We define biodiversity as the number, and the identity or composition, of biological units at different scales: the genetic, the species, the guild and functional group, and the ecosystem and landscape levels. Genetic diversity is the diversity of genetic characteristics or features within a biological species. Species diversity, the number and composition of taxa present at a site, is the most frequently considered aspect of biodiversity (although it can be difficult enough to assess, as we discuss below). Finally, the diversity of ecosystem types within the landscape represents another feature of natural biodiversity that affects ecosystem processes.
Genetic diversity within arid-region organisms has been widely recognized as a valuable commodity. Much research in" biotechnology is currently directed toward the transfer of genes (whether single-gene or more complex traits) for salt tolcrance, drought tolerance and other features that would have great commercial significance in the improvement of dryland crops. There is increasing recognition of the potential commercial value of the wide range of genetic diversity within desert genera such as Acacia (Thomson el al. 1994), Prosopis (Fagg and Stewart 1994) and Opuntia (Pimienta-Barrios 1994).
Some widespread desert plants are known to contain considerable genetic diversity. For example, North American I.arrea tridentata, the creosote bush, exists in three distinct forms (diploid, tetrapoloid and hexaploid chromosome races; Hunziker et al. 1977), and the major taxa of the US mtermountain semi-desert, including Artemisia and Chrysathamnus, contain significant ecological and genetic differentiation within species (West 1988). Schuster et at. (1994) documented significantly higher levels of genetic variation within populations of four North American desert plants than in published averages for perennial plants in general, and attributed this to the selective pressure of environmental heterogeneity in arid lands. Phenotypic plasticity and other forms of variability may exist even within genotypes, and be especially important in deserts; for example, Sayed and Hegazy (1994) discuss the fitness advantage of some annual plants being able to switch from C3 to CAM photosynthesis under field conditions.
Biodiversity, as reflected in species richness, appears to be moderately high in semi-arid regions and declines with increasing aridity for most taxa (Shmida 1985). For example, Pianka and Schal! (1981) demonstrated a decline in Australian bird species richness from about 300 species (in 240 km x 240 km grid cells) to 100, with mean annua! rainfall dropping from 160 mm to <20 mm. Davidson (1977) reported a positive association between rainfall and the diversity of granivorous ants and small mammals. O'Brien (1993) calculated a reduction in southern African woody plants from 567 species (in 10 000 km2) to 27, as average annual rainfall declined from 1300 mm to 55 mm. There arc very few cases, however, where study methods and sampling have been consistent enough to allow such geographic or large-scale comparisons. For plants, soil characteristics are an important influence on diversity within a region. The concentrations of particular mineral nutrients, and in some cases the ratios between nutrients (such as the Ca:Mg ratio), are positively correlated with plant species richness (e.g. El-Ghareeb and Hassan 1989; Ai-Homaid et al. 1990; Franco-Vizcaino et al. 1993). In genera!, soil texture, soil parent material and topographic position (which influences moisture and nutrient availability) are the primary correlates of spccies distributions and community gradients (e.g. Parker 1988; Palmer and Cowling 1994).
Certain taxa are diverse in arid lands relative to other biomcs; these include predatory arthropods, tenebrionid beetles, ants and termites, snakes and lizards, and annual plants (Cloudsley-Thompson 1975; Crawford 1981; Wallwork 1982; Pianka 1986; Ludwig et al 1988). However, there is substantial variation in the richness of particular taxa among the deserts of different continental areas, suggesting that biogeography and historical factors are as important as ecological factors. For example, predatory arthropods (spiders, scorpions, solpugids) are diverse in North America but not in Australia, while for ants that pattern is reversed.
Studies of functional group diversity in arid lands have focussed on the convergent evolution of organisms in similarly hostile environments (e.g. Mooney 1977; Orians and Solbrig 1977; Cloudsicy-Thompson 1993). For plants, identification of guild or functional group membership seems relatively straightforward, as growth form (e.g. shrub, grass, annual forb) is quite highly correlated with phenology and physiology (e.g. photosynthetic pathway). Thus whether one uses physiology, timing of activity, or structure/morphology, one is liable to end up with similar categories or groups of species. An analysis of the flora of the Jornada Long-Term Ecological Research site in the Chihuahuan desert, for example, showed a non-random association of life-form with photosynthetic pathway (L.F. Huenneke, unpublished data, 1995). On the other hand, relatively subtle differences in morphology or physiology wichin a growth form may have significant effects on the pattern of species coexistence. For example, differences among shrub species in seedling tolerance of drought result in individualistic patterns of recruitment in different microsites (Esler and Phillips 1994), and would presumably also result in different species becoming established in years with different rainfall patterns. Hence the shrub species are not truly substitutable for one another.
The extremely harsh ambient conditions in arid lands are associated with biological traits or behaviors that allow temporary escape or avoidance of those conditions. Many arid-land organisms have wide distributions and/or large individual home ranges, and can sometimes be considered ephemeral at any one site because they are migratory or nomadic (birds, kangaroos, antelope, locusts). Others are ephemeral in time or behavior; aestivation, cryptobiosis, drought deciduousness, seed dormancy and other mechanisms ensure that biological activity is pulsed to occur at the most likely time of resource adequacy (Louw and Sccly 1982). Variability in dormancy and germination characteristics undoubtedly accounts for a major portion of the diversity of desert plants, especially annual species (Venable and Lawlor 1980; Kemp 1989; Gutterman and Ginott 1994). Westoby (1972) and Noy-Meir (1973) formulated a model of pulse and reserve for biological activity in deserts; a discrete pulse of inputs (e.g. rainfall) triggers a burst of activity, some of which is translated into a long-lasting reserve (e.g. the soil seed bank or a population of aestivating animals). The pulsed nature of activity has implications for the difficulty of sampling or monitoring species composition and biological diversity at any one site or time.
Landscape diversity and the connections between landscape units are important to the biological diversity and ecosystcm function of arid lands. Schlesinger and Jones (1984) found that the diversity of perennial plants declined on desert piedmonts cut off from overland flow from adjacent upslope areas. Yair and Danin (1980) also found that the presence of rocky areas of limited infiltration increased species diversity downslope, where the runoff water could enter the soil profile. Recent work in India (Puri et al. 1994a,b) has explored the relationship between windbreak plantings of native leguminous trees and crops in agricultural fields; in some cases the effect on crops is negative, while for other tree species there is an ameliorating effect on crop environment and productivity, but in all cases there are strong interactions among landscape units.
5.3 BIODIVERSITY AND ECOSYSTEM PROCESSES 5.3,1 Production
Plant biomass and primary production range from moderate values in semi-arid ecosystems (400-800 g' year"1, average aboveground values) to virtually zero in hypcr-arid systems. Productivity and biomass are closely correlated with total precipitation and also with the predictability of precipitation, across the semi-arid/arid gradient (LeHouerou 1984; Evenari 1985; Dregne and Tucker 198&; LeHouerou et ai. 1988). Structurally, semi-arid vegetation usually comprises grasslands (actually diverse mixtures of perennial and annual grasses and forbs) with some presence of stem and/or leaf succulents, subshrubs and larger woody plants. More arid systems have lower abundance and cover of grasses, with a greater relative importance of woody plants and succulents. There may be a prominent arborescent component, as in the Sonoran Desert of North America. Relative abundances of different growth forms and different photosynthetic pathways vary with the seasonality of precipitation; e.g. summer rainfall favors the abundance of C4 pathway grasses (Louw and Seely 1982). Annual species occur in times of moisture availability, both in open areas and beneath the canopy of perennial plants. The prevalence of open areas increases with aridity, and many plants become increasingly restricted to water courses or other areas of run-on or moisture accumulation (as in Schlcsinger and Jones 1984). In hyper-arid situations, there is virtually no plant cover except in unusual years and in those sites where water can accumulate.
Noy-Meir (1973, 1985) thoroughly reviewed the theory and empirical evidence regarding primary productivity in arid lands. He concluded that variation among plant growth forms prevented any generalizations about the ratio of above-ground to below-ground productivity, or about the turnover rate of plant biomass in a community. In other words, the relative abundance of various growth forms (annuals, persistent subshrubs and succulents, woody trees) was critical in determining ecosystem structure and dynamics, and these relative abundances varied too greatly among desert locations to allow any simple generalizations or conclusions. The implication is that different growth forms contribute differentially to productivity, and that the removal of particular growth forms will affect ecosystem structure and production. Different growth forms do not provide "redundancy" for one another; the removal of shrubs will remove above- and below-ground biomass that simply cannot be replaced or replicated by grasses, say. Sala et al. (1989) found that shrubs and grasses in the steppe of Argentina do rely largely on different soil resources; removal of one group had little effect on production by the other.
One would infer that systems with all growth forms represented should have higher rates of production than those missing some growth forms, all else being equal (particularly climate). However, we know of few available data to test this prediction. There is less conceptual basis (and even fewer data) to support the notion that the presence of more species within a growth form will lead to higher productivity. Cowling et al. (1994) pointed out that functional "redundancy" within growth forms, where several species responded similarly to climatic factors, was usually explained by the species having different responses to rare catastrophic droughts. Thus species richness would be predicted to minimize fluctuations in plant cover and production over time. Diversity of photosynthetic processes might also be expected to affect productivity. However, in one cold-winter semi-arid region, there was no substantial difference in carbon fixation or water use between communities dominated by shrubs with either the C3 or C4 photo-synthetic pathways (Caldwell et al. 1977). It is unclear whether the two groups of species are "substitutable" for one another in a given environment, however.
Was this article helpful?