Juniper rings (Qilian Mountains, 1,400 km west of Erlitou): 1561-1546 BC was somewhat dry, with sharp dry years in 1558 and 1548 BC, from only 12 cores

scripts/xia_rings.py on Yang et al.'s raw measurements. Each core is compared with its own 31-year running mean and the cores are averaged. Check: after 1500 BC this simple index matches the published chronology at r = 0.97. Before 1500 BC only 11-12 cores cover each year. Against 1800-1300 BC: the 16 years 1561-1546 BC rank 50th driest of 486 (top 10 percent); 1558 BC (z -2.1) and 1548 BC (z -2.2) are among the 26 years (5 percent) at z -2 or lower; 1541 BC (z -3.4) was drier still. Eruption test over all 11 eruptions of 1678-1539 BC: rings are narrow two years after (mean z -0.59, chance 0.03), but seven lags were tried, so about 0.2 after allowing for that. Calibration: 1626 BC (z -2.5) is two years after Aniakchak, 1628 BC. Caveats: the files count years astronomically (they have a year 0), yet the paper calls the file's driest year, -1049, '1048 BC', so BC labels may be off by up to two years; these trees answer to local rain, which the authors say is not tied to the monsoon, so they cannot speak for the Yi and Luo valley.