Spectral composition priors & sources

每一个数字都有出处,或者明确写着「没有出处」。这是 AI 文献抽取层的第一个可核验产物:把论文里的「光谱类型 → 成分/密度先验」结构化,并与本项目的估值引擎逐项对齐。

这些是「类型先验」,不是对任何一颗具体小行星的实测。 绝大多数小行星从未测过成分。本站页面上的「Assumed water / metal content」来自这张表, 属于按光谱类型填的占位值。三种来源等级:verified(DOI 已用 CrossRef 核验标题)、 institutional(机构一手数据源)、 documented-only(站内方法页有记载但 DOI 未能核实,只作线索、不冒充引用)。 未核实到的 DOI 一律留空,绝不编造。

1 · 来源清单

键出处与核验状态
garenne2014 verified
Garenne et al. (2014), The abundance and stability of 'water' in type 1 and 2 carbonaceous chondrites, Geochimica et Cosmochimica Acta
doi:10.1016/j.gca.2014.03.034 ↗
CrossRef 核验:The abundance and stability of "water" in type 1 and 2 carbonaceous chondrites · Geochimica et Cosmochimica Acta · 2014
lee2023 verified
Lee et al. (2023), The water content of CM chondrites, Meteoritics & Planetary Science
doi:10.1111/maps.14099 ↗
CrossRef 核验:The water content of CM chondrites · Meteoritics & Planetary Science · 2023
carry2012 verified
Carry (2012), Density of asteroids, Planetary and Space Science, Table 3
doi:10.1016/j.pss.2012.03.009 ↗
CrossRef 核验:Density of asteroids · Planetary and Space Science · 2012
CORRECTED 2026-11-07: an earlier internal note pointed at 10.1016/j.icarus.2012.04.012, whose real title is about Io's thermal emission — unrelated. Verified via CrossRef before use.
elvis2013 documented-only
Elvis (2013), asteroid ore / PGM percentiles, Table 1 (as cited in METHODOLOGY.md)
无可核验 DOI — 仅列为线索
Cited in our methodology for iron-meteorite PGM percentiles (50th 6.8 ppm, 10th–90th 0.17–75) and the 4500 kg/m3 M-type working density. A verifiable DOI could NOT be found via CrossRef on 2026-11-07. Flagged as a lead, not a citation.
buseck1977 documented-only
Buseck (1977), pallasite metal fraction, via Elvis 2013 (as cited in METHODOLOGY.md)
无可核验 DOI — 仅列为线索
Pallasite total iron 50 wt% (38–70). Not independently DOI-verified.
kargel1994 documented-only
Kargel (1994), PGM = 7 x Ir relation (as cited in METHODOLOGY.md)
无可核验 DOI — 仅列为线索
wustl institutional
Washington University / RELAB spectral library and meteorite metal grades
https://speclib.rsl.wustl.edu/ ↗
nasaAntmet institutional
NASA Antarctic Meteorite Collection (curation data)
https://curator.jsc.nasa.gov/antmet/ ↗
usgsMcs institutional
USGS Mineral Commodity Summaries 2025 (Earth metal prices)
https://pubs.usgs.gov/periodicals/mcs2025/ ↗

2 · 按光谱型的密度先验

ρ_bulk,单位 g/cm³,方括号内为范围。来源见上表。

光谱型陨石类比默认 [范围]来源
C CI/CM chondrite1.4 [1.2–2.9] carry2012
B CO/CV/CK/CR chondrite2.4 [2.1–2.9] carry2012
D no firm analog1.5 [1.2–2] carry2012
P no firm analog1.8 [1.2–2.6] carry2012
S, Q ordinary chondrite2.7 [2–3.2] carry2012
V HED achondrite3 [2.5–3.6] carry2012
M iron meteorite4.5 [3.5–4.9] elvis2013
X M/E/P degenerate3 [1.8–4.9] carry2012
E enstatite achondrite2.6 [2.3–3] carry2012
A pallasite3.7 [3.2–4.2] carry2012
U unknown2 [1.2–4.9] 无先验

3 · 成分品位先验

「跨带」= 引擎用一个跨越 CI 与 CM 的单一区间,而非逐子类给值;这是方法页写明的设计,不是口径漂移。

光谱型资源类比默认 [范围]来源
C water (跨带) CI chondrite 15 [10–20] wt% garenne2014
C water (跨带) CM chondrite 7 [1.9–10.5] wt% lee2023
B water carbonaceous-dry 2 [0–5] wt% 无 DOI
S, Q metal ordinary chondrite (H/L/LL) 10 [3–20] wt% wustl
M metal iron meteorite 90 [70–95] wt% wustl
X metal M/E/P degenerate 45 [0–95] wt% wustl
A metal pallasite 50 [38–70] wt% buseck1977
M pgm_ppm iron (IIIAB) 6.8 [0.17–75] ppm elvis2013
X pgm_ppm X-complex 3 [0–75] ppm elvis2013

4 · 文献约束实例(有返回样品的天体)

把已发表的样品分析结果叠加在类型先验之上,会改变什么?答案:改变证据等级和定性结论,绝不产生新的百分比或金额。样品论文证实含水/有机质存在,但没有给出可替换查表值的整体品位;把「检出 11 种蒸发盐矿物」换算成「含水 X%」会是编造。详见 Bennu / Ryugu 页面上的对照框。

天体证据等级返回样品改变没有改变
Bennu L1 → L3 OSIRIS-REx, delivered 2023-09-24, 121.6 g Evidence tier L1 → L3; water confirmed as a real (not assumed) resource; organic chemistry added as a new dimension; valuation policy unchanged (engine still refuses a water price). No water/metal percentage, no notional dollar figure, no engine input. The 8% / 3% placeholders are explicitly retained and labelled.
Ryugu L1 → L3 Hayabusa2, returned 2020-12-06, 5.4 g Evidence tier L1 → L3; the spectral-class family assignment is now confirmed rather than assumed; organic chemistry added; valuation policy unchanged. No water/metal percentage, no notional dollar figure, no engine input.

机器可读契约:literature-constraints.json。未被收录的天体(如 Itokawa 无可核验成分论文、Psyche 2029 年才到)在契约里明确写了排除理由。

5 · 明确不填的类

D, T, O — Only ~12 of 24 Bus-DeMeo classes have a credible mineralogical interpretation. These have no defensible composition prior and are NOT filled with a guess.

6 · 下载与引用

机器可读版本:spectral-priors.csv(20 行,含 DOI、来源等级、置信度与备注)。 JSON 契约:/public/data/spectral-priors.json。

本表只适用于已有实测光谱型的天体。没有光谱型的 13.8 万颗如何给区间、反照率能推断到什么程度,见 反照率判据方法页(71.6% 准确率、M 型零样本,只降权不认定)。

引用:Helios Prospects (2026). Spectral composition priors with verifiable provenance. https://www.heliosprospects.com/priors.html

与引擎的一致性由 _priors-verify.mjs 锁定:改动任一侧而不同步,门禁即失败。方法说明见 methodology。