Caracterización del comportamiento espacial de las propiedades hidrofisicas del suelo con cultivo de cacao, su incidencia en su capacidad productiva.

Abstract

Agriculture is an important pillar in social and economic development, so it involves an increasing exploitation of soil resources, regardless of the spatial behavior of the quality of it, causing degradation of the structure and generating lower yields. This relationship between production and the spatial variability of hydrophysical properties is a starting point for exercising precise, sustainable and long-term land management systems. However, most farmers are unaware of the current state of their soils. Due to the previous antecedents, the present investigation was carried out associating statistical and geostatistical techniques in the cocoa area of the Faculty of Agricultural Sciences belonging to the Technical University of Machala, with the purpose of identifying the spatial behavior of the hydrophysical properties and delimiting homogeneous zones that allow incorporating a zoned property management optimizing the water, soil, economic and social resources. The study was carried out in two depths, 0-10cm being the HA and 10-20cm the HB with a total of 20 sampling points for the analysis of twelve hydrophysical properties such as Sand, Silt, Clay, CC, PmP, PT, Da, Dr, MO, IEE, Ib, Ks. The analyses were carried out at the Laboratory of the Technical University of Machala and the physical tests of water infiltration with the use of the Munz rings. The correlation between the variables was determined by means of Pearson's method in the software Rstudio 3.6.3. Later, the semivariograms were made in the program Surfer16, determining the geostatistical parameters as Sill, Plateau and Range, two theoretical models of adjustment were used as the Spherical and Gaussian of which the one of better adjustment was selected by means of the coefficients of determination, Nash and average quadratic error. Likewise, the spatial variability was demonstrated by means of ordinary Kriging interpolations in the Arcgis 10.3 program with the previously calculated geostatistical parameters. Homogeneous zones were identified by means of the cluster analysis carried out in the software Rstudio 3.6.3, determining the optimal number of groups by means of the silhouette method and corroborating it by the location of the cluster in such a Euclidean distance. Finally, the delimitation of the zones was carried out in the program Arcgis 10.3 by means of simple interpolations of the sampling points with the assignment of the belonging cluster. The results showed a greater variability in the spatial behavior in the first 10cm of study, generating the formation of four homogeneous zones in HA and two in HB through the cluster analysis with a Euclidean distance of 50 and 60 respectively, in both study profiles, the properties presented weak spatial dependence with respect to the theoretical adjustment models Spherical and Gaussian, except for the apparent density and total porosity that were classified as strong and moderate respectively. Through the characterization of the properties, it was identified a soil of frank textural class, structurally stable with high organic matter contents that, however, do not affect the production of the property, since this is very low as a consequence of the lack of cultural practices and hydric deficit.