"""Build SOHO/VIRGO TSI charts from the full-mission Version 8.0 daily file. Source: NASA SOHO archive VIRGO_TSI_daily_V8.txt (1996-2021) Reference: Finsterle et al. 2021, Sci Rep 11, 7835 https://doi.org/10.1038/s41598-021-87108-y Attribution: VIRGO Experiment on the cooperative ESA/NASA Mission SoHO, VIRGO Team / PMOD/WRC, Davos, Switzerland """ from pathlib import Path import matplotlib.dates as mdates import matplotlib.pyplot as plt import pandas as pd BASE_DIR = Path(__file__).parent INPUT_FILE = BASE_DIR / "tsi_raw" / "soho_virgo_v8" / "raw" / "VIRGO_TSI_daily_V8.txt" OUT_DIR = BASE_DIR / "tsi_satellite_sources" OUT_PNG = OUT_DIR / "soho_virgo_v8_tsi_daily_monthly_annual.png" OUT_CSV = OUT_DIR / "soho_virgo_v8_tsi_daily.csv" OUT_MANIFEST = OUT_DIR / "soho_virgo_v8_tsi_manifest.txt" def load_daily() -> pd.DataFrame: """Parse VIRGO V8 daily file; use col 8 (fused new SI-traceable scale, ~1361 W/m²).""" rows = [] for line in INPUT_FILE.read_text(encoding="latin-1").splitlines(): line = line.strip() if not line or ";" in line: # skip blanks and all comment/header lines continue parts = line.split() if len(parts) < 8: continue try: date = pd.to_datetime(parts[0].split("T")[0], format="%Y-%m-%d") tsi_new = float(parts[7]) # col 8: VIRGO A+B fused, new SI scale tsi_org = float(parts[4]) # col 5: VIRGO A+B fused, original WRR scale unc = float(parts[5]) # col 6: uncertainty of fused series except (ValueError, IndexError): continue if not (1200.0 < tsi_new < 1500.0): continue rows.append((date, tsi_new, tsi_org, unc)) frame = pd.DataFrame(rows, columns=["date", "tsi", "tsi_orig", "uncertainty"]) return frame.sort_values("date").reset_index(drop=True) def build_chart(frame: pd.DataFrame) -> None: monthly = frame.set_index("date")["tsi"].resample("MS").mean().reset_index(name="tsi") annual = frame.set_index("date")["tsi"].resample("YS").mean().reset_index(name="tsi") annual["plot_date"] = annual["date"] + pd.offsets.Day(181) plt.style.use("seaborn-v0_8-whitegrid") fig, ax = plt.subplots(figsize=(12.5, 6.8), dpi=170) ax.scatter(frame["date"], frame["tsi"], s=6, color="#67e8f9", alpha=0.30, edgecolors="none", label="Daily values (V8 fused, new SI scale)") ax.plot(monthly["date"], monthly["tsi"], color="#0891b2", linewidth=1.8, label="Monthly means") ax.plot(annual["plot_date"], annual["tsi"], color="#155e75", linewidth=2.4, marker="o", markersize=3.4, label="Annual means") ax.set_title("SOHO/VIRGO TSI — Version 8.0 Full Mission (1996–2021)", fontsize=15, pad=14) ax.set_ylabel("Total Solar Irradiance (W m^-2)") ax.set_xlabel("Year") ax.xaxis.set_major_locator(mdates.YearLocator(2)) ax.xaxis.set_major_formatter(mdates.DateFormatter("%Y")) ax.set_xlim(frame["date"].min(), frame["date"].max()) ax.legend(loc="upper right", frameon=True) note = ( "Source: VIRGO_TSI_daily_V8.txt — NASA SOHO archive (soho.nascom.nasa.gov)\n" "Plotted: col 8 — VIRGO A+B fused, new SI-traceable scale (~1361 W/m²)\n" "Attribution: Finsterle et al. 2021, Sci Rep 11, 7835 | PMOD/WRC / VIRGO Team" ) ax.text(0.015, 0.03, note, transform=ax.transAxes, fontsize=8.5, bbox=dict(boxstyle="round,pad=0.35", facecolor="white", edgecolor="#cbd5e1", alpha=0.96)) fig.tight_layout() fig.savefig(OUT_PNG, bbox_inches="tight") plt.close(fig) def write_outputs(frame: pd.DataFrame) -> None: OUT_DIR.mkdir(parents=True, exist_ok=True) daily = frame.copy() daily["date"] = daily["date"].dt.strftime("%Y-%m-%d") daily.to_csv(OUT_CSV, index=False) monthly_rows = frame.set_index("date")["tsi"].resample("MS").mean().shape[0] annual_rows = frame.set_index("date")["tsi"].resample("YS").mean().shape[0] lines = [ "SOHO/VIRGO TSI Version 8.0 manifest", "", "Dataset: VIRGO Total Solar Irradiance daily, Version 8.0", "Instrument: VIRGO (PMO6-V + DIARAD, A+B fused) on SOHO", "Platform: SOHO (ESA/NASA cooperative mission)", "Publisher: NASA SOHO archive (soho.nascom.nasa.gov)", "Source URL: https://soho.nascom.nasa.gov/data/EntireMissionBundles/VIRGO_TSI_daily_V8.txt", "Raw file: tsi_raw/soho_virgo_v8/raw/VIRGO_TSI_daily_V8.txt", "", f"Date span: {frame['date'].min().date()} to {frame['date'].max().date()}", f"Daily rows (valid): {len(frame)}", f"Monthly bins: {monthly_rows}", f"Annual bins: {annual_rows}", f"TSI range (new scale, col 8): {frame['tsi'].min():.4f} to {frame['tsi'].max():.4f} W m^-2", f"TSI range (orig scale, col 5): {frame['tsi_orig'].min():.4f} to {frame['tsi_orig'].max():.4f} W m^-2", "", "Attribution (required):", " Finsterle, W., Montillet, J.P., Schmutz, W. et al.", " The total solar irradiance during the recent solar minimum period measured by SOHO/VIRGO.", " Sci Rep 11, 7835 (2021). https://doi.org/10.1038/s41598-021-87108-y", ' Acknowledge: "VIRGO Experiment on the cooperative ESA/NASA Mission SoHO', ' from VIRGO Team through PMOD/WRC, Davos, Switzerland"', "", "Scale note:", " Plotted series uses the NEW SI-traceable scale (col 8, ~1361 W/m^2).", " The original WRR scale (col 5, ~1365 W/m^2) is also exported in the CSV.", " Scales differ by ~0.34% / ~4.6 W/m^2 — see Finsterle et al. 2021 Table 2.", ] OUT_MANIFEST.write_text("\n".join(lines) + "\n") def main() -> None: if not INPUT_FILE.exists(): raise FileNotFoundError(f"Missing input: {INPUT_FILE}") frame = load_daily() if frame.empty: raise ValueError("No valid VIRGO V8 rows after filtering") build_chart(frame) write_outputs(frame) print("saved", OUT_PNG.name) print("saved", OUT_CSV.name) print("saved", OUT_MANIFEST.name) print("daily_rows", len(frame)) print("date_span", frame["date"].min().date(), "to", frame["date"].max().date()) print("tsi_range (new)", round(frame["tsi"].min(), 4), "to", round(frame["tsi"].max(), 4)) if __name__ == "__main__": main()