Back past 969, not through a Roman cloud
The 2025 paper starts in 969 CE. The Mann AMO file starts in 500 CE, so the same cloud equation can be run for those extra centuries. It cannot be run through the Roman Warm Period. That interval is taken here as 1–300 CE, and the Atlantic predictor does not exist yet. Solar irradiance does. A blank cloud cell is the honest comparison.
Earliest year with the Mann index. 500–968 is an extension. 969–2022 is the existing atlas, not a rerun.
No cloud fraction. Steinhilber solar anomaly 0.18 W m⁻², close to the modern era and a bit above the medieval anomaly.
Marine studies find warm Roman Mediterranean surface water. That is sea temperature, not this cloud model, and it was not the same in every province.
Cloud fraction from 500 CE
Solid line only where the AMO file exists. The gap before 500 is the Roman centuries, left empty on purpose. Shaded bands: Late Antique Little Ice Age, the paper’s medieval and Little Ice Age windows, and the modern era.
From 500 to 968 the cloud curve uses the same equation and the real Mann AMO and Steinhilber irradiance. Soil moisture is a bridge into the medieval envelope, not the Old World Drought Atlas. From 969 onward the curve is the existing atlas series, not a rerun.
Solar irradiance, including the Roman centuries
Steinhilber total solar irradiance anomaly, decadal means, from 300 BCE. The Roman Warm Period band is 1–300 CE. This is the series that actually crosses that period. It is not a substitute for cloud cover.
The comparison
Cloud and AMO are blank before 500. Solar is Steinhilber’s anomaly in W m⁻². Little Ice Age cloud is still the high mark of the years we can actually compute.
| Period | Years | Solar | AMO | Cloud |
|---|---|---|---|---|
| Roman Warm Period | 1 CE – 300 CE | 0.183 | — | — |
| No Atlantic index | 301 CE – 499 CE | 0.149 | — | — |
| Late Roman, before 536 | 500 CE – 535 CE | 0.246 | 0.058 | 46.3% |
| Late Antique Little Ice Age | 536 CE – 660 CE | -0.244 | 0.085 | 46.0% |
| Early medieval | 661 CE – 968 CE | 0.068 | 0.066 | 45.2% |
| Medieval Climate Anomaly | 969 CE – 1249 CE | 0.117 | 0.131 | 43.2% |
| Little Ice Age | 1250 CE – 1849 CE | -0.126 | -0.115 | 49.1% |
| Modern Warming Era | 1850 CE – 2022 CE | 0.161 | 0.024 | 43.4% |
Roman Warm Period
Usual extra-tropical window (about 1–300 CE). Mann’s AMO file has not started. No cloud fraction is computed.
No Atlantic index
Solar irradiance only. The AMO series used here begins in 500.
Late Roman, before 536
First decades of the Mann file. Same cloud equation as the atlas, with a late-antique moisture bridge. Not in the 2025 paper.
Late Antique Little Ice Age
Büntgen and others place a cold interval here, opened by the 536 and 540 eruptions. Those sulfate spikes are not in this volcanic list, so the cloud curve does not contain them.
Early medieval
Approach to the paper’s start in 969. Moisture is eased toward the medieval drought so the join is not a cliff.
Why the Roman sky stays blank
The cloud equation needs the Mann et al. 2009 AMO series. That file starts in 500 CE. The Roman Warm Period, taken here as 1–300 CE, is earlier. Solar irradiance (Steinhilber) does cover it. Inventing a cloud fraction there would be a guess, so it is left blank. Mediterranean marine syntheses, including Margaritelli et al. 2020, describe a warm Roman surface-water phase in this basin. That is a sea-temperature result, not a cloud reconstruction, and it is not uniform across the whole empire.
Paper window from 969 · Where the Atlantic index actually starts