Tropical environments

Two tropical climates: where they are and what they are like

Both climates are hot all year, so the difference between them is not temperature but how the rain is spread through the year. The humid tropical (rainforest) climate is found within about 10° of the equator, in the Amazon basin, the Congo basin, Indonesia, Malaysia, New Guinea and parts of Central America. The seasonally humid tropical (savanna) climate lies polewards of it, roughly between 10° and 20° north and south, in the Brazilian cerrado, the Llanos of Venezuela, the savanna belts north and south of the Congo basin, northern Australia and the drier interior of India.

In the rainforest every month is hot and every month is wet. Mean monthly temperatures sit at about 26-28 °C throughout the year, so the annual range is only about 2-3 °C - smaller than the difference between day and night in many other climates. Rainfall is above 2 000 mm a year in most areas and can exceed 3 000 mm on windward slopes, and it falls in every month, usually as heavy convectional afternoon storms with thunder.

The savanna has the same heat but splits the year in two. The wet season arrives when the rain-bearing low pressure belt moves overhead and brings 500-1 500 mm in a few months; the dry season follows when high pressure settles over the area and there may be no rain at all for several months. Temperatures are high throughout, and the hottest weeks usually come at the end of the dry season, just before the rains break, when the sun is overhead, the ground is bare and no cloud moderates the heat.

The two kinds of variation are examined as often as the totals. Diurnal range is small in the rainforest, about 7-10 °C, because cloud and humidity trap heat at night and reflect sunlight by day. In the savanna dry season it can exceed 20 °C: with no cloud to hold heat in, nights are surprisingly cold, while the unshaded ground heats fiercely by midday. The annual range is small in the rainforest and much larger in the savanna, growing larger still further from the equator and deeper into a continent.

Explaining the pattern: the ITCZ, subtropical anticyclones, monsoons and ENSO

The intertropical convergence zone (ITCZ) is a belt of low pressure near the equator where the north-east and south-east trade winds meet. The air there is heated and loaded with water vapour evaporated from warm seas, so it rises, cools, condenses into towering cumulonimbus cloud and produces heavy convectional rain. The ITCZ follows the thermal equator, the line of highest temperature, which shifts north and south through the year as the sun moves, lagging about a month behind it because land and sea take time to heat.

That migration explains both climates. Where the ITCZ stays overhead all year, as in the equatorial belt, rain falls in every month and you get rainforest. Where it moves away for part of the year, the rain stops while it is gone, and you get savanna: the wet season is the period when the ITCZ is overhead, and the dry season is the period when it has moved away. Because the ITCZ crosses the equator twice a year, equatorial places have two rainfall maxima near the equinoxes, while savanna places have one.

The subtropical anticyclones are belts of high pressure centred at about 30° north and south, where air that rose at the equator has cooled, become denser and is sinking. Sinking air warms as it descends, so its relative humidity falls and the sky clears - these are the belts of the world's great deserts. In the dry season the subtropical high expands equatorwards and takes over the savanna, giving months of clear, dry, sunny weather, which is also why the vegetation is dry enough to burn.

A monsoon is a seasonal reversal of the prevailing wind, driven by the different rates at which land and sea heat and cool, working with the ITCZ. In summer the Asian landmass heats far faster than the Indian Ocean, so low pressure forms over the land and draws moist air in from the sea: the south-west monsoon blows from the Indian Ocean between about June and September and supplies almost all the rain of the seasonally humid tropics of South and South-East Asia. In winter the land cools faster than the sea, high pressure builds over Asia and the wind reverses as the dry north-east monsoon, which is dry over most of India but picks up moisture crossing the Bay of Bengal and brings rain to the Coromandel coast. West Africa works the same way, with moist south-westerlies from the Atlantic bringing the wet season to the Sahel.

El Nino Southern Oscillation (ENSO) is an irregular oscillation of pressure and sea-surface temperature across the tropical Pacific that can override the normal pattern for a year or more. In an El Nino year the trade winds weaken or reverse, warm surface water spreads eastwards and the main area of rising air moves with it, so Indonesia, Papua New Guinea, northern Australia and north-east Brazil become unusually dry while Peru and Ecuador get torrential rain and floods. In a La Nina year the trade winds strengthen, warm water piles up in the west and the pattern reverses, with heavy rain and flooding in Indonesia and eastern Australia and drought on the Peruvian coast. For this topic the drought side matters most: El Nino years raise the fire risk in the Amazon and Indonesia and shorten the wet season in southern Africa.

• ITCZ: low pressure where the trades converge; air rises, cools and rains. It follows the thermal equator and lags the sun by about a month.
• Rainforest: the ITCZ is overhead all year, so rain falls every month. Savanna: the ITCZ is overhead for only part of the year, so there is one wet and one dry season.
• Subtropical anticyclone: high pressure at about 30°, with subsiding air, clear skies and no rain. It expands equatorwards in the dry season.
• Summer monsoon: low pressure over the heated land draws moist onshore air - the south-west monsoon of Asia, the south-westerlies of West Africa.
• Winter monsoon: high pressure over the cooled land gives dry offshore winds, except where they cross a warm sea and pick up moisture.
• El Nino: weak trades, warm water in the eastern Pacific, drought in Indonesia, northern Australia and north-east Brazil, floods in Peru and Ecuador. La Nina reverses it.

Weathering processes and their role in shaping tropical landforms

Weathering in the humid tropics is fast because all three of its requirements are present at once: high temperatures, which speed up chemical reactions; abundant water, which is both reagent and transport; and dense vegetation, whose roots and decaying litter add acids and carbon dioxide to the soil water. The result is not just a weathered surface but a thick layer of rotted rock, the regolith or saprolite, tens of metres deep in places - and that deep regolith is the raw material of most of the landforms in this topic.

Hydrolysis is the reaction of minerals with water. In granite the feldspar is attacked first, because it is unstable at the surface: water breaks it down into kaolinite clay and soluble ions of potassium, sodium and calcium, which are carried away in solution, while the quartz is left behind as sand. This is why deeply weathered granite becomes a soft, clay-rich material that can be dug with a spade, even though fresh granite is hard. Carbonation is the reaction of minerals with carbonic acid, which forms when carbon dioxide dissolves in rainwater. It is the key process on limestone: calcium carbonate is converted to soluble calcium bicarbonate and carried away in solution as hard water. The reaction is fast in the tropics because warm water reacts quickly and because rainforest soil air can hold many times the carbon dioxide of the atmosphere.

Pressure release (dilatation) happens when the weight of overlying rock and soil is removed by erosion. The rock below expands and cracks, opening sheeting joints roughly parallel to the surface, and slabs between them peel away in sheets - one reason granite domes have smooth, rounded backs. Salt crystal growth occurs where salt solutions in pores and cracks are concentrated by evaporation until crystals form; the growing crystals push the rock apart and flaking and granular disintegration follow. It is most effective in the seasonally dry tropics and near the coast, where salt and strong evaporation coincide.

Chelation is the removal of metal ions by organic acids. Lichens, roots and humus release acids that combine with iron and aluminium ions and hold them in solution so they can be washed out of the rock, which etches shallow hollows and pits into bare rock surfaces and helps break down what hydrolysis leaves behind. Vegetation root growth is also a physical process: roots prise joints open, and a tree blown over lifts a slab of rock and soil with its root plate. Roots also keep the regolith open so water and air penetrate, and weathering continues at depth.

Together these processes explain the shape of tropical landforms. Deep chemical weathering along joints rots the rock but leaves the most resistant masses, the core stones, sound inside it. Later, when the climate dries or the vegetation is cleared and running water strips the soft regolith away - a process called stripping or exhumation - those core stones are left standing as tors, kopjes and bornhardts. On limestone the same equation runs the other way: nothing resistant is left, the whole rock is soluble, and the result is the hollow-and-cone landscape of tropical karst.

Granite landforms: bornhardts, kopjes and tors

Granite is a coarse-grained intrusive igneous rock, and two of its properties matter here: it is made of minerals with very different resistance to hydrolysis (quartz resists, feldspar does not), and it is cut by sets of joints, usually roughly vertical and horizontal. Water moves down the joints, so weathering attacks the rock along them, from the surface downwards and from each joint inwards. The boundary between rotted and sound rock is the weathering front.

Because weathering works inwards from the joints, rock between widely spaced joints survives as blocks of sound rock surrounded by rotted material - the core stones. Where joints are closely spaced the whole mass rots; where they are widely spaced, large rounded blocks remain. If the climate then becomes drier, or the vegetation is cleared and the regolith is washed away, the soft material goes and the core stones are left exposed, rounded further by sheets spalling from their backs. This two-stage history, deep rotting under a wetter climate followed by stripping, is why these landforms are found today in the seasonally humid savanna as well as in the rainforest.

A bornhardt is the largest form: a great dome of bare granite, hundreds of metres across, rising from a plain. A dwala is the smooth, rounded, whale-backed dome surface of such a hill, shallow enough to walk across, and a ruware is the lowest form of the same family - a broad, gently convex slab of bare rock that barely rises above the plain, with soil and grass in the shallow hollows between slabs. A kopje (koppie) is a small rocky hill made of a cluster of rounded granite blocks and boulders standing above the surrounding plain, and a tor is an isolated group of jointed blocks left on a hilltop or slope. The names overlap because they are one process at different scales, so a clear description of shape and process scores better than a fussy definition.

Two controls decide where the forms occur. Rock structure decides the pattern: widely spaced joints give large domes and huge core stones, closely spaced joints give small blocks and coarse gravel, and one dominant joint direction can line up a whole series of kopjes. Climate decides the depth of rotting, and therefore the size of the forms: the very deep regolith needed for a bornhardt requires a long period of hot, wet conditions, which is why bornhardts are typical of the humid and seasonally humid tropics and rare in temperate latitudes. Their presence today in Zimbabwe and Namibia is evidence that those areas were wetter in the past.

Limestone landforms: cone, cockpit and tower karst

Tropical karst needs four things: thick, pure, well-jointed limestone; heavy rainfall; high temperature; and vegetation to enrich the soil water with carbon dioxide. Limestone is removed by carbonation, and the reaction runs far faster in the tropics than in Britain because warm water reacts quickly and because rainforest soil air is rich in carbon dioxide from roots and rapid decay. Water enters along joints, so joints widen into gullies, gullies into runnels and runnels into closed hollows; once a hollow reaches the water table the drainage goes underground and the surface becomes a maze of hills and depressions, with streams that vanish into swallow holes.

Cone karst is a landscape of many rounded conical hills, often aligned along the main joints, separated by closed depressions and star-shaped valleys, with little surface drainage. The cones are the residue left between depressions that are being dissolved and widened from below. Cockpit karst is the same process where the joint network is dense and the rainfall very high: dissolution is concentrated into many small centres, producing deep bowl- or star-shaped hollows called cockpits separated by steep-sided cones, so that from the air the surface looks like an egg box. The Cockpit Country of Jamaica is the type area.

Tower karst represents a later stage. The towers are steep-sided, often near-vertical pillars of limestone with flat or pointed tops, standing on a flat plain of alluvium - the landscape of Guilin and Yangshuo in Guangxi, southern China, and of Krabi in Thailand. They form when dissolution has gone on so long that almost all the limestone between the towers has been removed: vertical joints guide the water down, the bases of the towers are attacked laterally by rivers and by the damp tropical air at ground level, and the debris is carried away, leaving isolated towers on a planed surface. Where the sea has flooded the plain, as at Ha Long Bay in Vietnam, the towers stand as islands.

Tropical karst is found in a broad belt between about 30° north and south wherever thick limestone meets a wet tropical climate: southern China and South-East Asia, Indonesia and the Philippines, the Caribbean (Jamaica, Cuba, Puerto Rico), Central America, Madagascar and parts of India. It is a landscape of underground drainage, so it is also a landscape of practical problems - no surface streams, thin soils that disappear down cracks, flash floods when a swallow hole blocks, and water supplies that are easily polluted because whatever enters a swallow hole reaches the caves unfiltered.

Vegetation: rainforest, savanna and the climax communities

A climatic climax is the community that would develop and remain in equilibrium with the climate if people did not interfere. In the humid tropics that is evergreen rainforest. In the seasonally humid tropics it is some form of savanna woodland or dry forest, depending on the length of the dry season and the rainfall total. Everything else in this section is a community held back from its climax, and two words do that job.

A subclimax is a community held below the climatic climax by a persistent physical factor rather than by people - seasonal drought, waterlogging, a thin or toxic soil, salt, or repeated natural flooding. A plagioclimax (also called a disclimax) is a community held below the climax by human activity - clearance, burning, grazing or mowing - and kept there only while that activity continues. The difference is the cause, not the appearance: the same grassland can be a subclimax on waterlogged clay and a plagioclimax where forest has been cleared and burnt every year.

The rainforest is stratified: an upper layer of emergents at 40 m or more, a continuous canopy at 20-30 m that intercepts most of the light, a shrub and herb layer beneath it, and a ground layer of seedlings and ferns in deep shade. Competition for light produces the characteristic adaptations - buttress roots for support in shallow soil, lianas that climb to the canopy, epiphytes growing on branches, drip tips to shed water, and large thin leaves in the shade. Flowering and fruiting are often spread through the year rather than seasonal, and biodiversity is very high, with many species each at low density.

The savanna is a grassland with trees, and the balance between the two is set by the length of the dry season, the rainfall total, the soil and the fire regime. Where rainfall is higher and the dry season short, trees form a woodland; where the dry season is long and rainfall low, it is open grassland with scattered trees. The plants are fire-adapted: grasses dry out and burn but their roots and buds survive below ground and regrow within days of the first rain, while trees such as acacia and baobab have thick corky bark, deep roots and the ability to resprout. Many savanna trees are deciduous or carry small, thorny, waxy leaves to reduce water loss, and the grass is mostly C4, which is efficient in heat and bright light.

Climate sets the limits and people decide where inside those limits the community actually sits. Forest clearance followed by repeated burning converts rainforest into derived savanna, a plagioclimax that looks natural and is entirely man-made. Grazing holds grassland open by eating tree seedlings, while removing large grazers or suppressing fire lets scrub and trees invade. Because these communities are maintained by a disturbance, removing the disturbance does not restore the climax quickly, and in the short term it produces a different community altogether.

• Climatic climax: the community in equilibrium with the climate - rainforest in the humid tropics.
• Subclimax: held back by a physical factor such as seasonal drought, waterlogging, thin soil or natural flooding.
• Plagioclimax: held back by people - clearance, burning or grazing - and dependent on that activity continuing.
• Rainforest structure: emergents, canopy, understorey, ground layer; buttress roots, lianas, epiphytes, drip tips.
• Savanna: grasses plus scattered trees; tree cover increases with rainfall and decreases with a longer dry season.
• Fire adaptation: underground buds in grasses, thick bark and resprouting in trees; frequent burning maintains grassland.

Soils: oxisols, tropical red earths and tropical brown earths

Tropical soils are the product of the same conditions that drive the weathering: heat, heavy rain and a large supply of organic matter. Three processes dominate. Hydrolysis breaks down the primary minerals and leaves clay behind. Leaching removes the soluble products - silica, potassium, calcium and magnesium - downwards and out of the profile, because rainfall exceeds evaporation for most of the year. Laterisation (ferralitisation) is the accumulation of the insoluble residue, the oxides of iron and aluminium, which give these soils their strong red and yellow colours and their low nutrient status. Organic matter is broken down so quickly by bacteria and fungi that little humus accumulates, and the nutrients released are taken up by the vegetation almost at once.

Oxisols, called latosols in older textbooks and ferralsols in the FAO system, are the typical soils of the rainforest. They are deep, heavily leached, strongly acidic and dominated by kaolinite clay and iron and aluminium oxides. They have a stable crumb structure and drain freely, but they hold very few available nutrients and almost no weatherable minerals to replace them. Tropical red earths are the soils of the wetter savanna and the forest margins: less extremely leached, red to reddish-brown, with more clay movement down the profile and a higher base status, so they are more fertile than an oxisol. Tropical brown earths belong to the drier savanna: browner, less leached, with more organic matter from grass roots, a near-neutral pH and available calcium, which makes them the most fertile of the three and helps explain why the savanna has historically supported denser farming populations than the rainforest.

In profile, a rainforest oxisol has a thin O layer of rapidly decaying litter, a thin dark A horizon, and then a very deep B horizon of red clay with iron concretions, merging into pallid rotted saprolite and then fresh rock. There is no sharp boundary between the horizons. A savanna brown earth has a thicker, darker A because grass roots die and decay within the soil each year, a B with clay and, in drier areas, nodules of calcium carbonate, and it rests on weathered rock or on a hard ironpan.

People change these soils quickly. Clearance removes the litter supply and exposes the surface to intense rain, so the thin A horizon is washed away and the soil compacts; burning destroys organic matter and drives off nitrogen; continuous cropping without a fallow removes the small reserve of available nutrients; and heavy machinery compacts the surface and reduces infiltration. Where an iron-rich horizon is exposed and hardens irreversibly it forms a duricrust or laterite that will not take a plough. The dark, fertile terra preta soils created by pre-Columbian communities in parts of Amazonia show the opposite: with a continuous input of organic matter and charcoal, these soils can be made productive.

Nutrient cycling, energy flows and trophic levels

The Gersmehl model shows the nutrient cycle as three stores joined by flows: litter on the surface, soil beneath it and biomass (the living vegetation). Nutrients move from biomass to litter as litterfall, from litter to soil by decomposition, and from soil back to biomass by uptake. Inputs come from precipitation and the weathering of rock; outputs leave by leaching and runoff, by erosion, and - in a savanna - by volatilisation in fire. The size of each store and the speed of each flow is what differs between the two ecosystems, and that is what the diagram is for.

In the rainforest the biomass store is huge, the litter store is tiny and the soil store is small. Decomposition is almost instantaneous in the heat and humidity, so litter does not accumulate; the nutrients released are taken up again within days by a dense mat of shallow roots and by mycorrhizal fungi, and trees recycle nutrients internally by withdrawing them from leaves before they fall. The flows are therefore very fast, and outputs are large in absolute terms because rainfall is heavy, but the tightness of the cycle means very little is lost. The consequence is a nutrient-conserving forest sitting on a nutrient-poor soil: remove the trees and you remove the store.

In the savanna the soil store is proportionally much larger and the biomass store smaller, because the dry season stops growth for months and because fire returns nutrients to the ground as ash. Fire is the distinctive flow: it transfers nitrogen and sulphur to the atmosphere by volatilisation, where they are lost to the system, while the bases - calcium, potassium and magnesium - remain in the ash and return to the soil. Repeated burning therefore produces a soil that is base-rich but nitrogen-poor. Uptake is seasonal, concentrated in the first weeks of the wet season when the grass grows fastest.

Energy enters both systems as sunlight and is fixed by the producers - trees and shrubs in the forest, grasses and trees in the savanna - into net primary productivity, then passed up the trophic levels: producers, primary consumers (insects and grazers), secondary consumers (birds, reptiles, small predators) and top predators. Two things follow. Energy is lost at every transfer, so food chains are short and top predators are few for the area of land. And the pyramid of numbers can look inverted when one huge tree supports thousands of insects, even though the pyramid of biomass stays upright because the tree itself is so large.

People attack the cycle at its weakest point. Clearing removes the biomass store, so litterfall stops, decomposition cannot replace what uptake and leaching remove, and the soil is left exposed to erosion. Burning volatilises nitrogen and sulphur, kills much of the soil fauna that drives decomposition, and, if repeated annually, turns a nutrient-conserving system into a nutrient-exporting one. Deforestation on a large scale also cuts the input side, because much of the rainfall in Amazonia is recycled by the forest itself: water that would have been transpired and rained out again runs off instead. That is why the nutrient cycle and the water cycle fail together.

• Stores: litter, soil, biomass. Flows: litterfall, decomposition, uptake.
• Inputs: precipitation and weathering. Outputs: leaching, runoff, erosion, and volatilisation in fire.
• Rainforest: huge biomass store, tiny litter store, small soil store, very fast flows, tight recycling.
• Savanna: proportionally large soil store, seasonal uptake, and fire losing nitrogen and sulphur to the atmosphere.
• Energy: sunlight fixed by producers into net primary productivity, then lost at each trophic level, so food chains are short.
• Human impact: clearing cuts litterfall, burning volatilises nutrients, deforestation reduces rainfall recycling.

Change and challenge: pressure on the tropical ecosystems

The pressure on both ecosystems comes from three things: more people, more demand for land and more demand for commodities. Populations in tropical countries have grown quickly, and governments have often encouraged settlement of forest land to relieve pressure elsewhere - building roads into it, offering land titles to smallholders and granting incentives to ranching and plantation companies. Once a road exists the forest behind it becomes accessible, and clearance follows the road in a fishbone pattern of spurs that is visible from space.

Large-scale monoculture replaces a diverse forest with a single crop - soy, oil palm, sugar cane, rubber or eucalyptus for pulp - which supports far fewer species, uses herbicides and fertilisers, compacts the soil and removes the fallow that traditional farming relied on. Logging removes the most valuable timber, and even selective logging damages the canopy, opens the forest floor to drying and fire, and requires roads, while clear-felling removes the biomass store entirely. Mining for iron ore, bauxite, copper, gold and tin strips the surface, produces tailings and polluted water, and in artisanal gold mining releases mercury into rivers; mines also need roads, power lines and settlements, so the damage spreads far beyond the pit.

Climate change adds a threat the ecosystems cannot adapt to quickly enough. Temperatures in the tropics are rising, and projections point to a drier dry season in Amazonia and parts of the savanna and to more intense rainfall when it does rain, which means longer droughts, worse fires and more erosion within the same year. Wildfires increase because dry seasons are longer, because logged and fragmented forest is drier and more flammable, and because fire is used to clear land; a fire that escapes into standing rainforest kills trees that have no adaptation to it, and repeated fires can convert forest to scrub. The fear is a feedback loop: forest dieback reduces the rainfall the forest itself generates, which causes further dieback.

The Brazilian Amazon is the detailed example. It holds about 60% of the Amazon basin and the largest remaining area of tropical rainforest on Earth, and its main land uses are cattle ranching, soy, logging, mining and smallholder settlement. The issues are those above: clearance and fire, the loss of the biomass store and with it the nutrient cycle, loss of biodiversity, soil erosion, the risk of a drier and more flammable forest, and the social cost to indigenous and riverine communities whose land and livelihoods depend on the forest standing.

The management response has four parts. Protection: a large network of federal and state protected areas and indigenous territories, which consistently show lower deforestation than surrounding land, and the Forest Code, which requires landowners in the Amazon to keep 80% of their land as forest. Monitoring and enforcement: the national space institute INPE runs PRODES and DETER, which track clearance by satellite in near real time, and the PPCDAm action plan, from 2004, combined that monitoring with inspections, fines, embargoes on illegally cleared land and restrictions on credit. Market agreements: the soy moratorium, from 2006, under which traders agreed not to buy soy grown on recently cleared land, and similar agreements in the beef supply chain. Finance and restoration: the ARPA programme funding protected areas, REDD+ and carbon-credit projects, payments for ecosystem services, agroforestry and restoration, and, from consumer countries, rules such as the EU deforestation regulation requiring proof that imports are not linked to deforestation.

Evaluating this honestly means separating what worked from what did not. Satellite monitoring combined with enforcement did work: deforestation fell steeply between 2004 and 2012, and it fell again in recent years when monitoring and enforcement were restored after a period of weakening. Protected areas, indigenous territories and the soy moratorium have measurable effects and are the strongest parts of the strategy. But enforcement has been underfunded and politically vulnerable, so rates rose when it was relaxed; cattle rather than soy is the largest driver and the cattle agreements have been much weaker; land grabbing and illegal mining continue inside indigenous territories; fire is now a bigger source of loss than clearance in some years; and there is leakage, because protecting one area pushes pressure into another. Restoration is slow and expensive. The defensible judgement is that the strategy has been partly successful: it has slowed deforestation, and it has neither stopped nor reversed it, because the economic incentive to clear land has never been removed.