Estimativa da umidade do solo por meio da técnica TDR, utilizando a análise de resíduos

• Autor(a) principal: Hashiguti, Heraldo Takao
• Data de Publicação: 2013
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da Universidade Estadual de Maringá (RI-UEM)
• Texto Completo: http://repositorio.uem.br:8080/jspui/handle/1/1234
• topic: Irrigação; Solo; Água disponível; Ajuste de modelos; Constante dielétrica; TDR; Brasil.; Irrigation; Solo; Available water; Adjustment models; Dielectric constant; TDR; Brazil.; Ciências Agrárias; Agronomia

The Time Domain Reflectometry (TDR) technique has stood out due to the speed, non-disturbance of the medium and the supposed reliability of the estimate it provides. Though initially it was assumed that a single universal model would be appropriate to relate the soil moisture to the soil apparent dielectric constant (Ka), the evolution of knowledge around the subject brought the realization that calibration is essential in this case. To describe the quality of the adjusted model several studies have used statistical indicators, but despite all this action strategy, an assessment centered in model estimates residue analysis, in relation to the values measured in several works, has led to the understanding that considerable expression errors continue to occur. Thus, this work was carried out with the purpose of evaluating the quality of eight models, using data from soil moisture and its corresponding Ka, measured by means of TDR technique. Among these models, it has been the general model of Topp, the model developed by Gonçalves with the incorporation of soil bulk density (Bd), the model of Ledieu, three models "imported" from other works, ie, of the Tommaselli & Bacchi, Kaiser and Santos, and a polynomial model along with a root model fitted to the data. For this evaluation, obtained in a Dystroferric Red Ultisol under different values of soil bulk density (Bd). The quality evaluation was conducted according to the usual criteria and described by a residue analysis from the models. The results confirmed that local calibration is crucial, but even so, the models that allegedly showed adequate quality, when subjected to residue analysis, showed errors whose magnitude reaches order values greater than 100% of the available water in the soil. These results show that even under the best circumstances, the TDR can lead to mistakes incompatible with operational requirements, both at research level and in management activities in the field. The statistical indicators commonly employed to evaluate the quality of the adjustment of a model studied in this work, proved insufficient to attest this quality, since large estimation errors were not detected by them.