Most ecosystems of the tropical and subtropical latitudes are seasonally stressed by drought (Schimper, 1898; Köppen, 1931; Murphy & Lugo, 1986). Research on population and ecosystems dynamics, and conservation efforts, however, rarely address these ecosystems, but rather concentrate on what is usually understood as tropical wet forest or rain forest. There has been enormous scientific and public attention directed toward documenting the effects of destruction of wet forests on soil fertility, biotic diversity, and global biogeochemistry. These concerns are certainly justified as the rates of forest and species loss accelerate. In contrast, relatively little attention has been given to forests subject to prolonged dry seasons (Ridpath & Corbett, 1985), and to their changing status. Degradation and conversion of ‘dry forest’ is far more advanced than that of wet forest: only a small fraction remains intact (Murphy & Lugo, Chapter 2; Sampaio, Chapter 3; Menaut, Lepage & Abbadie, Chapter 4; Rundel & Boonpragob, Chapter 5; Gentry, Chapter 7), and the area explicitly conserved is hardly perceptible. This is unfortunate because the forests with prolonged annual drought occupy more area than wet forests, have been of greater use to humans, and are still poorly known over most of their distribution.

The extent of forest in the drier tropics, and even its character, are difficult subjects for debate and research. Particularly in Africa, India and Asia, the relations between savannas, woodlands and dry forests (of various leaf habits) are notoriously complex (Furley, Proctor & Ratter, 1992). Savannas and their degradation are certainly priority subjects of tropical and global ecology, but as with wet forests, they are well studied compared with dry forests.

Introduction

The production and emission of trace gases from tropical forests are of interest for calculating global budgets, for determining local ecosystem mass balances, and for the insight they provide into ecosystem processes. Globally, tropical dry forests are widespread and functionally different from other biomes, and hence need to be characterized directly. More-over, the extent and intensity of human modification of dry forests (in comparison to other tropical biomes: Murphy & Lugo, 1986) raises the possibility that land use change in this biome could be globally significant in the concentration or distribution of one or more gases.

At the ecosystem level, loss to the atmosphere can be a major pathway of transfer of nitrogen and sulfur from terrestrial ecosystems (Likens et al., 1977; Bowden, 1986), and hence a possible determinant of the long-term fertility and potential productivity of terrestrial ecosystems. Since gaseous emissions are not normally included in watershed-level measurements of nutrient inputs and outputs, they can provide a useful complement to watershed studies. Finally, on the process level the trace gases emitted from terrestrial ecosystems represent either end products or by-products of fundamental metabolic processes that occur within terrestrial ecosystems. Their magnitude, timing and regulation provide insight into the operation of those processes even where emissions are not quantitatively significant in either global or local element budgets (Matson, Vitousek & Schimel, 1989).

None of these reasons is specific to tropical dry forests; all apply to any biome with a large areal extent and significant human modification. However, the dry forest biome is particularly interesting for three reasons. (1) It occupies an intermediate position on the great moisture gradient that underlies variation among tropical ecosystems (Medina, Chapter 9).

Introduction

Conversion of tropical dry forest to agriculture and pasture is occurring at alarming rates, and entails nearly total destruction of forest structure and composition, and disruption of ecosystem functions. In this chapter the driving forces of this process are briefly discussed, the general pattern of the transformation is described, and the consequences of the conversion for ecosystem functioning are analysed in some detail. Particular attention is given to specific environmental conditions and management practices that determine or alter the extent of the perturbation.

Causes and patterns of forest conversion

Driving forces

Janzen (1986) considers dry forest as the ‘most endangered major tropical ecosystem’ because the distribution of these forests has been reduced to a small fraction of the original area. Less than 0.1% of the original dry forest has conservation status in Pacific Mesoamerica. The status of dry forest is just as critical, or worse, for regions of Australia, Southeast Asia, Africa, and major parts of South America (Janzen, 1986, 1988). Actual rates and the extent of dry forest conversion are analysed by Murphy & Lugo (Chapter 2). They have pointed out that we will never know the true original or potential extent of dry forest because many savannas and scrub or thorn woodlands are thought to be derived from disturbed dry forest (Murphy & Lugo, 1986; see Menaut, Lepage & Abbadie, Chapter 4).

There are ecological, social, political and economic factors involved in forest clearing for agricultural purposes. Often subsistence farmers are blamed for the massive deforestation in the tropics, but other groups are also involved in this process. According to Thapa & Weber (1990), those responsible for deforestation can be broadly categorized into three main groups.

Introduction

The dry forest and scrub vegetation in Brazil, generally called ‘caatinga’, covers an estimated area of 6–9 × 105 km2 in the northeastern region. It is conditioned by the prevailing semiarid climate, with high potential evapotranspiration throughout the year (1500–2000 mm y−1) and low rainfall (300–1000 mm y−1), which is usually concentrated in 3–5 months and is very erratic (Reddy, 1983). Drought years are common and severe droughts lasting 3–5 years have occurred every 3–4 decades.

The area has been inhabited for more than 10,000 years, mainly in the river valleys and humid mountains, but according to early colonial sources, population density was generally low. Cattle raising spread in the 18th century and still is the main economic activity. From that period on, population pressure has increased in more favorable areas, where subsistence agriculture is practised in fenced plots. Until the middle of this century, cattle roamed freely on the non-agricultural land, independently of land ownership, but most properties are now fenced.

Land productivity is low and since resources are limited and birth rates have been high, the area has been a center of continuous migration to more favorable places in the same region, mainly the coastal area, or to other regions in the country. Migration increases during catastrophic drought periods. Nonetheless, population has steadily increased in the area and most of it has remained at a bare subsistence level. Social and economic parameters are the worst in the country, from lowest per capita income to highest illiteracy.

Government efforts to foster economic development in the region have centered on the coastal area, except for large irrigation projects.

Introduction

The neotropics may sustain the largest number of living species of plants and animals on earth (Wilson, 1988; McNeely et al., 1990). In the vast neotropical region, however, there are extensive regions such as the Amazonian rain forests, the Venezuelan ‘tepuis’, and the dry forests, where there is little biological knowledge of ecosystem composition and ecological interactions.

Research in neotropical dry forests during the last two decades has provided the scientific community with surprises such as the discovery of an ‘extinct’ peccary, Catagonus wagneri, in the Paraguayan chaco in 1974 (Wetzel et al., 1975). It is, therefore, important to recognize patterns of biological diversity in dry forest to gain insights into causal processes in biogeography and ecosystem function, and to assign them appropriate conservation values.

In this chapter I describe the general patterns of terrestrial vertebrate diversity and conservation in dry forests, and contrast them with patterns in adjacent moist/wet forests. The chapter is divided into five sections. In the first section, a general description of the major neotropical dry forests is given, emphasizing size and degree of perturbation. The following section presents a detailed analysis of patterns of species richness and diversity. The third section is dedicated to a description of community structure and ecological responses to climate seasonality. In the fourth section the origins of the dry forest vertebrate faunas are discussed. Finally, the last section is devoted to conservation problems.

Introduction

Nutrient cycling processes have been well documented for tropical moist forest (Vitousek & Sanford, 1986; Bruijnzeel, 1991) but few comprehensive syntheses exist for tropical and subtropical dry and deciduous forests (Lugo & Murphy, 1986; Singh, 1989). Tropical dry forests are considered among the most threatened tropical ecosystems (Janzen, 1988) because they experience considerable exploitative pressure (Murphy & Lugo, 1986). In India, such pressures have been responsible for the transformation of vast areas of deciduous forest into savanna (Singh, 1989). The current rate of destruction of deciduous forest makes it imperative that we gain a thorough understanding of nutrient cycling in the remaining intact and successional forests. At present one of the principal agricultural practices relies on forest slash burning, resulting not only in recurrent nutrient losses thereby affecting the long-term productivity of the system but also substantially contributing to emissions of C and N to the atmosphere (Kauffman, Sanford & Sampaio, 1990; Maass, Chapter 17).

Seasonally dry forest production is controlled by the amount and distribution of annual rainfall (Martínez-Yrízar, Chapter 13), and this may explain why nutrients have not been considered in detail (Murphy & Lugo, 1986; Singh, 1989). For example in a recent review of forest nutrient cycling there was only one reference to tropical dry forest (Vogt, Grier & Vogt, 1986). If we assume that water availability alone limits primary production in tropical deciduous forest, we could conclude that nutrient limitation is not important. However, multiple resource limitation of plant growth is common in natural communities (Chapin et al., 1987).

Introduction

A large amount of work has been published on various aspects of the functioning at the ecosystem level in forests throughout the world. However, information for tropical dry forests is sparse. In comprehensive reviews of forest productivity studies (Bray & Gorham, 1964; Art & Marks, 1971; Murphy, 1975; Jordan & Murphy, 1978; Brown & Lugo, 1982; Proctor, 1984; Vitousek, 1984; Murphy & Lugo, 1986a; Vogt, Grier & Vogt, 1986), there are no more than ten references regarding studies on primary productivity of tropical dry forests. Subsequently a few more studies have been published, particularly from Brazil, México, Puerto Rico and India (Table 13.1).

This review of biomass distribution and primary productivity of tropical dry forests emphasizes the aboveground portion of the vegetation. There is no consensus on a precise definition of dry forest. Following Murphy & Lugo (1986a), my review includes tree-dominated sites in tropical regions in which at least the dominant trees are drought deciduous. These sites include those with strongly seasonal climate, that is, including a well-defined dry season. Mean annual rainfall is usually below 1500 mm and annual potential evapotranspiration:precipitation ratio is normally greater than one.

After a review of forest phytomass, I survey the process of litterfall and of litter decomposition, treating methodologies reviewed by Proctor (1983) and Wieder & Lang (1982). The significance of litterfall and litter decomposition in nutrient cycling in tropical deciduous forests is presented in detail by Jaramillo & Sanford (Chapter 14). Finally, I examine the few attempts that have been made to estimate net primary productivity (NPP).

Introduction

Many studies of neotropical dry forest have tended to treat them in a very broad context, typically focusing on how they relate to or are different from moist or wet forests (e.g. Holdridge et al., 1971; Rzedowski, 1978; Gentry, 1982a, 1988; Hartshorn, 1983). Others have taken them as a relatively tractable surrogate for the more diverse moist or wet forests (e.g. Janzen, 1983, 1984, 1988; Hubbell, 1979) where taxonomy often poses severe limitations for the resolution of biologically interesting questions. Other authors have concentrated on the interesting physiological adaptations of dry forest organisms to seasonal water stress (e.g. Medina, Chapter 9; Holbrook, Whitbeck & Mooney, Chapter 10 and included references) or on various aspects of nutrient flow and biomass (e.g. Lugo et al. 1978; Murphy & Lugo 1986a, b). In addition there have been floristic and community ecological studies of individual dry forests (e.g. Troth, 1979; Valverde et al., 1979; Thien et al., 1982; Hartshorn, 1983; Heybrock, 1984; Lott, Bullock & Solís, 1987; Kelly et al, 1988; Rico-Gray et al., 1988; Arriaga & León, 1989; Cuadros, 1990; Saldias, 1991; Dodson & Gentry, 1992; see also summaries for México in Rzedowski, 1991a, b, and for the chaco in Prado, 1993). However, there have been remarkably few attempts to focus on the distinctive floristic composition of dry forests as a whole or on how different dry forest plant communities differ from each other.

Introduction

Many neotropical dry forests are dominated by trees that shed their foliage and remain leafless for a substantial period each year. Because the majority of deciduous species drop their leaves during the dry season and renew their canopies with the onset of the rains, the question of how these trees cope with seasonal reductions in soil moisture and increases in evaporative demand is most simply answered by calling them ‘drought avoiders’ (sensu Levitt, 1972). This categorization, however, gives little insight into the conditions, constraints and consequences that accompany the deciduous habit in these forests. Furthermore, the coexistence of even a small number of evergreen species indicates that the deciduous habit is not unconditionally imposed by the environment and that patterns of leaf fall and renewal must be viewed as part of an integrated response to environmental conditions. Seasonality in water availability clearly plays a major role in structuring patterns of activity and growth in this biome, but there have been few studies of the dominant life form (and fewer of life-form diversity; see Medina, Chapter 9). In this chapter, we address characteristics of trees of tropical dry and deciduous forests that influence their patterns of water use.

Plants respond to changes in resource availability on several scales. Our review considers three such levels: ‘structure’ encompasses features that remain relatively constant throughout the life of a plant, such as rooting patterns or stem hydraulic properties; ‘physiology’ focuses on parameters that influence diurnal patterns of water use and gas exchange; while ‘phenology’ considers seasonal patterns of meristem activity.