Amazon plant diversity and climate through the Cenozoic

The modern Amazon lowlands date from the mid-Cenozoic, when the Andean orogeny defined an eastward, or northward draining basin of immense size. As the Andean orogeny continued throughout the Neogene and Quaternary periods, this system approached a steady state, with Andean flanks eroding to fill subsiding basins along the Andean foreland, followed by secondary erosion of sediments and transport in rivers across >3000 km to the Atlantic Ocean. Pollen data suggest that tropical forests were established on these lowlands at least by the end of the Miocene, with many modern Amazon genera already present. Such antiquity is consistent with hypotheses that explain high diversity of modern Amazon forests as the result of prolonged existence and minimal disturbance or extinction. These ‘museum’ models for Amazon diversity came to be challenged by the paradigm of ice-age aridity, with its claim that the forests of the Amazon lowlands were fragmented with each glacial advance into more or less tiny isolates (called refugia) separated by grass-dominated communities like savanna. For the latest Pleistocene, the arid Amazon with forest refugia model can be tested by pollen analysis of lake sediments. Interpretation of Amazon pollen data requires recognition both of the vast diversity of pollen types and an understanding of the dispersal of zoophilous pollen in the Amazon. Amazon pollen analysis has been codified into a pollen atlas and this work used to review the available pollen data. Within the huge expanse of the Amazon basin only three lacustrine records long enough and continuous enough to test this model have been published, those from Lake Pata, the Carajas Plateau and the Amazon fan. The review suggests community stability from all three data sets, with no evidence for forest fragmentation in the Amazon lowlands. The principal response of Amazon forest communities to climatic change is in the relative abundance of the more sensitive taxa, particularly in response to interglacial warming. A summary of geomorphological phenomena finds no evidence for glacial age aridity other than marginal reductions in precipitation within forest cover. The Amazon historical record is incomplete in time as well as in space, there being essentially no paleobotanical data for much of the Cenozoic, particularly for the early Pleistocene. The data so far available, however, suggest stability and continuity of the forest cover over the vast Amazon lowlands throughout the Cenozoic as the circumstance in which the high Amazon plant diversity evolved.