Characterization of regional production systems

In Valle de Santiago we worked with 55 maize cropping systems and in Salvatierra with 13 systems of the same type. Surveys were conducted in both municipalities to characterize the types of technical maize itineraries.

Sampling of cropping systems

a) Soil. In the selected cropping systems, 10 soil subsamples per hectare were taken, at a depth of 0 to 20 cm, after which a composite sample was taken for laboratory analysis. The properties analyzed were pH (1:2, soil:water), organic matter (Walkley-Black method), total nitrogen (Kjeldahl-N method), phosphorus (Olsen), interchangeable potassium (in ammonium acetate, pH 7.0) and cation exchange capacity (ammonium acetate, pH 7).0) with the methods specified in Mexican standard NOM-021- SEMARNAT-2000 (^{Secretaría del Medio Ambiente y Recursos Naturales [SEMARNAT], 2000}), and carbon content was estimated from the organic matter value obtained. Table 1 presents the average results for each municipality.

b) Biomass. To estimate the total aerial biomass, three sampling areas (3 linear meters each) were randomly selected in each cropping system in which the population density was measured. All plants in the sampling unit were cut at ground level and weighed in the field, then the cobs were separated and their total fresh weight was obtained. Three plants were randomly selected from the cut plants, weighed and dried in an oven at 70 °C for 48 h to determine moisture content and dry weight, after which total biomass and grain yield were estimated. Plant organs (leaves and stems) and grains were separated from these three plants and sent to the laboratory for analysis of their N, P and K content. Total nitrogen was analyzed with the semi-micro Kjeldahl technique (^{Bremner, 1965}), while P and K were determined by inductively coupled plasma spectrometry (Varian Liberty Series II ICP-AES, Varian®, USA) (^{Alcántar & Sandoval, 1999}).

Information analysis

a) Nutrient extractions. The N, P and K content in the vegetative biomass and grains of the three plants analyzed was multiplied by their respective biomass produced (kg DM·ha^-1), and then added to obtain the total extraction of nutrients.

b) Contribution of nutrients (N, P and K) from the soil and fertilizers. The availability (f) of N, P and K, and the potential soil contribution of these nutrients (SN, SP and SK, respectively) were calculated according to the Quantitative Evaluation of the Fertility of Tropical Soils (QUEFTS) model (^{Janssen et al., 1990}; ^{Sattari, van Ittersum, Bouwman, Smit, & Janssen, 2014}):

where C is the soil organic carbon (g·kg^-1), assuming that this represents 58 % of the organic matter content of the soil, N is the total nitrogen (g·kg^-1), P is the available phosphorus (mg·kg^-1), K is the interchangeable potassium (cmol·kg^-1) and the pH is that of the soil. The efficiency of fertilizer recovery used was: N = 0.4, P = 0.2 and K = 0.4 (^{Bruulsema, Fixen, & Snyder, 2004}; ^{Ghosh, Singh, & Mishra, 2015}; ^{Peña-Cabriales, Grageda-Cabrera, & Vera-Núñez, 2001}).

c) Mineralization of soil organic carbon (SOC) and residues. The mineralization coefficient of ^{Henin and Dupuis (1945)} (K₂) was used for clay soils; in the residues, it was assumed that 45 % of these were carbon. The isohumic coefficient (K₁) of 0.12 (^{Soltner, 1990}) was used to estimate organic matter humification. Based on the available residues, three retention scenarios were made: 100, 50 and 30 %.

Contributions for the analysis of N, P and K management in the soil

In the maize cropping systems of Valle de Santiago and Salvatierra, great variation in grain yield was found (between 7 300 and 15 800 kg·ha^-1), with no association between nutrient availability and grain yield (Figure 1). This questions the existence of a variety of limiting factors (abiotic, biotic and their interactions) that eventually act on biological growth for biomass production and yield.

Figure 1 Relationship between available nutrients and maize yield in farming systems in Valle de Santiago (2012, ○) and Salvatierra (2014, ●), Guanajuato, Mexico. 

The amount of available nitrogen varied widely (from 86 to 284 kg N·ha^-1) (Figure 1A), although most farming systems were concentrated in the range of 134 to 236 kg N·ha^-1. The application of this nutrient via fertilization ranged from 71 to 554 kg N·ha^-1, a situation that was more evident in Valle de Santiago. In the case of phosphorus (Figure 1B), its availability varied between 8 and 91 kg P·ha^-1; however, the highest availability was between 28 and 75 kg P·ha^-1, and only 10 % of the cropping systems were greater than 70 kg P·ha^-1. It is important to point out that the dose of phosphorus applied through fertilizers was on average 45 kg P·ha^-1, with a range of variation between 26 and 92 kg P·ha^-1.

Most cropping systems had high availability of potassium (Figure 1C), since this nutrient is abundant in the region (Table 1), which implies little need for its application. However, in 57 % of the cases evaluated, potassium fertilizers are applied, with doses varying from 18 to 50 kg K·ha^-1. It is important to note that the needs for this nutrient for maize cultivation are usually similar to those of nitrogen (^{Cruzate & Casas, 2003}), so these inputs contribute in a certain way to reducing the short- and medium-term decapitalization of potassium in the soil.

On the other hand, the balance of N, P and K (extraction vs. availability) showed different trends among them (Figure 2). In the case of nitrogen (Figure 2A), it can be seen that in 59 % of the cropping systems the extraction was lower than the availability, in 29 % of the cases the extraction was higher than the availability and in 12 % there was a balance condition. In addition, 90 % of the cropping systems had between 150 and 250 kg·ha^-1 of available nitrogen, and about 80 % of these extracted between 100 and 200 kg·ha^-1. The average dose of nitrogen applied was 309 kg·ha^-1, and the amount of unused nitrogen ranged from 7 to 195 kg·ha^-1. In general, 89 % of the nitrogen supply to the system corresponds to fertilizers, which indicates the low availability of this nutrient in the soil. It is estimated that the region consumes around 330 thousand tons of nitrogen per year, with a recovery efficiency of 20 to 40 % (^{Peña-Cabriales et al., 2001}). Considering the results and assuming a recovery efficiency of 40 %, losses may be associated with the form and timing of application, volatilization, immobilization and leaching (^{Grageda-Cabrera et al., 2011}; ^{Liu, Herbert, Hashemi, Zhang, & Ding, 2006}).

Figure 2 Relationship between available and extracted nutrients in maize cropping systems in Valle de Santiago (2012, ○) and Salvatierra (2014, ●), Guanajuato, Mexico. 

It is important to note that there were not enough technical elements or agro-economic information to define the dose strategies and the timing of fertilizer application according to the phenological and reproductive states of maize. In general, two fertilizations are made, one in the sowing and the other 30 to 40 days afterwards; however, this can generate losses because they do not conform to the highest nutrient demand stages of maize. In the cases where the extracted N was higher than the available N, it is believed that the recovery efficiency could be higher than 40 %, in addition to which there are contributions via rain and mineralization of nitrogen from the soil and residues. Various studies in the region have shown that crops do not recover about 260 thousand tons of nitrogen, where 20 to 30 % is denitrified as nitrous oxide and molecular nitrogen, between 20 and 30 % is leached as nitrates, and 10 to 18 % is volatilized as ammonia (^{Grageda-Cabrera et al., 2011}) and contributes 60 % of nitrogen dioxide (^{Montzka, Dlugokencky, & Butler, 2011}; ^{Reddy, 2015}).

Due to the nature of nitrogen fertilizers, technical strategies aimed at reducing losses and improving recovery efficiency must be implemented. In this sense, adequate fertilization doses must be defined according to the yield objective and the available nitrogen in the soil; in addition, according to the maize demand curve, the application of fertilizers must be multi-fractionated, preferably three applications via mechanical means or fertigation (^{Andraski, Bundy, & Brye, 2000}; ^{Cueto-Wong et al., 2013}; ^{Paredes, Alamilla, & Mandujano, 2014}).

In the case of phosphorus (Figure 2B), it was found that 87 % of the cropping systems had an extraction lower than the availability. Phosphorous recovery efficiency is of particular interest because it is one of the least available and mobile nutrients. Overall, annual recovery efficiency varies between 15 and 25 %; however, phosphorus tends to accumulate in the soil, mainly in cropping systems where annual applications of phosphorus fertilizers are made, a general situation in this case. In view of the above, recovery efficiency should be evaluated over the long term (^{Ghosh et al., 2015}; ^{Smil, 2000}).

In the evaluated systems, 20 % of the extracted phosphorus comes from fertilizers and the rest is obtained from the soil, which indicates that there are sufficient reserves of this nutrient, since the average values in both municipalities are higher than 11 mg·kg^-1 (Table 1). According to Mexican standard NOM-021-RECNAT-2000 (^{SEMARNAT, 2000}), which establishes the specifications for soil fertility, salinity and classification, this value is high. Thus, phosphorus is not considered a limiting factor in Valle de Santiago and Salvatierra; however, phosphorus management should focus on nutrition records in order to maintain reserves and provide more efficient and cost-effective management of phosphorus fertilizers.

On the other hand, the potassium balance (Figure 2C) also indicated that the extraction (100 kg K·ha^-1) was less than the availability, since on average 226 kg K·ha^-1 were not used. The application of potassium is an uncommon and relatively new practice, considering that it was not included at the end of the 20th century. In this sense, it was recorded that 62 % of the farmers use it; however, they pointed out that potassium was not considered in their maize nutrition plan because they were unaware of the role it has in the crop. In addition, the soils of this region are characterized by having a high exchangeable potassium content (Table 1), a condition associated with their nature (^{Ramírez-Barrientos, Figueroa-Sandoval, Ordaz-Chaparro, & Volke-Haller, 2006}), and it is sufficient to cover maize requirements. Several studies have observed a response in yield when applying this nutrient in potassium-rich soils (^{Aguado-Lara, Etchevers-Barra, Hidalgo-Moreno, Galvis-Spíndola, & Aguirre-Gómez, 2002}), which could be associated with nitrogen availability.

Considering the results obtained, it is necessary to establish mechanisms to determine the antagonistic and synergistic effects between potassium and nitrogen (^{Zhang et al., 2010}). In addition, in Mexico, and particularly for this type of soil, there is little information on potassium in soil-plant interaction, which suggests the need to investigate this and other elements in depth to improve the diagnosis of their presence in the soil (^{Aguado-Lara et al., 2002}) and generate regional empirical models that allow having technical bases for their adequate and efficient use. Additionally, it is important to note that most of the potassium extracted by crops is concentrated in their vegetative parts, which indicates that residues are an important source of potassium. For this reason, a key element that contributes to increasing potassium levels in the soil is to incorporate the residues, since they can significantly promote the recycling of this nutrient in the cropping system (^{Zhang et al., 2010}), with their consequent ecological and economic advantages in the durability of the regional system.

The behaviors recorded in Figures 1 and 2 show that there is differentiated management in the application of nutrients in the maize cropping system, which was also reflected in a generalized imbalance between the availability and extraction of N, P and K; this has implications for the reduction or depletion of essential nutrients (^{Cassman et al., 1996}). However, in the soil there are reserves of P and K that cover maize’s demand, which indicates that currently they are not limiting factors for the production of this grain. On the other hand, N is a limiting factor because the soil only contributes 11 % of the total available nitrogen, and the rest is supplied through nitrogen fertilizers. The fractionation of nitrogen fertilization and its application according to its demand in the different phenological stages of maize are determining factors for an efficient use of fertilizers (^{Ali et al., 2005}).

Contributions for the analysis of soil C management

Grain production ranged between 7 300 and 16 100 kg·ha^-1, and implicitly there is a similar production of residues, this considering that the average harvest index was 0.5. The residues generated in the evaluated cropping systems play a determining role in increasing the levels of organic matter, carbon and nutrients. In the scenario where 100 % residue retention was assumed, SOC mineralization fluctuated between 345 and 1 000 kg C·ha^-1·year^-1 (Figure 3A). In this case, 19 % of the plots were located in the upper part of the bisector, which indicates that the incorporated residue volumes improve the C reserves in the soil. However, even with this management there are challenges, since in most cropping systems the organic reserves and their mineralization in the soil are insufficient with respect to the C contributed by the residues.

Figure 3 C mineralization in soil and residues in three residue retention scenarios in maize cropping systems in Valle de Santiago (2012, ○) and Salvatierra (2014, ●), Guanajuato, Mexico: A) 100 %, B) 50 % and C) 30 %. 

In the 50 and 30 % residue retention scenarios (Figures 3B and 3C), it can be seen that the mineralization and contribution of C of the residues tends to be less than the C mineralization of the soil, which contributes to assuming its humic decapitalization, as well as indicating an imbalance between the two C sources. In the 50 % scenario (Figure 3B), C mineralization of residues ranged from 153 to 442 kg C·ha^-1 and only one case was in a balance condition. When reducing residue retention to 30 % (Figure 3C), C mineralization showed a greater imbalance, since the difference between C mineralization of the soil and residues was 509 kg C·ha^-1. This scenario illustrates the general trend of farming systems in the region, where residues are removed or burned to facilitate soil preparation work.

Several studies have shown that residue retention increases organic matter levels, mainly in regions where they are the only source of organic matter (^{Fregoso-Tirado, 2008}; ^{Flores-Sánchez, Groot, Lantinga, Kropff, & Rossing, 2015}; ^{Rusinamhodzi, Wijkc, Corbeels, Rufino, & Giller, 2015}). On the other hand, the burning of residues sends carbon emissions into the atmosphere and decapitalizes soil C, although the capture of atmospheric carbon by crops and its incorporation into the soil can mitigate CO₂ emissions. In addition, residue burning has reduced soil organic matter levels (from 2.6 to 0.6 %) in the last three decades, and has favored the increase in fertilization doses (^{Fregoso-Tirado et al., 2002}); consequently, soil compaction problems and decreased nutrient efficiency occur. In the regional context, a challenge is to improve soil quality and its efficient technical-economic management.

The incorporation of residues, as well as the efficient application of mineral and organic sources, are strategies that contribute to increasing organic matter and carbon reserves, releasing nutrients (^{Govaerts et al., 2009}), influencing yield and reducing mineral fertilizer needs in the long term (^{Kamkar, Akbari, Teixeira-da Siliva, & Movahedi-Naeni, 2014}). These strategies should be applied through a holistic-integral approach (^{Liu et al., 2006}), as they demand knowledge about local-scale decomposition processes and their potential for regional replication, which is essential for recycling, mineralization, availability of macronutrients and reduction of negative impacts on the environment (^{Powlson, Whitmore, & Goulding, 2011}). These options should be an essential component in strategies aimed at improving the productivity of farming systems, in addition to increasing resilience to climate change (^{Lal, 2009}).