Full text
| |
| Author(s): |
Cesar Nunes de Castro
Total Authors: 1
|
| Document type: | Master's Dissertation |
| Press: | Piracicaba. |
| Institution: | Universidade de São Paulo (USP). Escola Superior de Agricultura Luiz de Queiroz (ESALA/BC) |
| Defense date: | 2004-07-02 |
| Examining board members: |
Jose Paulo Molin;
Jose Ruy Porto de Carvalho;
Marcos Milan
|
| Advisor: | Jose Paulo Molin |
| Abstract | |
Precision Agriculture is a line of research relatively recent in the agricultural area and one that presents promising perspectives on the generation of new field management techniques that allow for cost reduction and input optimization. Techniques as variable rate application of fertilizers, agrochemical and better sensing of soil characteristics, and other production factors, give information that enables better management of fields being cultivated and result in greater economic revenue. For it to be possible to optimize agricultural production using these techniques some initial information are necessary. For one, there is the utilization of yield maps to evaluate the spatial variability of fields. Some researches demonstrated that, conjugated to those maps, other levels of information should be incorporated, permitting to reveal factors that explain spatial variability of yield in a field. Among these information there are physical-chemical soil characteristics maps. One of the concepts for mapping soil factors is sensing its electrical conductivity, which is related to soil characteristics such as texture, moisture, salinity and other. This research studied the measurement of the electrical conductivity of some types of Brazilian soils. Then, correlation of the electrical conductivity readings with soil parameters, obtained by laboratory analysis of soil samples, and with crop yields, cultivated on three experimental areas, was obtained. Besides that, it was verified the possibility of definition of soil management zones, using soil electrical conductivity, which can successfully represent homogeneous regions with regards to attributes that influence the development of crops. For accomplishing these objectives, several analysis techniques were used. After the descriptive statistics analysis of the data set, the geostatistical analysis was made for the evaluation of the variables degree of spatial dependence and then obtaining the maps with the spatial distribution of these same variables. In the possession of these maps, it was possible to verify Pearsons linear correlation among all variables, followed by the principal components analysis to infer which of them had greater influence on soil variability in each of the experimental areas. Thenceforth, fuzzy logic was used to define the soil management zones and an analysis of variance evaluated the importance of the zone divisions made. Results showed the usefulness of electrical conductivity, besides corroborating the efficiency of the proceedings used, because of its correlation with some physical-chemical properties of the soil, especially high correlation with clay content in one of the experimental areas. (AU) | |