面向物理系本科一年级 · 可执行教程
教程目标：理解 Grayson+2025 的 5 个 simulation feedback models（EAGLE, EAGLE-AGNdT9, EAGLE-NoAGN, SIMBA, SIMBA-NoAGN）如何通过 synthetic X-ray 观测转换成 Figure 3 上的 5 个离散 conditional template 点。 Tutorial goal: Understand how 5 simulation feedback models from Grayson+2025 (EAGLE, EAGLE-AGNdT9, EAGLE-NoAGN, SIMBA, SIMBA-NoAGN) become 5 discrete conditional template points on Figure 3 through synthetic X-ray observations.
目标读者：物理系本科一年级，已学完普通物理（电磁学/光学），了解基本的原子物理概念（能级、跃迁），但不需要天文观测经验。 Target audience: First-year physics undergraduates who have completed general physics (electromagnetism/optics) and basic atomic physics, without requiring astronomical observing experience.
核心问题：Grayson+2025 用 EAGLE 和 SIMBA 宇宙学模拟的 z=0.1 snapshots，通过 pyXSIM+SOXS 生成 synthetic eROSITA X-ray profiles，与 Zhang 的 observed stacks 比较。5 个模型的唯一区别是 AGN feedback 的开关和强度——同一个 galaxy sample，不同的反馈物理，预测的 X-ray 亮度差了一个量级以上。这些模型如何变成 Figure 3 上的 5 个点？
本教程只需要 Python 标准科学栈。所有数据文件已随教程提供。
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from pathlib import Path
plt.style.use('seaborn-v0_8-whitegrid')
plt.rcParams['font.size'] = 12
plt.rcParams['figure.dpi'] = 150
plt.rcParams['savefig.dpi'] = 300
# NOTE: do NOT add a CJK font fallback here — see SKILL.md
# All matplotlib text labels stay English-only.
DATA = Path("assets/data")
print("Environment ready!")Environment ready!星系中心超大质量黑洞（SMBH, \(M \sim 10^6 - 10^9 M_\odot\)）在吸积物质时会释放大量能量，称为活动星系核反馈（AGN feedback）。这股能量可以加热或驱散星系周围的 CGM（circumgalactic medium）气体，显著改变其 X 射线亮度。
Grayson+2025 使用两大宇宙学模拟框架：
| 框架 | 全称 | AGN Feedback 模型 |
|---|---|---|
| EAGLE | Evolution and Assembly of GaLaxies and their Environments | Single-mode thermal injection |
| SIMBA | Simulating the Intergalactic Medium with Black holes and AGN | Two-mode: wind + jet |
| 模型 | AGN Feedback 描述 | 预期 X-ray 亮度 |
|---|---|---|
| EAGLE (Ref) | \(\Delta T_\mathrm{AGN}=10^{8.5}\) K single-mode thermal feedback | 参考水平 |
| EAGLE-AGNdT9 | \(\Delta T_\mathrm{AGN}=10^{9}\) K — 更强的热反馈 | 更亮（更多热气体） |
| EAGLE-NoAGN | 关闭全部 AGN feedback | 更亮（无 feedback 驱散气体） |
| SIMBA | High-Eddington wind + Low-Eddington jet | 参考水平 |
| SIMBA-NoAGN | 关闭全部 AGN feedback | 极亮（气体未被驱散） |
关键洞察：关闭 AGN feedback 后，气体未被加热/驱散，冷却后密度更高 → X 射线亮度大幅上升。NoAGN 模型是 feedback sensitivity experiment——不能再现 stellar-mass function，但能告诉我们 feedback 对 CGM X-ray 亮度的敏感程度。
graph TD
A["EAGLE/SIMBA z=0.1 snapshots<br/>M31-mass stellar bin"] --> B["pyXSIM: 3D gas → photon list"]
B --> C["SOXS: eROSITA instrument simulator"]
C --> D["Synthetic 0.5-2 keV<br/>surface-brightness profile<br/>(arXiv v2 Fig. 7)"]
D --> E["Digitize 10-30 kpc<br/>intrinsic luminosity-density"]
E --> F["Distance-cancel<br/>surface-brightness conversion"]
F --> G["v19 APEC absorbed/intrinsic ratio<br/>(T_0.5-2.0 = 0.678988916)"]
G --> H["Figure 3: 5 discrete<br/>conditional template points"]graph TD
A["EAGLE/SIMBA z=0.1 snapshots<br/>M31-mass stellar bin"] --> B["pyXSIM: 3D gas → photon list"]
B --> C["SOXS: eROSITA instrument simulator"]
C --> D["Synthetic 0.5-2 keV<br/>surface-brightness profile<br/>(arXiv v2 Fig. 7)"]
D --> E["Digitize 10-30 kpc<br/>intrinsic luminosity-density"]
E --> F["Distance-cancel<br/>surface-brightness conversion"]
F --> G["v19 APEC absorbed/intrinsic ratio<br/>(T_0.5-2.0 = 0.678988916)"]
G --> H["Figure 3: 5 discrete<br/>conditional template points"]
从 m31_cgmsum_conditional_prior_ledger.csv 提取 Grayson+2025 的 5 个模型记录。
ledger = pd.read_csv(DATA / "m31_cgmsum_conditional_prior_ledger.csv")
# Filter for Grayson+2025 models
grayson_ids = [
"grayson2025_eagle_agndt9",
"grayson2025_eagle",
"grayson2025_eagle_noagn",
"grayson2025_simba",
"grayson2025_simba_noagn",
]
grayson = ledger[ledger["prior_id"].isin(grayson_ids)].copy()
print(f"Loaded {len(grayson)} Grayson+2025 models")
print(f"Columns: {list(grayson.columns)}")
grayson[["prior_id", "original_central", "original_low", "original_high",
"primary_central", "primary_low", "primary_high",
"figure_central", "figure_low", "figure_high"]]Loaded 5 Grayson+2025 models
Columns: ['prior_id', 'component', 'label', 'reference', 'doi', 'original_quantity', 'original_central', 'original_low', 'original_high', 'original_units', 'conversion_to_primary', 'primary_central', 'primary_low', 'primary_high', 'scope', 'caveat', 'provenance', 'parameter_error_low', 'parameter_error_high', 'north_primary', 'south_primary', 'conversion_check_low', 'conversion_check_high', 'primary_central_definition', 'primary_interval_definition', 'all_field_inverse_variance', 'figure_band', 'figure_central', 'figure_low', 'figure_high', 'figure_central_definition', 'figure_interval_definition', 'figure_all_field_inverse_variance', 'figure_parameter_error_low', 'figure_parameter_error_high', 'figure_north', 'figure_south', 'figure_conversion_check_low', 'figure_conversion_check_high', 'figure_context_low', 'figure_context_high']| prior_id | original_central | original_low | original_high | primary_central | primary_low | primary_high | figure_central | figure_low | figure_high | |
|---|---|---|---|---|---|---|---|---|---|---|
| 7 | grayson2025_eagle_agndt9 | 4.704477e+35 | 4.184141e+35 | 5.153521e+35 | 0.256815 | 0.228411 | 0.281329 | 0.225900 | 0.200914 | 0.247462 |
| 8 | grayson2025_eagle | 9.381776e+35 | 8.236145e+35 | 1.041199e+36 | 0.512147 | 0.449608 | 0.568386 | 0.450494 | 0.395483 | 0.499963 |
| 9 | grayson2025_eagle_noagn | 1.970985e+36 | 1.685815e+36 | 2.187420e+36 | 1.075952 | 0.920279 | 1.194103 | 0.946427 | 0.809495 | 1.050355 |
| 10 | grayson2025_simba | 3.405962e+36 | 2.838274e+36 | 3.982109e+36 | 1.859300 | 1.549402 | 2.173816 | 1.635475 | 1.362882 | 1.912129 |
| 11 | grayson2025_simba_noagn | 2.221874e+37 | 2.164747e+37 | 2.340690e+37 | 12.129117 | 11.817265 | 12.777728 | 10.668993 | 10.394682 | 11.239522 |
数据包含 5 个模型。每条记录有：
| 列名 | 含义 |
|---|---|
original_central |
digitized 的 10–30 kpc intrinsic luminosity-density（erg s\(^{-1}\) kpc\(^{-2}\)） |
primary_central |
经过 distance-cancel surface-brightness 转换后的值 |
figure_central |
再经过 v19 APEC absorbed/intrinsic ratio 后的 Figure 3 值（flux unit） |
conversion_to_primary |
统一转换因子（\(5.45896 \times 10^{-37}\)） |
Grayson+2025 的 Figure 7 展示了 eROSITA synthetic 0.5–2.0 keV surface-brightness profile（单位：erg s\(^{-1}\) cm\(^{-2}\) arcmin\(^{-2}\)）。对每个模型，取 10–30 kpc 环带的均值，得到该模型的 intrinsic luminosity-density（单位：erg s\(^{-1}\) kpc\(^{-2}\)）。
这一步的 digitization 已由项目完成，结果存储在 original_central 中。
models = grayson["prior_id"].str.replace("grayson2025_", "").tolist()
labels = ["EAGLE\nAGNdT9", "EAGLE\nRef", "EAGLE\nNoAGN", "SIMBA\nRef", "SIMBA\nNoAGN"]
colors = ["#2676b8", "#2676b8", "#2676b8", "#d47a2c", "#d47a2c"]
fig, ax = plt.subplots(figsize=(8, 5))
x = np.arange(len(models))
centrals = grayson["original_central"].to_numpy()
lows = grayson["original_low"].to_numpy()
highs = grayson["original_high"].to_numpy()
errors_low = centrals - lows
errors_high = highs - centrals
bars = ax.bar(x, centrals, color=colors, alpha=0.85, edgecolor="white", linewidth=0.5)
ax.errorbar(x, centrals, yerr=[errors_low, errors_high], fmt="none",
ecolor="#333333", capsize=5, capthick=1.5, linewidth=1.5)
# Annotate values
for i, (c, lo, hi) in enumerate(zip(centrals, lows, highs)):
ax.annotate(f"{c:.2e}", (i, c), textcoords="offset points",
xytext=(0, 8), fontsize=8, ha="center", va="bottom")
ax.set_xticks(x)
ax.set_xticklabels(labels, fontsize=10)
ax.set_ylabel("Intrinsic 0.5-2.0 keV luminosity density\n(erg s^-1 kpc^-2)")
ax.set_title("Grayson+2025: Digitized 10-30 kpc intrinsic luminosity-density")
ax.set_yscale("log")
ax.grid(axis="y", alpha=0.3)
plt.tight_layout()
plt.show()
观察： - SIMBA-NoAGN 的 intrinsic luminosity 比 EAGLE-AGNdT9 高约 47 倍（\(2.22 \times 10^{37}\) vs \(4.70 \times 10^{35}\)）。 - EAGLE 系列中，NoAGN 比 Ref 亮约 2.1 倍——关闭 AGN feedback 后气体未被驱散。 - SIMBA 系列中，NoAGN 比 Ref 亮约 6.5 倍——SIMBA 的 AGN feedback 对 CGM 的影响更大。
Grayson+2025 的 synthetic profiles 是 surface-brightness（单位面积立体角上的流量），而 Figure 3 需要的是 distance-canceled 量（与 Zhang stacks 统一）。
转换公式：
\[ S_{\text{primary}} = L_{\text{intr}} \times \frac{1}{4\pi d_L^2} \times \frac{1}{\Omega_{\text{arcmin}^2}} \]
其中： - \(d_L\)：luminosity distance（z=0.1 时约 450 Mpc，但距离在转换中约掉） - \(\Omega_{\text{arcmin}^2}\)：1 arcmin\(^2\) 的立体角 = \(1 / 11,818,102.86\) sr
最终转换因子（对所有 5 个模型统一）： \[ \text{conversion_to_primary} = \frac{1}{4\pi \, (1 \text{ kpc in cm})^2} \times \frac{11,818,102.86 \text{ arcmin}^2}{\text{sr}} = 5.45896 \times 10^{-37} \]
CONV = 5.458957861988933e-37 # from the ledger
for idx, row in grayson.iterrows():
manual = row["original_central"] * CONV
ledger_val = row["primary_central"]
print(f"{row['prior_id']}: manual={manual:.6f}, ledger={ledger_val:.6f}, diff={abs(manual - ledger_val):.2e}")
assert abs(manual - ledger_val) < 1e-10, f"Mismatch for {row['prior_id']}"
print("\n✓ All 5 models match the ledger to < 1e-10")grayson2025_eagle_agndt9: manual=0.256815, ledger=0.256815, diff=5.55e-17
grayson2025_eagle: manual=0.512147, ledger=0.512147, diff=0.00e+00
grayson2025_eagle_noagn: manual=1.075952, ledger=1.075952, diff=2.22e-16
grayson2025_simba: manual=1.859300, ledger=1.859300, diff=0.00e+00
grayson2025_simba_noagn: manual=12.129117, ledger=12.129117, diff=0.00e+00
✓ All 5 models match the ledger to < 1e-10Grayson+2025 的 synthetic profiles 是 intrinsic（无吸收），但 Figure 3 需要 absorbed（经过银河系中性氢吸收后的）值。吸收转换使用 v19 APEC 等离子体模型（\(kT=0.225\) keV, \(Z=0.3\) Z\(_\odot\), \(N_\mathrm{H}=6.7 \times 10^{20}\) cm\(^{-2}\)）。
关键假设：所有 5 个 simulation 模型使用同一个 v19 APEC spectrum，而不是各自 simulation 的 multiphase spectrum。这是简化假设，但使得 5 个模型之间的差异完全归因于 intrinsic luminosity 的不同。
v19 APEC 的 0.5–2.0 keV absorbed-to-intrinsic ratio： \[ T_{0.5-2.0} \approx 0.879618233 \]
注：该 ratio 与 Zhang+2024 M31-mass stack 使用同一 v19 APEC 模板，保证 Grayson simulations 与 Zhang observed stacks 在同一吸收校正下比较。
#| label: step3-verify
T_APEC = 0.879618232586871
for idx, row in grayson.iterrows():
manual = row["primary_central"] * T_APEC
ledger_val = row["figure_central"]
print(f"{row['prior_id']}: manual={manual:.6f}, ledger={ledger_val:.6f}, diff={abs(manual - ledger_val):.2e}")
assert abs(manual - ledger_val) < 1e-10, f"Mismatch for {row['prior_id']}"
print("\n✓ All 5 models match the ledger to < 1e-10")fig_centrals = grayson["figure_central"].to_numpy()
fig_lows = grayson["figure_low"].to_numpy()
fig_highs = grayson["figure_high"].to_numpy()
fig_errors_low = fig_centrals - fig_lows
fig_errors_high = fig_highs - fig_centrals
fig, ax = plt.subplots(figsize=(8, 5))
x = range(len(models))
bars = ax.bar(x, fig_centrals, color=colors, alpha=0.85, edgecolor="white", linewidth=0.5)
ax.errorbar(x, fig_centrals, yerr=[fig_errors_low, fig_errors_high], fmt="none",
ecolor="#333", capsize=5, capthick=1.5, linewidth=1.5)
# Annotate values
for i, c in enumerate(fig_centrals):
ax.annotate(f"{c:.3f}", (i, c), textcoords="offset points",
xytext=(0, 8), fontsize=9, ha="center", va="bottom",
fontweight="bold")
# Reference lines
ax.axhline(y=0.828, color="#b74842", linestyle="--", linewidth=1.5,
label="Zhang M31 stack = 0.828 (observed)")
ax.axhline(y=0.385, color="#6f7782", linestyle=":", linewidth=1.5,
label="H&S13 MW high-lat median = 0.385")
ax.set_xticks(x)
ax.set_xticklabels(labels, fontsize=10)
ax.set_ylabel("Absorbed 0.5-2.0 keV brightness\n(10^-15 erg s^-1 cm^-2 arcmin^-2)")
ax.set_title("Grayson+2025: 5 simulation models on Figure 3")
ax.legend(fontsize=9, loc="upper left")
ax.grid(axis="y", alpha=0.3)
plt.tight_layout()
plt.show()
# Print summary
print("\nFigure 3 values (flux unit):")
for i, row in grayson.iterrows():
print(f" {row['prior_id']}: {row['figure_central']:.4f} "
f"[{row['figure_low']:.4f}, {row['figure_high']:.4f}]")
Figure 3 values (flux unit):
grayson2025_eagle_agndt9: 0.2259 [0.2009, 0.2475]
grayson2025_eagle: 0.4505 [0.3955, 0.5000]
grayson2025_eagle_noagn: 0.9464 [0.8095, 1.0504]
grayson2025_simba: 1.6355 [1.3629, 1.9121]
grayson2025_simba_noagn: 10.6690 [10.3947, 11.2395]关键观察：
print("Full conversion chain for all 5 models:")
print("=" * 80)
print(f"{'Model':<25} {'Original':>12} {'× CONV':>10} {'× T_APEC':>10} {'Figure 3':>10}")
print("-" * 80)
CONV = 5.458957861988933e-37
T_APEC = 0.879618232586871
for idx, row in grayson.iterrows():
orig = row["original_central"]
primary = orig * CONV
figure = primary * T_APEC
name = row["prior_id"].replace("grayson2025_", "")
print(f"{name:<25} {orig:>12.3e} {primary:>10.6f} {figure:>10.6f} {row['figure_central']:>10.6f}")
assert abs(figure - row["figure_central"]) < 1e-9
print("-" * 80)
print("✓ Full conversion chain verified for all 5 models")Full conversion chain for all 5 models:
================================================================================
Model Original × CONV × T_APEC Figure 3
--------------------------------------------------------------------------------
eagle_agndt9 4.704e+35 0.256815 0.225900 0.225900
eagle 9.382e+35 0.512147 0.450494 0.450494
eagle_noagn 1.971e+36 1.075952 0.946427 0.946427
simba 3.406e+36 1.859300 1.635475 1.635475
simba_noagn 2.222e+37 12.129117 10.668993 10.668993
--------------------------------------------------------------------------------
✓ Full conversion chain verified for all 5 modelsSIMBA-NoAGN 在 linear 图上 off-scale，但 log scale 可以同时展示所有模型。
#| label: step6-log
#| fig-cap: "Grayson+2025 models on log scale — SIMBA-NoAGN is 47× brighter than EAGLE-AGNdT9"
#| fig-width: 8
#| fig-height: 5
fig, ax = plt.subplots(figsize=(8, 5))
# Use scatter for log scale
for i, (model, c, color) in enumerate(zip(models, fig_centrals, colors)):
ax.scatter(i, c, color=color, s=120, zorder=5, edgecolor="white", linewidth=1.5)
ax.errorbar(i, c, yerr=[[c - fig_lows[i]], [fig_highs[i] - c]], fmt="none",
ecolor=color, capsize=5, capthick=1.5, linewidth=1.5, alpha=0.7)
# Reference
ax.axhline(y=0.828, color="#b74842", linestyle="--", linewidth=1.5,
label="Zhang M31 stack = 0.828")
ax.axhline(y=0.385, color="#6f7782", linestyle=":", linewidth=1.5,
label="H&S13 MW median = 0.385")
ax.set_xticks(x)
ax.set_xticklabels(labels, fontsize=10)
ax.set_ylabel("Absorbed 0.5-2.0 keV brightness\n(10^-5 erg s^-1 cm^-2 arcmin^-2)")
ax.set_title("Grayson+2025: 5 simulation models (log scale)")
ax.set_yscale("log")
ax.legend(fontsize=9, loc="upper left")
ax.grid(axis="y", alpha=0.3, which="both")
plt.tight_layout()
plt.show()| 步骤 | 假设 | 来源 |
|---|---|---|
| 模拟 snapshot | z=0.1 EAGLE/SIMBA, M31-mass stellar bin | Grayson+2025 |
| X-ray 模拟 | pyXSIM + SOXS → eROSITA synthetic profiles | Grayson+2025 |
| Digitization | 10–30 kpc bin from Fig. 7 | 本项目 |
| 距离 | M31 distance assumptions cancel in the conversion | Zhang+2024 方法 |
| 光谱模型 | 单一 v19 APEC（\(k=0.65\) keV, \(Z=0.3\) Z\(_\odot\)），不是各自的 multiphase | 本项目简化 |
| 吸收 | \(N_\mathrm{H}=6.7 \times 10^{20}\) cm\(^{-2}\)，phabs | HI4PI 巡天 |
| NoAGN 模型 | 不能再现 stellar-mass function；仅作 feedback sensitivity experiment | Grayson+2025 |
| SIMBA-NoAGN | 极亮（~10.67 flux unit），在 linear 图上 off-scale | 本项目观察 |
feedback suppression factor = (NoAGN brightness) / (Ref brightness)。对 EAGLE 和 SIMBA 分别计算。
#| label: ex1
#| echo: false
# Solution
eagle_ref = grayson[grayson["prior_id"] == "grayson2025_eagle"]["figure_central"].values[0]
eagle_noagn = grayson[grayson["prior_id"] == "grayson2025_eagle_noagn"]["figure_central"].values[0]
simba_ref = grayson[grayson["prior_id"] == "grayson2025_simba"]["figure_central"].values[0]
simba_noagn = grayson[grayson["prior_id"] == "grayson2025_simba_noagn"]["figure_central"].values[0]
print(f"EAGLE: NoAGN / Ref = {eagle_noagn:.4f} / {eagle_ref:.4f} = {eagle_noagn / eagle_ref:.2f}")
print(f"SIMBA: NoAGN / Ref = {simba_noagn:.4f} / {simba_ref:.4f} = {simba_noagn / simba_ref:.2f}")
print(f"\nSIMBA AGN feedback suppresses CGM X-ray brightness by a factor of {simba_noagn / simba_ref:.1f}x")
print(f"EAGLE AGN feedback suppresses CGM X-ray brightness by a factor of {eagle_noagn / eagle_ref:.1f}x")选择 SIMBA，手动用 original_central、conversion_to_primary、和 APEC ratio 计算 Figure 3 值。
#| label: ex2
#| echo: false
simba_row = grayson[grayson["prior_id"] == "grayson2025_simba"].iloc[0]
manual_primary = simba_row["original_central"] * CONV
manual_figure = manual_primary * T_APEC
print(f"SIMBA Ref:")
print(f" original_central = {simba_row['original_central']:.3e} erg s-1 kpc-2")
print(f" × conversion = {CONV:.6e}")
print(f" = primary = {manual_primary:.6f}")
print(f" × T_APEC = {T_APEC:.9f}")
print(f" = figure_central = {manual_figure:.6f}")
print(f" ledger figure = {simba_row['figure_central']:.6f}")
assert abs(manual_figure - simba_row["figure_central"]) < 1e-10
print(" ✓ Match")graph TD
A["EAGLE/SIMBA z=0.1<br/>M31-mass stellar bin"] --> B["pyXSIM: gas → photons"]
B --> C["SOXS: eROSITA simulator"]
C --> D["Fig. 7: SB profile<br/>0-10 kpc"]
D --> E["Digitize 10-30 kpc<br/>intrinsic luminosity-density"]
E --> F["÷ 4π kpc_cm²<br/>× 11,818,102.86 arcmin²/sr"]
F --> G["× T_0.5-2.0 = 0.678988916<br/>(v19 APEC, single spectrum)"]
G --> H["Figure 3: 5 discrete<br/>conditional template points"]
H --> I["EAGLE-AGNdT9: 0.226"]
H --> J["EAGLE Ref: 0.450"]
H --> K["EAGLE-NoAGN: 0.946"]
H --> L["SIMBA Ref: 1.635"]
H --> M["SIMBA-NoAGN: 10.669"]graph TD
A["EAGLE/SIMBA z=0.1<br/>M31-mass stellar bin"] --> B["pyXSIM: gas → photons"]
B --> C["SOXS: eROSITA simulator"]
C --> D["Fig. 7: SB profile<br/>0-10 kpc"]
D --> E["Digitize 10-30 kpc<br/>intrinsic luminosity-density"]
E --> F["÷ 4π kpc_cm²<br/>× 11,818,102.86 arcmin²/sr"]
F --> G["× T_0.5-2.0 = 0.678988916<br/>(v19 APEC, single spectrum)"]
G --> H["Figure 3: 5 discrete<br/>conditional template points"]
H --> I["EAGLE-AGNdT9: 0.226"]
H --> J["EAGLE Ref: 0.450"]
H --> K["EAGLE-NoAGN: 0.946"]
H --> L["SIMBA Ref: 1.635"]
H --> M["SIMBA-NoAGN: 10.669"]
一句话总结：Grayson+2025 用 EAGLE 和 SIMBA 宇宙学模拟的 5 个 AGN feedback 变体预测 M31-mass galaxies 的 CGM X-ray 亮度。经过 synthetic eROSITA 观测、distance-cancel surface-brightness 转换、和统一的 v19 APEC 吸收校正后，5 个模型在 Figure 3 上形成 5 个离散点——亮度跨度从 0.226 到 10.669 flux unit，清晰展示了 AGN feedback 对 CGM X-ray 亮度的巨大影响。
教程结束 🎓 下一步：继续阅读 Henley & Shelton 2013 population tutorial，了解高银纬 MW 背景如何变成 Figure 3 上的 population median。