3-D Emissions Mapping and Social Cost Estimation for US Domestic Aviation at West Coast Hubs

Hesam Shafiei Nia; Don MacKenzie

cs.LGSep 16, 2026

3-D Emissions Mapping and Social Cost Estimation for US Domestic Aviation at West Coast Hubs

Authors: Hesam Shafiei Nia, Don MacKenzie

Organizations: Department of Civil and Environmental Engineering, University of Washington Seattle, WA, 98195

Abstract

Existing aviation emissions inventories lack accurate trajectory data for high-resolution social cost and health impact assessment. This paper develops a 3-D emissions map by reconstructing flight trajectories for US west coast hubs to estimate regional environmental and near-airport health impacts. A physics-informed autoencoder (AE) is applied to ADS-B trajectory records for January 2025 covering US west coast hubs. The encoder combines a Convolutional Neural Network (CNN), a Bi-GRU, and a 3-D CNN with skip connection; the decoder is a Temporal Convolutional Network (TCN). It is benchmarked against a baseline-AE and cubic spline interpolation. Emissions are mapped via EUROCONTROL Base of Aircraft Data (BADA) performance tables and ICAO Engine Emissions Databank (EEDB) emission indices, with altitude corrections via Boeing Fuel Flow Method 2 (BFFM2). Social costs are quantified for all flight phases, with health impacts assessed for Landing and Takeoff cycles within 50 km of each hub. The proposed AE model outperforms both a TCN-AE and cubic spline interpolation across 5% to 50% missing rates. Monetizing the emissions inventory shows NOx produces a small net cooling effect in direct climate forcing, while accounting for 99.7% of monetized air-quality and health cost despite being under 0.4% of CO2 by mass, making it the dominant health-cost driver. To our knowledge, this is among the first studies combining AE-based trajectory reconstruction with separate spatial-temporal feature encoding and altitude-based emissions modeling to produce a regional aviation emissions inventory for air quality, climate impact and population exposure. The resulting emissions map and social cost estimates provide quantitative context for environmental impact assessment and near-airport health policy evaluation for US domestic aviation.

Figures & tables

Figure 1: Proposed deep autoencoder architecture of the CBiG3D model.

Figure 2: Emissions estimation approach schematics.

		Value (USD/t)
Category	Species	LTO	Cruise	Reference
Climate (all altitudes, all flight phases)	$\mathrm{CO}_{2}$	190	190	U.S. Environmental Protection Agency (2023)
Climate (all altitudes, all flight phases)	$\mathrm{NO}_{x}$	$-$ 590	$-$ 940	Grobler et al. (2019)
Air quality & Near-airport health (LTO $\leq$ 50 km radius)	$\mathrm{NO}_{x}$ b	20,000	24,000	Grobler et al. (2019)
	HC (NMVOC) a	19,000	–	Grobler et al. (2019)
	CO	520	–	Grobler et al. (2019)

Table 1: Unit damage costs used to monetize climate and near-airport air quality/health impacts.

Attribute	Metric	Proposed AE a	TCN-AE	Cubic Spline
Overall	MAE	0.061 $\pm$ 0.056	0.184 $\pm$ 0.143	2.603 $\pm$ 53.833
Overall	RMSE	0.139 $\pm$ 0.093	0.274 $\pm$ 0.183	10.488 $\pm$ 221.482
Latitude (deg)	MAE	0.416 $\pm$ 0.929	2.896 $\pm$ 3.140	0.238 $\pm$ 1.118
Latitude (deg)	RMSE	0.507 $\pm$ 1.131	3.090 $\pm$ 3.300	0.298 $\pm$ 1.475
Longitude (deg)	MAE	0.274 $\pm$ 0.467	3.000 $\pm$ 2.894	0.181 $\pm$ 0.829
Longitude (deg)	RMSE	0.326 $\pm$ 0.582	3.201 $\pm$ 3.008	0.232 $\pm$ 1.146

Table 2: Imputation accuracy (mean

\pm

SD), by model and trajectory attribute.

		Overall		Position (km)		Altitude (m)		Groundspeed (m/s)
Rate	Model	MAE	RMSE	MAE	RMSE	MAE	RMSE	MAE	RMSE
5%	Proposed AE	0.050 $\pm$ 0.063	0.116 $\pm$ 0.103	44.07 $\pm$ 141.54	47.76 $\pm$ 149.04	135.7 $\pm$ 266.5	150.8 $\pm$ 301.0	14.11 $\pm$ 51.56	15.39 $\pm$ 53.75
	TCN-AE	0.185 $\pm$ 0.148	0.274 $\pm$ 0.193	464.46 $\pm$ 443.42	471.80 $\pm$ 447.84	2388.8 $\pm$ 3066.9	2423.7 $\pm$ 3070.4	26.50 $\pm$ 43.80	27.71 $\pm$ 44.48
	Cubic Spline	0.111 $\pm$ 0.317	0.314 $\pm$ 0.869	13.42 $\pm$ 73.89	16.61 $\pm$ 96.41	283.1 $\pm$ 4121.7	385.0 $\pm$ 5982.6	1.87 $\pm$ 7.83	2.23 $\pm$ 9.48
10%	Proposed AE	0.058 $\pm$ 0.055	0.134 $\pm$ 0.094	51.03 $\pm$ 110.73	59.77 $\pm$ 132.62	252.0 $\pm$ 373.1	289.5 $\pm$ 450.7	12.06 $\pm$ 40.08	14.77 $\pm$ 46.17
	TCN-AE	0.184 $\pm$ 0.145	0.275 $\pm$ 0.186	455.53 $\pm$ 403.09	471.35 $\pm$ 414.90	2499.1 $\pm$ 3115.9	2585.0 $\pm$ 3138.1	26.69 $\pm$ 43.00	28.83 $\pm$ 44.46

Table 3: Imputation MAE and RMSE (mean

\pm

SD) Synthetic and real missing-data rate analysis.

Figure 3: $\mathrm{NO}_{x}$ by altitude in the US west coast area.

Figure 4: $\mathrm{NO}_{x}$ by latitude and longitude across altitudes in the US west coast area.

Altitude band	Fuel	$\mathrm{CO}_{2}$	$\mathrm{NO}_{x}$	$\mathrm{CO}$	HC	nvPM	Fuel	$\mathrm{CO}_{2}$	$\mathrm{NO}_{x}$	$\mathrm{CO}$	HC	nvPM
	Fuel (t)	Total (kg)					Share of fleet total
LTO ( $<$ 914 m)	6,051.0	19,121,232.7	92,908.8	6,915.9	194.1	201.2	4.9%	4.9%	6.3%	17.5%	4.9%	3.7%
Low (914–3 km)	8,937.9	28,243,764.0	151,073.6	1,948.8	197.2	388.6	7.2%	7.2%	10.2%	4.9%	4.9%	7.1%
Mid (3–6 km)	12,286.6	38,825,561.2	187,137.9	2,549.3	299.9	540.6	9.9%	9.9%	12.7%	6.5%	7.5%	9.9%
High (6–9 km)	16,656.6	52,634,856.0	210,635.8	4,104.1	482.8	732.9	13.4%	13.4%	14.3%	10.4%	12.1%	13.4%
Cruise ( $>$ 9 km)	80,755.8	255,188,454.4	834,213.4	24,025.1	2,826.5	3,591.1	64.8%	64.8%	56.5%	60.8%	70.7%	65.8%

Table 4: Fuel burn and emissions by altitude, west coast January-2025 inventory.

Category	Pollutant	Tons	Unit price (USD/t)	Cost (USD)
Climate (all altitudes, whole area)	$\mathrm{CO}_{2}$	394,013.9	190	74,862,635
	$\mathrm{NO}_{x}$	1,476.0	$-$ 918	$-$ 1,354,893
	HC	4.0	0	0
	$\mathrm{CO}$	39.5	0	0
	$\mathrm{nvPM}$	5.5	0	0
	Subtotal			73,507,742

Table 5: Social cost of west coast January-2025 aviation emissions.

Trajectory Inference

Trajectory inference reconstructs continuous paths and underlying dynamics from sparse, noisy, or incomplete observations.

See all 37 papers

FlowATC: Aircraft Trajectory Prediction via Flow Matching
Sep 15, 2026Mathurin Petit, Emir Torun, Louis Brusset +2Probabilistic Trajectory Prediction Flow Matching
École Polytechnique, Palaiseau, 91128, France · Technische Universität Berlin, Berlin, 10623, Germany · Mines Paris–PSL University, Paris, 75006, France +2
ASCENT: Transformer-Based Aircraft Trajectory Prediction in Non-Towered Terminal Airspace
Mar 17, 2026Alexander Prutsch, David Schinagl, Horst PosseggerTrajectory Prediction Multimodal Trajectory Prediction
Institute of Visual Computing, Graz University of Technology, Austria
From Regional to Global: Transfer Learning for Atmospheric Transport Emulators
Sep 20, 2026Jeff Clark, Elena Fillola, Nawid Keshtmand +2Surrogate Modeling Task Transferability Estimation
School of Engineering Mathematics and Technology, University of Bristol, UK · School of Chemistry, University of Bristol, UK

Sep 15, 2026J/K move · Enter open · S save

FlowATC: Aircraft Trajectory Prediction via Flow Matching

Mathurin Petit, Emir Torun, Louis Brusset, Jordan Kam+1

École Polytechnique, Palaiseau, 91128, France · Technische Universität Berlin, Berlin, 10623, Germany · Mines Paris–PSL University, Paris, 75006, France+2

Probabilistic Trajectory Prediction Flow Matching Air Traffic Management Flow-Based Generative Modeling Trajectory Prediction Diffusion Models

Open paper AskPDF arXiv

Building accurate decision-support tools for next-generation air traffic control requires robust trajectory prediction models. We present a flow-matching architecture trained exclusively on historical aircraft trajectories, with no route labels or chart supervision. Trained on 1.15 million Automatic Dependent Surveillance-Broadcast trajectory windows collected over the San Francisco Bay Area, the model generates aircraft trajectory distributions that closely match historical traffic, reproducing known airspace structure around San Francisco Airport such as the shape of SFO's published NIITE FOUR departure procedure. Our model is trained directly on the native, irregular ADS-B sampling interval. Trajectory prediction is cast as sequence inpainting using a block-causal Transformer that denoises future state tokens conditioned on the observed history using Conditional Flow Matching or Denoising Diffusion Probabilistic Models. We compare our architecture against constant-velocity, deterministic-Long Short Term Memory, and Conditional Variational Autoencoders baselines. At matched parameter count, CFM outperforms DDPM by 11-26% in minADE@20, and both generative objectives surpass the CVAE baseline by 31-41%. We further show that the error degrades gracefully with prediction horizon, and the architecture remains effective when retrained on temporally decimated feeds. Lastly, we sample

K

independent completions, yielding spatial probabilistic occupancy estimates that can serve as input to downstream conflict-risk estimation.

Abstract

Trajectory Inference

Explore similar work

FlowATC: Aircraft Trajectory Prediction via Flow Matching

ASCENT: Transformer-Based Aircraft Trajectory Prediction in Non-Towered Terminal Airspace

From Regional to Global: Transfer Learning for Atmospheric Transport Emulators