To validate our implementation, we compared Earthster’s results against the EXIOBASE-ImpactWorld+ (IW+) benchmark. The EXIOBASE-ImpactWorld+ benchmark contains the official ImpactWorld+ characterization factors adjusted for EXIOBASE exchanges. The table below shows the relative difference between Earthster and EXIOBASE-ImpactWorld+.
Legend: Values represent the relative difference as a percentage (0% = identical results; 100% = maximum difference).
Indicator | Mean difference (%) | Median difference (%) | Maximum difference (%) |
Fossil and nuclear energy use | 79.9 | 85.9 | 99.8 |
Mineral resources use | 9.7 | 5.3 | 100 |
Climate change, long term | 0 | 0 | 0.2 |
Climate change, short term | 0 | 0 | 0.2 |
Terrestrial acidification | 0 | 0 | 0 |
Marine eutrophication | 0 | 0 | 0 |
Freshwater acidification | 0 | 0 | 0 |
Freshwater eutrophication | 0 | 0 | 0 |
Particulate matter formation | 0 | 0 | 0 |
Human toxicity non-cancer | 0 | 0 | 0 |
Human toxicity cancer | 0 | 0 | 0 |
Water scarcity | 0 | 0 | 0 |
Photochemical ozone formation | 0 | 0 | 0 |
Ionizing radiations | 0 | 0 | 0 |
Ozone layer depletion | 0 | 0 | 0 |
Analysis of Divergences
While the majority of categories show a near-perfect match (<1% difference), Earthster intentionally differs from the benchmarks in the Fossil and nuclear use and Mineral resources use categories.
1. Fossil Fuel Energy Content
The 79.9% average divergence in Fossil Energy Use is driven primarily by an error in the benchmark, whereas Uranium ore is being characterized as if it were pure Uranium. This leads by severe overestimation of impacts whenever the exchange Uranium is present. In addition to that, it is also due to differing assumptions regarding the energetic content of raw fuels, but on a much lower scale. The reference benchmark utilizes values that appear to be at the lower range of physical likelihood for several key resources.
Benchmark Values: The benchmark assigns 19.1 MJ/kg to bituminous and anthracite coal. This is significantly lower than the standard minimum thresholds usually observed for these fuel types in global energy reporting.
Earthster Approach: We utilize Higher Heating Values (HHV) that reflect more typical industrial ranges (e.g., 32.6 MJ/kg for anthracite). This prevents a systematic underestimation of the total energy extraction in the model.
Natural Gas Liquids (NGL): Additionally, Earthster characterizes NGLs, which are currently omitted in the reference benchmark results.
2. Mineral Mass-Balance Corrections
The 11.2% divergence (with a 100% maximum) is due to a significant "Ore vs. Metal" discrepancy in the reference benchmark’s logic.
Over Characterization in Benchmark: For Platinum Group Metals (PGMs), the benchmark erroneously applies pure-metal characterization factors to the total weight of the gross ore (like for Uranium above). This inflates the impact of these minerals by several orders of magnitude.
Earthster’s Physical Realism: We apply global average concentration factors to map only the actual metal content within the ore. This ensures that the resource scarcity results reflect the physical depletion of the metal rather than the weight of the surrounding rock.
Completeness: The benchmark omits Chromium, Manganese, and Fertilizer minerals n.e.c. Earthster includes these (using a Phosphate proxy for fertilizers) to ensure a complete representation of global material flows.