Introduction

In the temperate semiarid region of Aguascalientes, due to the importance of the production system of family dairy (small units with smaller herds of 30 cows in production) and the frequent incidence of intermittent and unpredictable droughts (^{Acosta-Díaz et al., 2003}), an annual deficit of more than 350 000 t of dry forage is estimated. The most critical deficiency occurs in the spring period, given the lack of rain and the scarcity of water (^{Carranza Trinidad et al., 2007}). Therefore, it requires the ability to evaluate dry matter yield and adaptation to environmental conditions in the region of fodder species; also accompany with some sowing collection system in situ rainwater, which carry the same techniques along with better use of the rain (because it increases the amount of water available to plants), they follow practices that help conserve ground, with resulting benefits (^{Martínez and Jasso, 2004}; ^{Osuna-Ceja et al., 2007}).

In the area, family milk production demand increase yield quality forage in the areas of rainfed. An alternative crop management to address it is sowing beds of triple row, with less conventional distance of 76 or 80 cm. This practice can increase plant density and increase dry matter yield per unit area, mainly due to increased interception of solar radiation during the growth cycle (^{Barbieri et al., 2000}; ^{Guevara-Escobar et al., 2005}; ^{Reta et al., 2007}; ^{Soltero et al., 2010}).

The highest yield of sowing a triple row has been linked to an increase in leaf area index and interception efficiency of solar energy per unit leaf area (^{Reta et al., 2007}). Thus, forage species of extended growth and high leaf area index, could cover better a groove and maximize the collection of solar energy and reduce direct evaporative water loss by covering the soil faster. On the other hand, upright or compact species, of low growth and low leaf area index, do not cover the entire area in wide rows, wasting solar and ground water; a situation that can be improved by reducing the line width of sowing and make more efficient use of land and rain.

Population densities and distribution of plants in the field have much to do with the characteristics of the cultivar development (height and branching of the plant) and environmental factors (soil, rainfall, temperature, etc.), which makes a distribution density and optimal plants for farming, not the best for another, especially if they differ in growth habit and earliness (^{Castellanos et al., 1990}; ^{Widdicombe and Thelen, 2002}; ^{Reta et at., 2003}; ^{Guevara-Escobar et al., 2005}).

It is essential to know the soil where plants grow and meet their physical condition, nutrient content, its electrical conductivity and cation exchange capacity. These factors are related to the ability of soil to provide plants the conditions required to obtain the nutrients they need to grow and produce the quantity and quality of forage as expected. Organic fertilizers not only improve soil physical conditions, but provide a significant amount of nutrients and reduction in production costs (^{Salazar-Sosa et al., 2007}; ^{Salazar-Sosa et al., 2010}).

One of the main problems in this region is shallow soils, with low nutrient and organic matter content, caused by overexploitation (^{Osuna et al., 2000}; ^{Osuna-Ceja et al., 2007}); ^{Del Pino et al. (2008)} reported that the application of manure increases the activity and amount of soil microbial biomass and are an alternative to reduce the use of chemicals, including fertilizers. Therefore, the use of organic fertilizers is proposed in addition to the nutritional requirements of the culture with mineral fertilizers in order to increase the yield and quality of the products.

With respect to the physical properties of the soil, usually the structure, porosity is enhanced, and increases the rate of infiltration, hydraulic conductivity, water retention and the apparent density is reduced. However, sometimes it is not possible to demonstrate its benefit in one or two crop cycle; this is especially true in soil treated with small or moderate amounts of organic waste (^{Fortis-Hernández et al., 2009}).

The aim of this study was to evaluate the combined effect of a triple row sowing, in situ rainwater capture and application of cattle manure; and the influence of the combinations between them on the production of maize, forage sorghum and chickpea rainfed and properties of soil fertility.

Materials and methods

This work was done during 2013, in the Experimental Site Sandovales, Aguascalientes, Mexico, of the National Research Institute of Forestry, Agriculture and Livestock (INIFAP), located at 22° 54' 16" north latitude and 102° 20' length west to 2045 m. The soil of the experimental area is planosol type of sandy loam with slightly acidic pH of 6.4 and less than 1% organic matter. The climate is semi-desert with summer rains (200-250 mm in the growing season). The average annual temperature is 16.2 °C, the maximum annual average is 20 °C and the minimum is 7.1 °C (^{Medina et al., 2006}).

Under these conditions of soil and climate, the response of three rainfed forage crops with different agronomic practices where the study factors were evaluated: a) sowing methods: conventional single row with a distance between rows of 76 cm, without attracting rainwater and triple beds of1.60 m rows, where the separation between them was 40 cm, with in situ capture of rainwater through "row diking" (a practice that was performed on the sides of the seedbed and consisted of ground lift a board 20 cm high at regular to store water and reduce soil erosion) and Aqueel (a practice that was done with a roller formed with individual gears that prints footprints on the soil without causing compaction during distances the process was passed over the bed at sowing for micro-reservoirs evenly, over the soil surface for capturing rain water in situ); b) improved varieties: maize (Zea mays L.) V-209; chickpea (Cicer arietinum) swineherd landrace and sorghum (Sorghum bicolor L. Moench) Sweet and; c) organic and chemical fertilization: the treatments were doses of cattle manure (0, 10 and 20 Mg/ha), chemical fertilizer 80-40-30 (N-P₂O₅-K₂O; kg ha^-1) and a control.

Manure has been applied since 2011, with the same dose per year under the same experimental plots. It was used as organic fertilizer because of its low cost and readily available in the study area. The characteristics of manure applied are: pH 7.5; EC 0.60 dS m^-1; MO 5.39%; total N 1.26%, ammonium 0.1130%, P 0.3533%, calcium 3.30, magnesium 0.69%, K 3.36%, sodium 0.95 mg/kg, molybdenum 555 mg/kg, iron 12 250 mg/kg, zinc 195 mg/kg, copper 45 mg/kg and boron 412 mg/kg. The manure had 3% moisture when applied. The experimental area was 3 000 m².

The distributions of treatments were made under a design randomized complete block arrangement of sub-divided plots with three replications. The large plot was the sowing method with and without water catchment, the median girl crops and organic fertilization treatments. The experimental plot consisted of four rows in conventional sowing three rows of beds of 5 m in length.

The terrain was prepared with vertical tillage (use of multi-plow) and pluming before sowing. The previous crop was rainfed maize. Sowing under rainfed conditions and in moist soil was made on 21 June; a seed was placed every 0.30 m in maize, at 0.14 m in chickpea and 0.08 m in sorghum, this allowed to distribute 3, 7 and 12 seeds per meter in these cultivars evaluated in two sowing methods, respectively; 17, 36 and 60 seeds per row 5 m long in conventional and 51, 108 and 180 seeds per bed of three rows, to handle a density of 66 000, 140 000 and 244 000 plants ha^-1 of maize, chickpea and sorghum, respectively. In conventional sowing, with the same distribution of seeds per meter a density of 43 600 was handled, 93 500 and 163 750 plants ha^-1 of maize, chickpeas and sorghum, respectively. Performing a weeding and agrochemicals were applied to control pests of the foliage.

The variables measured in climate, soil and plant were:

Climate: in the year of study, data of daily precipitation (mm) were obtained from the automated weather station Sandovales, located at a distance of 300 m from the experimental area.

Soils: profile depth was measured, bulk density (ρb), total porosity (ƒt), usable water sheet (L) and sorptivity (So), for the surface layer 0-15 cm. In all cases, three measurements were made for each sowing method, cultivation and fertilization. The method of cylinder of volume was used known for ρb (^{Jury et al., 1991}); ƒt, and La were estimated according to ^{Sustaita et al., 2000}; So was determined by the method of infiltration cylinder (^{Talsma, 1969}). Soil samples were taken only in the topsoil (0-15 cm) and were sent to the laboratory to determine its texture, moisture constant (CC) and (PMP) nitrate (NO3), organic matter (OM), pH and electrical conductivity (EC) of the experimental area before sowing. Texture (Hydrometer of Bouyoucos), CC and PMP (pot and pressure membrane), CE in extract; MO (Walkley and Black, 1934; ^{Page et al., 1982}), pH in a water ratio: soil 2.5: 1 and nitrates analyzed through colorimetrically (^{Page et al., 1982}).

Plant: crop harvest took place when the grain showed a milky-doughy state (to 97-100 days after sowing). The cutting height was 5 cm from ground level. The cut forage was weighed, subsequently obtaining a sub-sample, which was the laboratory for dried in a forced air oven at 60 °C until constant weight (4% moisture), to have the dry matter (MS) and transform the results on a dry basis (^{Reta et al., 2007}). The samples were processed in a Wiley mill with 0.5 mm mesh. The total N was determined by the Kjeldahl method; P by colorimetry and K by flame photometry (^{Chapman and Parker, 1986}). NPK total accumulated in the MS is calculated from the total weight of MS per unit area obtained by the nutrient concentration in the determined sample. The efficiency of utilization of the three minerals studied was determined by the MS division performance between the total accumulated nutrients in the same area.

Data were analysed using the SAS statistical package, version 8 (SAS Institute, 1999) and when significance was detected between treatments, the least significant difference test (DMS) was applied.

Results and discussion

The year 2013 was characterized by the higher amount of precipitation than normal for the region. During the first 100 days of crop development, rains occurred with some regularity, accumulating a total of 532 mm during the growing season; during the cycle several significant periods of precipitation occurred. In a period of 10 days rained 52.6 mm, in another, it rained 75.6 mm in 7days and a third 171 mm in 5 days (Figure 1).

Figure 1 Precipitation during the growing season of forage species (maize-sorghum-chickpea). Sandovales, Aguascalientes. 2013. 

The duration of the crop to the cutting step was 100 days from sowing to understand the milky doughy stage. The rain in this period was 372.8 mm. In Figure 1, the distribution of rainfall occurs during the growing season. It is observed that, the total rainfall 62.3% occurred in the first half of the cycle (vegetative, flowering and grain formation), which had only growing 37.7% Rain on stage realized grain filling; this indicates an erratic distribution to crop needs.

Therefore, rarely the moisture content through the soil profile reached the water storage capacity. This is because the rains are scarce and poorly regularly distributed during the stage of crop development. So that intermittent drought, unpredictable at any stage during the life cycle of the crop, is a constant threat to farmers, especially in this semiarid region of North Central Mexico. However, is noteworthy that in this cycle, there was enough retention of moisture in the root zone, so that the crop did not suffer from extreme drought during different stages of development. The rains continued through the experiment tested at a level of humidity higher than 40% of available moisture in production systems evaluated during the first 85 days of crop development treatments. The system of sowing beds in triple row had no moisture limitations, since plants showed no symptoms of water stress at any stage of development due to the reserve of soil moisture for the extra water retained by effect of recruitment practices in situ (Aqueel and row diking). It was obvious to note that rainwater was evenly distributed on the ground, avoiding concentration in the lower parts and runoff out of the ground; as in conventional sowing.

Physical quality of the soil and its relationship with organic matter

The integration of technological components in a production system to promote increased physical and chemical quality of agricultural soils and the reduced effect of moisture deficit in the performance of MS in various cultivars of time by withdrawal of water in situ adding manure in the surface layer (0-15 cm) led to the evaluation of their properties physical and chemical changes due to the combined effect of components (Table 1). In contrast to what was observed by some researchers (^{Osuna et al., 2006}, ^{Osuna-Ceja et al., 2007}) excelled the lack of significance of the apparent density (ρb), total porosity (ƒt) and usable water (La) between sowing methods and cultivars evaluated. In contrast, ρb decreased and ƒt increased significantly (up to 10%) and in proportion to the manure (Table 1), which may be encouraged airflow and water and root development of the plants (^{Salazar et al., 2007}).

Table 1 Variation and comparison of means of bulk density (Mg m^-3), total porosity (%), usable water (cm) and sorptivity (cm sec^1/2) obtained with two sowing methods, species and compost, Sandovales, Aguascalientes, Mexico. 2013. 

ρ_b= densidad aparente, ƒ_t= porosidad total, L_a= lámina de agua aprovechable, S_o = sortividad.

The highest La showed and apparent relationship with the addition of manure (Table 1). This is attributed to the development of a more porous structure favoured by the organic matter (OM), meaning higher capacity to store water. Instead, the soil without adding manure (control all 0 manure application) has a smaller La, probably due to the lower total porosity of the soil (Table 1). The average values of OM in the surface layer (0-15 cm) are statistically different (p≤0.05) in soil with manure treatments analysed. The highest content of soil presents MO 20 Mg/ha (2.05%) of cattle manure (EB) compared to the control (1.19%) absolute (TA); Eb/TA (2.05 / 1.19= 1.72) indicates that EB exceeded 72% at RT.

The results of sorptivity (property reflects the influence of the management of the soil matrix in the passage of water during the initial period of infiltration) observed in the treatments of 10 and 20 Mg/ha in the surface layer reflects the ability of manure to significantly increase the porosity and water movement into the soil in arid areas. The data were grouped according to treatment, i.e., the absolute control and chemical fertilization grouped sorptivity values (So) lower and higher values of ρb; while the manure treatments grouped the highest values of So, with lower values of ρb (Table 1). Thus manifesting an effect of MO on ρb and the other evaluated properties (^{Dexter et al., 2004}; ^{Osuna et al., 2006}).

Regarding the content of nitrates in the soil, it did not show any statistical difference (p≤ 0.05) between sowing and cultivation methods evaluated. However, at the beginning of the cycle significant differences (p≤ 0.05) were observed in the concentration of nitrates in the soil for compost treatments (Table 1). In the layer 0-15 cm depth in the control treatment, average values of 5.2 mg nitrate/kg were found and to 14.8 mg/kg for the treatment of 20 Mg/ha of manure. This result reflects the benefits of manure with increased mineralization, as evidenced by an increased amount of NO₃ on the soil as the application rate increased manure. Similar results reported by ^{Salazar-Sosa et al. (2007)}, who observed a significant effect to apply manure in more than 40 Mg/ha, increasing all macronutrients necessary for the development and production of the plant. The highest concentration of nitrates was found in the first 15 cm because that's where possibly the temperature conditions, aeration, soil and humidity favour microbial and enzymatic activity.

The results of the study showed the positive effect of manure, independent of the method of sowing and cultivating, as both evaluated proportions increased physical and chemical quality of the soil. This technological component, combined with a vertical tillage and row diking, allow the sustainability of these soils of semiarid temperate highlands, especially the vertical tillage that does not reverse the soil, leaving few crop residues as mulch on the surface and the uptake of rainwater in situ and, the application of organic matter on the consecutive years (^{Castellanos et al., 1996}; ^{Osuna et al., 2006}; Ceja-Osuna et al., 2007).

Forage production

It was found that there was a significant response (p≤ 0.05) for the effect of sowing method. The yield of green biomass (BV) higher than was obtained in the bed sowing method to triple row followed by the conventional method (Table 2). The very same was observed in the evaluated cultivars, and its significant differences (p ≤ 0.05) in the combined analysis. On average, the performance of BV on the plots of maize and sorghum were higher than in the plots of chickpea (Table 2). Organic fertilization treatments affected the production of green fodder which reflected statistical difference (p≤ 0.05). Mean test showed that production rates were BV 20.6 33.9 Mg/ha, and treatments of 10 and 20 Mg/ ha of manure higher yield obtained with 30.4 and 33.9 Mg/ ha, respectively. The lower yield of 20.6 Mg/ha was obtained by the control (Table 2).

Table 2 Effect of the study factors on the yield of green and dry matter biomass under two sowing methods, species and cattle manure, Sandovales, Aguascalientes, Spring-summer-2013 cycle. 

RMV= rendimiento de materia verde; RMS= rendimiento de materia seca.

Regarding MS, the highest yield was obtained on beds of 1.60 m triple row, which exceeded 54% with the conventional distance to 76 cm, respectively (Table 2), which is consistent with that reported by ^{Guevara-Escobar et al. (2005)} and ^{Reta et al. (2007)}.

Crop response was significant (p≤ 0.05), through organic fertilization and sowing methods. The highest dry matter yield was obtained with maize, sorghum and then finally in plot with chickpeas.

Regarding the application of manure it showed statistical differences (p≤ 0.05) between the treatments. The analysis showed that the mean DM production ranges were 6.27 to 16.47 Mg/ha, and the treatments of 10 and 20 Mg/ha of manure obtained better yields with 13.7 and 16.4 Mg/ha, respectively. The lower yield of 6.2 Mg/ha was obtained by the control (Table 2).

The treatments with the highest production were those who had manure, which is due to the residual effect of beneficial manure, since during three previous years was added to the soil in the same amounts in the same experimental plots.

Moreover, the treatment performance of chemical fertilizer also was below that obtained with the lower dose of manure application; similar results were found by ^{Aguilar-Benitez (2012)}. These results are probably due to the fact that in addition to providing nutrients to the crop, adding manure for three consecutive years has improved the structure and levels of soil organic matter (^{Julca-Otiniano et al., 2006}).

Nutrient concentration in dry matter (DM)

In relation to forage quality parameters, as for milk production in dairy systems familiar in this region of semiarid temperate highlands are of great importance to maintain the quality and quantity of this product, the results found on average per treatment are shown in Table 3. The concentration of NPK either in sowing methods and between cultivars showed no significant response (Table 3). However, the level of manure and chemical fertilization changed the N concentration in the MS (p≤ 0.05) of the three cultivars.

Table 3 Concentration and accumulation of N, P and K in the dry matter of maize, chickpeas and sorghum planted under two sowing methods, and compost, Sandovales, Aguascalientes, Mexico. 2013. 

In the treatment of 20 Mg/ha of manure it was observed an increase in the accumulation of N in MS compared to other treatments. The amount of P and K taken was the same for all the treatments.

In the treatment of 20 Mg/ha of manure accumulated over 74 kg N/ha than the control (Table 3), double of the 36 kg N ha^-1 reported by ^{Reta et al. (2007)}. Concentrations of P and K in the treatment of 20 Mg/ha of manure were similar to those observed in the other treatments, thus, increased extraction of these elements is due to higher dry matter yield (10.2 Mg/ha). The results suggest that, the positive effect of a triple bed sowing row on MS accumulation of aerial parts (Table 2) and the extraction of N (Table 3), was due to higher DM yield. These results were comparable with those reported in the literature for high density sowing (^{Guevara-Escobar et al., 2005}; ^{Reta et al., 2007}; ^{Salazar-Sosa et al., 2007}; ^{Soltero et al., 2010}; ^{Osuna et al., 2012}).

Conclusions

In the semiarid rainfed area in Aguascalientes, maize, sorghum and chickpea sowing for forage in beds of1.60 m at triple row increases the yield of BV and MS in 57 and 54% on average, compared to conventional sowing to 0.76 m.

Bed sowing at 1.60 m on triple row with row spacing of 0.40 m high plant density, is suitable for the production of fodder maize, sorghum and chickpea to capture rain in situ and vertical tillage under the conditions examined.

Continuous application of 10 and 20 Mg ha^-1 of manure had a positive effect on some physical and chemical properties of the soil, as ρb, ƒt, La, So, the content of MO and NO3. Maize, sorghum and chickpea established in soil with 10 and 20 Mg ha^-1 of manure showed a high yield of BV and MS. The addition of manure along with the acquisition in situ of rainwater modified the water-soil-plant system by providing nutrients and reducing negative effects of water deficit stress system, which could be observed by the positive effects on performance BV and MS.