TREATMENT METHODS FOR BULK SOIL GEOCHEMICAL PROXY APPLICATIONS: LESSONS LEARNED FROM EARLY MIOCENE PALEOSOLS OF THE LAKE VICTORIA REGION

In the past two decades numerous paleosol-specific proxies have been utilized for paleoenvironmental reconstructions. In particular there has been a diversification of elemental geochemical proxies for estimating paleoprecipitation (MAP). However, there is no standard treatment plan for the removal of calcite: some papers report removal of the mineral, whereas others include it before the proxy is applied. Here we present data from Calcisols (paleosols containing a carbonate horizon) and Protosols (paleo-Inceptisols) from the Early Miocene Wayondo and Hiwegi Formations on Rusinga and Mfangano Islands, Lake Victoria, Kenya that demonstrate how different treatments methods can dramatically affect MAP estimates. The Hiwegi Fm. Protosols have mostly diagenetic calcite, and the Wayondo Fm. Calcisols have a mixture of pedogenic nodules and diagenetic cement. Before determining elemental geochemistry, half of the matrix from the Hiwegi and Wayondo Fms. paleosols was treated to remove calcite and the other half was left untreated to retain all of the calcite. The CALMAG and CIA-K bulk-geochemistry proxies were then used to estimate MAP. MAP estimates for untreated Calcisols from the Wayondo Fm. using CALMAG ranged between 0-279 mm/yr, whereas estimates from the same Calcisols treated to remove calcite ranged between 315-472 mm/yr. Similarly, untreated MAP estimates using CIA-K were between 222-606 mm/yr, whereas those from the same paleosols treated to remove calcite were between and 731-881 mm/yr. MAP estimates for untreated Protosols from the Hiwegi Fm. using CALMAG ranged between 0-313 mm/yr, whereas estimates from the same Protosols ranged between 283 and 444 mm/yr when treated to remove calcite. MAP estimates from the same paleosols using CIA-K range between 225-642 mm/yr when untreated and between 826-867 mm/yr when treated to remove calcite. The considerable differences in estimates between untreated and treated paleosols result in vastly different paleoecological reconstruction. Thus, this study indicates the importance in assessing the treatment methods of paleosols in the development and application of bulk geochemistry proxies for MAP as well as the types of soils that appropriate for paleoprecipitation studies.