1. Introduction

Preprocessing

This report estimates the carbon footprint of the Aix-en-Provence conference based on travel survey responses from participants.

Geographic coordinates for cities are sourced from city_geographic_coordinates.csv, automatically populated by two preprocessing scripts that must be run before knitting this document:

  1. check_country_names.r — checks that country names are standardised across all source files
  2. update_coordinates.r — geocodes all cities via OpenStreetMap (no API key required) and maintains the coordinates database

Distance calculation

Each participant’s city of origin is geocoded to GPS coordinates via OpenStreetMap. The distance to the conference venue is computed as the great-circle distance (orthodromie) using the Haversine formula — i.e. the shortest path on the surface of the Earth, ignoring terrain and actual flight or road routes. This is a lower bound: real journeys are always longer, particularly for plane travel (detours, layovers). Distances are expressed one-way; emissions are computed for the full round trip.

Emission factors per transport type (kg CO2e)

Factors are sourced from ADEME Base Empreinte® 2024 and EEA 2023 (see table below). The plane factor (0.230 kg CO₂e/km) includes radiative forcing — the additional warming effect of contrails and high-altitude NOₓ emissions, estimated at 2–4× the CO₂-only impact. This is scientifically recommended for honest aviation accounting. The train factor uses the EEA European average (0.033 kg CO₂e/km) rather than the ADEME French value (0.009, largely nuclear-powered TGV), because many international participants cross multiple national rail networks with higher carbon intensities (Germany: 0.046, UK: 0.041).

Emission factors by transport mode (ADEME 2024 / EEA 2023)
Transport mode kg CO2e/km
Plane 0.230
Train 0.033
Gasoline car 0.218
Gasoline bus 0.103
Electric bus 0.014
Bike 0.000
Feet 0.000

Emissions per leg are calculated as:

\[\text{CO}_2\text{e (kg)} = \text{distance (km)} \times \text{emission factor (kg CO}_2\text{e/km)}\]

Round-trip strategy

The survey may collect either outbound travel only, or both outbound and return legs:

  • Two-leg mode — if the CSV contains to_city / to_country / to_transport columns, the return leg is computed explicitly and may differ from the outbound (e.g. train out, plane back).
  • Outbound × 2 mode — if only outbound columns are present, total emissions are approximated as outbound × 2, assuming a symmetric return journey.

The mode detected for this knit is indicated in the results.

Mean and median estimation

The distribution of individual emissions is strongly right-skewed: plane travellers emit an order of magnitude more than train travellers. This pulls the mean well above the median. Both are reported: the mean is used for extrapolation (it preserves total mass); the median gives a better picture of the typical participant’s footprint.

Naive vs. stratified extrapolation

Observed emissions (144 respondents) are extrapolated to all 171 participants using two approaches:

  • Naive extrapolation — the sample mean is multiplied by the total number of participants: \(\hat{E}_{\text{total}} = \bar{e} \times N\). Simple but sensitive to the skewness introduced by a few high-emission plane travellers.
  • Stratified extrapolation — participants are split into two groups (plane vs. other modes). Each group’s mean is extrapolated separately, then summed. This reduces variance and gives a more stable estimate when the proportion of plane travellers in the sample may not perfectly represent the full population.

A 95% confidence interval on the mean is computed via the t-distribution and propagated to the extrapolated total. It reflects sampling uncertainty but does not capture non-response bias (i.e. whether non-respondents travel differently from respondents).

Green grants analysis

Green grants are financial incentives awarded to participants who chose a low-carbon alternative (train, coach) over flying. For each recipient, the analysis estimates:

  • Actual emissions — computed from their declared origin city and transport mode, using the same distance and emission factor method as the main analysis.
  • Counterfactual emissions — what they would have emitted had they flown (same distance × plane factor).
  • Emissions saved — the difference between counterfactual and actual.
  • Additional travel time — estimated from average mode speeds (train: 120 km/h, car: 100 km/h, coach: 80 km/h) versus plane travel time (distance / 800 km/h + 3 h airport overhead), for a full round trip.

The impact is expressed both for the grant recipients alone and as a share of the conference’s total observed emissions.

2. Emissions summary

Résultats clés — 144 répondants sur 171 participants (84.2%)

  • Émissions totales observées : 65.3 tonnes CO₂e
  • Moyenne par participant : 453 kg CO₂e [IC 95% : 299 — 608 kg]
  • Médiane : 67 kg CO₂e

3. Extrapolation to 171 participants

Extrapolation à 171 participants

  • Extrapolation naïve (moyenne × total) : 77.5 t CO₂e [IC 95% : 51.1 — 104 t]
  • Extrapolation stratifiée (avion vs. autres) : 77.7 t CO₂e

4. Analysis by transport mode

Emissions by transport mode ( both legs )
Transport mode N participants Total emissions (kg CO2e) Mean emissions (kg CO2e)
Plane 58 31091 536
Train 61 1238 20
Gasoline car 7 428 61
Gasoline bus 15 112 7
Electric bus 1 0 0
Bike 1 0 0
Feet 1 0 0

5. Analysis by country of origin

Top 15 countries by total emissions
Country of origin N participants Total emissions (kg CO2e) Mean distance (km)
USA 5 9307 8093
Taiwan 1 2301 10004
Korea 1 2146 9329
Germany 25 1980 876
UK 8 1977 1075
Israel 3 1974 2861
Italy 18 1765 530
India 1 1746 7592
Finland 3 1596 2313
Canada 1 1497 6511
UAE 1 1128 4904
France 48 965 263
Poland 3 912 1322
Turkey 1 641 2786
Ireland 2 629 1368

6. Geographic maps

Map 1: Number of participants by city

Map 2: CO2 emissions by country

7. Distribution of individual emissions

8. Additional visualisations

Transport mode breakdown (pie charts)

Summary by transport mode ( both legs )
Transport N trips Total kg CO2e % trips % CO2
Plane 58 31090.9 40.3 94.6
Gasoline car 7 428.3 4.9 1.3
Gasoline bus 15 112.2 10.4 0.3
Train 61 1237.8 42.4 3.8
Electric bus 1 0.0 0.7 0.0
Bike 1 0.0 0.7 0.0
Feet 1 0.0 0.7 0.0

Breakdown by country: transport modes and emissions

Emissions distribution by transport mode

9. Green Grants impact

The conference awarded green grants (low-carbon travel subsidies) to encourage 18 participants to travel by train or coach instead of flying.

Impact des green grants — 18 bénéficiaires

  • Émissions économisées : 5.4 tonnes CO₂e (84.8% de réduction pour ces participants)
  • Temps de trajet supplémentaire total : 90 heures (moyenne 5 h/personne)
  • Impact sur le total de la conférence : 8.2%

Detail by recipient

Green grant impact per recipient (round trip)
City Country Transport used Distance (km) Actual emissions (kg CO2e) Emissions if flown (kg CO2e) Emissions saved (kg CO2e)
Berlin Germany Train 1161 77 534 457
Leipzig Germany Train 1013 67 466 399
Leipzig Germany Train 1013 67 466 399
Leipzig Germany Train 1013 67 466 399
Amsterdam Netherlands Train 985 65 453 388
Brussels Belgium Train 819 54 377 323
Brussels Belgium Train 819 54 377 323
Lille France Train 811 54 373 320
Frankfurt am Main Germany Train 773 51 355 304
Frankfurt am Main Germany Train 773 51 355 304
Mannheim Germany Train 702 46 323 277
Mannheim Germany Train 702 46 323 277
Paris France Train 639 42 294 252
Paris France Train 639 42 294 252
Paris France Train 639 42 294 252
Zurich Switzerland Train 491 32 226 194
Milan Italy Train 367 24 169 145
Milan Italy Gasoline bus 367 76 169 93

Breakdown by transport mode

Impact by alternative transport mode
Transport mode N recipients Total emissions (kg CO2e) Emissions saved (kg CO2e)
Gasoline bus 1 76 93
Train 17 882 5263

Grant vs no-grant comparison

Overall conference impact

Green grant impact on the conference carbon footprint
Metric Value
Green grant recipients 18 participants
% of participants 12.5 %
Emissions with green grants 0.96 tonnes CO2e
Emissions if flown 6.31 tonnes CO2e
Emissions saved 5.36 tonnes CO2e
Reduction for these participants 84.8 %
Impact on conference total 8.2 %

10. Summary

Key results

  • Response rate: 84.2% (144 / 171 participants)
  • Total emissions (144 respondents): 65.3 tonnes CO2e
  • Extrapolated total — naive (mean × total): 77.5 t CO2e [95% CI: 51.1 — 104 t]
  • Extrapolated total — stratified (plane vs. other): 77.7 tonnes CO2e
  • Mean per participant: 453 kg CO2e (median: 67 kg — gap reflects skewness from plane travellers)
  • Dominant transport mode: Plane (58 participants)

Equivalences

The stratified extrapolation (77.7 tonnes CO2e) is equivalent to:

  • Approximately 405 km driven by a petrol car
  • Or 39 round-trip Paris–New York flights
Carbon footprint summary
Metric Value
Survey respondents 144 participants
Response rate 84.2%
Total participants 171 participants
Total emissions (respondents) 65.3 tonnes CO2e
Extrapolated — naive (mean × total) 77.5 tonnes CO2e
95% CI on naive extrapolation [51.1 — 104] tonnes CO2e
Extrapolated — stratified (plane vs. other) 77.7 tonnes CO2e
Mean per participant 453 kg CO2e
Median per participant 67 kg CO2e

Note: Emission factors and distances are average estimates. A more precise analysis would require trip-specific data (actual routes, aircraft types, load factors, etc.).