Vertical distribution, nutrient concentration and seasonal changes of fine root mass in a semi-deciduous tropical dry forest and in two adjacent pastures in the Western Llanos of Venezuela

With the objective to contribute to a better understanding of ecological consequences of deforestation on the below-ground system in the Western Llanos of Venezuela, we evaluated the vertical distribution, nutrient concentration and seasonal changes of total fine root mass (FRM) (<2 mm diameter) in a semi-deciduous tropical dry forest and in 2 adjacent pastures of Cynodon nlemfuensis: a young pasture (YP, 5 years old) and an old pasture (OP, 18 years old) in the Obispo municipality, Barinas State. This evaluation included measurements at the end of the rainy season, during the dry season and during the subsequent early rainy season in 2005/2006. Highest FRM was recorded during the dry season, which probably indicates a plant water-stress response mechanism. The highest proportion (63‒88%) of FRM was concentrated in the 10–20 cm soil layer at all studied sites, probably due to a higher nutrient and moisture content at that depth. Non-significant differences (P>0.05) were observed in the total concentrations of organic carbon, nitrogen, phosphorus, calcium and magnesium in the FRM in soils supporting forest, OP and YP at the evaluated depths. Non-significant changes in the total FRM and nutrient concentrations were observed between the sampling periods and the 3 study sites. YP soils showed a slight increase in FRM that could be associated with the root growth of secondary vegetation, which is considered a weed and is periodically removed. Our results suggest that the land use change from tropical forest to pastures has not significantly affected the mass of fine roots and their carbon and nutrient concentrations. Further studies are needed to determine if these findings apply to other ecosystems.


Introduction
Fine root mass in tropical forests is the most important component of below-ground C dynamics and can contribute significantly to global net primary production (Malhi et al. 2011). In tropical dry forests, fine root production is high, constituting an important source of carbon (C) and nutrients in the soil (Fiala et al. 2017). It has been suggested that the roots in these forests provide more N, P and K than the above-ground biomass (Singh et al. 1989).
Several environmental factors affect production of fine roots in tropical dry forests, such as the marked climatic seasonality (Murphy and Lugo 1986;Singh et al. 1989), soil nutrients (Blair and Perfecto 2001) and forest disturbances and land use history (Castellanos et al. 2001;Jaramillo et al. 2003). The clearing of dry forests for the establishment of pastures can have important effects on the production and distribution of roots within the soil profile and on the contributions of C and nutrients (Jaramillo et al. 2003).
In Venezuela, large areas of tropical dry forests have been cleared and replaced by pastures. It is likely that this conversion has had impacts on the mass and distribution of fine roots within the soil profile. Studies on the replacement of tropical deciduous forests with pastures have shown important effects on C and nutrients in the soil profile (Crespo 2015), and pastures have the potential to store significant amounts of C in soils (Rodríguez et al. 2013;Crespo 2015). The dense and deep root system of grasses contributes to the formation of soil aggregates and provides protection to the soil C, making it least susceptible to oxidation and eventual loss to the atmosphere (Cambardella and Elliot 1993).
Considering that the rate of deforestation of tropical dry forests is increasing rapidly, it is important to improve the understanding of the ecological consequences of forest disturbance on the below-ground system. Detailed information on the changes that have occurred in fine root production once forests have been cleared and converted into pastures can be very useful in predicting the consequences of deforestation and to design effective pasture management strategies. The objective of this study was to evaluate the fine root mass and its nutrient composition in a tropical dry forest and in 2 adjacent pastures (5-and 18-years-old) planted following the logging and burning of forest in the Western Llanos of Venezuela.

Study site
The study area was in the El Mangón farm, located in the Obispo municipality, Hurtado sector of Barinas State in the Western Llanos of Venezuela (40º01'10"-40º59'10" N, 91º57'30"-91º25'18" W; 120 masl) (Figure 1). Average annual rainfall is 1,244 mm, with a rainy season from April to December and a dry season from January to March. Average annual temperature is 26.8 °C, and the relief is flat with a slope between 0 and 2% (Ewel et al. 1976;SIRA-INIA-CENIAP 2010). The landscape corresponds to an alluvial overflow plain, influenced by the fluvial dynamics of an old bed of the Santo Domingo River (Schargel and Strebin 1970). Table 1. Soil characteristics of the study sites (mean ± SD) (Source: González-Pedraza and Dezzeo 2011Dezzeo , 2014aDezzeo , 2014b.

Characteristic
Depth ( The soil parent material is of alluvial origin, a sandy clay loam with kaolinite as dominant clay mineral (Ewel et al. 1976). In general terms, the soils have been described as deep, moderately fertile, with moderate concentrations of organic matter, nitrogen and phosphorus, high cation exchange capacity and base saturation (Table 1), with clay loam and silty clay loam textures (Table 2) (Schargel and Strebin 1970;González-Pedraza and Dezzeo 2011, 2014a, 2014b. The main soil orders found in this area correspond to Inceptisols and Vertisols, among them Tropaquepts, Usterts and Aquerts (Mogollón and Comerma 1994).
According to González-Pedraza and Dezzeo (2011), soils at the study site are fine-textured with a predominance of silt. Forest and OP soils show similar clay and sand contents, while YP soils have significantly higher clay content (up to 18% more than forest and OP soils). In general, the soil properties at the studied sites tended to be similar ( Table 2).
The predominant vegetation in the region is semideciduous, periodically flooded forest (Ewel et al. 1976). These forests consist of 2 strata, the upper one consisting of deciduous and semi-deciduous trees with heights of 15-20 m, and the lower one consisting of deciduous and evergreen trees with heights less than 15 m (Table 3) (Aymard 2015).  In this region in the last 50 years, large areas of natural forest have been converted into pasture using the slashand-burn method. Estrella grass (Cynodon nlemfuensis Vanderyst) is grown for grazing by cattle (Truter et al. 2015).
Field work was carried out in an area of tropical dry forest with dominant deciduous vegetation and in 2 adjacent areas, where the original forest had been cut down manually and burnt, and Estrella grass had been planted (5-and 18-years-old, YP and OP, respectively). The pastures have never been fertilized, but have been mowed annually to control weeds and to promote grass growth. At sampling time, species from the original forest that could not be cut by hand, as well as vegetation of secondary growth, like palms and some species of legumes, were observed in the pasture. Cattle were introduced into YP for grazing during the dry season and early and late in the rainy season. On each occasion the cattle remained in this pasture until they consumed all available grass. Every 1-2 months cattle were introduced to the old pasture (OP) and remained for 3-7 days, while consuming all available grass.

Fine root sampling
At each selected site (forest, YP and OP), soil samples were taken from a 600 m 2 plot (20 × 30 m). The distance between the 3 sites was 1-3 km. At each plot, soil samples for determining the vertical distribution of the fine root mass (<2 mm) and their nutrient concentrations were collected at 12 points arranged along 3 transects with 4 sampling points per transect. The samples were collected with a soil core at 0-5, 5-10, 10-20, 20-30 and 30-40 cm depths at the end of the rainy season 2005 (November). For determining the seasonal changes in mass of fine roots, soil samples at 0-5 and 5-10 cm depths were collected at 12 points in each study plot along 3 transects on 3 occasions: end of rainy season 2005 (November), during the dry season (March 2006) and early rainy season (May 2006) (Figure 2).

Fine root mass
All collected soil samples were dried and passed through a 2 mm soil sieve. The fine roots were extracted from the soil fraction which passed through the 2 mm sieve. From each sieved soil sample, 2 × 200 g subsamples were taken. In these subsamples all fine roots (both living and dead) were manually extracted and weighed. Mean weight of fine root mass from the 2 subsamples was used to calculate the dry matter (DM) of fine root mass from each soil sample. The data were converted to t DM/m 2 considering the area of the core, before being converted to a hectare basis according to the following equation: Fine root mass (t DM/ha) = weight of fine root mass (t DM/m 2 ) × area of one hectare of soil (10,000 m 2 /ha).

Laboratory analyses
Total organic carbon in fine roots (TOCr) was determined using the Loss on Ignition (LOI) method (Schulte and Hopkins 1996). It was assumed that 58% of the organic matter (OM) was organic carbon (Nelson and Sommers 1996) and %TOCr was calculated according to the following equation: %TOCr = [pre-ignition weight (g)post-ignition weight (g)] × 0.58 × 100/pre-ignition weight (g).
Concentration of total organic carbon in the fine roots was multiplied by the respective root mass to give TOCr and the results were expressed in kg/ha based on dry weight and bulk density of the soil.
Total nitrogen (N), phosphorus (P), calcium (Ca) and magnesium (Mg) in the fine root mass was extracted by digestion with concentrated sulfuric acid and oxidation with hydrogen peroxide (Tiessen and Moir 1993). Total nitrogen in the roots (TNr) was determined by the colorimetric method of Keeney and Nelson (1982) on a Technicon Autoanalyzer II. Total phosphorus (TPr) was measured by the molybdenum blue method of Murphy and Riley (1962). Total Ca and total Mg were measured on an atomic absorption spectrophotometer (AA). Total N, P, Ca and Mg concentrations obtained in the fine root mass were multiplied by their respective root masses, and the results were expressed in kg/ha based on dry weight and bulk density of the soil.

Statistical analysis
Statistical analysis of data was carried out using one-way analysis of variance (ANOVA). Means were separated with Tukey's test when statistical differences (P<0.05) were observed. When necessary, the data were transformed to homogenize variances, and when that did not meet this assumption (P>0.05) according to Levenne's test, a non-parametric Mann-Whitney test was also applied. To determine relationships between variables at sites of interest, a simple linear regression analysis was used. All statistics were computed using STATISTICA for Windows 6.0 (Statistica 2001).

Vertical distribution of fine roots
No significant differences (P>0.05) between study sites (forest, YP and OP) were detected in the vertical distribution of the fine root mass. However, total mass of fine roots in the top 40 cm of soil was somewhat higher in YP (5.95 ± 0.57 t DM/ha) than in forest (5.27 ± 0.68 t DM/ha) and OP (4.57 ± 0.54 t DM/ha). Distribution of fine root mass was not uniform down the soil profile as root mass in the 10-20 cm horizon was generally greater than in the other horizons at all study sites (P<0.05; Figure 3).

Seasonal changes in fine root mass
Within each sampling period, no significant differences (P>0.05) between forest, YP and OP were observed in the mass of fine roots in both 0-5 cm and 5-10 cm horizons. Differences between seasons in mass of fine roots occurred in the 0-5 cm horizon, with root mass increasing during the dry season to values exceeding those in the early or late rainy season for all sites (P<0.05; Table 4).

Total carbon and nutrients in the fine root mass
The total concentrations of carbon and nutrients in fine roots in the top 40 cm of soil was determined only in soil samples collected at the end of the rainy season. No significant differences (P>0.05) in total organic carbon (TOCr) and total nitrogen (TNr) in fine roots were found between Forest, YP and OP except for the 30-40 cm soil depth, where YP contained significantly (P<0.05) more TOCr than forest soil (Table 5). TOCr was highest (P<0.05) in the 10-20 cm horizon in all soils. In contrast, TNr tended to decrease steadily with increasing soil depth, with levels in the top 5 cm exceeding those in the 30-40 cm horizon (Table 5).
No significant differences (P>0.05) between forest and pasture soils were found in total phosphorus, calcium and magnesium in the mass of fine roots in the top 40 cm of soil (Table 6). However, there were changes in concentrations down the profile. A significant increase (P<0.05) in total phosphorus in roots in the forest soil in the 10-20 cm horizon was observed. Total calcium and magnesium in fine roots tended to decline with depth but differences were inconsistent and rarely significant.
Significant positive relationships existed between fine root mass and nutrient content in 0-40 cm soil horizons in forest and pasture (Table 7).

Discussion
The total fine root mass in the top 40 cm of soil was higher than those reported by Jaramillo et al. (2003) for semideciduous dry forests (3.6-4.3 t DM/ha in the top 60 cm) and for pastures of Panicum maximum, Cenchrus ciliaris and Andropogon gayanus (3.1-3.7 t DM/ha in the top 60 cm). However, the values for YP and OP can be considered low compared with that reported by Crespo and Lazo (2001) for a pasture of Cynodon nlemfuensis (10 t DM/ha in the top 15 cm of soil).
While no significant differences in total fine root mass were found between sites, the top 20 cm of soil accounted for 72.5, 68.4 and 66.2% of the total in Forest, YP and OP, respectively. Accordingly the fine root mass declined with depth as reported by Du et al. (2019) in 4 vegetation types (grassland, shrubland, secondary forest and primary forest) in a karst area, Southwest China during vegetation restoration. In that study the fine roots were concentrated in the top 10 cm of soil, which accounted for more than 57% of the root biomass, and decreased with increasing soil depth (soil samples from 0 to 30 cm deep). In karst ecosystems, fine roots contribute to the regulation of nutrient cycling in terrestrial ecosystems and the high density of fine roots within the top few centimeters of soil is crucial for acquiring nutrients.
According to Du et al. (2019), although fine root biomass of all vegetation types decreased with soil depth, the decrease was more rapid in forests (especially in secondary forests) than in other vegetation types. They suggested that the vertical distribution patterns of fine roots showed a more rapid decline in species-rich communities than in species-poor communities, which probably reflected changes in soil water content, nutrient concentration and bulk density in the soil profile. Fiala et al. (2017) also measured a reduction in fine root mass (diameter <1 mm) with increasing soil depth in 6 Cuban forests (submontanous evergreen broad-leaved forest, submontanous evergreen narrow-leaved forest, semi-deciduous narrow-leaved forest, semi-deciduous broad-leaved forest and 2 species of mangroves, Rhizophora mangle and Avicennia germinans). Fiftyseven percent of the dry mass of fine live roots was concentrated in the upper 0-5 cm soil layer in the semideciduous forests. Semi-deciduous narrow-leaved forest had live (87 g DM/m 2 ) and dead (284 g DM/m 2 ) fine roots and semi-deciduous broad-leaved forest had live (200 g DM/m 2 ) and dead (210 g DM/m 2 ) (range 371-410 g DM/m 2 ) fine roots in the top 15 cm of soil, which exceeded those in evergreen forests. The authors concluded that fine root biomass is better predicted by nutrients in litterfall. The mangrove stands had 554 g live fine roots DM/m 2 (A. germinans) and 758 g DM/m 2 (R. mangle) in the top 15 cm of soil.
Variations in root dry mass and percentage distribution of roots of Florico grass (Cynodon nlemfuensis) in the 0-40 cm soil layer under 4 different grazing strategies and seasons were evaluated by Barros et al. (2017). High concentrations of roots were observed in the 0-10 cm layer (51.8-65.6%) for all grazing strategies in all seasons. This concentration of roots near the soil surface was explained by the branched architecture of the root system, common in forage grasses and the ability of the plant to acquire water and nutrients. According to Barros et al. (2017) the root system has less need to go deeper to acquire water and nutrients when grazing is less severe.
Similarly, Rodríguez et al. (2013) reported more than 80% of the below-ground root biomass was present in the 0-5 cm soil layer in different grasslands of Mayabeque province, Cuba.
Despite C. nlemfuensis being easy to establish, persistent, highly productive and adapted to different climate and soil conditions, this species does not withstand high-intensity grazing for long periods. After defoliation by grazing the plants consume organic reserves for restoration of tissues lost and then physiological activity is adjusted as the stocks of reserves are gradually restored (Sollenberger 2008;Truter et al. 2015).
The significant increase in root mass detected in the 10-20 cm soil horizon at all study sites could be associated with the highest nutrient concentrations in soil at that depth (González-Pedraza and Dezzeo 2011;2014a;2014b). A further hypothesis may be that plant roots explore deeper soil in search of water, particularly during the dry season, in order to survive during periods of water stress, as shown by Snyman (2009) in pastures under different drying conditions in semi-arid regions of southern Africa. A high concentration of roots at this same depth of soil was also reported for savannas in tropical dry forest areas of India (Tripathi et al. 1999).
In this region the soils have been described as deep, moderately fertile, with moderate concentrations of organic matter, nitrogen and phosphorus, high capacity for cation exchange (CEC) and saturation with bases, with clay loam and silty clay loam textures (Schargel and Strebin 1970;González-Pedraza and Dezzeo 2011). Schargel and Strebin (1970) described a buried horizon at a depth of close to 50 cm, where increases in the percentages of clay, CEC, calcium, potassium, carbon and nitrogen were observed.
Fine roots are usually quick to respond to changes in vegetation type, soil temperature, moisture and nutrient content (Du et al. 2019). The increase in fine root mass during the dry period at 0-5 cm soil depth (Table 4) is contrary to the results found by Singh et al. (1989), Kummerow et al. (1990) and Trujillo et al. (2006). However, other studies have reported small increases in root biomass with decreased water availability in the soil (Castellanos et al. 2001;Snyman 2009;Barros et al. 2017;Du et al. 2019), and this has been considered as an adaptive strategy of plants under significant water stress (Snyman 2009).
For example in temperate grasslands, Walter et al. (2012) demonstrated that root length was highest when rainfall variability was intermediate and was 43 and 24% shorter with extreme and low rainfall variability. According to this author, although enhanced root growth under drought conditions is viewed as an adaptive feature of many species, sometimes grassland roots may not respond to dryness with enhanced root growth.
During the dry season, Barros et al. (2017) reported a higher concentration (65.6%) of fine roots of C. nlemfuensis near the surface (0-10 cm) in more severe grazing treatments than in less severe ones. This may be associated with the renewal of the root system, also known as "turnover", during the favorable dry season growth period and with the accumulation of organic reserves. The lower densities of root dry mass in winter, relative to the other seasons, would reflect the fast recovery of the aerial parts of plants in the rainy season and redirection of organic reserves to restore the root system reserves.
Pastures under grazing tend to accumulate more reserves in their roots, as a mechanism of adaptation to defoliation (Barros et al. 2017). This aspect, combined with the water shortage during the dry season, could be contributing to the increased mass of fine roots during this time of year. On the other hand, the significant reduction in the mass of fine roots with increasing soil depth that occurred during the dry season could be related to a higher concentration of nutrients in the first 5 cm of soil, as was noted by Castellanos et al. (2001) for tropical dry forests and pastures.
The TOCr and TNr results obtained in the forest and pastures of this study are similar to those reported by Trujillo et al. (2006), Jaramillo et al. (2003) and Crespo and Lazo (2001) for other tropical dry forests and pastures. The similarity between the studied forest and pastures in the C and N concentrations in the mass of fine roots is also consistent with the results found by Jaramillo et al. (2003). The high values for TOCr and TNr in fine roots within the 10-20 cm soil horizon are closely related to the high root mass present at that depth of soil. A positive and significant correlation (P<0.05) was found between the mass of fine roots and the amounts of TOCr and TNr throughout the soil profile (Table 4).
While the sites were not contiguous, soils on which the various sites were located were very similar. Our data suggest that conversion of forests to pasture did not change carbon and nutrient levels in the soil to any significant degree. Several factors can influence these outcomes such as the land cover, post-conversion land management, climate and soil type and texture (Dengiz et al. 2019). Dengiz et al. (2019) analyzed the spatial variability of soil organic carbon density under different types of land cover within different soil types in a subhumid terrestrial ecosystem and found that soil type and land cover were critical factors influencing spatial variation of soil organic carbon (SOC) density. Land cover was the primary factor affecting variation in SOC density, while soil properties like texture, genetic horizons and soil depth also had an important influence. The observation that organic carbon concentration decreases with increasing soil depth under all land cover types has been generally observed in most situations.
Soil fertility may have a direct influence on fine root mass production. In this study, soil fertility affected the quantities of fine roots in these soils (high correlation between fine root mass and nutrient content in the top 40 cm of forest and pasture soils). Reynolds and D´Antonio (1996) indicated the possibility that nutrient availability has direct effects on root mass because fine roots are generally plastic organs and plants can deploy photosynthate below-ground to gather growth-limiting resources. Similarly, Hutchings and de Kroom (1994) said that root proliferation in fertile soils promotes extensive exploration of the soil and increases root surface area for greater water and nutrient uptake.
Tropical pastures can accumulate large stores of C in the soil. Chaplot et al. (2016) evaluated the potential of grassland rehabilitation through high density-short duration grazing to sequester atmospheric carbon and found a significant increase in SOC stocks with increasing grass biomass and grass cover to rates as high as 12%/year. Amongst the proposed explanations were increased root biomass production, greater soil aggregate stability and associated greater organic matter protection from decomposers.
Our findings suggest that the quantities of fine roots, soil carbon and other nutrients in soils supporting tropical pastures can be equal to or exceed those in soils supporting tropical forest trees. However, the values reported here were recorded at 3 specific times of the year from 3 relatively similar sites but not strictly the same soils. Nevertheless, the data do add to the increasing database of soil C storage information under various types of vegetation.

Conclusions
According to the results of this study, the land use change from forest to pastures has not significantly affected the mass of fine roots (<2 mm) and their carbon and nutrient concentrations in the soil. Additionally, the changes in distribution of fine root mass down the soil profile were closely related to the changes in the nutrient content of the soil at the considered depths.
The moderate increase in the mass of fine roots in the young pasture (YP) may be associated with additional root inputs from the growth of secondary forest vegetation, which is considered a weed and therefore removed periodically. Further studies in other regions are needed to determine if these findings apply to other ecosystems.
Further studies are needed to clarify the effects of climate seasonality and soil nutrients on magnitude and distribution of fine root mass within the soil profile and its contribution to C storage in soil under both forest and pastures.