3 Results – Blue Whales Blue Skies Emissions Analysis

3.1 Program Impact

This analysis shows a net global increase of 22,066.08 mt of CO2e emissions. However, this corresponds to only a 0.6% increase relative to the counterfactual. The increase is driven by spillover occurring outside VSR zones, where emissions rise by 1.3%. Meanwhile, the program leads to a -25.5% reduction in emissions inside VSR zones (@tbl-program-summary, Figure 3.1).

Table 3.1: Summary of the 2024 BWBS program analysis (100-year GWP). The global counterfactual is the sum of the regional counterfactuals, ensuring Inside + Outside = Global.

Metric	Global	Inside VSR zone	Outside VSR zone
Actual CO2e (mt)	3,452,922	66,960	3,385,963
Counterfactual CO2e (mt)	3,430,856	89,887	3,340,969
Net Impact CO2e (mt)	+ 22,066	-22,928	+ 44,994
% Change	+ 0.6%	-25.5%	+ 1.3%

Figure 3.1: Summary of CO2e emissions impact by spatial domain (100-year GWP). Negative values indicate emission reductions compared to the counterfactual. The global impact represents the net effect combining regional (inside VSR) and spillover (outside VSR) impacts.

Net global increases are also observed for all other pollutants analyzed, although these remain small, at below 0.7%. Figure 3.2 further shows that the program achieves substantial reductions in local pollutants, with increases limited to CH4, CO, and VOCs. The increases in these three pollutants are consistent with a decrease in combustion efficiency at reduced speeds, as described in Chapter 4 and detailed in Appendix A regarding how our model captures these nonlinear combustion dynamics.

Figure 3.2: Program impact across multiple pollutants by spatial domain. Negative values indicate emission reductions.

Substantial reductions in emissions inside the VSR zones are associated with changes in emissions intensity, with lower emission intensity observed during the within-season period (Figure 3.3). Since vessels travel relatively similar distances during the within-season and outside-season periods, reductions in emission intensity can be explained, as expected, by a reduction in speed during the season inside the VSR (Figure 3.4).

Figure 3.3: Comparison of baseline emission intensity (before season) vs. observed within-season emissions for participating vessels (100-year GWP). Points below the 1:1 line indicate vessels that emitted less than expected based on their baseline intensity.

Figure 3.4: Vessel-level distributions of distance traveled and average speed by period and domain. Diamonds indicate fleet means. Dashed line marks the 10.4-knot VSR compliance threshold.

Despite no clear increase in average speeds outside VSR zones during the season, and in fact a slight decline in global average speeds from baseline levels (Figure 3.4), the reductions in speed within VSR zones and the associated emission reductions were offset by increased emissions elsewhere (Figure 3.1). Vessel-level analysis reveals that a subset of vessels drives this spillover effect, exhibiting increased emission impacts outside VSR zones. As shown in Figure 3.5, this pattern may reflect ‘catch-up’ behavior whereby vessels compensate for reduced speeds within VSR zones by increasing their speeds outside these areas.

Figure 3.5: Left: Comparison of emission impacts inside vs. outside VSR zones for each vessel (100-year GWP). Right: Relationship between speed changes inside vs. outside VSR zones. Yellow points indicate potential cause of spillover.

3.2 Vessels driving spillover

To better understand the spillover effect, we split the fleet into two groups based on their speed behavior: vessels that slowed inside VSR zones but sped up outside (“catch-up” vessels) and all other vessels. Figure 3.6 compares the CO2e emissions impact, while Figure 3.7 presents the multi-pollutant impacts for each group side by side. Together, these figures show clear evidence that catch-up vessels are driving the spillover effect.

Figure 3.6: CO2e emissions impact (100-year GWP) by spatial domain for non-catch-up vessels (left) vs. catch-up vessels (right).

Figure 3.7: Multi-pollutant emissions impact (% change) for non-catch-up vessels (left) vs. catch-up vessels (right), by spatial domain.

3.3 Spatial patterns of speed and emissions changes

To visualize where speed and emissions changes occur geographically, we aggregate ping-level data for compliant vessels to a 0.25° × 0.25° grid. For each vessel and grid cell, we calculate aggregate speed (i.e., from total distance / total hours by vessel/pixel) and CO2e emissions intensity (i.e., from total CO2e / total distance by vessel/pixel) separately for trips during and outside the VSR season. We then compute the within-vessel difference (i.e., during season - outside season) and take the median across vessels for each grid cell. We present results for all compliant vessels, as well as separately for catch-up vessels (those that slowed inside VSR zones but sped up outside) and non-catch-up vessels.

Figure 3.8 and Figure 3.9 show mixed patterns in both speed and intensity. When representing across catch and non-catch vessels Figure 3.10 shows how most of the increases are associated with catch-up vessels.

Figure 3.8: Median change in aggregate speed (knots) from outside season to during the VSR season across all compliant vessels, by 0.25° grid cell. Negative values (blue) indicate speed reductions during the season, while positive values (red) indicate speed increases.

Figure 3.9: Median change in CO2e emissions intensity (mt CO2e per nautical mile, 100-year GWP) from outside the season to during the VSR season across all compliant vessels, by 0.25° grid cell. Negative values (blue) indicate reduced emissions intensity during the season. Positive values (red) indicate increased intensity.

Figure 3.10: Median change in aggregate speed (knots) from outside the season to during the VSR season for non-catch-up vessels (top) vs. catch-up vessels (bottom). Catch-up vessels are those that slowed inside VSR zones but sped up outside.

A closer look at activity within the West Coast EEZ shows that, while both groups of vessels reduce emissions inside the VSR, catch-up vessels account for a larger increase in speeds and intensities outside these areas, both within and beyond the EEZ (Table 3.2; Figure 3.11).

On average, non catch-up vessels show reductions in speed and intensity across all areas, despite some localized increases observed. In contrast, catch-up vessels exhibit average reductions only within VSR zones, with slight increases outside. In terms of spatial proximity to the VSR, this later group appears to increase speeds around the boundaries of the zones, particularly just outside the Southern California VSR zone, predominantly to the south of it, and in the area between the Northern and Southern California areas (Figure 3.11). In fact, catch-up vessels tend to concentrate these increases within the EEZ, compared to areas outside the EEZ (Table 3.2).

Table 3.2: Speed and emissions intensity changes outside VSR zones by vessel group, split by inside vs. outside the U.S. West Coast EEZ. Weighted means use the number of vessels per grid cell as weights.

Location	Grid cells	Mean speed change (kn)	Wtd. mean speed change (kn)	Mean intensity change (mt CO2e/nm)	Wtd. mean intensity change (mt CO2e/nm)
Non catch-up
Inside VSR zone	59	-3.128	-2.240	-0.111006	-0.078879
Inside EEZ (excluding VSR)	739	-0.496	-0.693	-0.028465	-0.031209
Outside EEZ	8833	-1.091	-0.973	-0.072311	-0.059352
Catch-up
Inside VSR zone	58	-2.100	-1.818	-0.070057	-0.056812
Inside EEZ (excluding VSR)	937	0.111	0.199	0.005315	0.008295
Outside EEZ	11146	-0.009	0.069	-0.008630	-0.000045

Figure 3.11: Spatial distribution of median speed changes (knots) from outside to during the VSR season, with U.S. West Coast EEZ and VSR zone boundaries overlaid. Panels show non catch-up vessels (left), catch-up vessels (right).

3.4 Compliant vs. non-compliant vessels

As an additional comparison, we next look at how the BWBS season impacted CO2e emissions from non-compliant vessels, and compare these impacts to those from compliant vessels. Figure 3.12 shows the relative (%) emissions impact for each group of vessels and each of the three spatial domains (inside VSR zones, outside VSR zones, and global). The left panel is simply the same as the bottom panel from Figure 3.1; the right panel shows similar information, but for non-compliant vessels. This indicates that even for non-compliant vessels, there is an adjustment in activity outside the VSR zones between during and outside the season, which could reflect a background variability between the baseline and the season that is not fully captured in the current analysis. This information helps contextualize the observed net global emissions increases of the compliant fleet. This suggests that some level of spillover may occur even in the absence of full behavioral adjustment, and that spillover is even higher for non-compliant vessels. This indicates that the program may still yield a net benefit in terms of net global impact once this background effect is accounted for. Future work aimed at producing a robust causal inference assessment of the program may be able to account for this background level of adjustment.

Figure 3.12: CO2e emissions impact (% change, 100-year GWP) by spatial domain for compliant vessels (left) vs. non-compliant vessels (right).

3.5 Company level results

We can explore which companies accounted for most of the program benefits in terms of global emissions reductions and reductions within the VSR zones, as well as which contributed to emissions spillover. Table 3.3 shows, for each company, the participating vessels that were matched to those available in our dataset and that had activity in both periods and both inside and outside the VSR, allowing baseline and counterfactual scenarios to be constructed. Among these, the table details which vessels were compliant (i.e., navigated below 10.4 knots within the VSR during the season and reduced their speed relative to baseline) and which were non-compliant. Percentage values in brackets represent each group’s share of the vessels included in the analysis. The table also reports how many compliant vessels per company exhibited non–catch-up behavior (i.e., did not increase speed outside the VSR) versus catch-up behavior (i.e., increased speed outside the VSR, contributing to spillover). Companies are sorted by the highest percentage of compliant vessels among those included in the analysis.

Table 3.3: Number of vessels listed in the program per company, and detail of those included in the analysis by compliance and catch-up status.

		Analysis			Compliant
Company	Listed	Matched	Non-compl.	Compliant	Non-catch-up	Catch-up
COSCO	24	15	0	15 (100%)	3 (20%)	12 (80%)
Dockendale	8	4	0	4 (100%)	2 (50%)	2 (50%)
Marathon Petroleum	7	6	0	6 (100%)	3 (50%)	3 (50%)
OOCL	15	8	0	8 (100%)	3 (37.5%)	5 (62.5%)
Polar Tankers	5	5	0	5 (100%)	1 (20%)	4 (80%)
Scorpio Group	16	2	0	2 (100%)	2 (100%)	0
Swire Shipping	3	2	0	2 (100%)	0	2 (100%)
Temm Maritime Co Ltd	1	1	0	1 (100%)	1 (100%)	0
Toyofuji	1	1	0	1 (100%)	0	1 (100%)
Wallenius	37	7	0	7 (100%)	2 (28.6%)	5 (71.4%)
ONE	52	36	3 (8.3%)	33 (91.7%)	12 (36.4%)	21 (63.6%)
Maersk	30	26	3 (11.5%)	23 (88.5%)	9 (39.1%)	14 (60.9%)
CMA CGM	70	34	5 (14.7%)	29 (85.3%)	17 (58.6%)	12 (41.4%)
OSG Ship Management, Inc.	8	6	1 (16.7%)	5 (83.3%)	3 (60%)	2 (40%)
Wan Hai	17	6	1 (16.7%)	5 (83.3%)	1 (20%)	4 (80%)
Hyundai Glovis	40	16	3 (18.8%)	13 (81.2%)	10 (76.9%)	3 (23.1%)
Yang Ming	24	17	5 (29.4%)	12 (70.6%)	5 (41.7%)	7 (58.3%)
MSC	69	25	8 (32%)	17 (68%)	6 (35.3%)	11 (64.7%)
D'Amico Group	6	3	1 (33.3%)	2 (66.7%)	1 (50%)	1 (50%)
Hyundai (UCC)	21	6	2 (33.3%)	4 (66.7%)	4 (100%)	0
Pacbasin	30	3	1 (33.3%)	2 (66.7%)	2 (100%)	0
NYK-RoRo	67	29	10 (34.5%)	19 (65.5%)	10 (52.6%)	9 (47.4%)
Mol ACE	33	15	6 (40%)	9 (60%)	5 (55.6%)	4 (44.4%)
Evergreen	31	13	6 (46.2%)	7 (53.8%)	3 (42.9%)	4 (57.1%)
K Line	57	22	11 (50%)	11 (50%)	2 (18.2%)	9 (81.8%)
Swire Bulk	12	3	2 (66.7%)	1 (33.3%)	0	1 (100%)
Anglo-Eastern Shipmanagement	16	2	2 (100%)	0	0	0
Cetus Maritime (Shanghai) Ltd.	3	1	1 (100%)	0	0	0
HMC Shipmanagement Co Ltd	5	1	1 (100%)	0	0	0
Hapag-Lloyd	9	2	2 (100%)	0	0	0
Pasha Hawaii	5	3	3 (100%)	0	0	0
Unisea Shipping Ltd	3	1	1 (100%)	0	0	0
Other (unmatched vessels)	32	0	0	0	0	0
Total	757	321	78 (24.3%)	243 (75.7%)	107 (44%)	136 (56%)

Figure 3.13 shows, for each company, the contribution of compliant vessels to emissions changes across the different spatial domains. This reveals a generalized decrease in emissions within VSR zones and shows that, for many companies, compliant vessels contributed to a net global reduction in emissions relative to their baseline. Fewer than half of the companies exhibit spillover effects driven by vessels engaging in catch-up behavior.

Figure 3.13: CO2e emissions impact (% change, 100-year GWP) by company and spatial domain for compliant vessels (those that navigated below 10.4 knots and reduced their average speed during the VSR season within VSR zones). The notation n=x indicates the number of compliant vessels included.

However, given that the definition of compliance is specific to this analysis, based on average speeds and including a safety margin of 0.4 knots around the 10 knot target speed, which differs from alternative definitions used in other studies such as Blue Whales Blue Skies (2025), overall company performance may vary. Thus, to better assess overall company impacts, Figure 3.14 includes all available vessels in the analysis for each company, rather than only the compliant subset. This provides a more complete picture of fleet-level effects. Under this broader view, a majority of companies still contribute to global emission reductions, although these are more modest than when considering only compliant vessels. Also we can see a larger change in emissions from companies with net emission increases with respect their baseline. Additionally, two companies show increased emissions within VSR zones when considering the entire matched fleet.

Figure 3.14: CO2e emissions impact (% change, 100-year GWP) by company and spatial domain for all vessels. Compliant vessels navigated below 10.4 knots and reduced their average speed during the VSR season within VSR zones; non-compliant vessels did not meet both criteria. The notation n=x/y indicates the number of vessels included in the analysis out of the total listed in the program for each company.

These results provide insights into the level of company participation in the program and how compliance with its targets can translate into measurable global emission reductions, helping to identify the most engaged participants. However, important limitations remain. The analysis includes only vessels with activity in both periods and both inside and outside the VSR, which restricts the ability to capture the full scope of each company’s fleet. Some vessels may be complying during the season and contributing to real emission reductions, but without corresponding baseline trips to construct counterfactuals, they cannot be included in the analysis. We also use a different definition of compliance in our analysis than the BWBS program as a whole; we assess compliance at the vessel-level, whereas the BWBS program assesses compliance at the company level. As a result, this company level assessment should be taken as illustrative, but not comprehensive or conclusive.