Assessing the culture of fruit farmers from Calvillo, Aguascalientes, Mexico with an artificial neural network: An approximation of sustainable land management

Abstract

The need to promote a sustainable land management (SLM) approach for land degradation neutrality (LDN) is a global challenge, and we must rely on local assessments of how rural participants are conducting their work. Therefore, the objectives of the present research are as follows: 1) evaluate the fruit culture of the guava producers (Psidium guajava, L.) in Calvillo, Aguascalientes, Mexico; and 2) to show whether it is possible to detect key variables in SLM by using an artificial neural network to improve productivity. The study unit (SU) was located in the community of San Tadeo (21.917 °N and 102.701 °W) in the municipality of Calvillo, Aguascalientes. We considered a total of 430 ha established with this fruit and nine orchards that consisted of 9.05 ha, which is equivalent to 2.09% of the total area. Each hectare was randomly sampled, with a significance of 95% (α ≥ 0.05). The information was collected by using empirical field techniques and a questionnaire sent to the producers. The results revealed that the productive success of the SU is based on the knowledge and ability of the fruit farmers to take advantage of bioclimatic and biophysical conditions for guava cultivation. It was demonstrated in approximately 90% of the analyzed cases that the producer’s SLM was between regular and excellent. This result was confirmed by the index of edaphic adequacy (Iea), which was created by nine of thirteen variables measured in the field, that indicated in 47% of the orchards (OR), the edaphic condition is optimum (Iea = 0) or close to optimum (-0.2 ≥ Iea ≤ 0.2). In addition, by using a Multilayer Percetron (MLP), it was possible to detect two key aspects in the producers’ work: cava design (DC) and soil management (SM).