Introduction

In the drylands are characterized by rainfall of short duration and high intensity (^{Rango et al., 2006}; ^{Wei et al., 2007}), which produce a strong erosive effect on the fragile soils and high evapotranspiration, and contribute to determine the type and condition and development of agriculture adapted to these constraints (^{Ries and Hirt, 2008}). In these regions, ecosystems are fragile, vegetation is sparse and therefore erosional processes occur quickly after heavy precipitation. Maintenance of native vegetation and any mechanism to reduce runoff, increasing infiltration, allow better conditions for the establishment of agriculture in arid and semiarid areas, which deserve the various alternatives studied and disseminated.

^{Fortanelli et al. (2008)} note that, globally, there are approximately 49 million square kilometers considered arid, semiarid and sub-humid. These comprise a complex group of natural regions located in various areas of the planet, interconnected with other ecosystems, because of its natural and cultural potential for sustainable development (^{FAO, 2010}). In tropical America, about 15 million km² are affected by dry weather that shape the peri-caribbean arid belt, including northern Venezuela and Colombia, the Caribbean islands, much of Central America, especially Mexico and Guatemala, and northwestern Brazil (^{Diaz, 2001}).

In Venezuela, semiarid ecosystems occupy 41 023 km², which are distributed from the Guajira to the Gulf of Cariaco, in addition to the high plateau of Barquisimeto and arid tables of the Andes. Additionally, it has been established that 11 420 km² of soils have problems of acidity, salinity or sodicity. The highest percentage of these soils is located in the coastal area of Sucre state, as well as in the states of Anzoategui, Falcon, Carabobo and Lara (^{Matteucci and Colma, 1997}). Given this, the development of agro-industrial activities in arid and semiarid ecosystems of Venezuela has been limited because it involves facing a number of adverse factors, such as lack of water, low vegetation cover, low production profile and low content of matter organic, among others (^{Pedroza et al., 2004}).

One of the problems that most often affects the quality of soils in arid areas is erosion, so alternatives must be sought for reduction. Among the proposed alternatives is the use of live barriers, which are dense obstacles to the level of the soil surface formed with plants (grasses, shrubs), in order to modify and reduce the speed and shear runoff in a slope (^{Dabney et al., 1993}; ^{Van Dijk et al., 1996}). Hedgerows have been used successfully with species such as Gliricidia sepium L., which have proven useful in reducing runoff and promote the gradual formation of terraces.

Another practice used for erosion control are torobas, which are anti-erosion devices built with wooden stakes and plant debris placed perpendicular to the direction of the runoff. The aim of the toroba is to reduce the erosive effect of the runoff and contribute to a better distribution of water to building up and infiltrate. There are different types of torobas: one, whose main objective is to divert the main channel for distribution to the plots, is called toroba bypass.

One of the most important aspects arising from the use of torobas long-term training miniterrazas and the anthropic horizon, which will undoubtedly lead to improved soil fertility, given the accumulation of sediments rich in minerals and organic matter provided by plant residues.

The research hypothesis is that after 20 years of implementation of the torobas, an improvement will be seen in conditions, physical and hydrological soil, also considering that torobas promote the formation of mini terraces, this led to the accumulation of matter organic and nutrients in the soil, therefore it is expected that in soils where the toroba device was implemented, the physical and hydrological and fertility conditions are higher than soil under natural conditions, where heavy rainfall and low coverage plant entailed a loss of soil by erosion, therefore the objective of this research was to evaluate the effect of toroba on improving soil quality in terms of fertility, resulting from the accumulation of nutrients from mineral sediments and organic matter.

Materials and methods

Description of the study area

The study was conducted in the towns of Mide municipality Urumaco, Falcon state, between the coordinates 11° 00' and 11° 04' north latitude and between 70° 13' and 70o 17' west longitude and Zamurito municipality Buchivacoa, Falcon state with coordinates 11' 03 north latitude and 70' 33 west longitude, both located in the Venezuelan semiarid area to the northwest of Falcon state (Figure 1).

Figure 1 Location of sampling sites for performance evaluation of torobas in sectors Mide and Zamurito in Falcon, Venezuela.  

The area Mide is typical of tropical dry forest, with an average rainfall rate of 400 mm/year characteristic of falconiano semiarid, distributed in two peaks of precipitation: the first from may to june and the second from September to November (^{COPLANARH, 1975}). The predominant production systems in the region are extensive goat rearing and subsistence farming under dry conditions to produce corn, millet and beans.

Soils Mide sector they belong taxonomically classified as Calciorthents, located from the geomorphological point of view on the alluvial terraces (T) of the Lagarto river, the terrain is flat and convex, have generally slope between 1 and 2.5%, with areas with slopes of 3.5% (^{COPLANARH, 1975}). The soils are constantly renewed by contributions from the Lagarto River, they have a sandy to moderaa permeability loam (Fa) to clay loam (FA) quickly, moderate to rapid infiltration, with a weak structure on the surface and moderate in depth, fertility is moderate to low, they are highly susceptible to erosion, especially in rills and gullies (^{Martínez et al., 1989}).

The Zamurito area is located in the area corresponds to a thorny tropical forest, has a average temperature of 29°C, an evaporation of 1 708 mm and an average annual rainfall of 490 mm. The sector Zamurito soils were taxonomically classified as Haplargids with general slopes between 0.5 and 1.5% to 3% and locally located from the geomorphological point of view within a glacis. What soils are characterized by clay loam to sandy loam with low organic matter content, pH of 7.2, high phosphorus, potassium and calcium means values of 122, 496 and 2 000 mg kg^-1, respectively, as which Mide the main agricultural activity is the goat farm (^{Sanchez et al., 2005}).

Sampling design. The experiment was conducted following a randomized block design; for this, a transect perpendicular to the arrangement of the torobas was established and six points were taken in the direction of the transect (Figure 2). Each site was evaluated separately and within them, the treatments were arranged in factorial form, where the use of antierosivo device (toroba and natural forest) and the position toroba (high, medium and low) was evaluated. Each position was considered in the two transects and two conditions were evaluated sites (toroba and forest), for six treatments for each sampling site (Mide and Zamurito) for a total of 12 experimental units.

Figure 2 Sample design for performance evaluation torobas in Mide and Zamurito sectors in Falcon, Venezuela.  

Variables evaluated.Asample of dried and sieved soil (2 mm) were determined distribution of particle size (Bouyucos, 1936), organic carbon (^{Walkley and Black, 1934)}, pH (^{Miller and Kisell, 2010}), electrical conductivity (^{Gavlak et al., 2003}), nitrogen (^{Kjeldahl, 1883}); phosphorus content (^{Olsen et al., 1954}), potassium (^{Pratt, 1951)}, calcium and magnesium (^{Doll and Lucas, 1973}).

To determine the saturated physical attributes distribution of particle size, bulk density (Da), porosity (EPT), macroporosity (fa), microporosity (fw), and hydraulic conductivity (Ks) soil soil samples were used not disturbed, obtained from a sampler type Uhland mark Humboldt model H4203.3, following the methodology described by ^{Pla (1983)}, the infiltration rate was determined by the method proposed by ^{Doran (2000)} method.

Statistic analysis. Statistical analysis consisted of an analysis of variance according to the sampling performed by comparing the positions within the toroba and in each location separately. The applied statistical model corresponds to a factorial experiment with two factors: position within the toroba (high, medium, low) and device (toroba and natural forest). For each combination of device 6 position and sampled. The statistical model was applied:

Where: F_i= the effect due to the position within the toroba, P: the effect due to the use of device, (FP)_ij= the effect of interaction between the position and device presence, and ε_ijk= the error. Where it was necessary comparisons Tukey (p< 0.05) they were performed. Statistical analysis was performed using the statistical package ^{Infostat computerized version 1.1 (2008)}.

Results and discussion

Soils Zamurito industry had a higher content of very fine particles (Table 1), thus a progressive deterioration of his physical condition was observed, which led to problems of compacting and worse hydrological conditions, which enhance the risk of water erosion . In the Mide in the field of physical deterioration problems were minor, given a lesser proportion of very fine particles, besides the proper functioning of the toroba, which reduced the potential for erosion risk.

Table 1 Distribution of particle size in sectors Mide and Zamurito. 

Arena muy gruesa de 2-2 mm; gruesa 1-0.5 mm, media 0.25-1 mm; fina 0.25-0.1 mm y muy fina 0.1- 0.05 mm.

Apparent density. In the Mide sector no significant differences (p< 0.05) were observed when comparing the toroba with the natural forest in any of the positions evaluated along the transect, the same behavior was observed in the Zamurito (Table 2) sector. The bulk density tends to be higher in the Zamurito sector compared to the Mide sector with a value of bulk density 1.54 Mg m³, compared to 1. 51 Mg m³, reported in Zamurito. This indicates a greater tendency to compaction Zamurito soils, due to the predominance of finer particles which promote compaction processes and restrict entry and flow of water in the soil.

Table 2 Apparent density in sectors Mide and Zamurito the use of soils under toroba device. 

According to the results the apparent bit density between sites and no changes between the system under toroba was observed compared to the system under forest, these results suggest that the system toroba possible to maintain the bulk density values similar to what the forest, however the values are above the critical values established by ^{Florentino (1998)} for a similar soil textural class, so a compacting process is evident in both conditions. The results are similar to those reported by ^{González-Pereza y Dezzeo (2011)}, who found no differences when comparing the density of a Venezuelan tropical dry forest with young and old pastures, these authors expected high values of apparent density high grasslands overgrazing product, which is one of the main causes of deterioration of physical properties in dry regions, especially by grazing goats (^{Sharrow et al., 2007} y ^{Geissen et al., 2009}).

Porosity. In Mide, the values of pore space and micropores tend to be higher under natural forest compared to toroba, even in the off position the pore space was significantly higher in natural forest compared to toroba (p< 0.05) while in the middle position micro porosity values it was higher in natural forest (Table 3).

Table 3 Porosity Mide sector in the soils under toroba use device. 

In Zamurito no changes (p< 0.05) for the variable porosity, macroporosity and microporosity when comparing the toroba with the natural forest (Table 4) were observed, no changes were observed along transect. However the trend shown is that the toroba device tends to improve soil physical conditions with respect to initial conditions, however implementation time is not enough for these changes are significant.

Table 4 Porosity in the industry Zamurito of toroba device. 

The highest values of total pore space corresponded to Mide sector with 45.19%, which was higher than in the Zamurito found sector, which presented values of 40.19%. In Zamurito higher bulk density values were reflected in a reduction of porosity, which indicates that this area compaction processes were more severe. When the distribution of macro and micropores was analyzed, it was found that, in effect, the best conditions for infiltration, aeration and water flow are presented in Mide; where there was a greater predominance of macro pores compared to Zamurito.

However, Mide higher micropore values were observed, which can cause problems of water movement, which promotes runoff and thus erosion (^{Ferreras et al., 2000}; ^{Jalota et al., 2001}; ^{Bravo et al., 2004}), these processes may be enhanced in the area due to low infiltration rate and the occurrence of high intensity rainfall.

Hydrological variables evaluated. In Mide were observed in the high position the infiltration rate was significantly higher (p< 0.05) in the toroba compared to natural forest (Table 5) for the lower and middle positions no significant differences were observed. With respect to the hydraulic conductivity this was higher in natural forest in relation to toroba in the upper and middle positions to the lowest position, no significant differences were observed, although it was noticed a dramatic decrease in hydraulic conductivity values.

Table 5 Hydrological variables evaluated in the soils in Mide sector under use toroba device. 

In Zamurito the infiltration rate was higher in the toroba compared to natural forest in the middle position, while in the highest position the highest values corresponded to the natural forest (Table 6) to the down position no significant differences (p< 0.05).

Table 6 Hydrological variables evaluated in the soils in Zamurito sector under use toroba device.  

Generally speaking, the higher speed of infiltration was reported on the floor of Mide, which has better structural conditions, while Zamurito, deterioration of physical conditions favored processes soil sealing, drastically reducing the infiltration rate. ^{Prieto et al. (2009)} corroborate these findings and point out that the increase in infiltration rate is counteracted by a major alteration of soil structure by increasing the intensity of the rain. The fine particles separated from soil aggregates increased as the kinetic energy of the rain was increased (proportional to the intensity), so that the pores are sealed and decreases the rate of infiltration.

Lower values of hydraulic conductivity were reported in soils Zamurito sector where the use of toroba, was not effective in controlling erosion, while the highest values were observed in soils of Mide, both soil Ks values were higher in the top point of transect, while lower values correspond to the lowest point of the transect, because in this the finer sediments accumulate. These marked differences but reflect changes in soil quality product management system, also they were influenced by the high variability that is variable in its determination (^{Johnston et al., 2009}).

Aggregate stability. The content of aggregates with smaller diameter of 0.25 mm was found in the Zamurito sector (Table 7) due to the disintegration of larger aggregates, which could represent a potential increased risk of erosion since, according to ^{Ramirez et al. (2006)}, the prevalence of aggregates with sizes of 0.05 to 0.25 mm reflects that soils are most susceptible to water erosion. An important aspect to is that in both cases the device toroba improved soil aggregation, and therefore reduced the risk of erosion, when an increase of aggregates larger than 2 mm.

Table 7 Size distribution of water stable aggregates and percent diameter in the Mide and Zamurito sectors. 

This behavior may be due among other reasons to the predominance of fine particles in Zamurito, especially silt and sand very thin, which favor the structural deterioration of the soil. ^{Pla et al. (1982)}; ^{Bravo (1999)} and ^{Pulido et al. (2009)} agree that the soils with predominance of very fine, fine silt and sand have low structural stability and susceptibility to separation when hit by raindrop, which is reflected in the rapid formation of the seal surface, accompanied by a sharp decrease in saturated hydraulic conductivity authors also report that high erosion soils have the highest percentage of aggregates in the smaller diameters.

Conclusions

The soils located on Mide presented better soil physical conditions compared to Zamurito sector, which resulted in higher values of infiltration, hydraulic conductivity and lower density. The recovery of soil physical properties was associated with the efficient use of torobas, while in Zamurito adverse soil and climatic conditions have not allowed the improvement of soil physical quality.

Torobas implementation of improved soil aggregation; in Mide, predominating aggregates larger than 2 mm, whereas in Zamurito predominated minor additions to 0.25 mm reflecting a susceptibility more be eroded soil.

Improving soil fertility was observed in Zamurito to observase values of organic matter, nitrogen, phosphorus and potassium from the accumulation of nutrients in the lower positions.