The Type Of Forest

Almost all Amazonian pollen records that extend back into the Late Pleistocene reveal higher inputs of pollen from montane forest taxa—for example, Podocarpus, Alnus, and Hedyosmum. Available data suggest that these forests lack exact modern counterparts. Lowland species persisted alongside montane species during the coldest episodes, but for much of the glacial period there may have been minimal penetration of the lowlands by montane taxa. These data argue that lowland Amazon forests of the Pleistocene were mesic systems periodically stressed by strong cooling and drought.

An emerging view supports the presence of forest cover across Amazonia throughout the Quaternary, but raises the possibility that the forests were much drier than those of today and may have supported less biomass (Prado and Gibbs, 1993; Pennington et al., 2000,2004; Cowling et al., 2004; Mayle and Beerling, 2004). A suggestion for a dry arc of vegetation connecting southern Amazonian savannas to those of Colombia (Pennington et al., 2000) has been revised in the light of new paleoecological data from lowland Bolivia (Mayle et al. 2000; Burbridge et al. 2004; Pennington et al., 2004).

The finding that dry forest did not expand as predicted in the Bolivian ecotonal region investigated by Mayle's team was a major obstacle to the corridor concept. Similarly, misidentified stone lines within the Belterra clays that were taken as a signal of aridity, have been re-evaluated and are now seen to support continuous Pleistocene edaphic moisture (Colinvaux et al., 2001).

Pennington et al. (2000) suggested that the Hill of Six Lakes data could equally represent dry forest as represent wet forest. However, this suggestion was based on the abbreviated list of species presented in the first reporting of the Lake Pata record (Colinvaux et al., 1996). The full species list of >300 pollen and spore types includes many species that are strong indicators of cool mesic forest—such as Cedrela, Cyathea, Podocarpus, Ilex, Brosimum, and Myrsine. Indeed, the most plausible connection from south to north across Amazonia is via the eastern corridor of low precipitation that extends from southern Guiana to eastern Brazil (Figure 3.3). Given the large river barriers, complex orographic features of western Amazonia and the overall climatic heterogeneity of the basin, it is more appropriate to consider Amazonia as a series of patches and connectivity as a shuffling within the patchwork rather than migration in corridors.

Although the dry forest flora is a minority component of the lowland Neotropical system, it is worth exploring the potential for connection between isolated sites as the lessons learned are applicable to the wetter systems as well. A first observation is that many of the sand savannas peripheral to, and within Amazonia contain endemic species and probably have not been completely connected recently, if ever (Pennington, pers. commun.). However, migration through the Amazonian landscape may not have been through white sand soils, as many of the dry forest species demand better soils (Pennington et al., 2004), and their migrational corridors may have been through riparian settings.

A significant problem in this arena is the iconography of biome-watching. Concepts such as "caatinga", "semi-deciduous tropical forest", or "wet forest" are at best the abstractions of mappers. These broad vegetation classifications can help us to think about a system, but we should not believe that they are sharply distinct in their ecology. Many of the same species that occur in a lowland semi-deciduous forest are found at the edge of their ecological tolerance within mesic forest. Nuances of a month more or less of dry season may substantially alter the dominant taxa, though perhaps resulting in much less change in the long tail of rare species (sensu Pitman et al., 1999). Tuomisto et al. (1995) recognize >100 biotopes of tierra firme forest types in Peruvian Amazonia alone, and Duivenvoorden and Lips (1994) some 20 plant communities within a small rainforest near Araracuara, Colombia, and 0lson et al. (2001) recognize 35 types of dry forest. Even within this wealth of forest types a given species may be facultatively deciduous in one setting, or in one year.

The dry valleys scattered along the eastern flank of the Andes are probably the result of interactions between prevailing winds and topographic blocking (Killeen, in press). As such these features are probably relatively fixed and have offered the potential for "island-hopping" around western Amazonia for several million years. The most probable path for migration through Amazonia would be where vegetation is potentially susceptible to fluctuations in precipitation.

A corridor of low precipitation crosses north-south across Amazonia at c. 50° W (Figure 3.4). Based on the seasonality and totals of precipitation, Nepstad et al. (1994) suggest that in this region evergreen forest trees are reliant on deep soil water when seasonal drought deficits occur in the surface soils. If this reliance on hydrology is true, any factor that lowers water tables beyond rooting depth could lead to a substantial

Figure 3.4. Precipitation data for Amazonia based on satellite monitoring (TRMM) showing the corridor (dotted lines) identified by Nepstad et al. (1994) where evergreen trees are dependent on deep soil moisture.

Figure 3.4. Precipitation data for Amazonia based on satellite monitoring (TRMM) showing the corridor (dotted lines) identified by Nepstad et al. (1994) where evergreen trees are dependent on deep soil moisture.

change in the flora. This same corridor was suggested by Bush (1994) to have been the section of Amazonia most sensitive to reduced precipitation. Whether this region experienced drought synchronously—or whether portions of it were wet while other portions were dry—would not inhibit the eventual migration of a species north to south through this region.

Finally, the distribution of dry forest arc species show some disjunctions across the Andes. As it is implausible that continuous habitat for these species spanned the mountains during the Quaternary, we are left with three possibilities. Congeners might have been mistaken for conspecifics. Or the species in question might be capable of dispersing across a major biogeographic barrier such as the Andes, in which case dispersal across lesser barriers within Amazonia obviates the need for any kind of habitat continuity. The third possibility is that these biogeographic patterns are relictual and derived from the pre-Andean biogeography of South America—that is, at least mid-Miocene in age (Hoorn et al., 1995). We consider the latter two of these options to be the most probable.

The message emerging from this analysis is that simplistic notions requiring the whole of the vast Amazon lowlands to have a single climate subject to uniform change are bound to fail. Pleistocene forests were shaped by cooling, low atmospheric CO2 concentrations, and at differing times by lessened rainfall. There is a strong suggestion of no-analog communities based on the prevalence of montane taxa interspersed with a full suite of lowland taxa. Some species that we now consider to be dry forest species may also have been able to survive within this forest, especially if it was structurally somewhat more open than modern forests. Thus, asking if a "dry forest arc" once existed may be much less relevant as a question, than "to what extent (if any) would conditions have to change to allow species to migrate through or around Amazonia?" The answer to this latter question may be "surprisingly little".

Was this article helpful?

0 0


  • jonas
    What causes the montane and lowland change in rainforest type?
    14 days ago

Post a comment