Yield simulation

It was performed using the simulation model SWAT with its extension ArcSWAT within the ArcGIS 9.3 interface. Because SWAT is a hydrologic model that works at basin level, it was necessary to condition the study area to perform the simulation at the level of this hydrologic unit (^{Inurreta et al., 2013}), the basin to simulate was divided into sub-basins using topographic information from a digital elevation model (DEM) and a map of surface runoff. In each subbasin were identified areas with the same interval slope, soil type and current land use, called HRU's. The slope map is generated by SWAT from DEM, however edaphological and current land use maps were taken from INEGI with a 1: 250 000 scale and fed into the system.

In the model were used climate, soil, coffee physiology and management practices information to simulate the coffee growth, estimate its yield at HRU level and cartographically express the result. This model consists of eight components: a) climate; b) hydrology; c) nutrients and pesticides in soil; d) soil erosion; e) plant growth and ground cover; f) management practices; g) processes in the main drainage channel; and h) water bodies (storage).

Climate information was extracted from a database of 1 145 stations from the SMN, which had full climate information of at least 10 years. With this information and using the weather generator from EPIC model (^{Sharply and Williams, 1990}), generated weather statistics and subsequently daily data for the period 1912-2010. The edaphological information to characterize soil subclasses from the INEGI map was obtained of a report from INEGI with data field from 1 247 agrologic wells. The missing information was generated from the following sources: the hydraulic properties and soil erosion factor (universal soil loss equation) were calculated in relation to texture according to ^{Colin et al. (2013)} and ^{Ramírez et al. (2009)}, and the albedo was calculated from the organic matter in the soil, applying a regression equation obtained from the original information of SWAT with 202 albedo soils and organic matter. Crop physiological parameters required by the model are presented in Table 1. Agronomic management was prepared according to (^{Méndez, 2011}; ^{Themis et al., 2011}).

Estimation of economic potential

The methodology to evaluate long term agricultural projects (^{Gittinger, 1982}) through the estimation of profitability indicatiors: cost/benefit ratio (RB/C), net present value (VAN), internal rate of return (TIR) at HRU level, whose mathematical expressions are presented below (^{Coss, 1984}).

Where: FC is the flow of production costs, FI is the revenue flow, i is bank rate and n plantation age.

FC and FI were estimated from the technical coefficients per hectare from the technological package proposed by ^{López (2011)} and Themis et al. (2011), considering the crop horizon of 15 years using the following formulas:

Where: TCoE₁ is the total cost of plantation establishment in year 1, calculated by the sum of inputs amount used for the establishment of one hectare of coffee by their respective average market price in the states of Chiapas, Veracruz , Puebla and Oaxaca in 2014; CoFP in the fixed cost of production from year 1 to year 15, calculated by the sum of depreciation costs of the plantation (TCoE minus the salvage value, by the 15-year life) and management costs (3% of income from coffee sales) and CoVP is the variable cost of production of year 1 to year 15, calculated by the sum of the different amounts of inputs for operation and maintenance of the plantation by their respective average market price from the states of Chiapas, Veracruz, Puebla and Oaxaca, during 2014.

Where: Pxkg is the market price of coffee cherries from year 3 to 15, RCxha is the coffee cherry yield per ha, Pxmad is the market price of wood in year 15 and Qmad is the amount of timber sold in year 15. the market price of coffee cherries was estimated with data from ICO (2016), whose original values are given in cents per pound of green coffee, therefore converted to Mexican pesos per kilogram of coffee cherries, applying the conversion factor 0.1841 kg of green coffee per kilogram of coffee cherry, timber prices are current average prices from the states of Chiapas, Veracruz, Puebla and Oaxaca during 2014.

Since establishment, production costs and income correspond to the lifetime of the project; these were taken to their current value using a bank rate of 10.4%, proposed by the World Bank as the opportunity cost of capital for public investment projects in Mexico (^{Coppola et al., 2014}). The study of economic potential was complemented with a sensitivity analysis, considering three scenarios: i) decrease in coffee price, considering the international price, assuming that in the long term the domestic market price decrease; ii) increase in fertilizer price, as one of the inputs of higher percentage in the structure of variable cost, caused by a general increase in prices according to the trend of inflation at national level; and iii) the combination of the above two situations, decline in coffee prices and increase of fertilizer price which is considered the most critical case.

Productive potential of crop of coffee

The surface with potential to produce the crop is presented in Figure 1, in the green, yellow, orange and red area could be produced coffee cherries with higher yield than the national average, which for the period 2000-2014 was 2 000.2 kg ha^-1 of coffee cherry (^{SIAP, 2016}) or 368.23 kg ha^-1 of green coffee, since a kg of coffee cherries is equal to 0.1841 kg of green coffee (^{Flores, 2015}). Although the competitive potential is to produce coffee placing areas above the world average, for the same period was 691 kg ha^-1 of green coffee (^{FAO, 2016}), equivalent to 3 753.4 kg ha^-1 of coffee cherry, in Figure 1, it is observed that there are about 1.896 million ha with competitive potential, located in the states of Veracruz, Puebla, Chiapas and Oaxaca, an area 2.5 times higher than that planted of coffee now throughout the country, which is 737 376.5 ha (^{SIAP, 2016}).

Figure 1 Surface with potential to produce coffee in the SouthSoutheast of Mexico. 

The distribution of the surface by yield level of coffee cherry on the main states with potential is shown in Table 2, where can be seen that on 32% of the regional surface can produce coffee, being Guerrero where greater surface was detected, although with lower yields than other states. Veracruz has the lowest percentage of surface with potential, however this state has the regions with higher yields, therefore despite occupying the third place in planted area, ranks second in domestic production (^{SIAP, 2016}), and the other state that follows in surface with potential for showing high yields is Puebla, therefore ranks third in domestic production. Globally the highest coffee yields are obtained by Vietnam and Brazil with an average of 10 and 6 t of coffee cherries respectively (^{Flores, 2015}), yields that can be obtained in Chiapas, Oaxaca, Puebla and Veracruz.

If Mexico wants to use the comparative advantages that some regions represent, it should produce coffee in 1 million ha where could obtain yields greater than 6 000 kg of coffee cherries ha^-1, to obtain similar yields to those from Brazil which is the main producer in the world (^{Flores, 2015}). But it is not enough to be competitive, but to conserve natural resources, to do so it is proposed to discard the 641 000 ha whose current land use is forest (Table 3) and to establish coffee in the remaining 381 000 ha.

Watersheds with greater length with potential to produce coffee are the Grijalva and Lacantun River, both in the state of Chiapas, comprising 578 HRU and Papaloapan River, comprising the states of Puebla, Oaxaca and Veracruz, with 158 HRU (Table 4). The basin where the highest yields can be obtained 9 000 kg ha^-1 is in the Jamapa River and others, followed in importance by the Astata River and others and Nautla River and others.

The results presented show the usefulness of estimating yields and thereby locate surface with productive potential, as ^{Rivera et al. (2012)} did, for the cultivation of cassava (Manihot esculenta Crantz) in Tabasco, where, by the variables: climate, soil, temperature, rainfall, altitude, photoperiod and growth period, were able to locate 171 121 potential hectares in Huimamguillo 70 386 in Balancan and 41 337 in Macuspana, relevant information for the design of policies to support the development of a culture in a region.

Economic potential of the crop

Obtaining high yields is not enough condition to take the decision to produce, it is also requires crop profitability, to know this information the technical coefficients from the technological package were valued, Tables 5 and 6 show the investments and operation costs of an hectare of coffee during the lifetime of a plantation respectively. As seen in Table 5, it is required to invest more than $26 000 pesos per hectare to establish a plantation, being the vegetative material the highest value with 57%.

As for operating costs, Table 6 shows that in year 1 costs are lower, mainly because the crop is growing and there is no production, also the amount of fertilizer applied is lower due to plantation size, costs start increasing each year, to reach the value of $24 136.28 in year 6, and they continue until year 15, the highest concept of cost is harvest, followed in importance by fertilization, this cost structure is similar to that reported by ^{López and Caamal (2009)}. To operation costs were added $1 392.00 per year for plantation depreciation, plus administration costs that vary according to yield.

To complete the estimate of profitability indicators is necessary to estimate the income of the plantation, which are determined by two variables: yield and price of coffee, the first is taken from the estimated productive potential, whose value depends on the agro-ecological region as can be seen in Figure 1. in the case of price, there is no single value, because coffee is not an homogeneous product, as shown in Figure 2, where can be seen the price behavior from the last 16 years, of the three predominant types of Arabica coffee in the world, plus the indicator price indicated by ICO (2016), as an indicator price of all coffee types, in the same Figure 2 are presented the average rural price of coffee in Mexico, where it can be seen that it has a similar behavior to international price, although are lower by 30% on average, as mentioned by ^{CEDRSSA (2014)}, so coffee price considered in the study is $6.08 per kilogram. With cost and revenue information estimated and updated the cash flow for 15 years, which in turn allowed estimating the R B/C, VAN and TIR indicators.

Fuente: estimación con datos de ICO, 2016 y ^{SIAP, 2016}.

Figure 2 Coffee cheery price in Mexico and the world. (In Mexican pesos). 

The results show that the minimum yield that a coffee plantation requires for a producer to start having profits is 4 500 kg of coffee cherries ha^-1, with which a R B/C of 1, a VAN of 0 and a TIR of 10.4%, equal to the bank rate, which are the conditions of equilibrium (Gittinger, 1974), therefore, plantations earning returns above this amount are profitable, situation that is not currently happening since the national average was just 1 670 kg ha^-1 in 2014 (^{SIAP, 2016}), this result confirms the approach from ^{CEDRSSA (2014)}, stating that coffee production has decreased its yield by disease problems and high production costs.

Table 7 presents the financial indicators assessed in six basins with the largest area with potential to produce coffee, it shows that in the basins from Jamapa river and others and from the Nautla river and other have higher potential yields, thus financial indicators are the most valuable, although in all basins these indicators are favorable, so if coffee plantations were established in these basins producers would get 30 to 48 cents for each peso invested, it would also allow them to hire a credit and pay an interest rate between 16% and 28% higher than the opportunity cost of capital.

As mentioned previously, the area with potential for coffee production is 381 000 ha (72% agricultural and 28% livestock) but varies by state, the main agricultural crops currently grown are corn, sugar cane, beans and coffee the beef production in the case of livestock (^{SIAP, 2016}), so the producer decision to change the current land use would be based is an expectation of higher gain to that currently obtained. In a study conducted in Mexico by ^{Domínguez et al. (2010)} mentions that grains are less profitable than fruits and vegetables, however it also requires less investment, so the risk is also lower.

In another study where coffee production both in a traditional system and organic production, coincides that the benefit cost is negative in the traditional system and 1.15 if organic coffee is produced in the state of Chiapas (López and Caamal 2013), confirming the need to establish plantations in productive areas and applying a technological package that ensures returns above the national average. Finally ^{Carrera et al. (2013)}, studying the behavior of beef production during the period from 1980 to 2009, they found that this activity has become less competitive. The latter allows placing coffee production as a viable activity for producers in the region studied, although it is necessary to establish a supportive policy, aimed at reducing the financial cost of the investment and promote intensive labor as suggested by ^{Barrera et al. (2011)}, in a similar study.

Table 8 shows the results of the sensitivity analysis under three scenarios, a) decrease in the price of coffee of 25%, which is the lowest value that coffee producers have obtained in Mexico in the last 10 years (Figure 2); and b) an increase in the price of fertilizer by 25%, which was the highest value obtained over a 10 year period analyzed (^{SE, 2016}) and c) a combination of the two. Financial indicators obtained show a high sensitivity to variations in coffee prices since a fall in price of 25%, as happened in late 2010, some watersheds are no longer profitable, situation that is exacerbated when combined with an increase in inputs price.

In 2014, 670 316 ha of coffee were planted in the southsoutheast of the country, producing 1 084 249 tons with a yield of 1.6 t ha^-1 (^{SIAP, 2016}), so that the area of 381 000 ha with potential has been reported in the four states, which would produce more than 6 t ha^-1, it would generate more than twice the current production, under profitable conditions and thereby achieving also to increase the income of coffee farmers in the country.