Methodological and Uncertainty-Focused Evaluation of Tiered Approaches for Maritime Black Carbon Inventories in the Philippines

Abstract

Black carbon (BC) is a short-lived climate pollutant with substantial warming and health impacts, yet its contribution from maritime activities in data-limited regions remains poorly constrained. This study conducts a methodological and uncertainty-focused evaluation of tier-based emission inventory approaches from the European Monitoring and Evaluation Programme/European Environment Agency (EMEP/EEA) Guidebook, examining fuel-based (Tier I) and activity-based (Tier III) methodologies using national fuel statistics, port call activity, vessel registry data, and an operational Philippine Coast Guard dataset. Monte Carlo uncertainty analysis, spatial mapping, and hotspot intensity analysis are applied to evaluate how each tier responds to data limitations and parameter uncertainty rather than to reconcile absolute emission magnitudes. Results indicate that Tier I provides scalability for national reporting but exhibits substantial uncertainty for gasoline-dominated segments due to reliance on particulate matter-based proxies, underscoring the role of Tier II as a targeted refinement option. Tier III applies an activity-based formulation using fuel consumption resolved by operational phase and phase-specific emission factors, consistent with EMEP/EEA Tier III guidance. These findings are integrated into a decision-oriented synthesis to support informed selection and combination of tiered emission approaches under data-limited maritime conditions aligned with national and international climate commitments.

1. Introduction

Black carbon (BC) is one of the potent short-lived climate pollutants (SLCPs) with strong radiative forcing, atmospheric warming effects, and documented impacts on air quality and human health. Maritime shipping has been recognized as one of the significant sources of BC, particularly in busy coastal and port regions where emissions influence both regional climate forcing and population exposure [1,2]. While climate mitigation strategies have historically emphasized long-lived greenhouse gases such as carbon dioxide (CO₂), increasing attention is now being directed towards SLCPs like BC due to their high near-term warming potential and the co-benefits associated with reducing particulate pollution [3,4].

Advances in BC assessment methods, including onboard instrument techniques, remote sensing, and non-instrumental inventory approaches, have improved the capacity to estimate emissions across diverse operational conditions [1,2]. Empirical studies including those synthesized by the International Council on Clean Transportation (ICCT), demonstrate that BC emissions are strongly influenced by engine load, fuel type, and aftertreatment performance, with emission factors generally decreasing at higher loads and reductions of 20–40% achievable through particulate control technologies [1,5]. These findings underscore the importance of accounting for operational behavior when quantifying BC emissions from ships.

The EMEP/EEA’s Emission Inventory Guidebook provides a tiered methodological structure (Tiers I–III) designed to accommodate varying levels of data availability, ranging from fuel-based national estimates to detailed activity-based calculations. Higher tier (Tier III) maritime black carbon inventories adopt activity-based formulations that resolve emissions by operational phase, typically underway, docking/undocking, and berthing, using fuel consumption and phase-specific emission factors, as defined in the EMEP/EEA’s emission inventory guidebook. This structure enables improved spatial and operational resolution where detailed activity data are available. However, practical evidence comparing how these tiers perform under real-world data constraints, particularly in terms of uncertainty, scalability, and policy relevance, remains limited, especially outside well-instrumented maritime regions [6,7]. As a result, many developing maritime nations rely on Tier I approaches despite recognized limitations for SLCP assessment, while the operational feasibility of higher-tier methods is often unclear [8].

These challenges are particularly pronounced in the Philippine maritime sector, which is characterized by a highly heterogeneous fleet that includes commercial and passenger vessels, small gasoline-powered craft, and government-operated assets such as those of the Philippine Coast Guard (PCG). National fuel statistics are reported in aggregated form, vessel-class specific fuel consumption data are limited, and technology-specific BC emission factors, particularly for gasoline engines below three gross tons, remain incomplete. Consequently, few studies have examined how different EMEP/EEA tiers behave within a single national context under such constraints, despite increasing regional interest in SLCP mitigation in Southeast Asia [4,9].

Based on these considerations, this study undertakes a methodological and uncertainty-focused evaluation of EMEP/EEA’s Tier I and Tier III approaches within the Philippine maritime context. Rather than presenting a new operational case application, the analysis evaluates how fuel-based and activity-based methods respond to data limitations by examining uncertainty behavior, spatial patterns, and hotspot sensitivity using nationally available datasets and imported Tier III operational datasets for PCG vessels. Tier II is not operationally applied but is considered as a targeted refinement option for gasoline-powered vessels in response to elevated uncertainty observed under Tier I.

To support transparent interpretation of the results, this study develops a decision-oriented synthesis, summarized in a tabular form as a decision-oriented matrix, which consolidates observed differences uncertainty behavior, diagnostic resolution, and spatial representation across tiers. This approach provides structured guidance for selecting and combining tiered methods for maritime BC inventories in data-limited settings. While grounded in the Philippine context, the findings are relevant to other archipelagic and developing maritime nations seeking to strengthen BC reporting and align inventory development with emerging national and international SLCP commitments [4,8,10].

Guided by this aim, the study pursues the following research objectives:

• To evaluate the scalability, uncertainty behavior, and applicability of EMEP/EEA’s Tier I and Tier III black carbon emission estimation methodologies under Philippine maritime data constraints.
• To assess the policy relevance and operational suitability of tiered methodologies for developing a preliminary national maritime BC inventory across domestic, government, and law enforcement fleets.
• To develop a decision-oriented synthesis that identifies appropriate tier selection and combination strategies under varying data, operational, and policy conditions.

2. Materials and Methods

2.1. Research and Analytical Process

The analytical workflow of this study as shown in Figure 1 was designed to evaluate how different EMEP/EEA tier-based methodologies behave under national-scale data constraints, with particular emphasis on uncertainty, scalability, and spatial interpretability rather than absolute emission magnitude.

The process begins with the definition of a comparative methodological objective: to assess the operational suitability of fuel-based (Tier I) and activity-based (Tier III) approaches for maritime black carbon (BC) inventories in the Philippine context. Parallel data inputs representing the minimum information typically available in data-limited maritime contexts were integrated to support fuel-based emission estimation and comparative processing analyses.

Pre-processing consists of two distinct steps. First, Tier I fuel consumption data reported in million tonnes of oil equivalent (MTOE) were converted to metric tonnes using standard international energy conversion factors from the Organization for Economic Co-operation and Development/International Energy Agency (OECD/IEA) energy statistics manuals [11]. Second, Tier I BC emissions were estimated using fuel-specific emission factors and BC mass fractions consistent with EMEP/EEA’s guidance, including PM_2.5-based BC proxies for gasoline due to the absence of dedicated direct BC emission factors for small gasoline-powered vessels [9,12]. Detailed formulations and parameter values are provided in Section 2.3. Tier III BC emissions were not recomputed in this study; instead, an existing operational dataset derived from a prior analysis of PCG vessels was imported to preserve its activity-based resolution and methodological integrity [13]. This dataset applies EMEP/EEA’s Tier III formulations that estimate emissions using fuel consumption resolved by operational phase and phase-specific emission factors. The imported Tier III emissions were used exclusively for comparative processing analysis, enabling a transparent comparison of Tier I and Tier III behavior without additional computational assumptions. Although vessel-specific engine load information underpins fuel consumption estimates in the original dataset, Tier III emissions are represented in this study using the standard EMEP/EEA’s fuel-by-phase formulation.

Three processing analyses were applied to both Tier I and Tier III outputs to evaluate methodological behavior under uncertainty and data constraints. First, Monte Carlo uncertainty analysis was used to characterize the propagation of uncertainty associated with fuel statistics, emission factors, BC fractions, and operational parameters. Uncertainty ranges were assigned based on parameter variability reported in the EMEP/EEA’s guidebook and prior maritime emission uncertainty studies, and repeated random sampling was performed to generate probabilistic emission distributions for each tier [6,9]. Second, spatial mapping was conducted to examine how differences in methodological structure influence the spatial distribution of BC emissions. Tier I emissions were spatially allocated using domestic port call intensity as a proxy for shipping activity, an approach commonly applied in activity-based maritime inventories where vessel-level fuel consumption data are unavailable [7,14]. Tier III emissions retained their original spatial resolution derived from vessel operations, allowing direct comparison of spatial specificity between tiers. Third, hotspot intensity analysis was applied to identify geographic areas and operational contexts associated with elevated BC emissions. This analysis supports the evaluation of each tier’s ability to inform localized mitigation and operational planning, consistent with previous studies examining emission variability under real-world shipping conditions [15].

The combined outputs of uncertainty characterization, spatial mapping, and hotspot intensity analysis enable a structured comparison of Tier I and Tier III methodological behavior, highlighting differences in uncertainty propagation, spatial resolution, and operational interpretability. These comparative insights are subsequently interpreted to identify methodological implications related to data requirements and scalability under national-scale constraints. Based on this interpretation, a decision-oriented synthesis is constructed to synthesize observed trade-offs and support informed tier selection and combination. Tier II was not implemented as an independent calculation but is discussed as a targeted refinement option for gasoline-powered vessels in response to elevated uncertainty observed under Tier I.

This structured workflow enables transparent evaluation of tiered methodologies and supports informed selection and combination of approaches for developing maritime BC inventories in data-limited national contexts.

2.2. Overview of Data Sources

This study integrates multiple nationally available datasets to support a methodological and uncertainty-focused evaluation of EMEP/EEA’s Tier I and Tier III approaches for estimating maritime black carbon (BC) emissions under Philippine data constraints. The Tier I and Tier III datasets represent different fleet segments and are not intended to describe identical vessel populations. Tier I estimates are derived from aggregated national fuel consumption statistics encompassing the full domestic maritime fleet, including commercial, passenger, fishing vessels and small gasoline-powered craft. In contrast, the Tier III dataset represents a subset of government-operated Philippine Coast Guard vessels with documented operational activity and fuel use resolved by operational phase. Rather than constructing a fully instrumented activity-based inventory, the selected datasets shown in

Table 1. Overview of data sources, inputs, and analytical roles used in the maritime black carbon assessment.

Data Sources	Data Type	Analytical Role in This Study	Tier Association
Department of Energy Energy Situationer *	Fuel Statistics	Fuel-based BC estimation and uncertainty analysis	Tier I
Philippine Ports Authority (PPA) Port Calls *	Activity proxy	Spatial allocation of fuel-based emissions	Tier I
Maritime Industry Authority (MARINA), MARINA Registry *	Fleet composition	Vessel category characterization	Tier I
IMARC Fuel Market Report *	Fuel distribution	Proxy allocation by vessel class	Tier I
PCG Tier III Dataset Imported activity-based emission estimates (not recomputed in this study) *	Operational Activity	Comparative uncertainty, spatial mapping, and hotspot intensity analysis	Tier III
EMEP/EEA Guidebook *	Methodological reference	Emission factors, BC fractions, uncertainty ranges	Tier I (applied) Tier II (contextual guidance) Tier III (methodological reference)
International Energy Agency (IEA); United Nations Statistics Division (UNSD); and Organization for Economic Co-operation and Development/International Energy Agency (OECD/IEA)	Conversion factors	Standardized unit conversion for fuel-based emission calculations	Tier I

* Overview of data sources, input types, and analytical roles used to support the methodological and uncertainty-focused evaluation of EMEP/EEA’s tiered approaches for maritime black carbon estimation in the Philippines.

represent the minimum information typically accessible in data-limited maritime contexts.

National fuel consumption statistics by fuel type (gasoline, marine diesel oil, and heavy fuel oil) were obtained from the Philippine Department of Energy (DOE) Energy Situationer [16]. These data provided the primary input for Tier I fuel-based emission estimation and uncertainty characterization. International energy conversion references from the OECD/IEA and related sources were used to convert reported energy units into metric fuel mass for emission calculations [11]. Domestic maritime activity was represented using port call statistics from the PPA [14], which were applied as a spatial allocation proxy for Tier I emissions in the absence of vessel-level fuel consumption data. Vessel registry information from MARINA [17] was used to characterize fleet composition and vessel categories relevant to national shipping activity. Market-based fuel distribution information from IMARC Group fuel market reports [18] was incorporated to address gaps in category-specific fuel use not resolved in national energy statistics. Tier III emissions were represented using an existing operational dataset for PCG vessels derived from prior activity-based analysis [13]. This dataset includes fuel consumption resolved by operational phase (underway, docking/undocking, and berthing). Tier III emissions were not recomputed in this study; instead, the dataset was imported to preserve its original activity-based resolution and was used exclusively for comparative processing analyses, including uncertainty characterization, spatial mapping and hotspot intensity analysis. Methodological guidance, emission factors, BC mass fractions, and uncertainty ranges were sourced from the EMEP/EEA air pollutant emission inventory guidebook [9]. These references provided the basis for Tier I emission estimation, uncertainty assignment, and consistency across analytical steps.

Together, these datasets support a transparent comparison of how fuel-based and activity-based tiered methodologies behave under realistic national data constraints, forming the foundation for the comparative processing analysis and the development of a decision-oriented matrix.

2.3. Tiered Emission Estimation Methods

2.3.1. Tier I Fuel-Based Emission Estimation

Tier I BC emissions of marine diesel oil and heavy fuel oil were estimated using a fuel-based approach consistent with the EMEP/EEA’s air pollutant emission inventory guidebook for navigation [9].

$$E_i={\sum }_m({FC}_m\times {EF}_{i,m})$$

(1)

where

$E_i$ is the total black carbon emissions (t),

${FC}_m$ is the mass of fuel type m sold in the country for navigation (t),

${EF}_{i,m}$ is the emission factor by pollutant i and fuel type m (kg/tonne), and

m is the fuel type (marine diesel oil, gasoline, heavy fuel oil).

Prior Tier I emission estimation on this study, national fuel consumption data reported in MTOE referred to in this study were converted to fuel mass using standard international energy conversion factors [11]. This converted fuel mass was utilized to represent ${FC}_m$ in the Tier I fuel-based approach.

$${Mass}_{fuel}={\displaystyle \frac{{Energy}_{MTOE}\times \left(41.868\times { 10}^6\frac{GJ}{MTOE}\right)}{{Calorific Value}_{fuel}\left(\frac{GJ}{tonne}\right)}}$$

(2)

where

${Mass}_{fuel}$ is the estimated mass of fuel consumed (tonnes),

${Energy}_{MTOE}$ is the reported national fuel consumption expressed in million tonnes of oil equivalent (MTOE),

$41.868 \times { 10}^6 {\displaystyle \frac{GJ}{MTOE}}$ is the standard energy conversion factor from MTOE to gigajoules, and ${Calorific Value}_{fuel}$ is the lower heating value of the specific fuel type, applied separately for gasoline, marine diesel oil, and heavy fuel oil.

For gasoline-powered vessels, BC emission factors were derived using PM_2.5-based proxies and BC fractions due to the absence of dedicated BC emission factors for small gasoline marine engines in the EMEP/EEA guidebook [9,12]. The algorithm used was as follows:

$$E_{BC}= {\sum }_f({FC}_f \times {EF}_{f }\times {BC}_{frac,f })$$

(3)

where

$E_{BC}$ is the total black carbon emissions (t),

${FC}_f$ is the fuel consumption by fuel type ƒ (t),

${EF}_f$ is the particulate matter (PM_2.5) emission factor, and

${BC}_{frac,f}$ is the BC mass fraction of PM_2.5.

This methodological limitation contributes to elevated uncertainty for gasoline-dominated segments and motivates the consideration of Tier II as a targeted refinement option.

2.3.2. Tier III Activity-Based Emission Estimation (Imported Dataset)

Tier III black carbon (BC) emissions were not recomputed in this study. Instead, an existing operational emissions dataset for PCG vessels derived from prior activity-based analysis was imported [13]. The original Tier III calculations followed the EMEP/EEA’s navigation algorithm, in which emissions are estimated on a trip-by-trip basis and resolved by operational phase (underway, docking/undocking, and berthing):

$$E_{TRIP}= E_{UNDERWAY}+ E_{BERTHING}+ E_{DOCKING/UNDOCKING}$$

(4)

For each trip, emissions of pollutant i were computed as follows:

$$E_{Trip, i,j,m}= {\sum }_m({FC}_{j,m,p} \times {EF}_{i,j,m,p})$$

(5)

where

${FC}_{j,m,p}$ represents fuel consumption by engine type j, fuel type m, and operational phase p, and

${EF}_{i,j,m,p}$ is the corresponding emission factor as defined in EMEP/EEA’s Guidebook [9].

The imported PCG Tier III dataset includes vessel-specific fuel consumption by operational phase, engine type, and fuel type, consistent with this formulation. To preserve the original activity-based resolution and methodological integrity, Tier III emissions were used exclusively for comparative processing analysis, including uncertainty characterization, spatial mapping, and hotspot intensity analysis, and were not recalculated within this study.

2.4. Uncertainty Analysis, Spatial Mapping, and Hotspot Identification

2.4.1. Monte Carlo Uncertainty Analysis

Uncertainty in black carbon emission estimates was quantified using Monte Carlo uncertainty analysis applied separately to Tier I fuel-based emissions and imported Tier III activity-based emissions. This approach is widely used in emission inventory studies to propagate uncertainty in input parameters through emission calculation equations and is consistent with recommendations in the EMEP/EEA air pollutant emission inventory guidebook [9].

For Tier I, uncertainty propagation accounted for variability in national fuel consumption conversions factors, emission factors, BC mass fractions, and fuel allocation proxies. Probability distributions were assigned to each uncertain input parameter based on uncertainty ranges reported in the EMEP/EEA guidebook and supporting maritime emission literature [9,15,19,20]. Random sampling was performed using 10,000 iterations to generate probabilistic distributions of total of BC emission estimates by fuel type.

For Tier III, Monte Carlo uncertainty analysis was applied to the imported PCG operational emissions dataset to characterize variability associated with operational parameters, including fuel consumption by operational phase, engine type, and activity duration. Tier III emissions were not recalculated in this study; instead, uncertainty propagation was conducted on the existing activity-resolved emission outputs to ensure consistency in uncertainty characterization across tiers [13].

2.4.2. Probability Density Functions and Quantile-Based Uncertainty Metrics

Uncertainty results were represented using probability density functions and corresponding quantile statistics. For each emission estimate, the 5th percentile (P05), median (P50), and 95th percentile (P95) were extracted to describe the range of probable outcomes. Quantile-based uncertainty metrics are commonly used in emission inventory analysis to summarize probabilistic outputs where input uncertainties are non-Gaussian or asymmetrically distributed [9,19].

Probability density curves were generated to visualize the distribution of simulated emission values for each fuel category (Tier I) and operational phase (Tier III). For Tier III visualizations, emissions were displayed on a logarithmic scale to improve interpretability of distributions spanning multiple orders of magnitude. Quantile markers were extracted to enable comparison of uncertainty ranges across tiers without reliance on single-point estimates.

2.4.3. Spatial Mapping Allocation of Emissions

Spatial analysis was conducted to examine how different emission estimation approaches translate national or activity-based data into geographic distributions. For Tier I, national fuel-based emissions were spatially allocated using domestic port call intensity as a proxy for shipping activity, consistent with established approaches used when vessel-level fuel consumption data are unavailable [7,14]. Emissions were distributed proportionally across port locations based on reported port call frequency.

Tier III emissions retained their original spatial resolution derived from PCG operational records. These data reflect actual vessel activity locations as reported in operational logs and were used without additional spatial redistribution.

2.4.4. Hotspot Identification Procedure

Hotspot identification was applied to both datasets to identify geographic areas associated with elevated emission intensity or operational concentration. For Tier I, hotspot intensity reflects concentrations of spatially allocated emissions associated with high port call frequency. For Tier III, hotspot intensity reflects concentrations of operational deployment frequency derived from activity-based records. Hotspot identification was performed using relative intensity ranking rather than absolute thresholds, enabling consistent comparison of spatial concentration across datasets with differing resolution and methodological structure [15,21].

3. Results

3.1. Fuel-Dependent Structure of Tier I Black Carbon Estimates

The Tier I results in Figure 2 reveal an apparent fuel-dependent emission structure that reflects methodological sensitivity rather than proportional combustion behavior. Gasoline contributes a dominant share of national Tier I black carbon estimates despite its lower energy density, indicating the strong influence of proxy-based emission factor construction.

The prominence of gasoline within Tier I results reflects both fleet structure and methodological constraints rather than verified emission intensity. National gasoline consumption is reported only in aggregated form and largely represents numerous small recreational and auxiliary craft registered under MARINA, for which vessel-level gross tonnage and engine power data are not systematically available. Consequently, Tier I gasoline emissions capture the combined contribution of a number of small, heterogenous vessels using proxy emission factors, based on an EMEP/EEA Tier I approach, combined with a fixed black carbon fraction [9], rather than emission attributable to specific ship classes or engine characteristics, which amplifies sensitivity to parameter assumptions relative to fuels with direct BC emission factors [9].

Market-based fuel distribution data were used to contextualize gasoline use across vessel categories; however, these sources do not resolve fuel consumption to individual recreational craft or engine specifications, reinforcing the reliance on Tier I proxy-based estimation for this segment.

In contrast, marine diesel oil (MDO) and heavy fuel oil (HFO) exhibit more moderate contributions, reflecting the application of fuel-specific BC emission factors that impose tighter bounds on estimated emissions. Previous studies have noted that the use of proxy-based BC derivation for gasoline can introduce systematic bias and elevated uncertainty compared to fuels with directly measured BC emission factors [1,2].

Overall, the Tier I fuel profiles indicate methodological sensitivity embedded within the emission factor structure rather than differences in operational combustion characteristics.

3.2. Operational Phase Differentiation in Tier III Emissions

Analysis of the imported PCG Tier III emission dataset reveals in Figure 3 a phase-resolved emission structure that differs fundamentally from fuel-aggregated Tier I estimates. Within this dataset, emissions are unevenly distributed across operational phases, with berthing contributing a disproportionately large share relative to underway and docking/undocking activities. Similar phase-dependent emission patterns of particularly elevated black carbon emissions during low-load engine operation have been documented in previous activity-based shipping inventories and measurement studies [15,17].

The presence of distinct emission signatures across operational phases in the Tier III dataset highlights the informational content embedded in activity-resolved emission outputs. Prior studies have shown that activity-based approaches are capable of isolating emissions associated with operational duration and load conditions that remain obscured in fuel-based inventories [6,7].

These observed phase-resolved characteristics illustrate the type of diagnostic insight that activity-based Tier III inventories can provide when such datasets are available, even when they are not recalculated within the scope of the present study.

3.3. Diagnostic Uncertainty Behavior in Tier I Estimates

Uncertainty distribution associated with Tier I domestic black carbon estimates exhibit strong asymmetry across fuel types, indicating that uncertainty propagation is structurally embedded within the Tier I formulation. As shown in Figure 4, gasoline-related emissions display a broad and skewed uncertainty envelope, while MDO and HFO show comparatively narrower and more centrally clustered distributions.

This divergence demonstrates that Tier I uncertainty is not a uniform scaling property applied across fuels but arises from the interaction between fuel type and emission factor construction. For gasoline, uncertainty compounds across multiple parameters, including PM_2.5 emission factors, BC mass fractions, and fuel allocation assumptions, consistent with uncertainty behavior described in inventory guidance and prior assessments [3,9]. In contrast, fuels with direct BC emission factors propagate uncertainty through fewer parameters, resulting in more constrained distributions.

These results indicate that Tier I inventories may yield relatively stable central estimates for certain fuels while simultaneously producing highly uncertain outcomes for others within the same national inventory. This fuel-selective uncertainty behavior limits interpretability at finer categorical resolution and underscores the importance of uncertainty-aware interpretation when Tier I methods are applied to heterogeneous fuel mixes.

3.4. Constrained Uncertainty Structure in Tier III Emissions

Tier III distributions, shown in Figure 5, exhibit markedly reduced dispersion relative to Tier I, even when emissions are resolved across distinct operational phases. The narrower uncertainty envelopes observed for berthing, underway, and docking/undocking activities indicate that activity-resolved inputs constrain parameter variability during uncertainty propagation.

Notably, uncertainty compression is observed despite substantial variation in emission magnitude across operational phases. This pattern suggests that reduced uncertainty in Tier III does not arise from aggregation effects but from the structural resolution of activity-based inputs. Vessel and phase-specific fuel consumption estimates constrain both operational duration and engine load, limiting the range of plausible emission outcomes [7,19,22].

This behavior demonstrates that Tier III inventories reshape the uncertainty structure itself rather than merely redistributing uncertainty across categories. When sufficient operational data are available, Tier III approaches produce emission estimates that are more tightly bounded and operationally interpretable, consistent with findings from high-resolution shipping emission studies [9,15].

3.5. Spatial Divergence Between Fuel-Based and Activity-Based Representations

The spatial distributions derived from Tier I and Tier III inventories reveal fundamentally different representations of maritime activity intensity, as illustrated in Figure 6. Tier I spatial patterns closely align with domestic port-call density, producing concentrated emission signals around major commercial hubs and inter-island shipping corridors. This pattern reflects the reliance on port-call frequency as a proxy for spatial allocation in the absence of vessel-level trajectory or fuel-use data, a common practice in fuel-based inventories [7,10].

In contrast, Tier III spatial patterns emphasize operational deployment rather than commercial throughput. High-intensity zones correspond to concentrated PCG activity, which does not consistently overlap with major domestic ports. This divergence reflects the activity-based logic embedded in Tier III formulations, where emissions are spatially linked to operational presence rather than traffic volume.

The spatial mismatch between Tier I and Tier III results demonstrates that spatial resolution is not solely a function of map granularity but also of the underlying activity logic used to assign emissions. Consequently, the two approaches encode fundamentally different spatial meanings within the same geographic domain, highlighting the importance of aligning spatial interpretation with methodological assumptions when comparing tiered inventories.

4. Discussion

4.1. Scope Comparison of the EMEP/EEA Tiered Approaches

The tiered structure of the EMEP/EEA methodology reflects deliberate trade-offs between data availability, analytical resolution, and interpretive capacity rather than a linear hierarchy of methodological quality. As summarized in

Table 2. Comparative scope of EMEP/EEA Tiers I–III black carbon estimation methods *.

Factor	Tier I	Tier II	Tier III
Fuel data basis	Aggregated national fuel consumption statistics	Vessel-class and engine technology specific fuel use	Vessel-specific fuel consumption resolved by operational phase
Activity representation	None; emissions scale with total fuel	Limited differentiation by technology class	Explicit resolution by operational mode (berthing, underway, and docking/undocking)
Spatial representation	Indirect, via proxy allocation	Indirect, via proxy allocation	Direct, based on reported vessel activity locations
Operational resolution	Not resolved	Partially resolved through technology differentiation	Fully resolved by activity phase
Uncertainty structure	Broad and fuel-dependent due to aggregation and proxy factors	Intermediate; reduced relative to Tier I when technology data are available	Constrained by vessel and phase-specific activity inputs
Primary analytical role	National-scale baseline emission	Targeted refinement for selected vessel classes	Activity-resolved diagnostic assessment

* This table summarizes the structural scope and resolution characteristics of EMEP/EEA Tiers I–III black carbon estimation approaches. It describes data inputs, activity representation, and analytical roles without reference to study-specific results.

, Tier I is designed to operate under highly aggregated national fuel statistics, enabling rapid baseline estimation but without resolving operational modes or activity heterogeneity. Tier II introduces limited differentiation through vessel class and engine technology, while Tier III explicitly resolves emissions by operational phase using vessel-specific activity data. These distinctions define the analytical scope of each tier and establish the boundary conditions within which results should be interpreted.

Within this structure, Tier I and Tier III serve fundamentally different analytical roles. Tier I emphasizes scalability and comparability across large fleets and reporting systems, whereas Tier III prioritizes operational specificity and diagnostic clarity. Importantly, the tiers are not interchangeable: differences in uncertainty structure, spatial meaning, and interpretability arise from their underlying formulations rather than from differences in emission magnitude. This reinforces the need to evaluate tier performance relative to study objectives and data constraints, rather than treating higher tiers as universally superior substitutes for lower-tier approaches.

4.2. Result-Derived Implications for Tier Selection Under Data Constraints

The comparative results of this study demonstrate that tier choice has a direct influence on the structure of uncertainty, the meaning of spatial outputs, and the interpretability of emission estimates. As synthesized in

Table 3. Decision-oriented matrix of observed Tier I and Tier III performance under data-limited maritime conditions *.

Criteria	Tier I (Fuel-Based)	Tier III (Activity-Based)	Result-Derived Implication
Uncertainty behavior	Fuel-selective and asymmetric, particularly for proxy-based fuels.	Narrower and phase-constrained across operational modes.	Uncertainty structure is governed by tier formulation rather than emission magnitude.
Primary sensitivity drivers	Emission factor formulation and fuel aggregation.	Operational duration and activity mode.	Activity resolution reveals emission-generating behavior obscured in Tier I.
Spatial meaning	Represents commercial traffic intensity via proxies.	Represents operational deployment patterns.	Spatial outputs encode different activity logics, not merely different in resolution.
Interpretability at sub-categories	Limited for fuel or phase-level diagnostics.	High for phase-specific diagnostics.	Tier III enables diagnostics interpretation beyond aggregate totals.
Data dependency	Operates with minimal input data accumulates structural uncertainty.	Requires detailed activity records to constrain estimates.	Tier selection reflects data availability rather than analytical preference.
Analytical use within this study	National-scale aggregation and comparative scaling	Diagnostic analysis of operational patterns	Approaches are complementary rather than interchangeable.

* This table synthesizes result-derived differences between Tier I and Tier III performance based on observed uncertainty distributions, spatial patterns, and phase-resolved diagnostics presented in Section 3.1, Section 3.2, Section 3.3, Section 3.4 and Section 3.5.

, Tier I estimates exhibit fuel-selective and asymmetric uncertainty behavior driven by proxy-based emission factor formulations, whereas Tier III estimates display narrower, phase-constrained uncertainty distributions resulting from activity-resolved inputs. These differences indicate that uncertainty behavior is governed by methodological structure rather than by total emission levels, with important implications for how results can be interpreted and compared.

Beyond uncertainty, the results show that Tier I and Tier III encode fundamentally different representations of maritime activity. Spatial patterns derived from Tier I reflect commercial traffic intensity through proxy allocation, while Tier III spatial outputs represent operational and activity logic. Similarly, Tier III enables phase-specific diagnostics that are inaccessible in fuel-aggregated approaches, even when overall fuel use is comparable. Taken together, these findings indicate that tier selection under data-limited conditions is best guided by the diagnostic questions being asked and the structure of available data, supporting a complementary use of tiered methods rather than a prescriptive hierarchy.

4.3. Implications Derived from the Observed Results

4.3.1. Fuel-Dependent Uncertainty Concentration

The dominance of gasoline in Tier I uncertainty envelopes in Section 3.3 indicates that uncertainty in national BC inventories is not evenly distributed across fuels. This pattern suggests that inventory confidence is disproportionately influenced by proxy-based gasoline estimation, limiting interpretability for small-vessel segments where direct emission factors are unavailable.

4.3.2. Operational-Phase Diagnostics in Tier III

Phase-resolved Tier III results in Section 3.2 demonstrate that emission intensity is unevenly distributed across operational modes, with berthing contributing disproportionately relative to underway operation.

This diagnostic structure is absent in Tier I outputs and confirms that activity-based inventories encode information about emission-generating processes rather than aggregating fuel use alone [22,23].

4.3.3. Spatial Non-Equivalence of Proxy and Activity Representations

Tier III results demonstrate that vessel- and engine-specific characteristics, including load factors and operational modes, are critical determinants of BC emission intensity. This supports vessel-level abatement strategies such as optimizing propulsion efficiency, maintaining engines to minimize incomplete combustion, and adopting retrofits such as diesel particulate filters where feasible. Studies show that BC mitigation effectiveness increases significantly when interventions are tailored to engine type and operating conditions rather than applied uniformly [1,2]. Given the operational variability observed in Philippine Coast Guard missions, vessel-level diagnostics and performance optimization could yield substantial reductions while improving fuel efficiency and operational readiness.

4.4. Limitations of This Study

This study does not include direct measurement-based validation of BC emissions. Uncertainty ranges were derived from literature-reported parameter variability and propagated through Monte Carlo analysis, consistent with established inventory practice. As such, results should be interpreted as comparative diagnostics of methodological behavior rather than absolute emission verification.

These constraints limit but do not affect the comparative findings regarding uncertainty structure, spatial behavior, or tier applicability.

4.5. Recommendations Anchored in Observed Results

4.5.1. Improve Vessel-Class Fuel Reporting

High gasoline uncertainty in Tier I results supports prioritizing vessel-class fuel reporting to reduce proxy dependence.

4.5.2. Expand Activity-Based Datasets for Government Fleets

The constrained uncertainty and operational diagnostics observed in Tier III outputs justify extending activity-based datasets for enforcement and government vessels.

4.5.3. Apply Tier II Selectively to Gasoline-Fueled Small Vessels

Fuel-selective uncertainty amplification in Tier I gasoline estimates supports targeted Tier II application where engine characteristics can be identified [9].

4.5.4. Integrate AIS Where Available

Spatial mismatches identified in Section 3.5 indicate that AIS data would improve route-level allocation beyond port-call proxies.

4.5.5. Develop Local Emission Factor Measurements

The reliance on default emission factors constrains uncertainty characterization, particularly for gasoline engines, reinforcing the need for local measurements.

4.5.6. Institutionalize Tiered Inventory Practice

Results show that no single tier is universally optimal; institutionalizing tier selection based on data availability and uncertainty behavior would improve inventory robustness.

4.5.7. Target Mitigation Using Tier-Appropriate Spatial Signals

Differences in domestic and PCG hotspot structures indicate that mitigation strategies should be spatially aligned with the tier used.

4.5.8. Prioritize Berthing-Related Controls Where Tier III Data Exist

Consistent dominance of berthing emissions in Tier III results supports prioritizing auxiliary load reduction and shore power at identified high-intensity locations.

5. Conclusions

This study examined how EMEP/EEA Tier I and Tier III black carbon emission estimation approaches behave under data-limited maritime conditions, using the Philippine domestic fleet and an operational PCG dataset as an empirical comparison. Rather than reconciling absolute emission magnitudes, the analysis focused on how methodological structure shapes uncertainty behavior, diagnostic resolution, and spatial representation.

The results show that Tier I fuel-based inventories provide scalable national-level baselines but exhibit fuel-selective and asymmetric uncertainty, particularly for categories reliant on proxy emission factors. In contrast, Tier III activity-based inventories produce more constrained uncertainty envelopes and resolve emissions by operational phase, enabling diagnostic interpretation of emission-generating behavior that is not accessible through fuel aggregation alone.

Spatial analysis further demonstrates that Tier III encodes fundamentally different activity logics: port-call based allocation reflects commercial traffic intensity, whereas activity-based mapping captures operational deployment patterns. As a result, spatial outputs from different tiers are not directly interchangeable and must be interpreted in relation to their underlying estimation logic.

Based on these findings, this study develops a decision-oriented synthesis that supports informed selection and combination of tiered emission estimation under data-limited maritime conditions. The results indicate that tier choice should be guided by analytical intent, data availability, and interpretability requirements rather than by a prescriptive methodological hierarchy. While grounded in the Philippine context, these insights may be relevant to other developing and archipelagic maritime states seeking to strengthen black carbon inventories in the absence of comprehensive vessel-level data.

Future research should expand activity-based datasets for PCG and government vessels, institutionalize tiered inventory practice, and explore feasibility and policy options for shore power adoption, auxiliary load reduction, and phased implementation planning.