Apportionment of Okinawa fall
The aerosol of Okinawa fall is also apportioned satisfactorily by two threads that are short, similar, and arrive at the same solution. There are shown in the table below and illustrated on this page.
| Okinawa fall, 16 elements | ||
| Step 0 | Nothing chosen | |
| Step 2 | Sea and crust chosen, fit 3.45 | |
| Kenting thread | Mallipo thread | |
| Step 3 | Kenting added, fit 0.48 | Mallipo May added, fit 0.73 |
| Step 4 | Lechang added, fit 0.28 | Kenting added, fit 0.45 |
| Step 5 | Lechang added, Mallipo May del., fit 0.28 | |
Step 0, nothing chosen, shows generally high F-ratios in East Asia, the Atlantic, and parts of the Pacific. The highest value is for Kenting.
Step 2, after sea and the best crust (FEC) are chosen, shows the highest F-ratios consistently in East Asia and Japan/Korea. The highest value remains Kenting. This step shows that the aerosol of Okinawa fall clearly comes from East Asia.
Step 3 of the Kenting thread adds Kenting. The fit improves to a respectable 0.48. Lechang has the next-highest F-ration in east Asia. The other high ratios in North America and Europe are not to be considered because they correspond to negative coefficients in the apportionment.
Step 4 of the Kenting thread adds Lechang. This improves the fit to a very good 0.28. The thread stops here because the other contributions are negative, as in the previous step.
The other thread begins with Mallipo May, the highest F-ratio in Japan/Korea. Adding Mallipo improves the fit to 0.73, which is not very good. No other good sources in Japan/Korea remain.
Step 4 in the Mallipo thread adds Kenting, the place with the highest remaining F-ratio. It improves the fit to 0.45. Again, the high F-ratios in North America and Europe are not to be considered because they correspond to negative coefficients. We also disregarded the highest value in EAs because it is for Shanghai, whose data we have reason to distrust.
Step 5 of the Mallipo thread adds Lechang, the second-highest remaining F-ratio in EAs, and drops Mallipo May because its F-ratio drops to a negligible 0.023. This makes the solution the same as the other thread: Kenting and Lechang.
