Mapeamento digital de solos e modelagem da recarga hídrica na Bacia do Rio Doce, Minas Gerais

• Ano de defesa: 2013
• Autor(a) principal: Souza, Eliana de
• Orientador(a): Fernandes Filho, Elpídio Inácio
• Banca de defesa: Santos, Gérson Rodrigues dos , Carvalho Júnior, Waldir de
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Viçosa
• Programa de Pós-Graduação: Doutorado em Solos e Nutrição de Plantas
• Departamento: Fertilidade do solo e nutrição de plantas; Gênese, Morfologia e Classificação, Mineralogia, Química,
• País: BR
• Palavras-chave em Português: Geoestatística; Mapeamento de solos; Recarga hídrica
• Palavras-chave em Inglês: Geostatistics; Soil mapping; Water recharge
• Área do conhecimento CNPq: CNPQ::CIENCIAS AGRARIAS::AGRONOMIA::CIENCIA DO SOLO
• Link de acesso: http://locus.ufv.br/handle/123456789/1642

The quantitative evaluation of pedological and hydric resources is important for adoption of sustainable and efficient practices and use management of resources. Climatologic and pedological data are usually collected in a discret manner and in most of the cases, the number of sites where the information is collected is usually not enough to fuel the predictive models. To overcome the shortage and spatial discontinuity of soil and water information models have been used to estimate certain properties and attributes, from other more readily available. Predictive models employ a variety of techniques and methods for generating information in discrete and continuous format. This study aimed to generate pedotranfer fuction for estimating soil bulk density and soil water content and evaluate the spatial prediction of soil attributes and hydric balance in the Rio Doce Basin, located in the State of Minas Gerais. In Chapter 1, we evaluated the spatial prediction of soil attributes: organic carbon, clay and cation exchange capacity (CEC), with the purpose of evaluating the performance of the prediction models from soil maps generated by both the conventional and with the Multinomial Logistic Regression methods, from a set of covariates consisting of maps derived from elevation model of the terrain, satellite images and maps obtained from conventional soil mapping The geostatistical model called Regression-Kriging (RK) was evaluated for prediction of the spatial attributes. In Chapter 2, we modeled the basin water recharge process for a period of two hydrological years (09/2007 - 09/2009). Hydric balance was calculated by means of the model BALSEQ (Portuguese acronym for Daily Sequential Hydric Balance). The estimated variables from the water balance; evapotranspiration, runoff, deep infiltration and precipitation were spatially predicted with the RK model. In Chapter 3, pedotransfer functions (PTFs) were used by means of Multiple Linear Regression model to predict soil bulk density, water content in four gravimetric potentials and available water capacity. The PTFs were generated for data grouped by horizon, soil class and textural groups. Comparisons between PTFs results obtained in this study with those compiled in the literature were performed. The results of Chapter 1 evidenced the importance of the conventional soil map when developing digital soil mapping. There is a need to increase the sampling size for clay and CEC, and to evaluate the covariates of higher correlation with these two attributes, in order to improve the performance of the spatial prediction models. Organic carbon was predicted with good performance by all evaluated models, presenting the best results when using the dataset that included the soil map obtained from the conventional method of survey combined with other covariates. The estimate and spatial prediction of hydric balance performed in Chapter 2 enabled better understanding of the recharge of the unconfined aquifer in the basin and the generated scenarios can be utilized in zoning the basin for use and management of its soils and water resources. In Chapter 3, we found good predictive power of the PTFs developed with the data collected in the basin; we also observed improved performance in relation to the functions compiled from the literature. The information generated with the developed PTFs are rarely found in current literature body and are of great importance as input variables for environment modelling studies, and can be used to fuel models of spatial prediction. In general, the contribution of this study is generating soil and hidric information for the Basin that enable to carry on climate change related studies from measurements of soil carbon stock, water availability and spatial distribution of its attributes and soil classes.