In the current analysis, alpha diversity shows broad correlations with both precipitation patterns and geology (Figure 10.5a-d, see color section). Total precipitation amount is positively correlated with diversity (Figure 10.5a) and average length of dry season negatively correlated (Figure 10.5d). Both reproduce the basic pattern of high average diversity in central and western Amazonia, and along the main stem of the Amazon, though with notable discrepancies. Strong gradients in alpha diversity are found across areas of high rainfall, particularly in northwest Amazonia, and, conversely, areas of low diversity are found in areas of high average annual rainfall
(Figure 10.5a). Manaus, in particular, has diversity as high as western Amazonian forests, but has much lower rainfall. High rainfall areas at the mouth of the Amazon and in Guianan forests show similar diversity to much drier forests in the western Amazon. Average dry-season length, while having its minimum at areas of highest diversity in western Amazonia, also has strong gradients in diversity across areas of short average dry-season length, particularly in northwest Amazonia (Figure 10.5c).
Total variability in monthly precipitation and diversity shows a closer correspondence to the basic patterns of alpha diversity (Figure 10.5b). Isoclines in total variability largely follow isoclines in alpha diversity, particularly in southern and northwest Amazonia. The transverse dry belt, separating central and western Amazonian forests from Guianan forests, is clearly visible in these images (Figures 10.5a-c) as an area of not only relatively low rainfall, but also an area of highly variable rainfall. Species diversity in this area, and to the east, is correspondingly lower than one would expect from a forest with similar total precipitation, but lower intra- and inter-annual variability.
Geologic age also shows broad correspondence to patterns of alpha diversity, with the highest diversity forests all falling on the Tertiary and Quaternary sediments west of Manaus, and the area of high diversity extending east along a narrowing tongue of Miocene and younger-aged Tertiary sediments (Figure 10.5d). Forests on the Proterozoic- and Archean-aged rocks of the Guianan and Brazilian Shields have relatively uniformly low diversity, even in areas of high rainfall and stable precipitation regime. Geologic age remains correlated with diversity, even after accounting for the association between total variability in precipitation and geologic age (residuals from local regression fit of total variability in precipitation and alpha diversity versus geologic age, Kruskal-Wallis x2 = 14.5509, df = 3, p = 0.002).
Because of the relatively homogeneous Middle- to Late Cenozoic sediments, patterns of diversity in central and western Amazonia present a test of precipitation's influence on diversity while minimizing variability in geologic age. Focusing on areas >60°W, where Amazonian tree plot density is highest, average alpha diversity of forest trees shows a broad peak from the equator south to with a nearly linear decrease to the south and a steeper decline to the north (Figure 10.6). Looking at the maximum diversity (sensu Ter Steege et al., 2003) one sees broadly the same pattern, the exception being a slower rate of decrease to the north of the equator. Both of these patterns correspond well to patterns of dry-season length and total precipitation variability in the southern hemisphere. North of the equator, maximum diversity follows precipitation, while average diversity decreases more rapidly than precipitation amount or either measure of precipitation variability. Another notable feature of this figure is the paucity of plots from 0 to 5°N and 6 to 10°S (see also Figure 10.2a).
Southern hemisphere vascular plant gamma (regional) diversity increases to ~4°S, then decreases slowly to ~12°S, falling off rapidly as one moves farther south. Gamma diversity remains high much farther south than alpha (local) diversity (Figure 10.6). While this trend appears to be discordant with predictions based on precipitation, it is understandable in terms of how underlying species abundances change as their ranges cross the precipitation gradient. Gamma diversity in this analysis is based on range data with a species only having to include a particular
Latitude (degrees S)
• Precipitation: Total variability
• Precipitation: Dry season length —•— Tree alpha diversity
Figure 10.6. Change in species richness (Wspecies) with elevation in four Neotropical inventories (Andes: Gentry, 1988; Costa Rica: Lieberman et al., 1996; Mexico: Vazquez and Givnish, 1998; western South America: Silman et al., unpublished). (a) 0.1-ha diversity results for woody plants >2.5 cm dbh. Mexican results are from a single elevational transect, while Andean plots range from 11°N to 16°S. Mexico and the Andes show similar rates of change in diversity with elevation (Mexico ~70 spp km-1; Andes ~60 spp km-1), with a distinct number of species in the source pools (intercepts) in the two regions. (b) Results for gamma diversity (all species whose ranges include a given elevation) in western Amazonian and Andean vascular plants with >10 collections (left axis) and tree alpha (local) diversity from a transect of 1-ha plots in Costa Rica (right axis; replotted from Lieberman et al., 1996, r2 = 0.99).
latitude/elevation combination in its range to be counted. The species can be present in the landscape, but be at low abundance near the edge of its range and therefore not likely to contribute to alpha diversity (sensu Holt et al., 1997). This influence on diversity will be particularly true when species outliers are found in local areas of suitable habitat outside its central range (Levin, 1995; Holt and Keitt, 2000).
Was this article helpful?