing at time of intervention. Sites C and F are single-day demonstrations; F carries no abatement and is excluded.

Portfolio totals (identified): approximately 78.2 million gallons of diesel fuel avoided, approximately 1.75 billion pounds (795,000 metric tons) of CO₂ emissions avoided, and approximately $255.0 million in cumulative identified economic impact — of which the realized / workpaper-backed foundation fleet accounts for ~76.4 million gallons, ~776,000 metric tons, and ~$247.6 million.

These numbers are reported to the nearest material precision. They are not predictive claims about future deployments — they are a retrospective accounting of measured intervention effects, calculated using the regulatory carbon balance method on per-engine emission profiles captured across the deployment portfolio.

5.2 Per-deployment commentary

The portfolio total is dominated by one deployment. Roughly 97% of the realized ~$247.6M / ~76.4M gallons comes from Site A alone; every other site below is a small identified pilot — a finding not yet fully abated — not a realized, workpaper-backed result. The impact is not spread evenly across fleets, and the two tiers are never blended.

Site A (open-pit commodity mine operation, Western U.S.) — realized / workpaper-backed. The foundation fleet and, by a wide margin, the dominant contributor to the portfolio. 76 vehicles powered by 78-liter / 85-liter V16 engines (CAT C175-16 / 3516C class), with 64 engines and 1,190 emission tests in the validated corpus across a multi-year operating window. Median per-engine CO₂ reduction of approximately 12% is modest at the engine level, but compounding across a large fleet operating thousands of hours per year per engine produces the largest aggregate impact in the portfolio. The deployment also demonstrates the durability question: engines returned to within-tolerance operation generally stayed within tolerance under periodic re-monitoring, and the small fraction that drifted out were caught before consuming significant excess fuel.

Site B (open-pit commodity mine operation, northern Africa) — identified (3 of 13 implemented). The highest per-engine percentage in the portfolio: a median of approximately 22%, with a single unit exceeding 29% under constant-load, upper-bound conditions as documented in the companion paper Section 6.3 — an upper bound on one engine, not a fleet-wide rate. The commercial significance is that these were brand-new, from-factory engines significantly out of emission specification — not units degraded by service life — which establishes a separate use case in engine commissioning and acceptance testing. Identified savings of approximately 951,630 gallons / ~$3,806,520 across the 13-vehicle deployment.

Site C (commodity mine, Western U.S. — demonstration) — identified (demo). A single-day demonstration of 5 vehicles (5 tests), not a sustained deployment. Identified savings of approximately 83,036 gallons / ~$332,144. Reported for completeness and included in identified totals, but flagged as a demonstration rather than an abatement engagement.

Site D (open-pit commodity mine operation, Central Asia) — identified (3 of 5 implemented). Small in absolute terms — approximately 40,596 gallons / ~$162,384 — but methodologically pivotal: this is the three-way ground-truth fill-to-fill validation site discussed in Section 4 of this paper and in the companion paper Section 7.2. The figures reflect a small fleet, one engine already within tolerance at baseline, and one engine that experienced post-intervention degradation requiring re-monitoring.

Site E (commuter rail, Western U.S.) — identified. A 19-vehicle transit fleet operating at fixed schedules and routes, with a median per-engine reduction of approximately 6% reflecting the operating discipline of a scheduled, typically better-maintained fleet. Identified savings of approximately 102,391 gallons / ~$409,564. The deployment demonstrates applicability beyond the off-highway mining context in which the methodology was originally developed.

Site G (open-pit commodity mine operation, Western U.S.) — identified. An identified pilot across 5 vehicles: approximately 689,511 gallons / ~$2,758,044. Included in the identified portfolio totals.

Site F (government / defense fleet) — excluded from impact totals. The application context is fleet readiness rather than fuel-cost optimization, and the deployment produced no abatement intervention; it is therefore excluded from the portfolio impact totals and described only at this level of generality. It is retained in the corpus as a readiness use-case.

Across the identified pilots (B, C, D, E, G), total identified savings are approximately $7.47M on ~1.87M gallons; the realized, workpaper-backed foundation (Site A) accounts for the remaining ~$247.6M on ~76.4M gallons. The two tiers are reported separately and are not blended.

5.3 Sensitivity and uncertainty

The aggregate totals carry the uncertainties inherent in any deployment-portfolio aggregation:

• Per-engine CO₂ measurement uncertainty is on the order of ±0.1–0.2 percentage points based on portable gas analyzer specifications (NDIR for CO₂, electrochemical for O₂); this propagates to a per-intervention savings uncertainty of approximately ±3–5% of the reported figure.

• Operating-hour estimates are sourced from customer-reported fleet utilization where available, and from typical-utilization assumptions for the engine class otherwise. Operating-hour uncertainty is the dominant uncertainty term in deployments where ECM hourmeter data is incomplete or unverified.

• Diesel price assumptions use regional retail/bulk pricing at the time of intervention. For deployments spanning multiple years, the price-weighted economic impact is sensitive to the timing of intervention vintages within the reporting window.

• Durability assumptions are conservative as noted in Section 2.3. To the extent that maintained engines continue to perform within tolerance beyond the most recent measurement, the reported impact understates the true cumulative impact.

Net of these factors, we view the headline numbers as accurate to within approximately ±10–15% at the portfolio level, with site-level numbers carrying somewhat wider uncertainty bands. The directionality is highly robust: the carbon balance method is the regulatory gold standard, and the per-engine measurements are anchored to certified engine baselines.

6. Contextualization

6.1 What the portfolio totals mean

To contextualize the portfolio totals against U.S. transportation sector emissions: the U.S. transportation sector produced approximately 455 million metric tons of CO₂ from diesel consumption in 2023. The realized foundation of approximately 776,000 metric tons of CO₂ avoided — with total identified impact of approximately 795,000 metric tons — represents on the order of 0.17% of one year’s U.S. transportation diesel emissions, from a single methodology applied across three continents.

Extrapolating from this aggregate to the global heavy-duty diesel fleet (estimated 30+ million engines worldwide in mining, construction, transit, marine, power generation, and on-road commercial applications), the addressable annual impact of systematic emissions-based maintenance is multiple orders of magnitude larger than what has been demonstrated to date.

6.2 Comparison to alternative emission-reduction strategies

Three broad strategies are available for reducing diesel fleet emissions:

• Fleet replacement with lower-emission engines or alternative powertrains. Capital-intensive, multi-year transition timelines, dependent on availability and cost of the replacement technology. Effective but slow and expensive.

• Fuel-side substitution (renewable diesel, biodiesel blends). Lower carbon intensity per gallon of fuel, but requires no changes to engine maintenance state — a poorly-maintained engine on renewable diesel still consumes excess fuel.

• Operational efficiency improvements, including emissions-based maintenance. Lower capital cost, shorter implementation timeline, applicable to the existing in-service fleet without replacement. The portfolio totals reported here demonstrate the effectiveness of this category.

These strategies are not substitutes — they are complementary. A fleet transitioning to renewable diesel still benefits from emissions-based maintenance (the maintenance-state improvement compounds with the fuel-carbon-intensity improvement). A fleet planning eventual electrification still benefits from operational efficiency improvements in the years before the transition completes.

6.3 Implications for carbon market participation

The portfolio totals translate directly into the units used in voluntary and compliance carbon markets. Only the realized tier is a credit candidate. The realized, workpaper-backed foundation corresponds to approximately 776,000 metric tons CO₂ — on the order of 776,000 potential credits at one credit per metric ton, subject to applicable registry methodology requirements and additionality / permanence / leakage analysis under the relevant standard (Verra VCS, Gold Standard, Climate Action Reserve, or comparable jurisdictional frameworks). The additional identified pipeline (bringing total identified to approximately 795,000 metric tons) is not yet credit-eligible, as those reductions are identified but not yet abated and verified.

We do not advocate for any specific registry treatment in this paper. We note that the underlying measurement-and-verification approach — measured pre-intervention and post-intervention CO₂ at the per-engine level, calculated using regulatory-endorsed carbon balance equations, anchored to certified engine baselines, and validated against fill-to-fill physical measurement — is structurally well-suited to the M&V rigor that high-integrity carbon registries require. Fleet operators and registry methodologists interested in evaluating applicability to specific protocols are referred to the companion technical paper for the methodology and to the authors for further discussion.

7. Limitations and Future Reporting

This paper reports retrospective aggregate impact. Several caveats and future-reporting directions are appropriate to note:

• The portfolio is biased toward off-highway mining applications (the mining deployments) where heavy-duty diesel engines operate at high load factors. Generalization to lower-load-factor applications (light-duty, intermittent-use) should be done with care.

• The reported numbers credit the methodology with the savings observed after each intervention. They do not credit it with savings prevented by avoided breakdowns or extended engine life — those effects are real but harder to quantify in a roll-up framework.

• Reporting at deployment-class anonymization (Sites A through J) limits external auditability. Future reporting under M&V-rigorous frameworks would benefit from per-engine and per-intervention disclosure, with operator consent and appropriate confidentiality protections.

• The portfolio reflects the methodology as deployed historically. Recent platform improvements — including continuous telematic monitoring using commodity O₂ sensors with stoichiometric back-calculation to CO₂, as described in the companion paper Section 6.4 — should improve both the granularity and the scale of future impact reporting.

Future work will include (a) per-vintage cohort analysis to characterize the time decay of intervention effects, (b) deployment of the methodology in additional geographies and application contexts, and (c) M&V-rigorous reporting in coordination with carbon registry methodologies as they evolve to address operational efficiency strategies for heavy-duty diesel.

8. Conclusions

Across a decade of deployments at thousands of heavy-duty diesel assets on three continents, emissions-based maintenance has produced:

• Realized, workpaper-backed (2011–2019 foundation fleet, 76 vehicles): approximately 76.4 million gallons of fuel avoided; approximately 776,000 metric tons of CO₂ avoided; approximately $247.6 million in workpaper-backed economic impact.

• Total identified (realized foundation plus platform-era pilots): approximately 78.2 million gallons; approximately 795,000 metric tons of CO₂; approximately $255.0 million in cumulative identified economic impact.

• These tiers are reported separately and never blended; projected/addressable figures appear only as declared-assumption models, not as realized results.

These results are calculated using the regulatory-endorsed carbon balance method codified in 40 CFR Part 1065, ISO 8178, and SAE J1003, with ground-truth validation against fill-to-fill physical fuel measurement at one of the deployment sites. They are anchored to a database of 24,079 EPA-certified engine families across 200+ manufacturers and reflect measured per-engine emission profile changes pre- and post-intervention.

The portfolio impact is large but the methodology is portable. The same approach applied to the global heavy-duty diesel fleet — conservatively, 30+ million engines worldwide — would yield additional impact multiple orders of magnitude beyond what has been demonstrated to date.

Operational efficiency strategies for heavy-duty diesel are complementary to fleet replacement and fuel substitution, not substitutes. As fleet decarbonization timelines remain measured in decades, the cumulative impact of in-service emissions-based maintenance over those decades is material to global transportation-sector emission trajectories.

References

• Companion technical paper: “Exhaust CO₂-Based Fuel Consumption Derivation for Diesel Fleet Maintenance Diagnostics,” Version 1.0, June 2026, Revealiency, an operating name of Emissions-Based Maintenance, LLC.

• U.S. Energy Information Administration (EIA). “Carbon Dioxide Emissions Coefficients by Fuel.” Available: https://www.eia.gov/environment/emissions/co2_vol_mass.php

• U.S. Environmental Protection Agency. 40 CFR §1065.655, “Carbon-based chemical balances of fuel, DEF, intake air, and exhaust.” Code of Federal Regulations, Title 40.

• U.S. Environmental Protection Agency. “Greenhouse Gas Equivalencies Calculator.” Available: https://www.epa.gov/energy/greenhouse-gas-equivalencies-calculator

• International Organization for Standardization. ISO 8178-1:2020 and ISO 8178-2:2021, “Reciprocating internal combustion engines — Exhaust emission measurement.”

• SAE International. SAE J1003, “Diesel Engine Emission Measurement Procedure.”

• Intergovernmental Panel on Climate Change (IPCC). 2006 IPCC Guidelines for National Greenhouse Gas Inventories. Volume 2: Energy.

• U.S. Energy Information Administration. Diesel retail and wholesale price datasets. Available: https://www.eia.gov/petroleum/gasdiesel/

• Verra, “Verified Carbon Standard (VCS) Program Methodology and Module Documents.” Available: https://verra.org/programs/verified-carbon-standard/

• Forbush, J. and Forbush, S. “Emissions-Based Maintenance System and Method.” U.S. Patent No. 10,718,284 B2. Issued July 21, 2020.

• Forbush, J. and Forbush, S. “Emissions-Based Maintenance System and Method.” Indian Patent No. 553,370. Granted October 2024.

Disclosure: The methodology referenced in this paper is implemented in a proprietary diagnostic platform protected under US Patent 10,718,284 B2 and Indian Patent 553,370, with additional applications pending. Aggregate results reflect the authors’ historical deployment work. Correspondence: jforbush@starboardresearch.org.

© 2026 Starboard Investment, LLC — Starboard Research, a segment of Starboard Investment, LLC