Portal de Periódicos da CAPES

Of orogeny, precipitation, precession and parrots

2005; Wiley; Volume: 32; Issue: 8 Linguagem: Inglês

10.1111/j.1365-2699.2005.01343.x

1365-2699

Mark B. Bush,

Species Distribution and Climate Change

Simple unifying explanations of Amazonian species diversity, such as stability begets diversity, or that allopatric speciation took place in glacial-age refugia, have fallen by the wayside. In their place a more complex story is unfolding. This new narrative spans millions of years, involves Earth's wobbling orbit about the sun and thrusting plates, but ultimately simplifies to the existence and persistence of niche space. Making inferences regarding past climates from modern biogeographical patterns is a mainstay of palaeoclimatic reconstructions, and yet can deceive. The construction of transfer functions that allow the translation of a species’ presence or abundance into climatic space relies on a lack of evolution. Over Quaternary time, it is plausible that the fundamental niche of many species is relatively constant. Hence, considerable confidence can be placed on inferring past conditions based on the modern bioclimatic envelope of a species, and even more when multiple species are used. However, the no-analogue communities of the past make comparisons of community abundance fraught with difficulty. From the palaeoecological data it is evident that while the fundamental niche probably changes little, the realized niche of many species may have been very different in the past. Such differences are exacerbated if there has been extensive extinction, such as the loss of megafauna, or landscape alteration by humans. Another way in which biogeographical data are used to infer past climate is by arguing from modern diversity to ancient disjunction or, its flip-side, from modern disjunction to ancient continuity. Haffer's (1969) refugial hypothesis is the most celebrated attempt to infer major biome change in Amazonia on the basis of modern species distributions. This elegant and readily falsifiable hypothesis espoused savanna expansion under arid glacial conditions, and consequent allopatry in forest fragments. Raised in the absence of well-dated palaeoecological or molecular data, this hypothesis is not faring well as the vacuum fills. The refugial hypothesis invoked Amazonian aridity to explain speciation patterns in birds, but molecular studies have shown that bird speciation does not conform to refugial expectations spatially or temporally. Ribas et al. (2005) revisit the phylogeny of Pionopsitta parrots and, in so doing, illustrate at least three major problems with the refugial hypothesis. First, any hypothesis that makes inferences based on modern biogeographical data makes an assumption that those data are good enough for such an analysis. Nelson et al. (1990) recognized that geographically uneven collecting intensity caused areas near to established research centres, or points of access, to appear to be unusually diverse. Ribas et al. point to an even more fundamental problem as they observe that the taxonomy of even well-known groups, such as birds, needs revision. The type species of Pionopsitta is revealed to be in a different (not even the most closely related) genus to the other eight species previously thought to form a monophyletic genus. These eight species are placed within a new genus, Gypopsitta. The second highlighted weakness is that the refugial hypothesis discounts alternative, more parsimonious, explanations of the pattern of speciation. For example, the basal division within the genus Gypopsitta comprises species with distributions west (Chocó and Central America) and east (Amazonian) of the Andes. Andean orogeny has been a continuous process for the last 20 Myr, but the functional separation of the Chocó from Amazonia may have occurred within the last 11 Myr. It is this orogeny, not precipitational change, which offers the most parsimonious explanation of the observed phylogeny. This pattern does not just split eight species of parrot, but was found to be the most basic biogeographical divide in an analysis of the distributions of 1717 Amazonian passerine birds (Bates et al., 1998). The third weakness that Ribas et al. accent is the timing of speciation. Rather than species radiation occurring within the Quaternary, the speciation of Gypopsitta, primarily took place in the Pliocene. The timing of their evolutionary spectrum from c. 8.4 to 0.6 Myr fits well with the observation by Moritz et al. (2000) that rain forest faunal divergences of sister taxa generally predate the Quaternary. A completely different line of evidence that illustrates another temporal problem with refugial-style speciation comes from the rapidly growing pool of South American palaeoclimatic data. Speleothem data from Botuverá Cave, in south-eastern Brazil, provide an isotopic record of fluctuations in the strength of the low-level jet, the windstream that moves tropical Atlantic moisture from equatorial Brazil along the eastern flank of the Andes and then south-eastwards back towards the Atlantic. Cruz et al. (2005) demonstrate that the low-level jet pulsed in strength with the 22,000 year rhythm of precessional orbital forcing. This flux in the amount of moisture entering Amazonia, provides a basic climatic teleconnection between locations along the route of the low-level jet. If the low-level jet is strong, sites that lie within its path will receive more moisture than when it is weak. Many data points are needed to test this rigorously, but this hypothesis is entirely consistent with the 180,000 year record of precessionally driven lake level fluctuations at Lake Pata, Brazil (Bush et al., 2002), in which high and low stands are synchronous with those of Botuverá Cave. These observations give rise to the simple prediction that western Amazonia was wet and cool during the peak of the ice age, and that periods of reduced precipitation, when they occurred, lasted c. 11,000 years. Thus the standard view of the last half million years, largely borrowed from the temperate zone, as being successive c. 100,000-year periods of glacial interrupted by 10,000 years of interglacial, is superceded for the most biodiverse sections of Amazonia by precessional rhythms. Temperatures rose and fell within the last Ice Age, and probably only reached their coldest extremes on about the same time-scale as there were wet and dry events, 11,000 years at a time (or less). Such a short timespan does not promote complete speciation. Some evidence for this period being too brief for bursts of widespread speciation is the lack of diversification among populations separated by 11,000 years of Holocene warmth. The ecological scale of the Pleistocene events also appears to be much smaller than Haffer and other refugialists envisaged. To date, only locations in modern savanna–forest mosaic landscapes have been shown to flip between states (Burbridge et al., 2004). Even just a few tens of kilometres from these border areas no evidence is found of biome-scale change (Pessenda et al., 1998). As an aside, while the western Amazon lowlands may have been oblivious to the larger cycles of ice ages, this was certainly not true of the Andes. The difference being that in the Andes the glacial cooling and changes in physical factors such as frost-elevations, height of cloud formation and intensity of ultra-violet light, caused many species to migrate down-slope. This movement resulted in changes in local productivity, and community composition. At the foot of the Andes, effects of temperature and altered CO2 concentrations simply had to be endured, for there was nowhere to go to escape them. But as the enormous diversity of these lowland forests testifies, those species did endure and survived both the cold of the coldest glacial time and the present unfamiliar warmth. That land–sea temperature differences, the configuration of the mountains and the influence of the great convective plain of Amazonia combine to set the precipitation pattern for the western half of the Amazon basin, must confer a degree of long-term climatic stability. Given these constraints, it is improbable that air-masses behaved very differently during the Pleistocene. As our understanding of Amazonian climate systems and linkages improves it shifts the onus of parsimony. Fragmenting Amazonian forests by savanna incursions is no longer the ‘easy answer’, but rather should be seen as the least viable of alternatives. While late Pleistocene temperature and precipitation fluctuations were insufficient to cause widespread biome transformation, they did cause individualistic migrations of species and re-assortment of taxa into communities without modern analogues. These observations undermine the argument that large landscape conversion from one biome to another is necessary to accommodate species movement between modern disjunct populations (e.g. Pennington et al., 2000). In eastern Amazonia, which is drier than the west, the possibility of periodic dry events creating a more seasonal forest is plausible. However, in addition to the east–west difference in climate there appears to be a north–south asynchrony as well. Thus, it remains to be demonstrated that a continuous corridor of drier-than-modern forest existed at any one time across eastern Amazonia. Such a temporally discontinuous pattern does not preclude the migration of taxa in stages across this area. Consequently, modern biogeographical patterns are not a test of past vegetation conditions. The lesson that should be learned from the falsification of the refugial hypothesis is that making predictions about past climate and community composition based on modern biogeographical patterns is perilous. The final flaw that all of these studies reveal is our basic ignorance of the autecology of any Amazonian species. Almost all our data draw on observations of realized not fundamental niches, and yet it is the continuity of the fundamental niches through time that has enabled the persistence of the species. Editor: Robert J. Whittaker