15.2.1.1 General characteristic of the soils of the Orinoco basin

Neotropical savannas are usually related to acid soils with high P-fixation capacity and consequently low P availability, low organic matter, nitrogen and exchangeable base contents (exchangeable bases <5 cmol kg⁻¹ soil), but high Al saturation (López-Hernández and Hernández-Valencia, 2008). An analysis of the main chemical and physical characteristics of different savanna soils in Venezuela confirms this assumption (Table 15.1). Soil pH ranged between 4.8 and 6.3 units, whereas the highest values recorded for nutrients were 5.3 mg P kg⁻¹ for available P, 1800 mg N kg⁻¹ for total N, 1.7% for organic C, and 1.2 cmol exchangeable bases kg⁻¹. The dystrophic nature of soils is a consequence of parent materials with low nutrient contents and/or strong weathering, which results in accumulations of resistant primary minerals (e.g., quartz), clays with low exchange activities, and the presence of amorphous and crystalline Fe and Al sesquioxides with strong P-fixation capacities.

Soil texture in Trachypogon savanna is usually dominated by sands (43–96%) (Table 15.1) and coarse gravels but an increase in clay content could be found in the subsurface due to the accumulation of clays and the formation of argilic, kandic, oxic, or cambic horizons. On the other hand, a hardpan of laterite or plinthite could be present in soils with alternating periods of flooding and drought. Precipitation is seasonal and has an influence on soil water content with periods of excessive wet and eventually anaerobic conditions, alternated with periods where the soil becomes dry and below the permanent wilting point. Anaerobic conditions are usually found in subsuperficial layers where water drainage is poor due to high clay and silt content, or the presence of a lateritic layer or concretions (López-Hernández and Hernández-Valencia, 2011).

According to the Soil Taxonomy classification system (USDA, 2010), the soils of Trachypogon savannas are mainly identified as oxisols, ultisols, in those with argilic, oxic, kandic, and cambic subsurface horizons, with or without a lateritic hardpan, while entisols are related to sandy soils from alluvial sediments or eolian deposits (e.g., dunes). For the World Reference Base for Soil Resource (IUSS, 2007), soil types are mainly identified as ferralsols, acrisols, and lixisols.

Physical and chemical features of savanna soils have an influence on plant physiognomy and, consequently, on the use of water, nutrients, and, finally, on primary production. Savannas are dominated by grasses. Most herbaceous species have an intensive superficial root system, whereas trees have extensive roots able to exploit water and nutrients from deeper soil layers (Medina, 1993). February et al. (2013) found for a broadleaved woody savanna in Australia that regardless of the type of grass, both tree and grass roots are concentrated in the top 20 cm of the soil. While trees have greater root production and contribute with a finer root biomass, grass roots contribute with a disproportional amount of N and C to the soil relative to total root biomass. They postulate that grasses maintain soil nutrient pools and provide biomass for regular fires that prevent forest trees from establishing, while savanna trees are important for increasing soil N content, cycling, and mineralization rates.

Open savannas are related to superficial hardpans, which interfere with root tree development, or with flooded soils. On the other hand, tree density increases when the hardpan is absent or deeper and/or water availability increases in the tree root zone, but flooding is not present (Medina, 1993). A similar relationship between physiognomy and soil fertility has also been observed in Brazilian cerrado where open savannas are associated with less fertile soils, especially those with low P and calcium (Ca) availabilities and high Al percent saturation (Lopes and Cox, 1977). Both physical (fast drainage, presence of hardpan) and chemical (low nutrient availability and acidity) features are important constraints for agricultural production, and must be overcome by using machinery and great amounts of fertilizers and limestone (Lozano et al., 2012).

15.2.1.2 Vegetation of savannas in the Orinoco basin

The species of plants of the savannas, whether herbaceous or tree components, are adapted to oligotrophic conditions. The floristic variation in the ecosystem depends in good part on the higher or lower contents of clay, which can affect the water conditions of the soils and the biological transformation of organic matter. Productivity may vary depending on small changes in the physical, chemical, and biological characteristics affecting soil fertility. The general low productivity in the Orinoco basin, at the same time, contributes to the low levels of organic matter found in its soils, with values around 1 to 2% or less (Hernández-Hernández and López-Hernández, 2002).

Grass species are the most important floristic component of tropical well-drained savannas of the Orinoco basin. Legumes and sedges also coexist in the herbaceous layer; however, their coverage is lower than 10% in comparison with grasses (Velasquez, 1965; Susach, 1984). San José et al. (1985) studied seven sites of Trachypogon savannas from low to moderate fertility and found that the above-ground biomass of grasses ranged from 80 to 100%, while legumes reached a maximum of 3.8% in the location with the highest pH (5.5) and base saturation (81%). The lower biomass of legumes in neotropical savannas has been related to the high content of Al and, in turn, low P availability which hinders the formation of effective N-fixing symbiosis (Medina, 1987).

In Trachypogon savannas, species richness is about 285 species of angiosperms belonging to 55 families, but dominance is restricted to only a few like Poaceae, Papilionaceae, Cyperaceae, and Asteraceae, with Poaceae being represented by Trachypogon sp., the genus with the higher coverage (Riina et al., 2007). Other species of grasses are Axonopus canescens, A. anceps, Andropogon selloanus, Leptocoryphium lanatum, Paspalum carinatum, Sporobolus indicus, S. cubensis, and several species of the genus Aristida; legumes of the genera Mimosa, Cassia, Desmodium, Eriosema, Galactia, Indigofera, Phaseolus, Stylosanthes, Tephrosia, and Zornia; and sedges of the genera Rhynchospora and Bulbostylis. As previously described, tree presence is variable from absent in open savannas to dense in savanna parkland where crowns are closer, but with minimal overlapping. Trees can also be found scattered or in groups, and most of them are pyrophyte species like Byrsonima crassifolia, Curatella americana, and Bodwichia virgilioides (López-Hernández and Hernández-Valencia, 2008).

15.2.1.3 Effects of fire on nutrient cycling in savannas

Fire is considered one of the most important factors when explaining the origin and maintenance of savannas, and acts as a selective force which favors the dominance of fire-tolerant species (Sarmiento, 1984). Vegetation is burned during the dry season in order to eliminate senescent leaves, as a plague control, and to stimulate the production of new and more palatable pastures for stock breeding. Loss of nutrients by fires is related to burning efficiency, which in turn depends on the amount and the heterogeneity of the accumulated above-ground biomass, its moisture content, the wind speed, and the oxygen diffusion to the combustion zone (Bilbao and Medina, 1996). Beside burning efficiency, the long-term loss of nutrients by burning depends on fire frequency, which is usually annual, biennial, or triennial (Coutinho, 1988). The accumulation of high amounts of fuel load increases burning efficiency, but when fires have been burning in the same year, recent burn scars are very effective at stopping fires (Bilbao et al., 2010).

Fires produce strong mineralization and loss of nutrients. In the case of gaseous elements, like N and S, losses occur by volatilization and the spread of ashes, whereas in the case of sedimentary elements, like P and bases (potassium (K), sodium (Na), calcium (Ca), and magnesium (Mg)), losses by fire are due just to the spread of ashes. Hernández-Valencia and López-Hernández (2002) assessed the nutrient losses in the herbaceous layer of Trachypogon savanna, comparing the amount of N, P, Ca, Mg, and K, before the burning of the above-ground vegetation and after the combustion in the remnant tissues and ashes. Results indicated that fires expelled to the atmosphere as gases and ashes as follows: 1.7 kg P ha⁻¹, 5.9 kg N ha⁻¹, 3.7 kg Ca ha⁻¹, 2.8 kg Mg ha⁻¹ and 5.0 kg K ha⁻¹, which represent more of 90% of the nutrient contents in the above-ground biomass, including litter. The high burning efficiency was related to the accumulation of dry foliages, its homogeneity, and the rapid fire spread enhanced by winds. In relation to their biogeochemical cycles, sedimentary nutrients like P, Ca, Mg, and K showed higher concentration in the ashes compared to the above-ground biomass; however, N was more diluted in the ashes due to its volatilization at high temperatures. Ash depositions returned to the soil between 21 and 34% of the sedimentary nutrients (P, Ca, Mg, and K), but only 1% N, and lowered the net losses to 1.1 kg P ha⁻¹, 5.8 kg N ha⁻¹, 2.4 kg Ca ha⁻¹, 1.9 kg Mg ha⁻¹, and 3.9 kg K ha⁻¹; however, those measures have uncertainties because of the suspension of ashes by winds and the high variability in ash depositions.

15.3.1.1 Atmospheric depositions (wet and dry deposition)

In savannas, ash produced by fires in situ or in nearby areas can be deposited as dry deposition or washed out from the atmosphere and dissolved in the rain. In this way, precipitation can return nutrients lost by fires and other volatilization processes to the soil. Inorganic inputs of N from bulk precipitation are relatively low in savannas, unless the savanna is located near an industrial or polluted area (López-Hernández et al., 2013). Mineral N inputs (NO₃+NH₄) by precipitation in Venezuelan savannas ranged from 2.2 kg N ha⁻¹ yr⁻¹ in savannas of Calabozo (Montes and San José, 1989) to 6.2 kg N ha⁻¹ yr⁻¹ in a place more distant from urban activities at Estación Experimental, La Iguana (Table 15.2) according to Chacón (1988). Sanhueza and Crutzen (1998) also reported wet deposition of mineral N from 1.12 to 4.6 ha⁻¹ yr⁻¹ for the Orinoco region. Those precipitation inputs in inorganic N are very similar to the value presented for savannas of Loudetia located at comparable geographical latitude at Lamto, Ivory Coast, Africa (1.3–2.3 kg NO₃ ha⁻¹ yr⁻¹ and 3.0 kg NH₄ ha⁻¹ yr⁻¹) according to Villecourt and Roose (1978), and to the value presented by Bustamante et al. (2006) in a cerrado area protected from fire in Central Brazil. Detailed information on organic N sources in Trachypogon savannas located in Central Venezuela is almost non-existent, although Pacheco et al. (2004) reported that soluble organic N inputs may be considered an important contribution, making up 75% of total N in rainfall; however, more work needs to be done to confirm this information. Nonetheless, in Lamto savannas (Ivory Coast), soluble organic N inputs have been reported to be as much as 14.5 kg ha⁻¹ yr⁻¹ (Villecourt and Roose, 1978; Abbadie, 2006). In general, typical mineral N inputs in Africa are reported to be very low (Giller et al., 1997).

As to dry deposition, Sanhueza and Crutzen (1998) presented information concerning the Venezuelan savannas during the wet and dry seasons with values ranging from 4.5 to 10.5 kg N ha⁻¹ yr⁻¹) (Table 15.2), most of them NH₄⁺+NH₃ forms. In the savannas of Lamto, Abbadie (2006) reported a dry deposition of 3.5 and 2.1 kg N ha⁻¹ yr⁻¹ for the ammonium and nitrate form, respectively.

15.3.1.2 Biological nitrogen fixation

In the savannas of the Orinoco basin, fires volatize every year up to 19–30% of the N required for the net primary production of savanna herbaceous vegetation (Chacón et al., 1992; López-Hernández et al., 2006). N volatilized by fires and N losses due to leaching and erosion are not compensated for by the scarce inorganic N inputs from the precipitation mentioned above. This situation could cause a progressive reduction in the potential productive capacity of savannas in the absence of additional mechanisms of N input, apart from precipitation (Abbadie, 1983; Chacón, 1988; Chacón et al., 1991; Cook, 1994). Consequently, N budgets in dystrophic savannas must be balanced by biological N fixation.

In general, N fixation can occur in savannas from three different sources: (1) rhizobium-legume symbiosis; (2) organisms located in the rhizosphere of grasses; and (3) microbial crusts on the soil surface formed by cyanobacteria.

15.3.2.1 Losses by fires

Burning of vegetation during the dry season is the rule in savannas (Vareschi, 1962; Vuattoux, 1976). This factor is supposed to be pivotal for the origin and maintenance of savannas (Beerling and Osborne, 2006) and behaves like a selective force which maintains the dominance of fire-tolerant species.

After the fire’s passage, more than 90–95% of the herbaceous aerial cover is lost, and most of the organic elements are transformed into CO, CO₂, NO, and NO₂ (Sanhueza and Crutzen, 1998). In unfertile neotropical Trachypogon savannas located at La Iguana Experimental Station, Venezuela, losses of 8.2 kg N ha⁻¹ yr⁻¹ have been measured (Table 15.2), whereas in the more fertile soils of Lamto Experimental Station in Ivory Coast, Africa, losses of 17 and 24 kg N ha⁻¹ yr⁻¹ have been reported for Loudetia and Andropogon savannas, respectively (Abbadie, 1983, 2006). The higher N losses from Lamto savannas are in good agreement with their higher above-ground primary production. Losses of N by biomass burning at La Iguana savannas are in accordance with more detailed information of gas emissions (NH₃, N₂, NOx, RCN) in other savannas of the Orinoco basin presented by Sanhueza and Crutzen (1998) and Bustamante et al. (2006), ranging from 8.8 to 26.5 kg N ha⁻¹ yr⁻¹.

15.3.2.2 Nitrogen losses by leaching

The loss of N by internal drainage is relatively small in well-drained savannas, which is understandable since the water-soluble levels of N forms are always scarce, except perhaps at the beginning of the rainy season when substantial organic matter decomposition occurs. The losses of N in the savannas of Estación Experimental La Iguana in Guarico Central were estimated at 2.1 kg N ha⁻¹ yr⁻¹ (Chacón et al., 1991; Table 15.2), whereas for a Loudetia–Andropogon savanna at Lamto, Ivory Coast, 5.6 kg N ha⁻¹ yr⁻¹ was reported to leave the system through internal drainage, mostly in organic form (Villecourt and Roose, 1978; Abbadie, 2006) coming from microorganisms or recent dead plant material. In savanna ecosystems, information on nutrient losses by leaching is scarce and perhaps related to methodological inconveniences for designing and installing an effective system to catch leachates (López-Hernández and Hernández-Valencia, 2011). As a consequence, more accurate measures are still needed.

15.3.2.3 NH₃ losses

The mineralization of the soil organic matter in native soils might produce ammonium and at high pH some of the NH₄⁺ exists as NH₃. However, in acidic, well-weathered savanna ecosystems, soil emissions of NH₃ are considered negligible (Abbadie, 1983; Sanhueza and Crutzen, 1998; López-Hernández, 2013).

NH₃ emissions in the budget presented in Table 15.2 are not included since animal waste and fertilizer applications are negligible, particularly in the case of the experimental site located at La Iguana Experimental Station.

15.3.2.4 Denitrification and emissions of NO and N₂O

Losses by denitrification in savannas of the Orinoco basin should be insignificant due to the low levels of nitrate and organic matter and the acidic pH of the soils. Similar information was presented by Abbadie (2006) in Lamto savannas, who reported that denitrification is highly variable in the space between soils. NO₃ pools in the soils of the tropical savannas of South America are low, which means that the rates of nitrification are also negligible; nonetheless, the levels of the intermediate NH₄ are, in general, significant.

In the savannas of the Orinoco region, Sanhueza and Crutzen (1998) have estimated an annual production of 0.18–0.63 kg ha⁻¹ and 0.36–6.3 kg ha⁻¹ for N₂O and N₂, respectively. However, in Orinoco savannas, soil emissions of NO are largely uncertain, ranging from 0.3 to 3.0 kg ha⁻¹ yr⁻¹ depending on the amount and distribution of rainfall (Cárdenas et al., 1993); all together N₂O, N₂, and NO losses in Trachypogon savannas account for about 7.5 kg ha⁻¹ yr⁻¹ (Table 15.2).

15.4.2.1 Losses by fires

In tropical savannas, burning of vegetation produces a strong mineralization pulse and loss of nutrients once ashes are spread outside the ecosystem. Studies in Calabozo savannas in Estación Biológica de los Llanos, comparing the P content in the above-ground biomass before and after the burning of the vegetation, indicated that about 1.7 kg P ha⁻¹ was expelled to the atmosphere (Table 15.3). This amount represents an important part of the above-ground biomass, including litter (Hernández-Valencia and López-Hernández, 2002). Although temperature was not recorded during the experimental period, the high burning efficiency (more than 90%) seems to be related to the accumulation of dry foliage and the rapid fire spread that was enhanced by winds. Ash deposition returned about 21% of the P transferred to the atmosphere by fires to the soil and lowered the net losses to 1.1 kg P ha⁻¹. However, those measurements could have significant uncertainties, because of the suspension of ash by winds and the high variability in the measured ash depositions. These results demonstrate the role of fire as net exporter of P and its contribution to the global balance of other elements in the atmosphere (i.e., C, S, and N). Also the need to manage fire to reduce carbon and other nutrient emissions from savanna ecosystems is emphasized. Reducing fire frequency could be an easy way to address this situation; if fires were produced on a biennial or triennial basis, the associated N and P losses could be compensated by inputs through precipitation, as was shown for the Brazilian Cerrado (Coutinho, 1988).

15.4.2.2 Losses by leaching

Nutrient losses from soil by leaching are likely in well-drained savannas due to high permeability of the soil. Leaching is usually restricted to the wet season and when an excess of water drains out of the rooting zone (Hernández-Valencia, 1996). For well-drained savannas the information on P losses by leaching is scant. At Estación Biológica de los Llanos, P concentrations were below 0.1 μg L⁻¹ in leachates sampled at 1 m depth; these values are typical for tropical weathered soils with concomitant high P-sorbing capacities (López-Hernández et al., 1981). Drainage out of the rooting zone (>1m) was recorded in three of the six rainy-season months and P concentration did not present any seasonal pattern. The amount of P that is lost by leaching ranged between 0.23 and 0.29 kg P ha⁻¹ yr⁻¹ (Table 15.3).

15.5.1.1 Nutrient uptake, translocation, and biomass production

The mechanisms involved in the adaptation of savanna plants to soils of low nutrient availability have scientific and practical importance (Medina and Bilbao, 1991). Natural selection favors species with competitive ability to overcome oligotrophic conditions. Most of them usually have low nutrient demands, low nutrient contents in their tissues, and, in turn, a low organic matter production. Since savannas are areas of stock breeding, intensive animal production is restricted by the poor production and palatability of forages (López-Hernández and Hernández-Valencia, 2008).

A review of 15 studied sites corroborates the mean low productivity values for Trachypogon savannas (López-Hernández and Hernández-Valencia, 2008), where above-ground production (3.78 mg ha⁻¹ yr⁻¹) surpassed mean values for below-ground production (2.64 mg ha⁻¹ yr⁻¹) and finally the overall primary production reached 6.74 mg ha⁻¹ yr⁻¹, values that are within the lowest in the world production of savanna ecosystems (Sarmiento 1984; López-Hernández et al., 2012).

Organic matter production by the herbaceous layer shows a seasonal trend correlated with precipitation and soil water availability (Hernández-Valencia and López-Hernández, 1997). The burning of the vegetation during the dry season stimulates plant growth and the demand for nutrients increases. As a consequence, due to their higher metabolic activity the youngest tissues reach the highest nutrient concentrations at that growth period. As long as the rainy season proceeds, biomass and nutrient storage increases and nutrient concentration decreases due to dilution or translocations from old to young leaves. Thus, the translocation to young leaves and roots can operate as efficient mechanisms for nutrient conservation. In the dry season, organic matter productivity and nutrient uptake decrease, therefore the nutrient contents in foliage reach their lowest concentrations. This behavior has been found for N, P, and S, while Ca and Mg show the opposite behavior (Medina et al., 1978; Hernández-Valencia and López-Hernández, 2000). In general, low levels of N, P, Ca, and S are commonly found in the foliage of grasses in well-drained savannas in South America. The maximum amounts recorded after fires (dry season) ranged between 6.2–8.9 mg N g⁻¹ and 1.8–3.1 mg P g⁻¹ while the maximum at the standing crop (wet season) ranged between 0.9–8.1 mg N g⁻¹ and 0.9–2.0 mg P g⁻¹ (Medina et al., 1978; Sánchez et al., 1985; Hernández-Valencia and López-Hernández, 2000).

Information about N and P uptake for primary production is scarce when compared to productivity data. Comprehensive studies on N cycles estimated values from 27.0 to 48.0 kg N ha⁻¹ yr⁻¹ and from 29.6 to 30.3 kg N ha⁻¹ yr⁻¹ for the unburned and burned savannas, respectively (López-Hernández and Hernández-Valencia, 2008), while in the case of P, reported phosphorus uptake was about 5.9 to 8.8 kg P ha⁻¹ yr⁻¹ in burned savanna (Hernández-Valencia and López-Hernández, 2000). According to Medina (1987), the ability of savanna grasses to extract nutrients from poor soils is probably more related to a dense root system for exploring the soil volume and, therefore, a higher root/shoot ratio than more fertile soils. Data from seasonal savannas of Venezuela indicate that the root/shoot at maximum standing crop ranged from 0.22 to 1.75 (Table 15.4). Unburned savannas showed lower root/shoot ratios due to the accumulation of above-ground biomass.

There is no information about nutrient uptake of savannas trees in the Orinoco basin; however, several patterns have been distinguished among grasses and trees (Medina, 1993). Grasses and woody plants from semi-arid savannas in general have higher nutrient contents than those from dystrophic humid tropical savannas; however, trees demand more nutrients than xeromorphic C₄ species, and thereby are associated with soils with better nutritional status. Evergreen trees are more frequent in dystrophic savannas, especially if fire is frequent, and produce leaves with lower nutrient contents than deciduous trees. Leguminous trees have usually higher N and P content and consequently higher N and P demands than non-legumes, which explains the low coverage of legumes observed in most P-depleted savannas. It has also been demonstrated that trees improve nutrient contents in the soil beneath them in comparison with the soil associated with the herbaceous layer of a treeless savanna (Table 15.5; Hernández-Hernández, 2008). The input of litter with a richer nutrient content has been argued as a mechanism to explain this finding (Kellman, 1979), but also it must be considered that trees can concentrate nutrients beneath the soil by pumping them from the subsurface to the topsoil, by acting as traps of dust and ashes suspended in the air and by using perching bird droppings.

15.5.1.2 Role of microbial biomass

The microbial biomass (MB) mediates the transformation (mineralization–immobilization) of biogenic nutrients (C, N, P, and S) between the inorganic and organic forms; therefore, MB plays an important role in nutrient cycling. It also acts as an important sink for plant nutrient conservation and utilization in tropical savannas. In the soils of the Orinoco savannas, in general, the contents of soil microbial biomass carbon (C-MB) are low (100–200 mg kg⁻¹), which is in agreement with the low soil organic matter content of the soils (López-Hernández and Hernández-Valencia, 2008).

15.5.1.3 Role of pedo-fauna in the generation of biogenic “hotspots” in Trachypogon savannas

In savanna ecosystems, termites and earthworms are among the most conspicuous components of soil macro-fauna and play an important role in the decomposition processes of organic matter and nutrient cycling (Lavelle et al., 1994; López-Hernández, 2001; Chapuis-Lardy et al., 2011).

Invertebrates belonging to macro-fauna are key species of soil functioning. They participate in litter decomposition, mix organic and mineral matter, create and maintain soil structure by digging burrows and modifying aggregation, and regulate microbial diversity and activity (Lavelle et al., 1994; Renard et al., 2013). The gut and biogenic structures (burrows, casts, nests) of earthworms and termites are specific habitats where soil microbial activities are either stimulated or attenuated (Martin, 1991). Earthworms are important actors in the regeneration of compacted soils by reduced tillage systems (Capowiez et al., 2012) and freshly egested earthworm casts, i.e., the by-products of gut passage, are generally characterized by an intense mineralization of organic matter and the release of nutrients for plants (Lavelle et al., 1994; López-Hernández et al., 1993; Ngo et al., 2012). Conversely, the mineralization in old casts is reduced, which allows for carbon and nutrient storage in soil in the long term (Martin, 1991; Lavelle et al., 1994). Thus, biogenic structures produced by earthworms (casts), ants, and termites have been considered as “hotspots” for nutrients and places to potentially start a successional development from savannas to a “micro-forest” (López-Hernández et al., 2012; Erpenbach et al., 2013).

15.5.1.3.2 Phosphorus contents in biogenic structures of earthworms (casts) and termite mounds and adjacent soils

Earthworm activity creates a series of geochemical and biological effects that are especially important for the cycling of P in soils (Chapuis-Lardy et al., 2011). The availability of P is strongly influenced by the physical adsorption or fixation of P in soil, which is much related to soil type (Figure 15.3). López-Hernández et al. (1993), using isotopic techniques, reported higher P availability (P in solution) and E₁ (P isotopically exchanged in 1 min) in fresh casts of P. corethrurus for two soils, one from the savanna of Lamto, Ivory Coast, with a moderate to low P-sorption capacity, and the other from Laguna Verde, Veracruz, Mexico, with a high P-sorption capacity. These changes in P availability were ascribed to (1) the greater pH of the gut content, (2) changes in sorption complexes induced by competition for sorbing sites between orthophosphates and carboxyl groups of a mucus glycoprotein produced by the earthworm in its gut, and (3) an increase in microbial activity during digestion (López-Hernández et al., 2012).

Earthworm casts usually contain larger amounts of organic P (e.g., for P. corethrurus; Chapuis-Lardy et al., 1998), probably derived from a selective ingestion of soil particles (López-Hernández et al., 1993; Chapuis-Lardy et al., 1998). Several studies have reported increased enzymatic activity and the stimulation of microbial activity in earthworm casts (Sharpley and Syers, 1976; López-Hernández et al., 1993; Brossard et al., 1996; Le Bayon and Binet, 2006). Devliegher and Verstraete (1996) showed that the increase of inorganic P in soils in the presence of earthworms was due to the production of alkaline phosphatases by earthworms (Satchell and Martin, 1984; Park et al., 1990) and a stimulation of microbial acid phosphatase production by earthworm activity (Satchell and Martin, 1984).

A similar role in the decomposition processes of organic matter for tropical and subtropical environments has been ascribed to termites in Africa and Australia; however, in neotropical savannas the importance of termites is considered secondary (Chapuis-Lardy et al., 2011). In a comprehensive P fractionation study done in soil and associated termite mounds of Nasutitermes ephratae, a common plant debris-feeding termite from savannas of the Orinoco Llanos (López-Hernández, 2001), it was found that all P concentrations were significantly higher in the Nasutitermes mounds than in the adjacent soils. Evidence suggests that nutrients accumulate in the mound from the nearby areas; therefore, the abundant termite populations could play an important role in controlling nutrient cycling in tropical savannas, in which available nutrients can often be a limiting factor for plant growth and development.

15.5.2.1 Mycorrhizal association

Mycorrhizal association is a symbiosis established since ancient times between fungi of the phylum Glomeromycota (Schüßler and Walker, 2010) and the roots of plants; 80% of the land plants possess such association (Smith and Read, 2008; Brundrett, 2009). These fungi are symbionts that are forced and require a source of carbon from the host to complete their life cycle; the resulting structure in the root is known as arbuscular mycorrhiza (AM), frequently present in terrestrial ecosystems. Only some families have lost the ability to form this symbiosis (10%), e.g., Chenopodiaceae and Brassiceae, among others (Smith and Read, 2008; Brundrett, 2009), consequently, AM presence is widely extended in nature.

The AM exert a beneficial effect on plant growth and development, because the fungi develop an extensive network of mycelium out of theroots, able to take elements like P, usually depleted in the rhizosphere due to slow diffusion in soil solution and fast root uptake (Smith and Read, 2008; Brundrett, 2009). Thus, a mycorrhized plant has an additional mechanism of nutrient incorporation than non mycorrhized plant. For this reason, some authors consider this symbiotic association as a biofertilizer and recognize its usefullness in farming systems, remediating contaminated soils, among other applications (Barea et al., 2002a, 2013).

Symbiotic AM can contribute greatly to plant P nutrition and may be induced by P deficiency and the presence of organic matter in soil (López-Gutiérrez et al., 2004 a,b). AM contribution may be indirect due to AM hyphae exploring a larger soil volume and inducing physiological changes that favor the establishment of P solubilizing and mineralizing microorganisms in the mycorrhizosphere (Pineda et al., 2013). Alternatively, it may be directed as a consequence of the production of extracellular phosphatases such as acidic phosphatase activity (APA) and the access to distant P sources otherwise not available to the host plant (Li et al., 1991; Tarafdar and Marshner, 1994; Joner and Johansen, 2000; Shen et al., 2011).

It seems possible that some specific mechanisms must be operating in the plant rhizosphere to allow plant growth in savannas. Thus, AM, soil microorganisms, and APA in the rhizosphere are considered as crucial mechanisms for P availability and plant P uptake. López-Gutiérrez et al. (2004a,b) focus on the biochemical processes occurring in the rhizosphere of native plant species (Trachypogon plumosus) of P-deficient savanna soils. Therefore, microbial activity, as measured by viable plate counts and dehydrogenase activity (DHA), mineralization activity (APA), and AM dynamics was studied in the rhizosphere of T. plumosus growing on three seasonal savanna soils differing in P content and soil order (Table 15.6).

15.5.2.1.1 Mycorrhizal status in the studied savanna soils

AM colonization, expressed as a percentage of AM-colonized root length (%CRL) in T. plumosus, was higher during the rainy season in entisol and vertisol but not significantly different among seasons in ultisol (López-Gutiérrez et al., 2004b; Figure 15.4). The higher AM colonization as well as the increase in AM spore number during the rainy season (data not presented) concurs with the higher plant P demand. Nevertheless, seasonality, intrinsic characteristics of the plant–fungus combination and also soil properties can alter these parameters (Smith and Read, 2008). Moreover, AM spore number is relatively high in these soils since 3.6 spores g⁻¹ dry soil is considered to be a high density (López-Gutiérrez et al., 2004b). Therefore, T. plumosus shows a particularly high mycorrhizal status for a grass in terms of AM infective potential spore density and root colonization when compared with reports from savannas of the Venezuelan Amazonia, which include values of 54.7% of CRL for Panicum pilosum (Poaceae) and only 29.8% for Andropogon bicornis (Poaceae) (St. John and Uhl, 1983). It has been reported that grasses with fine roots, abundant root hairs, and rapid growth are less sensitive to AM colonization (St. John, 1980).

The fact that AM colonization is, in general, high and increases during the rainy season (Figure 15.4), when P demand by plants also increases, suggests the importance of AM symbiosis as a mechanism for P uptake of T. plumosus in savanna ecosystems. This is emphasized when considering that one of the main strategies of grass species for P uptake in acid soils is a preferential distribution of resources to the roots and T. plumosus possesses a dense and fibrous radical system that allows an intensive exploration of the soil substrate (Medina and Bilbao, 1991).

Regardless of the presence or not of seasonal changes in the different mycorrhizal parameters studied, the sustained high level of AM colonization throughout the different seasons indicates that T. plumosus requires AM symbiosis to achieve an adequate P nutrition in savanna ecosystems.

15.5.2.1.2 Enzymatic activities in the studied savanna soils

Studies on the enzymatic activities of savanna soils in two contrasting seasons indicated that APA significantly increased during the rainy season in all soil orders examined (López-Gutiérrez et al., 2004a). This increase, however, was proportionally higher for entisol (Figure 15.5). These results suggest that P mineralization, as expressed by APA, increases during the rainy season, when plant nutrient uptake and nutrient leaching are higher.

Dehydrogenases are not expected to be free in soil; therefore, DHA expresses living cell activity and that is why it is an indicator of biological activity (López-Gutiérrez et al., 2004a,b). DHA significantly decreased during the rainy season in all soil orders (Figure 15.6). Even though several soil factors such as soil texture affect DHA, this enzymatic activity can be used to detect soil seasonal changes. Furthermore, DHA can correlate with changes in the microbial biomass due to long-term soil amendment. In Trachypogon savannas, DHA activities (Figure 15.6) appear to be well related to organic matter contents, which, in turn, control biological activities, thus DHA in entisol with poor soil carbon content (0.54%, Table 15.6) values was very low; on the contrary values increased in vertisol and ultisol with a higher soil carbon content (5.1 and 2.36%, respectively). In other studies done in ultisol and oxisol, extremely poor and acid soils of Uverito savannas evidenced an increase of APA, β-glucosidase and protease B-AA activity during the rainy season (Gómez and Paolini, 2011).

Furthermore, microorganisms in the rhizosphere may contribute to P nutrition through the synthesis and release of phosphatases when P is not available (Hayat et al., 2010; Shen et al., 2011). The decrease in DHA and bacterial counts during the rainy season, however, suggests that the increase in APA during that period is not mediated by rhizospheric microorganisms, i.e., soil extracellular phosphatases may have been already present in the soil or have plant root origin (Shen et al., 2011). López-Gutiérrez et al. (2004a) noted that in savanna ecosystems, as well as other tropical ecosystems affected by climatic seasonality, the APA appears to be an important mechanism of adaptation for the acquisition of the P. During the rainy season, the APA increased, but the rhizosphere microorganisms do not appear to be involved in the mineralization activity, but seem to contribute to the cycling of P by the immobilization and release P during the dry and rainy seasons, respectively. Native plants such as T. plumosus were highly colonized by mycorrhiza, suggesting that this may constitute an important adaptive mechanism of native species to soils deficient in P. A summary of the interrelationships among AM colonization, microbial biomass, and mineralization (APA production) in Trachypogon plumosus rhizosphere during wet and dry seasons is presented in Figure 15.7.

15.5.2.2 Releasing of organic acids and competition for phosphorus sorbing places in the rhizosphere

Many soil and rhizosphere microorganisms have the ability to solubilize phosphate (Richardson, 2001; Hayat et al., 2010). As early as the 1950s, Sperber (1958) found that organic acids, such as lactic, glycolic, citric, and oxalic, produced by microbial activities in soils were able to solubilize insoluble sources of phosphates. Moreover, it was noted that there is a relationship between the amount of phosphate solubilized and the soil production of organic acids (citric and oxalic) produced by the fungi Aspergillus niger, A. terreus, and Sclerotium rolfsu. Microorganisms can lead to the dissolution of insoluble phosphates by the production of inorganic or organic acids and/or the pH decrease, thus freeing available phosphate. In addition, these organic acids excreted by the microorganisms can act as chelating agents of metals present in insoluble sources as phosphate rock, also releasing soluble phosphates. The interaction between organic acids and sorbing places in soil materials has been very well studied by using the model of the adsorption isotherms (López-Hernández, 1977). Particular attention has been given to the competition between orthophosphate and organic anions in sesquioxides and clay materials (Nagarajah et al., 1968; Shen et al., 2011) and in soils (López-Hernández et al., 1986).

The rock phosphates are natural phosphate sources that are a more economical option for agriculture use, particularly in countries where the availability of raw material for the production of soluble phosphate is complicated (Toro et al., 1997). Leon et al. (1986) showed an increase in the number of soil microbial solubilizers after fertilizing with rock phosphate. Toro et al. (2008) also obtained a similar effect by applying phosphate rocks in a system managed with conservation techniques in savanna soils. These results suggest that the use of rock phosphate as fertilizer is favored by the proliferation of microbiota with an ability to solubilize P-insoluble sources. The interaction of phosphate microbial solubilizers and the arbuscular mycorrhizal fungi is positive and helps plant nutrition; when the source used is the rock phosphate, both microorganisms act synergistically (Toro et al., 1997, 1998) emphasizing the promising use of these organisms for agricultural management (Barea et al., 2002b; Richardson et al., 2009; Bhattacharyya and Jha, 2012).

In a study to characterize the ability of phosphate solubilizing bacteria to dissolve insoluble phosphate compounds and rock phosphate from native acidic soil of the savannas of Guárico State, Venezuela, it was found that these bacteria showed high ability to solubilize phosphate from Fe and Al compounds, and, to a lesser extent, from phosphate rock (Osorio, 2007). The species of bacteria found in the rhizospheres studied were mainly: Bacillus pumilus, Burkholderia cepacia, Bacillus circulans, and Pseudomonas fluorescens. By using the technique of high resolution liquid chromatography it was noted that the most active bacteria in the solubilization process generated acetic, tartaric, malic, butyric, and oxalic acids.