Abstract. The study of past sedimentary records from coastal margins allows us to reconstruct variations in terrestrial input into the marine realm and to gain insight into continental climatic variability. There are numerous organic proxies for tracing terrestrial input into marine environments but none that strictly reflect the input of river-produced organic matter. Here, we test the fractional abundance of the C 32 alkane 1,15-diol relative to all 1,13-and 1,15-long-chain diols (F C32 1,15 ) as a tracer of input of river-produced organic matter in the marine realm in surface and Quaternary (0-45 ka) sediments on the shelf off the Zambezi and nearby smaller rivers in the Mozambique Channel (western Indian Ocean). A Quaternary (0-22 ka) sediment record off the Nile River mouth in the eastern Mediterranean was also studied for long-chain diols. For the Mozambique Channel, surface sediments of sites most proximal to Mozambique rivers showed the highest F 1,15−C32 (up to 10 %). The sedimentary record shows high (15-35 %) pre-Holocene F 1,15−C32 and low (< 10 %) Holocene F 1,15−C32 values, with a major decrease between 18 and 12 ka. F 1,15−C32 is significantly correlated (r 2 = 0.83, p < 0.001) with the branched and isoprenoid tetraether (BIT) index, a proxy for the input of soil and riverproduced organic matter in the marine environment, which declines from 0.25 to 0.60 for the pre-Holocene to < 0.10 for the Holocene. This decrease in both F C32 1,15 and the BIT is interpreted to be mainly due to rising sea level, which caused the Zambezi River mouth to become more distal to our study site, thereby decreasing riverine input at the core location. Some small discrepancies are observed between the records of the BIT index and F C32 1,15 for Heinrich Event 1 (H1) and the Younger Dryas (YD), which may be explained by a change in soil sources in the catchment area rather than a change in river influx. Like for the Mozambique Channel, a significant correlation between F C32 1,15 and the BIT index (r 2 = 0.38, p < 0.001) is observed for the eastern Mediterranean Nile record. Here also, the BIT index and F C32 1,15 are lower in the Holocene than in the pre-Holocene, which is likely due to the sea level rise. In general, the differences between the BIT index and F C32 1,15 eastern Mediterranean Nile records can be explained by the fact that the BIT index is not only affected by riverine runoff but also by vegetation cover with increasing cover leading to lower soil erosion. Our results confirm that F C32 1,15 is a complementary proxy for tracing riverine input of organic matter into marine shelf settings, and, in comparison with other proxies, it seems not to be affected by soil and vegetation changes in the catchment area.