ABSTRACT Most Brazilian municipalities do not have regulated areas for solid waste disposal in civil construction. Usually, residues are disposed of vacant lots and dumps, posing risks to the population health and the environment. Soils are the primary means for the disposal or recycling of waste, highlighting the importance of well-characterized soils and their respective landscape. This study aimed to establish a land suitability system for solid residues in civil construction and apply such information in a case study in Southeastern Brazil. An unprecedented digital soil map with a resolution of 30 m was created using the random forest classifier algorithm and soil field prospection information. A guide listing favorable soil-landscape attributes that most prevent soil erosion, water bodies or water table contamination was elaborated and discussed. Thus, such information was linked through a suitability system to classify areas with potential for receiving waste on a daily volume basis as follows: large size: >500 m³ day-1, medium size: >100 m³ day-1 and <300 m³ day-1, and small size <100 m³ day-1. Topography and soil depth were the most limiting factors of the areas in the case study. The proposed attributes as criteria for the suitability system complement the current state legislation. A total of 236 ha closer to the urban perimeter connected by roads in good condition were classified as suitable for managing medium- and small-scale daily volume, whose destination might reduce transportation and installation costs in the study area. Usually dumps environment wellcharacterized well characterized landscape Brazil 3 soillandscape erosion discussed Thus follows 500 >50 day1, day1 day 1, 1 100 >10 day- 300 <30 <10 day1. 1. legislation 23 smallscale scale area 50 >5 10 >1 <3 <1 2 5 > <

RESUMO Mapas de solos existentes (dados legados) associados a técnicas de mapeamento digital são alternativas para obter informações a menores custos, entretanto, alguns testes são necessários para se fazer isso de forma mais eficiente. Este estudo objetivou comparar diferentes métodos para extrair informações de mapas de solos em escala detalhada usando árvores de decisão para mapear classes de solos em duas sub-bacias hidrográficas de Minas Gerais, validar os mapas em campo e usar o melhor método para extrapolar informações para áreas maiores, também validando estes mapas. Mapas de solos detalhados das sub-bacias do ribeirão Vista Bela (VBW) e ribeirão Marcela (MCW) foram utilizados como fonte de informação. Sete métodos para extrair informações de mapas foram comparados: polígono todo, eliminando 20 e 40 m das bordas dos polígonos, e com buffers em torno dos pontos amostrados com raios de 25 m, 50 m, 75 m e 100 m. O algoritmo Classification and Regression Trees (CART) foi utilizado para gerar árvores de decisão e permitir a criação de mapas de solos. A acurácia dos mapas foi avaliada através da acurácia global e índice Kappa. O melhor método foi utilizado para extrapolar informações para áreas maiores e esses mapas foram validados. Os melhores métodos para VCW e MCW foram, respectivamente, eliminando 20 m das bordas dos polígonos e buffer com raio de 25 m ao redor dos pontos. Mapas para áreas maiores foram obtidos usando esses métodos. Remoção de áreas de incerteza de mapas legados de solos contribuem para uma melhor modelagem e predição de classes de solos. A utilização de informações geradas neste trabalho permitiu a extrapolação validada de mapas de solos para regiões do entorno das sub-bacias hidrográficas.

ABSTRACT Existing soil maps (legacy data) associated with digital mapping techniques are alternatives to obtain information at lower costs, however, tests are required to do it more efficiently. This study had as objectives to compare different methods to extract information from detailed scale soil maps using decision trees for mapping soil classes at two watersheds in Minas Gerais, validate these maps in the field and use the best method to extrapolate information to larger areas, also validating these maps of larger areas. Detailed soil maps of Vista Bela creek (VBW) and Marcela creek (MCW) watersheds were used as source of information. Seven methods to extract information from maps were compared: the whole polygon, eliminating 20 and 40 m from the polygon boundaries, and with buffers around the sampled points with radii of 25 m, 50 m, 75 m, and 100 m. The Classification and Regression Trees (CART) algorithm was employed to create decision trees and enable creation of soil maps. Accuracy was assessed through overall accuracy and kappa index. The best method was used to extrapolate information to larger areas and maps were validated. The best methods for VCW and MCW were, respectively, eliminating 20 m from polygon edges and buffer of 25 m of radii from points. Maps for larger areas were obtained using these methods. Removing uncertainty areas from legacy soil maps contribute to better modeling and prediction of soil classes. Information generated in this work allowed for validated extrapolation of soil maps for regions surrounding the watersheds.