RANDOM SELECTION OF SAMPLE ELEMENTS AS A RESOURCE TO CONSTRUCT THEMATIC MAPS FOR THE SOIL RESISTANCE TO THE PENETRATION

• Autor(a) principal: de Oliveira, Márcio Paulo
• Data de Publicação: 2013
• Outros Autores: Dalposso, Gustavo Henrique, Vieira, Maycon Daniel, Secco, Deonir
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Engenharia na Agricultura
• Texto Completo: https://periodicos.ufv.br/reveng/article/view/337
• topic: dependência espacial; exatidão global; índice Kappa; krigagem; matriz dos erros.

The study investigated the effect of reduction of the sample elements number used in the representation of the distribution of spatial variable of soil resistance to the penetration (RSP) on layer 0-0,1m in the thematic maps. The Grid100, with one hundred sample elements marked every 10m is used as reference set and the reduction of the number of samples is considered as Grid 25. In other words, 25 elements sample marked every 20 m plus 21 elements sample that is neighbor of three elements sample selected by random process within Grid25. The random process efficiency was verified by comparison of maps of spatial variability of soil resistance to penetration on layer 0-0,1m. The first map constructed for a region of 10.000 m2 considered a sample set with 100 points marked every 10m, called reference map with the Grid100. The second map constructed to the same region, but using the sample set reduced, the Grid46, called model map. The comparison between the thematic maps was realized by error matrix with the index values of global accuracy and Kappa of 0,87 and 0,78, respectively. The index showed that the thematic maps have a very good similarity. This form to select the sampling grid is a resource to reduced number of sample to minimizate the field operations and the costs including the laboratory labor to measure other soil properties. Thus, the reduction of the number of sample grid, called Grid46, is suggested to sampling the soil properties in the region of study with purpose to reduce the number of samples to be collected and the costs to measure it.