Synchrony of Regeneration, Gaps, and Latitudinal Differences in Tree Species Diversity

Abstract

Treefall gaps are necessary for establishment and/or growth to maturity of many forest trees. It therefore seems reasonable to hypothesize that characteristics of gaps influence both persistence of populations and coexistence of species in forests. The purpose of this essay is to suggest one way that gap characteristics might result in differences in tree species diversity between tropical and temperate zone forests. Latitudinal gradients in species diversity seem unlikely to be a result of differences in the physical or distributional characteristics of gaps in different forest types. Gap formation rates (fraction of ground area converted to new gaps per year or fraction of canopy trees or canopy area dying per year) are similar for a variety of forests in tropical (Brokaw 1985a, Denslow 1987) and temperate (Runkle 1985) zones. Gap size varies widely within each region, but average sizes and the range of sizes are similar (Runkle 1982, 1985, Denslow 1987). In both regions gaps occur most years and throughout the year, although they may be more common some months than others (Brokaw 1982b, Romme and Martin 1982). Latitudinal differences might exist in the ability of species to respond to gap conditions. That is, two gaps with the same size, shape, and mechanism of formation represent different regeneration opportunities in temperate zone and tropical forests. One mechanism, discussed by Poulson and Platt (1989), is that within-gap heterogeneity in light levels, especially in large gaps, will be greater in temperate than tropical zones, resulting in greater potential opportunities for withingap specialization for different light levels in temperate than tropical forests. I argue that winter imposes a synchronization of responses that reduces the potential diversity of species possible in temperate zone forests. I will discuss this relationship using the model of Chesson and Warner (1981), Warner and Chesson (1985), and Chesson (1986). The storage effect model, as presented in the above papers, may be described as follows. Species can persist in a community if their adults are long lived and if their reproductive success is such that population increases in scattered good years offset declines suffered in more frequent bad years. The regenerative consequences of good years are stored in long-lived individuals that survive through many years with little successful reproduction. A low-density species must experience good years frequently enough that extinction does not occur. High variance in reproduction over time, as a result of environmental variability or intrinsic factors, increases the chance that favorable reproductive periods will be encountered, at least occasionally, by rare species. The model also assumes that species at low densities can increase greatly in good years but decline only slightly in bad years, whereas abundant species can decline to a greater extent in any given year. Tropical forests are characterized by a broad interplay among seasonal variations in gap characteristics and climate. Thus arrival in gaps (before or soon after formation) as well as survival and growth to maturity (see Martinez-Ramos et al. 1989, Schupp et al. 1989) have to be evaluated relative to seasonal phenology of tree species. In tropical forests, seasonal variation occurs in arrival of seeds at a site. The seasonal rhythm of both fruitfall and seed germination has been well documented (Foster 1982, Garwood 1983). Differences among species in these properties could create some of the variability predicted by the storage effect model to result in enhanced species diversity. Being present on a site before or soon after a gap is created results in a large advantage over later arrivals. Growth rates are higher and mortality lower for plants establishing before or soon after treefall (Denslow 1987). Uhl et al. (1988) found that 4 yr after gap formation, advance regeneration accounted for 97% of all trees 1 m tall in five artificial single-treefall gaps and 83% of all trees in a multiple-treefall gap in Venezuela. Brokaw (198 5b, 1987) found that successful stems all were established prior to or within 1-3 yr following gap formation. Despite the importance of new seed dispersal into gaps and the production by pioneer species of lightrequiring seeds (Whitmore 1989), the great majority of tropical rain forest woody species are present to some extent in the understory before gaps form (Hubbell and Foster 1986a, b, Brokaw and Scheiner 1989). Although such advance regeneration is present contin-