Study area

The study area is located south of the village “San José de La Parrilla”, Nombre de Dios, Durango, Mexico, which is 76 km from the city of Durango (Figure 2). The geographical coordinates are: 23° 44’ 14.54” N and 104° 7’ 32.37” W and the altitude is 2,130 m. The studied area is 12 ha and is part of a private property.

Figure 2 Geographical location of the study area in Durango, Mexico. 

Figura 2. Ubicación geográfica del área de estudio en Durango, México. 

General site description

The site is used for beef cattle production and has an average slope of 14 %. According to ^{COTECOCA (1979)}, the vegetation corresponds to a medium grassland with shrubs and trees (Cb(B)35), which is composed of perennial, medium, and tufted grasses that are associated with shrubs and low-growing trees (less than 5 m). The soil is of colluvial origin and is derived from igneous rocks; it has a shallow depth (less than 25 cm) and medium internal drainage, but when it rains, the surface runoff is rapid. This corresponds to the leptosol soil group, which is characterized by having little development, high stoniness, rocky outcrops, and a large amount of calcareous material (^{SEMARNAT, 2014}).

The area was overgrazed for over 50 years, but since 2002, it has been fenced off and managed with light to moderate grazing. Despite this, the effects of improper management, eroded areas, presence of calcareous outcrops, and invasion of undesirable plant species are still evident.

The land’s hydrology has not allowed for soil improvement due to runoff flowing from the ridges to the gullies, causing loses of water, soil, and nutrients through streams in the area. All this is increased by the lack of soil conservation works.

Soil physicochemical characteristics

To characterize the soil’s physicochemical properties, 18 samples were randomly extracted at a depth of 0 to 20 cm, with which a composite sample was established. The sample’s physicochemical parameters that appear in Table 1 were reviewed in accordance with the methods indicated in NOM-021-RECNAT-2000 (^{SEMARNAT, 2002}).

Climate

The climate is semi-dry temperate with rains in summer and an average annual rainfall of 400 to 600 mm in the July-September period, and an average annual temperature of 16 to 20°C (^{INEGI, 2017}). In order to obtain climatic data for the site and study period, the temperature and relative humidity were recorded from June 2017 to December 2018 by using a data logger (HOBO). The minimum temperature recorded was -7 °C, the maximum was 32 °C and the average was 18 °C. The minimum relative humidity in the summer was 18 %, the average was 62 % and the maximum was 86 % in both years. Annual precipitation at the site was 485 mm in 2017 and 395 mm in 2018.

Site preparation

The area was surveyed with a fixed-wing drone (model eBee). With the information obtained, a plan was drawn up with level contour lines one meter apart (Figure 3a), where the HKD was established according to the methodology described by ^{Gras (2012)}. Two keylines were identified to change the site’s hydrology (Figure 3b). These were designed as level canals to retain and direct the rain runoff from the gullies (where it normally accumulates and runs) to the ridges (generally the driest parts), and thus distribute the water throughout the land and also infiltrate it.

Figure 3 Sketch of the study area: (a) topographic plane of contour lines; (b) hydrological keyline design, treatments and sampling sites. 

Figura 3. Croquis de la zona de estudio: a) plano topográfico de las curvas de nivel; b) diseño de líneas clave hidrológicas, tratamientos y sitios de muestreo. 

The foregoing was done in order to reduce the runoff speed, allow water to remain on the site, reduce the dragging of particles, and finally favour the regeneration of soil and vegetation on the site.

In July 2017, the HKD was implemented. The keylines were marked on the ground with points 15 m apart, by using two 2-meter-long graduated rulers and a transparent hose with water acting as a level. The canals were built following the outline of these keylines with a D-4 Bulldozer using a 3-meter-wide blade. Subsequently, the tillage pattern was made according to the design with a farm tractor and a Yeomans-type chisel plough (Figure 3b). The canals were used as guides to design the tillage pattern that the Yeomans-type cultivator should follow, which according to ^{Ruiz (2013)} is made in a parallel and progressive manner below and above the canals. However, it was delineated smoothly higher on the gullies, and lower on the ridges (Figure 3a). The ploughing was done where the topography and vegetation allowed the entry of agricultural equipment (Figure 3b).

Treatments

According to the previous description, the implemented and evaluated treatments were: T1) Control, without HKD; T2) HKD, part of the area below the canals; and T3) HKD+Yeomans (Figure 3b).

Variables evaluated

Six sampling sites (repetitions) were established for each treatment, where humidity, soil compaction, soil erosion, vegetation cover, and aerial phytomass yield were quantified (Figure 3b).

Soil

Humidity (%): To measure it, gypsum blocks were used according to ^{Florentino (2011)} and with the help of a multi-meter, the soil’s electrical resistance emitted by the blocks was measured, which is inversely related to humidity. The soil moisture curve determined by gravimetry was used for its calibration. The measurements began in October 2017 by placing sensors in each treatment. The readings were grouped into three periods: end of the rainy season in 2017, the dry season in 2017-2018, and end of the rain season in 2018.

Compaction (kg cm ^-2 ): Compaction was recorded 18 months after the study was established; a Dickey John brand compaction tester was used. The meter’s tip was placed on the ground and then uniform and constant pressure was applied to slowly penetrate the tip until it could no longer enter; this allowed to register the resistance to penetration.

Erosion (ton ha ^-1 ): Erosion was evaluated one year after the trial was established and after the rainy season. Erosion was estimated according to the ^{UNESCO method (2017)}; for this, 25-cm-long nails were placed and buried 15 cm apart with washer as a stop. Soil loss or accumulation was measured with a digital vernier to the nearest hundredth of a millimetre.

With the data obtained, the erosion was estimated by adding the readings of the nails that lost soil, and those that had sedimentation were given a value of zero. Then, they were divided by the total number of nails. The sedimentation average was quantified with the same criteria.

At the same time, the net erosion was determined, which is the difference between the removed soil and the sedimented one (exit and entry on the surface of the treatments). After adding them, the mobilized soil was obtained according to the procedure indicated by ^{Pizarro and Cuitiño (2002)}. The results were extrapolated to tons per hectare.

Vegetation

Vegetation cover (%): Six and 18 months after the experiment was established, the vegetation cover was determined with the use of Canfield Lines (^{Silva-Piña et al., 2018}); for this, 5-meter-long lines were established, where the following was estimated:

Aerial phytomass yield (kg ha ^-1 ): Six months into the design, the phytomass yield was estimated in the HKD and Control treatments. It was not quantified at that time in the HKD+Yeomans, because the vegetation cover was disturbed by the passage of the Yeomans-type cultivator, so at the moment, there was no effect revealed. Finally, at 18 months, the yield was estimated in the three treatments.

In rectangular 0.5 m² plots (quadrants of 1.0 x 0.5 m), all the plants were cut flush with the ground. Subsequently, they were dried in a forced air oven at 55°C until the weight was constant. Finally, the dry weight and yield were determined, and based on this, the kilograms per hectare were extrapolated (^{Gutiérrez-Arenas et al., 2018}).

Statistical analysis

The research work was carried out according to a completely randomized design. The data were subjected to a normality analysis using the Shapiro-Wilk test (P < 0.05). When the variables met the normality requirement, analysis of variance was performed and in cases where there were significant differences, Tukey’s mean comparison tests were performed (P < 0.05). When the data did not present normality, the Kruskal-Wallis non-parametric test was used. Statistical analyses were carried out through the use of the STATISTICA program.

Soil

Humidity. The differences between treatments were not significant at the end of the period in 2017 (P > 0.05). However, in the dry period (Nov. 2017 to Jun. 2018), the humidity differed between treatments (P < 0.05), where HKD+Yeomans kept the humidity above the permanent wilting point (3.9 %). Equally, in the final rainy season of 2018, the HKD+Yeomans treatment statistically surpassed the control (P < 0.05) with 2.9 % (Table 2).

Compaction. The soil’s resistance to the penetrometer was similar in the three treatments (P > 0.05; Table 3). A resistance of 23 kg cm^-2 means that compaction is high in the first 20 cm of the soil profile (^{Unger and Kaspar, 1994}; ^{Singh et al., 2015}).

Erosion. The area with HKD had the least erosion, followed by HKD+Yeomans, and the control. In the latter, soil loss was higher than treatments two and three, with 126 and 117 ton ha^-1, respectively (Table 4).

Vegetation

Vegetation cover. Six months after the experiment was established, low herbaceous cover (13 %, on average) was found with no statistical difference between treatments (P > 0.05), but with a statistical difference among components (P < 0.05) (Figure 4a), where the highest component was the mulch with around 60 % in both treatments.

Figure 4 Effect of the treatments on the different components of the vegetation cover; (a) year 2017; (b) year 2018.  

Figura 4. Efecto de los tratamientos sobre los diferentes componentes de la cubierta vegetal; a) año 2017; b) año 2018. 

At 18 months, there were important changes in the vegetation cover (Figure 4b). In the HKD and HKD+Yeomans, the percentage of soil without cover was lower than the control with 20 and 35 percentage points, respectively. The herbaceous cover was higher in the HKD with 32 percentage points and HKD+Yeomans with 38 units with respect to the Control (P > 0.05).

Aerial phytomass yield

Six months after the experiment was established, differences were observed between treatments (P < 0.05), where HKD+Yeomans surpassed the Control with 753 kg ha^-1 and HKD with 360 kg ha^-1 (Table 5, P < 0.05). At 18 months, HKD+Yeomans and HKD exceeded the Control with 770 and 454 kg ha^-1, respectively (Table 5, P < 0.01).

Soil

Humidity. At the end of the 2017period, higher humidity was expected in the treatments where HKD was applied, but it is difficult for humidity to be conserved due to the soil’s characteristics in the experiment due to the depth of 0-20 cm, sandy loam texture, 14 % slope and compacted soil. These results are in agreement with those obtained by ^{Lal and Shukla (2004)}, as well as with those reported by ^{Toro-Mujica (2023)}, in the semi-arid region of Chile.

In the dry period, significant differences were observed in humidity between the treatments. In this regard, ^{Almaraz and Gras (2012)} explain that the Yeomans underground subsoiler creates canals with little soil disturbance, where water is easily transported, runs slowly, infiltrates and remains in the ground longer.

This is repeated, in the final 2018rainy season, where the HKD+Yeomans treatment surpassed the Control. This confirms that the hydrological keyline design accompanied by the Yeomans underground cultivator favours water retention. Studies by ^{Oscanoa and Flores (2019)} describe that the introduction of soil conservation techniques such as furrows and holes. influence the increase in soil humidity by up to 6 %.

Compaction. High resistance values observed in soil penetration, are influenced by the soil’s sandy loam texture as it is susceptible to compaction especially in the upper layers (^{Oscanoa and Flores, 2019}). This condition can affect the establishment and development of vegetation since root growth decreases when compaction is greater than 20 kg cm^-2 (^{Unger and Kaspar, 1994}). In addition, the pore space is reduced, and the water infiltration rate is affected (^{Lal and Shukla, 2004}).

^{Ruíz (2013)} explains that the use of the Yeomans plough in grasslands, in addition to help water transport without disturbing the soil, it cuts the compacted layers and improves their depth. Therefore, compaction is expected to decrease in the medium term in the areas where the implement is used consecutively.

Erosion. According to the amount of soil removed in the different treatments, the presence of the HKD reduced the erosion range from highest (> 150 ton ha^-1) to lowest (< 50 ton ha^-1) according to ^{Montes-León et al. (2011)}.

Regarding the accumulation of sediments, the HKD treatment retained a greater amount; this indicates the dragging of external materials to the treatment area, resulting in negative net erosion and a high volume of mobilized soil (Table 4). On the other hand, the Control presented the lowest values in sediments; this may be due to the lack of structure to retain the runoff and dragging of particles, which increased net erosion. ^{Ponce-Rodríguez et al. (2019)} and ^{Giambastiani et al. (2023)}, explain that the presence of conservation works established with contour lines, changes the terrain’s hydrology and runoff is reduced and distributed by the established infrastructure. Likewise, conservation works are reported to reduce erosion (^{Oscanoa and Flores, 2019}).

Grassland areas with high erosion rates are much more difficult to recover by simply excluding them and moderating grazing (^{Distel, 2013}). Thus, other practices are required as is the case in the study area where the grassland has minimally recovered in almost two decades.

The amount of soil eroded was equal to the amount of sedimented soil in the HKD+Yeomans (Table 4), which equates to a net erosion of zero; this is the lowest volume of soil mobilized among the three treatments. This shows that the canals created by the Yeomans subsoiler produce detachment and dragging of soil particles; however, these are retained by HKD+Yeomans reducing erosion as mentioned by ^{Buitrago (2013)}. A higher rate of soil infiltration allows for less soil removal.

Vegetation

Vegetation cover. The low percentages of vegetation cover estimated six months after the study was established, reflect the grassland’s degradation due to improper use. However, the high fraction of mulch is a contribution of organic matter and seeds, which will promote the regeneration of soil and vegetation (^{Massara et al., 2013}).

Regarding the changes reflected in components evaluated at 18 months, it is expected that the presence of the hydrological keyline design favours the development of vegetation, which is in line with the results of ^{IMTA (2013)}, who reported an increase of 13 % in the forage sorghum yield. Similarly, ^{Valdez and Aramayo (2018)} achieved positive results in the growth of grasslands vegetation in Patagonia.

Aerial phytomass yield. The higher forage yield in treatments with keyline can be explained on the amount of available water was higher, especially in the HKD+Yeomans, where it was above the permanent wilting point in the dry period (Table 2). This shows that the use of the Yeomans plough integrated with HKD in grasslands favoured the humid condition for plant growth.

According to ^{Ruiz (2013)}), the technique leaves roots in the soil and helps to conserve organic carbon, which in turn will conserve more water (144,000 L ha^-1 for every 1 % of organic carbon), which favours soil fertility as reported by ^{Hishe et al. (2017)}) for other conservation works, and therefore helps vegetation development (^{IMTA, 2013}; ^{Valdez and Aramayo, 2018}; ^{Toro-Mujica, 2023}).