Pre-european vegetation map 14 Mb PDF

Legend Vegetation immediately prior to European colonization in and around the Chapada Diamantina National Park, N.E. Brazil

Roy Funch

Brazil is the fifth largest country in the world (more than 8,500,000 km²), and by far the largest tropical nation (approximately 5° N to 34° S and 33° to 70°W), straddling the equator and encompassing a wide range of climatic, geographical and geomorphological conditions that help account for its enormous diversity of ecosystems/habitats and its tremendous biodiversity that includes more than 56,000 plant species.

The northeastern sector of Brazil (approximately 4° to 16° S and 35° to 46° W) has a semi-arid climate, but with large temporal and spatial variations in rainfall. In general, this NE region of Brazil experiences low monthly rainfall rates and a long dry season, lasting from six to nine months (MME, 1981). The rainy season coincides with the astral summer months, and bi-modal peaks of precipitation are often observed, one occurring between November and January and the other from March to April. The climate at the study site is strongly influenced by altitudinal and rain-shadow factors, and will be discussed in more detail in the text.

Embedded within this semi-arid region of NE Brazil is the Espinhaço mountain chain. The northern sector of the Espinhaço mountain chain is known as the Chapada Diamantina, and is completely contained within the state of Bahia . The Sincorá Range , focus of the present work, constitutes the eastern flank of the Chapada Diamantina, and more than half of this range has been incorporated into the Chapada Diamantina National Park.

The landscape prior to arrival of Europeans

The dating of the human occupation in the inland regions of northeastern Brazil is currently much debated but certainly initiated many thousands of years ago (Dillehay, 1996), with Europeans only arriving in the areas in significant numbers after the 1700s. The impact of their occupation will be discussed in greater detail in the sections that follow).

The inland regions of northeastern Brazil were occupied by nomadic Amerindians tribes, such as the Janduís and Paiacus in present day Bahia State at the time of European settlement. Very few archaeological studies have been undertaken in the region surrounding the Serra do Sincorá, available information is essentially limited to the locations of sites with rock paintings. These Amerindians were essentially semi-nomadic hunters and gatherers with a very primitive agriculture restricted to growing limited crops, and they possessed no domesticated animals or cattle. An overwhelming percentage of all known archaeological sites are along river courses in the drier caatinga environments to the west of the Sincorá Range with few sites being recognized in the mountains themselves (personal observation), indicating that more permanent settlements were generally restricted to the low-land riparian zones. This may be attributable to the fact that the acidic, sandy mountain soils are much less fertile than those found in the clay/limestone bottom land of the local caatinga , and were therefore much less attractive to agricultural efforts. Additionally, forest clearing there would have been difficult using only stone tools, while the prolonged dry season in the caatinga zone would facilitate the use of burning for opening and preparing agricultural plots. As such, we may assume that the Sincorá Range was probably only visited for hunting purposes and would only have been slightly altered from its primeval state even after long contact with native populations.

The arrival of European colonists

When Cabral and his men first anchored off the coast of Brazil in A.D.1502, at what later became known as Bahia , they must have marveled at the lush evergreen forests which at that time dominated the landscape along the Atlantic coast, and extended far into the interior. The history of the destruction of these forests initiated with the harvesting of a valuable red dye from Brazil Wood or Pau Brasil ( Caesalpinia echinata Lam.). In succeeding years, first the Portuguese, and also the French, Dutch and English carried away thousands of tons of this, but even after the species was almost eliminate, the process of forest destruction continued, until today the Atlantic forests are only reduced to less than 10 % of their former extent. The consequent elimination of biodiversity, which such wanton destruction entailed, and the presumed extinction of many species of plants and animals before they could be recorded for posterity, is a dismal chapter in the history of European colonization of Brazil .

The interior of Bahia , on the other hand, remained more or less unknown to the European colonists for much longer. Indeed, until the early 1800s, there is no evidence of significant European colonization in the Sincorá range itself, although there were limited settlements on the plains to both the east and west (IPAC-BA, 1980). This is very different from the history of settlement in more western parts of the Chapada Diamantina, where gold was discovered in 1719, and even before that, there are records of settlement at least as far back as 1687 (Harley & Giulietti, 2004).

Until 1844 there was apparently no significant European habitation in the Sincorá Range itself, although there were limited settlements on the plains land to both the east and the west (IPAC-BA, 1980). Little historical information is available concerning the eastern settlements near to the Sincorá Range , except that they included at least one early “quilombo” (a settlement of escaped slaves) and were most likely limited to subsistence agriculture, hunting, and fishing. There was neither easy communication nor any significant market economy with the major cities along the coast (at more than 300 kilometers distance). The drier western region was known for its open-range cattle ranching, and although cattle were historically brought overland to the coast from reasonably long distances, communication and economic contact was probably more frequent by way of the navigable São Francisco River , 200 kilometers to the west than overland across the Chapada range.

With the discovery of diamond in 1844, the Serra do Sincorá was invaded by thousands of miners and adventurers, often accompanied by their families and slaves, which resulted in the founding of towns and villages, the extraction of wood products for construction and energy, and the opening of agricultural areas throughout the mountains. As the forests at first were almost impenetrable and the diamonds were panned directly out of the rivers, the miners initially spread through the mountains principally along the water courses. But with the arrival of more and more prospectors, mining proceeded onto the hillsides where diamond-bearing gravel could be encountered at the contact between the overlying soil and the sterile bedrock. Mining on the hillsides required the elimination of all vegetation (accomplished by cutting and burning) and the subsequent erosion and removal of all loose material (soil, sand, clay, gravel, larger rocks, etc.) down to the bedrock/sterile layer. The resulting denuded land was subsequently abandoned and simply left behind. As the diamonds were, in general, more plentiful on the lower slopes, the waves of mining and erosion swept the hills from bottom to top (Torreno, 1926).

Local topography and drainage patterns directed the material eroded from the hillsides almost exclusively into the Paraguaçu River and its principal upper course tributaries (the São José and Santo Antônio Rivers ). Calculations (RR Funch, 2004) have demonstrated that truly enormous quantities of soil were washed from the hillsides and re-deposited in the river valleys at the foot of the mountains – at least 48 million cubic meters in just one section of the mountains, material sufficient to cover the 30 km² worked by the miners to a depth of 1.5 meters . Essentially all of the eastern slopes of the Sincorá Range between the Piaba and Lençóis Rivers were mined, representing an area of well over 300 km² essentially totally denuded of its vegetation and soils.

Allied with the physical destruction of the landscape, fire stands out as the most widespread agent of habitat degradation. Burning in the mountains was absolutely commonplace. Retired miners relate that they and their fellow workers set fire everywhere, year-round, in order to maintain trails open, facilitate the search for new mining sites and the functioning of working mines, and to eliminate snakes and other wildlife that were considered noxious. Additionally, the owners of cattle and other domestic animals would regularly burn the high mountain plains to improve grazing conditions, and hunters still set many fires to remove cover and likewise facilitate grazing by a sought after rodent herbivore ( Kerodon rupestris ) (see RR Funch, 2004).

The result of all this anthropogenic aggression on the flora within the Sincorá Range has been enormous, and includes:

1) Removal of soil and native vegetation from extensive areas in the mountains;

2) Destruction of forests in huge areas of the mountains, and;

3) The spread, into newly cleared or damaged areas, of vegetation adapted to open habitats, consequent on forest destruction and soil erosion, as well as the destruction of other more closed plant communities.

The sudden burst of activity occasioned by the diamond boom in 1844 lasted only a few decades at best, and the region soon lapsed back into a subsistence agricultural mode (although with a continuous low level of mining activity) (Sampaio, 1936). However, accelerated development has occurred in the past two decades, in great part due to the opening of new heavily mechanized agricultural areas, the settling of landless farmers, a short burst of mechanized diamond mining (now prohibited by law), and tourism, resulting in a general economic expansion of the region. These activities have also resulted in altered patterns of land and water use, increasing human population densities, land clearing and deforestation (mostly for agricultural purposes), and the intensive use of agro-chemicals in the headwaters of the regionally important Paraguaçu River Basin.

With this knowledge of some of the major environmental and landscape alterations resulting from mining activities and other human uses of the regional natural resources, and with the completion of a detailed vegetation map of the Sincorá Range (which includes information on edaphic conditions), it is possible to look backwards in time and attempt to visualize the types of vegetation that once occupied the area in and around the Sincorá Range.

While attempting to describe the past landscapes in the region, efforts to discover and examine the natural and anthropogenic factors that impacted a given vegetation form during historical times often opened the way to a better understanding of the factors that determine its current regional distribution, as well as their role in the evolution of the regional landscape. Retrospective landscape ecology also has applications for conservation planning for the Chapada Diamantina National Park (which occupies approximately half of the entire Sincorá Range ). One of the key (and most costly) management strategies of the Park administration is to prevent and combat wildfires. If this strategy can function well it will become important to know what types of changes will be seen in the vegetation cover of the Park, and how will these changes affect wildlife (including the re-introduction of locally extinct species), tourist use, and the local hydrology.

Description of the area

The study area included the entire legal limits of the Chapada Diamantina National Park (CDNP) as well as areas up to 50 km beyond its boundaries, in the northern sector of the Sincorá Range , Bahia State , NE Brazil (approximately 12° 11´- 13° 32´S x 40° 53´- 41° 42´W).

While the Sincorá Range does not have precisely defined limits, it is roughly bounded by the Sincorá River to the south and the town of Afrânio Peixoto ( municipality of Lençóis ) to the north, where its easily recognizable scarps blend back into the high, dissected plateau that constitutes the bulk of the Chapada Diamantina. An almost continuous fault scarp approximately 250 m high and 80 km in extension sharply defines the western border of the Sincorá Range, while the eastern flank rises up quickly from the flat plain that stretches almost uninterrupted to the Atlantic Coast, about 300 km further east. This range has a strong N-S orientation, being approximately 160 km long, and from 20 to 30 km wide, covering a total area of approximately 4000 km² (Funch 2004). The Chapada Diamantina National Park (1,520 km²) occupies most of the northern half of the Sincorá Range .

Almost the entire central area of the Sincorá Range and the Chapada Diamantina National Park is composed of extremely old (1.7+ b.y.) Precambrian sandstones, conglomerates, and quartzites and is part of the ancient Brazilian Shield (CPRM, 1994) with elevations averaging about 900m above sea level, and with peaks up to 1700m.

The climatic variability experienced in the region is attributable in part to seasonal and annual changes in atmospheric circulation over tropical South America and the tropical sector of the South Atlantic (the Hadley-Walker circulation cell) as well as to the high albedo of the northeastern region that reflects more radiation than neighboring areas (Amazonia and the Atlantic Ocean) and inhibits cloud formation. As such, years with greater precipitation occur when warmer waters of the tropical South Atlantic Ocean (and correspondingly cooler waters of the tropical North Atlantic ) generate strong North-Eastern trade winds that force the Intertropical Convergence Zone (ITCZ) towards the south, where it covers the Brazilian northeast. On the other hand, “drought” years occur in this region when the waters of the tropical South Atlantic Ocean are cooler (with correspondingly warmer waters in the tropical North Atlantic ). This situation generates stronger South-Eastern trade winds that force the ITCZ now towards the north, resulting in less cloud cover and less rainfall in NE Brazil (Nimer, 1977; Uvo and Nobre 1989). Additionally, the El Niño/Southern Oscillation (ENSO) phenomena, originating in the eastern Pacific (Senado Federal, 1997), has significant effects on this same region in Brazil , being associated with unusually strong episodes of drought. The climate in the Sincorá Range itself is seasonally humid and mesothermic, with mild, wet summers and cool, dry winters. Average monthly temperatures range from about 18° to 22°, although these vary considerably with altitude. Freezing temperatures have never been reported in the region. The period of heaviest rainfall is normally from November to March, with an average yearly rainfall of 1192 mm , although total precipitation is extremely variable (e.g. 357 mm in 1993 and 1721 mm in 1989), as is its distribution throughout the year (Harley, 1995; Seixas, 2004). The dry season usually lasts from July to early November, but periods of prolonged drought are not uncommon.

The eastern mapping region is a generally flat but deeply dissected plane, covered principally by distrophic latosols. Elevations there range from approximately 400 m to 600 m a.s.l., with an annual average rainfall at the extreme eastern edge of the mapping area of 580 mm . The western sector of the study area is a large, flat plain at approximately 900 m a.s.l., covered principally by distrophic argisols and latosols over deep Tertiary-Quaternary deposits (Rocha et al. , 2005). The annual average rainfall at the extreme western edge of the mapping area is 797 mm .

Present day vegetation types

The present day vegetation forms the starting point for any attempts to recreate the past. A map indicating the current distribution of the vegetation forms present in and around the CDNP is available from the home page of this site, and must reflect the primary plant communities extant before the arrival of Europeans in Brazil .

Vegetation types

A diverse series of vegetation types and landscapes were recognized, which are listed and briefly described below.

1. Campo rupestre (CR)

2. Sandy sedge meadows (M)

3. Latosolic forests (F1)

4. Montane lithosolic forests (F2)

5. Riparian forests (F3)

6. Wetlands (W)

7. Transition/Mosaic areas (T)

8. Cerrado (Brazilian savannas) (C)

9. Caatinga (semi-arid vegetation) (caat)

10. Capitinga (open vegetation on deep sands) (S)

11. Other forests (F4)

Two categories that were included on the current vegetation map of the region, alluvial sands and anthropogenic areas, are not strictly relevant in a consideration of past ecosystems. This topic will be discussed in greater detail in the following sections.

Campo rupestre – is an open, low vegetation form found on nearly continuous rock surfaces on mountain tops (usually above 800 m a.s.l.) and is primarily composed of sclerophyllous, evergreen, small trees, shrubs or sub-shrubs (Harley and Simmons, 1986). It is not a single physiognomic vegetation type, but varies according to the substrate in a continuous mosaic of micro-habitats and floristic groupings, where certain families are characteristic, notably: Orchidaceae, Cactaceae, Bromeliaceae, Velloziaceae, Asteraceae, Cyperaceae, and Melastomataceae. Other areas with blocks of loose stone or deep cracks that can collect slightly greater amounts of soil and humidity will harbour more shrubs and small trees, represented by families such as: Clusiaceae, Verbenaceae, Asteraceae, Begoniaceae, Melastomataceae and Aquifoliaceae (Conceição et al. , 2005).

Sandy sedge meadow (M) vegetation is a variation of campo rupestre that occupies relatively flat landscapes in the mountains on extremely thin, sandy, litholic neosols that are often saturated for fairly long periods of times during the rainy season. The same type of vegetation can also occur within high mountain valleys that have deep sandy soils but are subjected to frequent burning (which removes the forest cover). These very low and open sedge grasslands are dominated by Cyperaceae, Gramineae, and Eriocaulaceae (Harley and Simmons, 1986). Grasses predominate where the sandy soils are thinnest (together with many species of Cyperaceae, Leguminosae, and Melastomataceae), and these lands are often used for pasturing domesticated animals during the regional dry season.

Sub-montane to montane semi-deciduous seasonal forests on deep latosols (Latosolic forests) ( F1 ) – these are semi-deciduous forests that occupy also the entire eastern border of the Chapada Diamantina at altitudes between approximately 400- 800 m a.s.l. and are subject to seasonal variations in rainfall (and, to a lesser extent, temperature). The landscape is flat to deeply dissected, and is covered with latosols that range in color from yellow to orange to dark red. These soils are generally very deep, well drained, dystrophic, highly acidic, and demonstrating a high aluminum content and little organic matter (MME, 1981; CPRM, 1994) (although there is usually a thin leaf litter layer).

Sub-montane to montane semi-deciduous seasonal to evergreen forests on litholic neosols (Montane lithosolic forests) ( F2 ) - cover the river valley slopes beyond the extent of the riparian forests in the Sincorá Range itself. These slopes can be boulder-strewn and sometimes exceeding steep. Due to the near ubiquitous presence of sandstone rocks in the area of the Park, the soils of these valley sides are sandy, usually thin and stony, dystrophic, and become progressively drier at greater distantes from the river (this effect can be modified depending on exposure, aspect, or the presence of natural seeps). As such, and although predominantly evergreen, these montane forests demonstrate a greater degree of deciduousness in the dry season than the gallery forests (LS Funch et al. , 2002). At altitudes near 1000 m and above, however, environmental conditions become increasingly more humid due to longer daytime cloud cover, dew and more rainfall. These forests are also generally found in more protected sites in the mountains such as narrow valleys or deep crevices in the sandstone rock that are sheltered from both drying and wildfires. The soils are sandy, rocky, and not especially deep, but due to the high humidity and lack of fires there is a significant accumulation of damp organic material on the forest floor. Forests become increasingly more humid and evergreen as they occupy higher sites in the mountains.

Sub-montane to montane evergreen riparian forests ( Riparian forests) ( F3 ) –are seasonal evergreen forests that follow river courses throughout the region. In the Sincorá Range itself the rivers margins are strewn with boulders, but the sandy substrate that collects between the rocks is continually humid, although dystrophic, and has a very low organic material content. As these forests are only found adjacent to perennial water courses, they are usually of very restricted width (mostly less than 10 m ) and will merge rapidly into sub-montane to montane semi-deciduous seasonal forests as the soil becomes drier and the inclination of valley sides become more pronounced. At higher altitudes, they become continuous with the montane evergreen forests

Wetlands ( W ) – are permanently inundated or boggy areas.

Transition/Mosaic Areas ( T ) – As a result of many interacting ecological factors related to altitude, exposure, slope, local geology, soil composition, soil depth, microclimate, etc., the contact between two (or more) vegetation forms will result in a species-rich transition zone (often a mosaic) composed of elements from both communities.

Cerrado (Brazilian savannas) ( C ) vegetation is characterized, in general, by the presence of both a ground and an arboreal layer. The ground layer is continuous in all but the pure forest physiognomy (called cerradão ), and is composed principally of grasses, sedges, and numerous subshrubs (which can appear to be herbaceous but possess an often well developed root systems with xylopods), acaulescent palms, with very few annual species. The arboreal layer is likewise generally discontinuous and up to 10 m tall, with twisted branches, thick bark, and leaves that are (generally) perennial, large, and thick.

Cerrados demonstrate wide and continuous variations in their physiognomy, ranging from open savannas with predominantly herbaceous plants to the forest formations. They have been traditionally classified into five basic classes: “campo limpo” ( C1 ), open savannas with predominantly herbaceous plants; “campo sujo” ( C2 ), open savannas with predominantly herbaceous plants with low ( £ 3m), well-spaced shrubs; “campo cerrado” ( C3 ), with taller ( £ 5m), more closely spaced shrubs; “cerrado sensu stricto ” ( C4 ), with taller shrubs and small trees more densely spaced, but still with an open canopy layer; and “cerradão” ( C5 ), with a forest physiognomy and an essentially closed canopy (Coutinho, 1978).

Cerrado vegetation typically grows on deep, well drained, dystrophic soils, with low pH and consequently very high levels of exchangeable aluminum. They are subject to a markedly seasonal climate of heavy (up to 1500 mm yr -1 ) summer rains (October to March) followed by a long drought period during the winter months. As a consequence of the prolonged dry period, fires are common in this biome and the plants have many adaptations to this fire regime, including thick bark, bud protection, xylopods that aid in root-sprouting, and some plants have fire-responses that synchronize flowering and induce fruit dehiscence.

Caatinga – (caat) vegetation occupies most of the extensive semi-arid region of NE Brazil . Rainfall levels are low (usually between 500–700 mm yr -1 ) and extremely variable, with an extended dry period of up to 9 months. The vegetation is highly xeromorphic, the plants having generally small leaves, spines, thick bark, and a well-developed root system, often with tubers. With few exceptions (such as palms and Zizyphus joazeiro Mart.) , the plants are deciduous. Families such as Leguminosae (especially Acacia ), Euphorbiaceae, Cactaceae, and spiny Bromeliaceae predominate.

Capitinga ( CAP ) – Small, scattered, and isolated areas of capitinga (up to ca. 5ha) exist throughout the region around the CDNP. These are small but unique and isolated patches of deep, extremely well-drained, fine sandy soils. These sandy areas seem to be associated with the direct on-site weathering of friable sandstone outcrops (personal observation), completely surrounded by latosol landscapes. While they show a vegetation similar in physiognomy to the coastal restinga vegetation and the dune areas bordering the São Francisco River, Bahia, the local denomination for these sites has been retained as they are well know components of the local environment and used for harvesting mangabá fruits ( Hancornia speciosa Gomez), their species composition is quite different from these other sites (unpublished data), there is no saline component to the environment, and the sands were not derived from wind or water transport with subsequent re-deposition.

The vegetation in these areas of capitinga is low and open, with many shrubs, acaulescent palms, scattered and infrequent small trees, and cacti; annual plants are rare and there is a high percentage of bare ground.

Other forests (F4) – Significant areas of forest cover in the southwest section of the mapping area were identified on satellite images and during field investigations. One area (F4a) is located in the south-central section of the mapping zone on a flat landscape (~1000 m) of yellow/sandy argisols. The second forest area (F4b) is found slightly more to the south, at 1200- 1300 m , on an extremely irregular landscape of steep but low (~100m) hills. The soils there are usually dark red latosols. Both areas have been extensively altered by anthropogenic processes

The Map

A map of the presumed distribution of vegetation prior to 1700 in the Sincorá Range and neighboring areas is presented in here, with accompanying legend. This reconstruction of the vegetation forms and their extension in pre-European times is based on two principal assumptions:

•  the climate of 300 years ago is basically the same as that experienced today;

•  the fire regime in the region has been significantly modified during this same period.

These assumptions will be analyzed in greater detail in the Discussion section

Overview

A comparison between the current vegetation map of the region and that previous to intensive human intervention reveals numerous significant differences (which will be discussed in greater detail below in the section on Vegetation Changes):

1) sub-montane to montane semi-deciduous seasonal to evergreen forests (F2) occupied the deep sandy soils in the wide valleys above 900m in the Sincorá Range, which represents most of the area currently dominated by sedge meadows derived from burning (M2);

2) the same sub-montane to montane semi-deciduous seasonal to evergreen forests (F2) also occupied the litholic neosols throughout the range, limiting the campo rupestre vegetation to reasonably large “islands” of rocky surfaces found generally near or at the very summit of major peaks, and to a micro-mosaic of smaller exposed rocky sites on the lower slopes;

3) wide swathes (5-25m) of riparian forests bordered the rivers throughout the entire region in response to an infrequent fire regime (and the more perennially humid conditions that resulted);

4) wetlands were less extensive at the foot of the mountains before mine tailings clogged and widened the river valleys below, but were more frequently encountered in the mountains themselves before being drained and disturbed by mining activities;

5) arboreal formations of cerrado ( cerrado s.s. and/or cerradão ) previously occupied most of the western-central region contiguous with the Sincorá Range, where the landscape is more dissected and has deeper hollows, with a less arboreal physiognomy ( cerrado s.s.) on the flatter landscape to the south. The more open forms of cerrado ( campo sujo and/or campo cerrado ) occupied only the convex tops of the rolling plains;

6) the extent of coverage of the caatinga vegetation has not changed significantly, although its previous physiognomy was probably more arboreal;

7) the overall extent of coverage of latosolic forests (F1) has not significantly changed in the region as a whole, although it originally occupied more small sites on the lower eastern slopes of the Sincorá Range before being eliminated through mining activities;

8) transition zones have very much changed in nature and in location in the mapping region;

9) alluvial/sandy areas were not present at the base of the mountains (for they are the result of mining activities);

10) capitinga areas have not demonstrated significant alterations over time.

Discussion

Climatic variations

Any attempts to project historical vegetation cover confront uncertain variables such as long and short-term climatic changes and anthropogenic influences. While rainfall can be extremely variable from year to year in the region mapped during the present work (Funch et al. , 2002; Nolesco, 2002), meteorological records dating back to 1911 (MME, 1981), and historical records (Sampaio, 1936) sustain the assumption that the local climate of 1700 was probably not significantly different from that of the present. The question of anthropogenic influences is considerably more complex, however.

Anthropogenic influences

Amerindians

There is no doubt that burning was a widespread practice by essentially all Amerindians, only the degree of that fire use is in question. The dense Amazon forest, for example, once considered pristine (and perhaps even immune to large-scale alteration by human cultures not dominating metal technology) has yielded evidence of extensive anthropogenic burning, soil management, and planting of useful plant species (Colinvaux et al. , 1996; Behling, 1997; Lehmann, et al. , 2003). Dendrochronology has likewise demonstrated that Amerindian-set fires were common in the Mid-Willamette Valley in Oregon (USA). Early settler's accounts indicated that the now forested valley was then covered by extensive grasslands with few large isolated trees, a situation maintained by these native peoples long before European colonization (Halbeck, 1961). Similarly, accounts by travelers in Mato Grosso State in Brazil in the early 20 th century vividly describe intensive burning attributed to the local Amerindians (Fleming, 1957).

While the degree of influence of fire on the early vegetation of the mapping region in and around the Sincorá Range cannot be definitively ascertained as no records exist that in any way document their frequency or extent, it must be assumed that the native Amerindian populations were responsible for at least occasional burning in the region. However, it also seems reasonable to assume that with native population densities two or three orders of magnitude lower than modern (European) levels in the region, and with their settlements being concentrated in the more fertile depressions outside of the mountain range (in the caatinga and cerrado biomes), fires would probably not have been prevalent in the highlands of the Sincorá Range itself.

European influences

The region surrounding the Sincorá Range was only sparsely populated by Europeans prior to 1700. Due to the navigability of the São Francisco river to the west, the suitability of the dry land vegetation there to open-range cattle, and the earlier cycle of gold mining (~1700) to the southwest in the region of Rio de Contas, there were small early European settlements throughout the region west of the Sincorá Range. The presence of farmers and their cattle in the cerrado and caatinga regions to the west of this range brings forward the probability of periodic burning in these regions. However, the analysis presented here assumes an overall fire frequency much less than that occasioned in the last few recent decades, due to the much lower population densities of these early settlers and the absence of evidence of habitation within the mountains of the Sincorá Range itself.

It must also be noted that fire has historically been a management tool used by agricultural (and, later, cattle-raising) communities principally for the purpose of optimizing gains from the local vegetation, in contrast to the post-1844 mining culture that used fire indiscriminately and constantly in an attempt to maintain the mountains as free as possible of any and all vegetation in order to facilitate prospecting and related mining activities (RR Funch, 2004). This topic will be discussed again in the section below on forests.

Reconstruction of historic vegetations

The reconstruction of historical vegetations has been undertaken in many different areas using a variety of techniques. For the more remote past, it is necessary to examine archeological and archaeobotanical evidence, including the fossil record, pollen analysis, C14 dating of plant materials, and phytolith identification (Behling, 1997; Colinvaux, 1996; Rosen, 2005). For more modern epochs, written accounts of explorers, early newspapers, personal letters and diaries available after European colonization (Halbeck, 1961; Frelich, 1995), archaeobotanical evidence such as plant remains found in adobe bricks in Spanish missions in California (Tellman, 1996), and even more recent detailed notations from public land survey records (Manies and Mladenoff, 2000) have been used to obtain glimpses of previous vegetation coverage.

Vegetation changes

(1) Forest formations and sedge meadow vegetation

The pre-1700 vegetation map of the Sincorá Range demonstrates significant extensions (approx. 130 km²) of forest on deep sandy soils on the floor of the anticlinal valley running along its central N-S axis, a landscape currently dominated by open sedge meadows. Climatic, edaphic, and cultural evidence points to the fact these sedge meadows are not the original vegetation form of this area, but that these fields were in large part opened and are currently maintained by the frequent fires that scorch the mountain range. These areas of sedge meadows rarely pass more than a few years without burning (as they have traditionally been used as native pasture and are burned annually to improve grazing).

The situation in the past, however, was very different. Many of the sedge meadows seen today occur at altitudes from 900 to 1200m where water resources are more than sufficient to support forest formations. The forests themselves would also create a situation of positive feedback, maintaining local humidity and ground water reserves, and at least half of the total land area currently occupied by sedge meadows vegetation has soil sufficiently deep to support arboreal forms. As such, the montane semi-deciduous seasonal to evergreen forests on deep sandy neosols were almost certainly the dominant vegetation form on the non-rocky surfaces of the high altitudinal areas of the Sincorá Range .

It might be expected that under more constant conditions of higher humidity, shading and reduced evapotranspiration, some sections of these high mountain valley floors might drain very slowly, reducing soil oxygen levels, and favoring a sedge meadow vegetation rich in herbs. However, most of the high valley areas now given over to sedge meadows (in response to repeated burning) do not have flat (concave) valley floors, but rather repeated, wide convex profiles, with narrow drainage channels between the rises and very deep soils. The deep soils and convex topography favors relatively fast drainage, and boggy meadow areas are not very common. The original situation can still be seen in small remnant forest sections that have resisted repeated burning. These areas still retain a thin strand of riparian vegetation lining the relatively shallow drainage channels, while lithosolic forests occupy the higher ground.

There was, however, ample coverage of sedge meadow vegetation within the Sincorá Range during historic times in the high and open mountain valley areas above 900 m with extremely thin and poorly drained soils (drainage being impeded by the underlying rock). With less frequent fires, however, this sedge meadow vegetation would have been dominated by shrubs and sub-shrubs and would have had less of the open-meadow physiognomy common today.

(2) Forest formations and campo rupestre vegetation

Forest formations were the dominant vegetation form on the lower eastern slopes of the Sincorá Range in historic times, but recent (post-1844) anthropogenic activities have allowed the campo rupestre vegetation forms to expand throughout the Sincorá Range at their expense. Significant areas of sub-montane to montane semi-deciduous seasonal to evergreen forests on litholic neosols (F2) and on deep latosols (F1) on the lower eastern slopes of this range were completely eliminated as a result of cutting, burning, and subsequent erosion of the soils that supported them during the course of diamond mining activities. With the cessation of mining disturbances, the now exposed rock surfaces became occupied by many campo rupestre species (even at altitudes less than 600 m ), such as Orthophytum burle -marxii L.B.Sm.&R.W.Read Cyrtopodium edmundoi Pabst, Cattleya elongate Barb.Rodr., Pleurothallis ochreata Lindl. Anthurium affine Schott, Vellozia spp . , and Neoregelia bahiana (Ule)L.B.Smith, among many others. Arboreal species, especially Simarouba amara Aubl. , Tapirira guianensis Aubl. , Maprounea guianensis Aubl. , Clusia nemorosa G.Mey. , and shrubs such as Tibouchina pereirae Brade&Mgf., Chamaecrista spp., Byrsonima spp., and Pavonia spp. re-populated the area between the boulders and open rock surfaces where sandy soils could accumulate, resulting in a mosaic of vegetation cover on the steep rocky and boulder-strewn landscape (Fig. 8). Frequent fires generally maintain these proto-forests in early recovery phases. As such, forest formations were, in fact, the dominant vegetation form on the lower eastern slopes of the Sincorá Range in historic times.

During pre-European times, the dense montane neosolic forests (F2) on the higher slopes of the Sincorá Range left only the most compact rock surfaces available to the more heliophilous campos rupestre vegetation. While these forests were less damaged directly by mining activities at a later date, they suffered extensively from frequent anthropogenic burning. In continuous feed-back cycles of burning, invasion by more flammable species, drying due to less plant cover, and repeated burning, these montane neosolic forests were driven back to only the most protected and humid sites within the mountains. The elimination of these forest formations allowed the rupicolous forms of campos rupestre vegetation to expand onto the now more exposed smaller rock surfaces, while many of the shrub and sub-shrub species of this same vegetation form occupied the areas with deeper soils between the boulders and rock slabs (together with other true invasive and fire-loving species such as Panicum sp ).

(3) Riparian forest

Dense sub-montane to montane evergreen gallery forests bordered essentially all of the rivers in the region in pre-European times in response the humid conditions promoted by the forests themselves, the more perennial nature of the mountain streams (in large part due, again, to the presence of these same forests), and an infrequent fire regime.

(4) Wetlands

Wetlands were slightly less extensive in the mapping region surrounding the Serra do Sincorá before the mining era.

The Santo Antônio River at the eastern edge of the Sincorá Range (in its 30 km course between meeting the Utinga River and draining into the Paraguaçu River) certainly formed limited meanders and wetlands before diamond mining enveloped the area, for it displays wide banks, low inclination, and would experience flooding conditions during the rainy seasons. However, its course has more recently been clogged and widened by the deposition of enormous quantities of mine tailings from the mountains, and the now extensive wetlands are largely of anthropogenic origin.

The upper course of the Paraguaçu River as it enters the Sincorá Range itself is of naturally low inclination, and wetlands there were most likely present in historic times as there is no clear evidence that mining activities have affected its drainage regime.

Small swamps and bogs were almost certainly more frequently encountered in the mountains of the Sincorá Range before being drained and disturbed by mining activities; but these were never very extensive due to the steep local topography, and these have not been mapped due to their generally very reduced size.

(5) Cerrado

It can be readily seen in the field, as well as on satellite images, that the cerrado zone in the mapping area is divided into two principal but contiguous geomorphological areas. The smaller, very northern sector is deeply dissected, with many narrow valleys containing permanent water courses. These less exposed and generally more humid conditions that would have favored arboreal forms of cerrado vegetation ( cerradão ) in pre-European times (in addition to riparian forests).

The bulk of the cerrado zone to the south, however, is a relatively flat plain dissected by a dendritic drainage network, with extremely deep soils that range from sandy argisols to white, yellow, or (more rarely) red clays (latosols). Much of this drainage network (which ranges from the relatively shallow intermittent water courses to slightly deeper valleys with perennial streams) still retains a relatively dense vegetation cover (currently from shrubby to arboreal, depending on their depth and the local disturbance regime). An oblique view of the area reveals that the inter-drainage knolls have very low, convex profiles, and the arboreal vegetation occupying the drainage channels forms a low fringe that often completely surrounds this slightly higher ground. This is different from other areas of cerrado where the soils are often boggy and the more arboreal forms of that vegetation will grow selectively on areas of raised ground. As only a few meters of elevation define the summits of these knolls (which are invariably covered by the open campo sujo vegetation), it is reasonable to assume that in the absence of frequent burning these open knolls would be almost completely surrounded by more shrub/arboreal forms of cerrado ( campo sujo, campo cerrado ) that would grade downward into truly arboreal coverage ( cerradão ) in the deeper depressions/drainage courses (with gallery forests along the river banks). But even in the absence of human use or a rigorous fire regime, the low rainfall, extremely deep soils and open exposure of these elevated and convex inter-drainage knolls would favor much the same open (non-arboreal) vegetation that is encountered today.

It is also possible to consider that fire may never have been a principal (frequent) controlling factor in this cerrado zone, for the cerrado occupies only a relatively small area (~240 km²) in the region as a whole that is completely surrounded (buffered) by vegetation forms generally less susceptible to burning ( caatinga vegetation to the north and west, and more humid forest zones to the east and south).

While the physiognomy of the cerrado vegetation has been modified to a certain degree by anthropogenic forces, the total area it occupies apparently did not expand nor contract significantly since historic times. This is because this vegetation zone is sharply defined by climatic/geographic constraints: it is in the rain-shadow of the Sincorá Range in a region of deep latosols, constrained to the north by edaphic factors (thin neosols), the mountain range to the east, dry low-lands on thin neosols to the west, and irregular low mountainous areas to the south that are 100 to 200 m higher than the cerrado plain and apparently locally diminish the rain-shadow effect (there are also many areas with deep latosols that better retain soil moisture).

(6) Caatinga

Caatinga vegetation is found in the extreme NW sector of the mapping area. A comparison of the historic and present vegetation maps reveals little or no alteration in the overall distribution of this vegetation type. This stasis has been maintained because the caatinga zone (much like the neighboring cerrado located at exactly the same longitude) is sharply defined by climatic/geographic and edaphic constraints (in the rain-shadow of the Sincorá Range in a region of shallow litholic neosols).

The best preserved areas of caatinga in this region (in a steep, rugged landscape with no signs of habitation, farming, or free-ranging cattle) have an arboreal physiognomy, and this probably represents the historic condition of this vegetation type in the area.

(7) Latosolic forests

Continuous, dense and tall s ub-montane to montane semi-deciduous seasonal forests on deep latosols (Latosolic forests) (F1) stood to the east of the Sincorá Range befor modern encroachment. Amerindians inhabiting the region probably interacted with the forest much as other tribes that inhabited the coastal forests to the east, using fire as an agricultural tool when and where possible. But the lack of any metal tools would necessarily limit the scope of their impact. European-derived settlements (including the quilombos ) would have been limited to slash-and-burn agricultural systems on small plots, leaving the bulk of the regional forests essentially intact. As such, the overall extent of coverage of these latosolic forests has not significantly changed in the region as a whole, although it originally occupied more small sites on the lower eastern slopes of the Sincorá Range before being eliminated through mining activities, as described above.

(8) Transition/mosaic zones

In pre-European times, as outlined above, forest formations densely covered large portions of the lower eastern slopes of Sincorá Range , and had two basic components - latosolic and neosolic forests (F1 and F2 forests, respectively). The latosolic forests covered all of the plains to the east of the Sincorá Range, but were also were present in localized patches on the lower slopes of the mountains themselves, up to altitudes slightly over 800 m . Likewise, at neighbouring sites at the same altitude, montane neosolic forests (F2 forests) were established where the local topography had favored the deposition of relatively deep sandy neosolos, derived from the erosion of the sandstones and conglomerate rocks that compose the Sincorá Range itself.

However, there were other sites on the lower slopes of these same mountains between about 500 and 800 m where deep soils had not accumulated, usually where the land was more steeply inclined, that constituted transition zones (T1) between the forest formations and the campo rupestre vegetation generally found above 900 m . Importantly, these were transition zones not only in the sense of presenting a floristic composition with elements from the neighboring ecosystems, but they also represent what could be called temporal transition zones. During extended periods with more consistently humid weather, these shallow-soiled sections of the lower slopes could support low forest formations, with arboreal species derived from both the lithosolic and latosolic forests (F1 and F2 forests). During prolonged drought phases, however, these shallow soiled areas would be much more susceptible to drying and burning, which would tend to remove the forest cover, and leave it with a more open and herbaceous/shrub physiognomy (with only a few “testament” trees) as seen today, with a predominance of drought-resistant species such as Himatanthus articulatus (Vahl)Woodson, Syagrus harleyi Glassman, Hancornia speciosa Gomez, Tibouchina pereirae. The unnaturally high frequency of wildfires resulting from human use and occupation of the lower slopes now overwhelms any possible response (i.e. forest growth) to temporarily more humid climatic conditions.

These transition zones may have functioned as source areas for species that now occupy the formerly heavily forested eastern slopes that were mined and denuded during the last one hundred and sixty years. The removal of the original forest cover, together with the thick soils that supported them, created conditions very similar to the naturally thin soils occurring in this transition zone, and many of the same species found in these transition areas (cited above) can also be encountered within the areas devastated by mining activities.

Regions of contact between neosolic forests and campo rupestre vegetation on talus slopes high in the mountains (T1) are mosaic zones more than transition zones, in the sense that the two vegetation types do not share the same habitat requirements (forests on deep soils, campo rupestre on exposed rock surfaces). Additionally, these areas actually represent “remnant” conditions of much greater contact between neosolic forests and campo rupestre in pre-European times, when there was less fire (and more forest) and these two vegetation forms were more integrated within the mountains. With the loss of most forest cover on the rocky landscape, the otherwise ubiquitous mosaic of forest and campo rupestre became restricted to cliff edges where the trees were protected enough to survive repeated burning.

Likewise, the T2 ( campo rupestre /sedge meadows) and T5 ( campos rupestre+ sedge meadows + vestiges of fire-disturbed forest) transition zones are anthropic, in the sense that the sedge meadows have replaced the previously more extensive forest formations in their encounter with campo rupestre formations on these slopes.

The T4 zone with cerrado and caatinga (in the rain-shadow of the Sincorá Range) represents a true transition contact between cerrado vegetation to the east occupying deeper soils with slightly more rainfall and caatinga vegetation to the west on thinner and drier soils.

The T9 and T3 mosaic zones between campo rupestre with caatinga or cerrado vegetation, respectively (also in the rain-shadow of the Sincorá Range) represent unique and natural mosaic contacts between campo rupestre vegetation occupying mostly rocky surfaces with caatinga or cerrado species encountered on sections with deeper soils.

(9) Alluvial/sandy areas

Locally extensive alluvial fans of almost pure sand (derived from intensive mining activities in the recent past) are now found at the mouths of rivers which drain the diamond mining zones of the Sincorá Range . None of these alluvial fans existed before the mining era (with the possible exception of the margins of Paraguaçu River as it enters the Sincorá mountain range near the very center of the mapping area.

(10) Capitinga

Capitinga vegetation occupies extremely small and isolated sections of land along the eastern border of the Sincorá Range . As these areas are of reduced size and have no economic value, they have been little disturbed and have not demonstrated significant alterations over time.

Conclusions

We can therefore visualize a pre-European vegetation composed of dense forests on the deeper soils at low altitudes immediately to the east of the Sincorá Range merging seamlessly with the riparian and montane forests on thinner neosoils up to nearly the tops of the highest mountain peaks, limited only by the availability of sufficient soil to hold them against the wind and seasonal water shortages. Campo rupestre vegetation would be isolated on large to very small rock outcrops scattered through the Sincorá Range , a factor that would have favored the rise of the many endemic species observed from the region and the uneven distribution of species composition observed over very short distances. To the west, arboreal/shrubby caatinga and various physiognomies of cerrado vegetation grew in the rain shadow of the mountains, with only a narrow transition/mosaic zone between them and the mountain vegetation due to the generally abrupt scarp edge that defines the western side of the range.

In can be seen, however, that over the last 300 years there have been dramatic changes wrought by the impact of European culture. Mining activities, anthropogenic burning, timber harvesting and agricultural clearing, and more recently urban development and tourism have all contributed or are contributing to the degradation of the natural environment, with the inevitable loss of both plant and animal biodiversity. A landscape of high altitude campo rupestre , montane forests, savannas and wetlands have been degraded into far more complex and disturbed, if less biodiverse, patterns of vegetation communities.

The neosolic montane forests have been eliminated by fire from over 470 km² (80 %) of their previous range, and these forests are now generally restricted to only the deepest and most protected canyons in the mountains. L atosolic forests have suffered severe damage from burning on these same lower slopes but, in contrast to the neosolic forests, have rarely been totally eliminated through this contact with fire. This difference may be related to the capacity of the dense latosols to harbor more moisture even in drought situations, thus favoring a partial resistance and recuperation of standing trees, root sprouting, and perhaps even the protection of some seeds or seedlings from fire destruction.

The retreat of the two principal forest types (F1 and F2) has resulted in the occupation of part of their former range by campo rupestre and the related sedge meadow vegetation within the higher elevation areas. On the lower mountain slopes, these formerly dense forest areas that were altered by mining have been replaced by primary succession trees, some campo rupestre species, a mixture of shrub and sub-shrub species from neighboring transition zones, as well as common invasive plants (such as Panicum sp ).

Just outside of the Sincorá Range the native vegetation (principally latosolic forests , caatinga , and cerrado ) has been severely impacted through human use and occupation, but there has been no major discernable shift in their distribution patterns since pre-European times.

The eventual success of conservation programs, and especially the suppression of wildfires, in the Chapada Diamantina National Park should result in: 1) the recuperation of large sections of montane semi-deciduous seasonal to evergreen forests on litholic neosols in all areas with sufficiently deep soils, with; 2) a reduction of the range of some strongly heliophilous plant species of the campo rupestre vegetation due to shading by the re-growth of this arboreal vegetation; 3) the re-growth of dense shrub and sub-shrub vegetation in areas now dominated by open sedge meadows in the highest regions of the Sincorá Range with thin neosols; 4) a significant recuperation of gallery forests that would provide important corridors for the dispersal of wildlife, and; 5) a positive feedback effect on water resources in the mountains, promoting the resurgence of perennial streams in the highland areas and perhaps the reduction of the now common flooding, this by reducing immediate runoff and promoting a greater retention of ground water.

It is still too early to say how much biodiversity has been or is being lost in the region, but what remains is increasingly vulnerable unless we can support programs of conservation and the development and protection of protected areas. The next paper in this series will describe in detail conservation recommendations for the Chapada Diamantina National Park and the neighboring caatinga and cerrado vegetations.          

Map legend , listing and describing the vegetation and landscape classifications used for mapping the pre-European vegetation of the Sincorá Range and bordering lands, Bahia, Brazil.

Behling, H. 1997 . Late glacial evidence of Podocarpus and probable indications of the first Amerindians in the Amazon mouth region ( Para State , Brazil ). Vegetation History and Archaeobotany 5 :241-246

Conceição AA, Rapini A, Pirani JR, Giulietti AM, Harley RM, Silva TRS, Santos

AKA, Correia C, Andrade IM, Costa JAS, Souza LRS, Andrade MJG, Funch RR, Freitas TA, Freitas AMM, Oliveira AA. 2005 . Campos Rupestres. In: Junca, F.A., Funch, L., Rocha, R. (org.), Biodiversidade e conservação da Chapada Diamantina . Brasília: Ministério do Meio Ambiente, 153-180.

Colinveaux PA, Oliveira PE de, Moreno JE, Miller MC, Bush MB. 1996 . A long pollen record from lowland Amazonia : Forests and cooling in glacial times. Science 274 :85-88

Coutinho, LM. 1978 . O conceito de Cerrado. Revista Brasileira de Botância 1 :17-23.

CPRM (Companhia de Pesquisa de Recursos Minerais). 1994 . Parque Nacional da Chapada Diamantina - BA. Informações básicas para a gestão territorial: diagnóstico do meio físico e da vegetação . CPRM, IBAMA, Salvador , Bahia , Brazil .

Dillehay TD. 1996 . Monte Verde: A Late Pleistocene Settlement in Chile . Vol 2: The Archaeological Context and Interpretation . Smithsonian Institution Press , USA  

Fleming P. 1957 . Brazilian Adventure . Penguin Books, Middlesex , England .

Frelich LE. 1995 . Old forests in the Lake States today and before European settlement. Natural Areas Journal 15 : 157-167  

Funch LS, Funch RR. Barroso GM. 2002 . Phenology of Riparian and Montane Forest in the Chapada Diamantina, Bahia , Brazil . Biotropica 34 : 40-50.

Funch RR. 2004 . A visitor's guide to the Chapada Diamantina . Gráfica Terra, Goias, Brazil.

Funch RR. 2006 . Avaliação dos limites do Parque Nacional da Chapada Diamantina através de análise da vegetação. Phd Thesis. Universidade Estadual de Feira de Santana, Bahia, Brazil.

Halbeck JR. 1961 . The original vegetation of the Mid-Willamette Valley, Oregon. Northwest Science Vol. 35, No. 2

Harley RM. 1995 . Introduction. In: B.L. Stannard ed. Flora of the Pico das Almas , Chapada Diamantina, Brazil . Kew , Royal Botanic Gardens , 27-63.

Harley RM, Simmons NA. 1986 . Florula of Mucugê, Chapada Diamantina, Bahia , Brazil . Kew , Royal Botanic Gardens .

Harley RM and Giulietti AM. 2004 . Wild Flowers of the Chapada Diamantina. Illustrated walks in the mountains of NE Brazil (Flores Nativas da Chapada Diamantina . Trilhas botânicas ilustradas nas montanhas do Nordeste do Brasil . Rima Editora, São Carlos, São Paulo, Brasil

IBGE (Instituto Brasileiro de Geografia e Estatísticas). 2006. Mapa de Biomas e de Vegetação . www.ibge.gov.br (12 June 2006)

IPAC–BA (Inventário de Proteção do Acervo Cultural da Bahia). 1980 Monumentos e sítios da Serra Geral e Chapada Diamantina Vol. IV . Salvador, Bahia.

      Lehmann J, Silva Jr. da, Steiner JP, Nehls C, Zech T, Glaser B. 2003 . Nutrient availability and leaching in an archaeological Anthrosol and a Ferrosol of the central Amazon basin: fertilizer, manure and chemical amendments. Plant and Soil 249 :343-347

Manies KL, Mladenoff DJ. 2000 . Testing methods to produce landscape-scale presettlement vegetation maps from U.S. Public Land Survey records. Landscape Ecology 15 :741-754

MMA (Ministério do Meio Ambiente). 2006 . Mapeamento dos ecorregiões brasileiras. www.mma.gov.br. (5 June 2006)

MME (Ministério das Minas e Energia). 1981 . Projeto Radambrasil, Folha SD.24 Salvador; geologia, geomorfologia, pedologia, vegetação e uso potencial da terra . Rio de Janeiro, Brazil.

Nimer E. 1977 . Clima. In: Geografia do Brasil - Região Nordeste . Vol. 2 IBGE. Rio de Janeiro.

Nolesco MC. 2002 . Registros geológicos gerados pelo garimpo, Lavras Diamantina – Bahia. PhD Thesis, Universidade Federal do Rio Grande do Sul.

Rocha WJS da F, Chaves JM, Rocha CC da, Lobão JB. 2005 . Unidades de Paisagem da Chapada Diamantina. In: Junca, FA, Funch L, Rocha R. org. Biodiversidade e conservação da Chapada Diamantina . Brasília: Ministério do Meio Ambiente, 47-64.

Rosen A. 2005 . Phytolith indicators of plant and land use at Çatalhöyük. In: Hodder I ed.  Çatalhöyük Project Volume IV: Inhabiting Çatalhöyük . Cambridge: MacDonald Institute, Cambridge University.

Sampaio T. 1936. O Rio São Francisco, trechos de um diário de viagem . (1 st ed. 1905) Livraria J. Leite, Rio de Janeiro, Brazil.

Seixas BLS. 2004 . Água: usos, características e potencialidades . Nova Civilização, Cruz das Almas, Bahia, Brazil.

Senado Federal. 1997 . Relatório da Comissão “El Niño”. Secretaria Especial de Editoração e Publicações. Brasilia, Brazil.

SUDENE. 1977 . Região Nordeste do Brasil – 1:100.000 . Ministério do Interior – Superintendência do Desenvolvimento do Nordeste, Brasília, DF, Brazil.

Tellman B. 1996 . Stowaways and invited quests: how some exotic plants reached the American southwest. Symposium Proceeding. California Exotic Plant Council , 1-7.

Torreno PC. 1926 . Pela Terra Diamantina – separata da Revista Brotéria 1924-1925. Imprensa Official do Estado, Bahia.

Uvo CRB, Nobre CA. 1989 . A zona de convergência intertropical (ZCIT) e a precipitação no norte do nordeste do Brasil. Part II. A influência dos ventos e TSM do Atlântico Tropical. Climanalise 4 :39-48.

^top