Study area

The study area is located in the micro-watershed of the La Venta River, in Copalita, state of Oaxaca, Mexico (16°12^’0” N, 96°25^’30” W), at an average altitude of 2 850 m (Figure 1). According to the Köppen classification as modified by ^{García (2004)}, the study area has a C (w) climate: semi-cold sub-humid, with a cool and long summer; winter rainfall less than 5 % of the annual rainfall, with an average annual precipitation above 1 300 mm, an average maximum of 215 mm during the month of July, and an average minimum of less than 10 mm in February. Average annual temperature of 18 to 20 °C, with a maximum of 24 °C in April and May and a minimum of 0 to 5 °C in December and January. The soil type is Luvisol; which have a high potential for agriculture, but also for temperate and tropical forests (^{INEGI, 2013}).

Figure 1 Location of the study area. 

Description of the sampling sites

The present research was carried out in five land uses with the Luvisol soil type (^{INEGI, 2013}):

a) Pasture. The area has different species of grasses ―Bouteloua repens (Kunth) Scribn. & Merr., Muhlenbergia robusta (E. Fourn.) A. Hitchc., Panicum bulbosum Kunth, Panicum mertensii Roth., P. parviglume Hack., and Paspalum convexum Humb, et Bonpl. ex Flüggé (^{Pacheco-Rivera and Dávila-Aranda, 2004})― that were naturally established on plots used for agricultural production for about 10 years, which were later converted to grassland currently, they have 32 years with the same land use.

b) Agricultural use. This occurs in areas of no more than 5 ha, where rainfed agriculture is practiced, with corn crops associated with broad beans and sometimes interspersed with potato crops, as they are species adapted to these conditions; no change of land use has been recorded in 42 years.

c) Pine tree plantation. Established in agricultural areas that were 10 years old, subsequently used as grassland for 17 years. Currently, the plantation is around 15 years old; the planting was staggered, with a density of 1 111 trees per hectare and an equidistance of 3 m × 3 m, the species used were Pinus ayacahuite Ehrenb. ex Schltdl., P. douglasiana Martínez and Abies religiosa (Kunth) Schltdl. et Cham.

d) Original pine forest (control). It consists of different taxa: P. ayacahuite, P. douglasiana, P. patula var. longepedunculata Loock. ex Martínez, P. leiophylla Schltdl. & Cham, and Abies religiosa.

e) Agricultural-livestock use. Farmers use these plots for grazing cattle, goats and sheep. These are agricultural areas that have been fallow for 1 to 2 years and have had intermittent agricultural activity for the last 20 years.

Soil sampling and analysis

In each soil use, four composite samples (from four subsamples) were collected at two depths (0-10 cm and 10-30 cm), for a total of 40, which were analyzed in the soil laboratory of the School of Forest Sciences of the Universidad Autónoma de Nuevo León. The samples were dried and sieved with a 2 mm metal mesh for the different analyses (Table 1).

Physical and hydrological properties

Table 1 shows the methodologies used to determine the physical and hydrological variables.

Statistical analyses

All variables were subjected to the Kolmogorov-Smirnov test with Liliefors and Levene's correction (^{Steel and Torrie, 1988}) to determine whether the distribution was normal and the homogeneity of variance. A logarithmic transformation (Log₁₀) was applied to the cumulative infiltration (I _b ), infiltration capacity (I _c ) and hydraulic conductivity variables, and an arctangent transformation was used for the initial infiltration variable (I _a ). In the end, only the variables Po, I _a , I _b , and I _c met the assumptions of normality and homoscedasticity. Subsequently, an analysis of variance (ANOVA) was performed for the different land uses. The Kruskal-Wallis test was used in the variables K, AD, MRP, S, Cl, Si, FC, PWP, and Aw at a depth of 0 to 10 cm, and in S, Cl, Si, FC, PWP, and Aw at a depth of 10 to 30 cm, as these do not meet the assumptions required for parametric analysis.

The Kruskal-Wallis Bonferroni post hoc test was also performed with pairwise comparisons for each land use. The relationship between the variables was analyzed using Spearman's correlation for the two depths sampled. Tukey's test (p=0.05) was used to compare means. Each of the statistical analyses was performed with a significance level () 0.05, in the statistical software SPSS ^® (Statistical Package for the Social Sciences) version 22.

Infiltration capacity and porosity

The results of Initial Infiltration (I _a ), Cumulative Infiltration (I _b ), and Infiltration Capacity (I _c ) for each land use are illustrated in Figure 2.

The plotted values represent the mean (n=X), ± standard error. Means with different letters indicate statistical differences according to Tukey's test (p =0.05).

Figure 2 Initial infiltration (a), Cumulative infiltration (b), Infiltration capacity (c), and Porosity (d) for different soil uses at the depth of 0 to 10 cm. 

The pine forest (control) exhibited the highest value of I _a (4 500 mm h^-1), followed by the pine plantation (2 100 mm h^-1), the agricultural area (760 mm h^-1), the grassland (500 mm h^-1) and the agricultural-livestock use (480 mm h^-1). The other components of infiltration (I _b and I _c ) had a similar behavior according to the land use.

The results of the analysis of variance showed that the variables I _a , I _b , I _c , and Po exhibited significant differences (p≤0.05) between land uses (Table 2).

The infiltration observed in the different kinds of land use, when compared to that of the pine forest (control), clearly shows the effect of the change in land use on this characteristic. Properties such as porosity, apparent density, mechanical resistance to penetration, and soil management practices cause modifications in the infiltration rates.

^{Luna et al. (2020)} evaluated infiltration in an Umbrisol type soil, subjected to different forestry treatments in a pine forest, and registered values below those obtained in the present research for the pine forest and in the pine plantation; the same applies to the study by ^{Chagoya et al. (2015)}.

Pine forest had the highest porosity (76 %), followed by grassland and pine plantation (70 %), agricultural-livestock (63 %), and the use with the lowest value was agricultural (61 %). The former data indicate the difference in infiltration and porosity between the pine forest (control) and the other evaluated uses. This shows that changing the use of forest land affects the different physical and hydrological variables. Soil porosity in the pine forest and pine plantation is associated with the incorporation of organic matter, while that of the grassland is associated with the large number of roots of herbaceous plants and grasses that are added each year and penetrate the surface, promoting aeration and infiltration of water into the soil.

Figure 3 shows the average infiltration behavior for each assessed land use during 210 minutes; it shows the difference between pine forest and pine plantation in relation to agricultural, grassland and agricultural-livestock uses. From the 90^th minute on, infiltration stability was observed for all kinds of land use.

The plotted values represent the average (n=x).

Figure 3 Average infiltration curves for different land uses. 

According to the Kruskal-Wallis test for the K, AD, MRP, S, FC, PWP and Aw variables, the null hypothesis of equality of centrality parameters for the different kinds of land use in the 0 to 10 cm depth was rejected (Table 3); while, at soil 10 to 30 cm deep, only the variable Aw registered significant differences (p≤0.05) between kinds of land use (Table 4).

Post Hoc Kruskal-Wallis tests were also performed for independent samples with pairwise comparisons and Bonferroni correction for land use. These showed significant differences for the variables K, AD, FC, PWP, S in the kinds of use of the Agriculture-Forest combination, as well as for K, PWP, and Aw the Forest-Agricultural-Livestock use at a depth 0-10 cm, while Aw had differences between Pine Plantation and the Agricultural-Livestock use at a depth 10-30 cm (Table 5).

Each row tests the null hypothesis that land use distributions 1 and 2 are equal. MRP = Mechanical resistance to penetration; K = Hydraulic conductivity; AD = Apparent density; Aw ¹ = Available water (0-10 cm); Aw ² = Available water (10-30 cm); FC = Field capacity; PWP = Permanent wilting point; S = Sand. Asymptotic significances (bilateral tests) are shown. The significance level is 0.05.

Hydraulic conductivity

The saturated hydraulic conductivity (K) showed significant differences among the kinds of land use that were analyzed. The mean K values were 0.001 cm s^-1 for agricultural and agricultural-livestock uses; 0.002 cm s^-1 for pasture; 0.008 cm s^-1 for pine plantation, and 0.022 cm s^-1 for pine forest. The low value of K for agricultural-livestock use is due to the increase in apparent density and the decrease in porosity. The static and dynamic pressure exerted by livestock on the surface soil causes compaction and mechanical resistance. The results for K are similar to those obtained in other studies (^{Quichimbo et al., 2012}; ^{Jaurixje et al., 2013}; ^{Novillo-Espinosa et al., 2018}).

Bulk density

The mean AD values for depth 0 to 10 cm were 0.63 g cm^-3 for pine forest, 0.78 g cm^-3 for pasture, 0.79 g cm^-3 for pine plantation, 0.98 g cm^-3 for agricultural-livestock, and 1.03 g cm^-3 for agricultural use. The low DA in the pine forest is due to its porosity and mechanical resistance, in addition to its sandy loam textural class. This shows that land use change increases DA, which also results in an increase in RMP and a decrease in K.^{Cruz-Ruiz et al. (2012)}, ^{Carlos-Gómez et al. (2014)}, ^{Chagoya et al. (2015)} and ^{Alejandro-Martínez et al. (2019)} document similar values of bulk density and porosity, and also indicate that AD tends to increase at greater depths and Po decreases in loam, sandy loam and clay textures.

Mechanical resistance to penetration of the soil

The MRP of the soil at a depth of 0 to 10 cm exhibited significant differences between kinds of land use. Grassland recorded the highest value for hardness (0.42 Mpa), followed by the agricultural-livestock use (0.27 Mpa). While, agricultural uses, pine plantation and pine forest had a MRP (0.15 Mpa). The high MRP value observed in the grassland is due to soil compaction caused mainly by livestock, in addition to increased AD and decreased Po; the exposure and lack of cover favored a greater loss of soil moisture. ^{Yáñez-Díaz et al. (2019)} cited similar results for mechanical resistance to penetration in a Vertisol soil, in which hardness is common when it loses moisture. Conversely, ^{Medina-Guillén et al. (2017)} indicated lower MRP values that may vary according to the type of vegetation, management and stocking rate present.

Soil Texture

At the 0-10 cm depth, only the proportion of sand had significant differences between land uses, ranging from 62 to 72 %. The proportion of clay and silt was in the range of 5 to 8 % and 22 to 30 %, respectively. For these two variables, no significant differences were observed between the kinds of land use.

At the depth of 10 to 30 cm, none of the texture variables showed significant differences, and the percentages of sand ranged from 67 to 73 %; silt ranged from 23 to 27 %, and clay, from 4 to 6 %. The texture assessment for the 0-10 cm depth corresponded to the classifications sandy loam (grassland, agricultural, pine plantation, and agricultural-livestock) and loamy sand (pine forest). The 10-30 cm depth corresponded to sandy loam (pasture, agricultural, agricultural-livestock) and loamy sand (pine plantation and pine forest). This means that the pine plantation and forest have lower apparent density and mechanical resistance to penetration but higher porosity; this favors the hydraulic conductivity and infiltration of the soil, which contribute to the recharging of aquifers and reduces the risk of erosion.

The results of the present research coincide with those obtained by ^{Lozano-Trejo et al. (2020)}, who evaluated a Luvisol soil in the Copalita River sub-basin.

Soil water characteristics

Field capacity (FC). The FC at a soil depth of 0 to 10 cm exhibited significant differences between the kinds of uses under study; values ranged from 33.59 to 51.86 % water; the pine forest (51.86 %) and the agricultural use (33.59 %) had the highest and lowest percentage, respectively, of water retention at field capacity. The FC of the remaining uses were: agricultural-livestock (39.64 %), grassland (42.77 %) and pine plantation (43.57 %).

No significant differences between kinds of use were detected at a depth of 10 to 30 cm. The FC ranged from 32.81 to 42.70 %, with agricultural use having the lowest value (32.81 %), and grassland, the highest (42.77 %); pine plantation registered 36.44 %; the agricultural-livestock use, 39.49 %, and pine forest, 40.22 %.

Pine forest was the land use with the highest FC value, probably because the organic matter content in the soil favors moisture retention, and due to its low density and mechanical resistance to penetration, all of which allows water infiltration.

Field capacity is affected by the change of use; thus, agricultural use had the lowest FC. This is due to its porosity and texture (^{FAO, 2005}); in addition, the high apparent density, lower organic matter content, lack of cover, low hydraulic conductivity and soil compaction increase the mechanical resistance to penetration, which, therefore, does not allow a larger amount of water to be retained (^{FAO, 2009}).

Permanent Wilting Point (PWP). The PWP of the 0 to 10 cm depth of the Luvisol showed significant differences between the kinds of land use, with a variation between 16.77 % and 39.72 %. Agricultural use had the lowest value (16.77 %), and the highest value corresponded to pine forest (39.72 %). The agricultural-livestock, pine plantation and grassland types of use registered: 19.02, 25.94, and 26.76 %, respectively. The PWP range for the depth of 10 to 30 cm was 18.70 % to 29.15 %; the agricultural sector had the lowest value (18.70 %), followed by agricultural-livestock (20.76 %), pine plantation (24.91 %), and grassland (27.09 %); the highest value, of 29.15 %, was determined in the pine forest. In general, a low clay content was observed in the Luvisol, which consequently exhibited lower moisture retention.

These results are similar to those obtained by ^{Muñoz-Iniestra et al. (2013)}. ^{Silva et al. (2011)} state that, by maintaining a cover, an agroforestry system allows for greater soil moisture.

Available water (Aw). The Aw content exhibited significant differences between the land uses at the two sampling depths (0-10 cm and 10-30 cm). At the depth of 0 to 10 cm, it ranged between 12.14 and 20.62 %; pine forest exhibited a lower percentage of Aw (12.14 %), followed by pasture (15.71 %), agricultural (16.83 %) and pine plantation (17.66 %) and agricultural-livestock use registered the highest percentage (20.62 %). The Aw at the depth of 10 to 30 cm fluctuated between 11.08 and 19.13 %; likewise, the pine forest had the lowest percentage, of 11.08 %, followed by pine plantation, with 11.54 %; agriculture, with 12.13 %, and grassland, with 15.69 %; while the agricultural-livestock use had the highest percentage, of 19.13 %.

The lowest percentage of available water was obtained in the pine forest and pine plantation, since these were the uses with a textural class (loamy sand) with a high percentage of sand, mainly at the depth of 10-30 cm. The higher the number of macropores, the faster the water seeps into the subsoil, as gravitational water.

These available water values are higher than those documented by ^{González-Nivia (2014)} and ^{La Manna et al. (2018)}, but lower than those indicated by ^{Béjar-Pulido et al. (2020)} in andosols in Uruapan, state of Michoacán, because sandy soil has a higher water infiltration, and, therefore, the percentage of available water there is lower. ^{Bachmair et al. (2009)} point out that land use change has an impact on the vertical and horizontal movement of water.

Correlation of physical and hydrological variables

Table 6 shows Spearman’s correlations (r _s ) for the physical and hydrological variables evaluated at the 0 to 10 cm depth interval. The AD correlated significantly, but negatively, with the Po, FC, and PWP. The MRP exhibited a negative and highly significant correlation with I _a and I _b . At the depth of 10-30 cm, only the field capacity (FC) had a significant positive correlation with S, Aw, and PWP.