ABSTRACT Pedosphere-hydrosphere interface accounts for the association between soil hydrology and landscape, represented by topographic and Remote Sensing data support and integration. This study aimed to analyze different statistical radiometric and spectral data selection methods and dimensionality reduce environment-related data to support the classification of soil physical-hydric properties, such as soil basic infiltration rate (bir) and saturated hydraulic conductivity (Ksat); as well as to act in data mining processes applied to hydropedological properties digital mapping. Accordingly, research integrated information from Visible to Infrared (VIS-IR) spectral indices and Sentinel’s 2A mission Multispectral Instrument (MSI) sensor bands, terrain numerical modeling and aerogeophysics set to model soil-water content in two soil layers (0.00-0.20 m and 0.20-0.40 m). Pre-processed data were subjected to statistical analysis (multivariate and hypothesis tests); subsequently, the methods were applied (variation inflation factor - VIF, Stepwise Akaike information criterion – Stepwise AIC, and recursive feature elimination - RFE) to mine covariates used for Random Forest modeling. Based on the results, there were distinctions and singularities in spectral and radiometric data selection for each adopted method; the importance degree, and contribution of each one to soil physical-hydric properties have varied. According to the applied statistical metrics and decision-making criteria (highest R2 and lowest RMSE / MAE), the chosen methods were RFE (0.00-0.20 m layers) and Stepwise AIC (0.20-0.40 m layers) - both concerned with the assessed variables (bir and Ksat). This approach captured the importance of environmental variables and highlighted their potential use in hydropedological digital mapping at Guapi-Macacu watershed.

Abstract Natural events associated to environmental disasters has increased with climate changes. Understanding the watershed behavior allows the managers to execute an efficient land use planning. By using as a study area the municipality of São Luiz do Paraitinga, the study’s goal was apply the Height Above the Nearest Drainage Model, which allows categorizing areas based on simulations of water level variations, to evaluate flooding risks at the municipality. The data were processed using ArcGIS Desktop v. 10.3, System for Automated Geoscientific Analysis and TerraHidro. The flood susceptibility map was generated with spatial resolution of 30 m. It was simulated water level variations of 7, 9, 12 and 15 meters and, according to the model, areas with high or very high flood susceptibility cover approximately 13% of the study area (81 km2). In general, the methods used afforded coherent results given the resolution of source data and available information.

ABSTRACT Notwithstanding the importance of soil surveys, advances in digital soil mapping have mainly focused on mapping soil attributes or properties rather than developing digital maps of soil units or soil classes. The purpose of this research was to develop digital soil unit maps based on primary soil data collection in areas without previously collected soil information. The covariate variability, the random effect across the data subset and the map outputs were the focuses of this study. We used five datasets with four models (Random Forest - RF, Gradient Boosted Machine - GBM, C5.0, and multinomial log-linear model - MLR). The covariates were grouped into five datasets, where four were grouped by Region Of Interest per Class (ROIC) and one was not grouped by ROIC. To evaluate the random effect to split the dataset, we ran each model 50 times and observed the overall accuracy (OA) and kappa index, and uncertainty, majority and variety maps. The OA of Dataset01 to 04 was lower than to Dataset05 accuracy. However, map outputs of RF and GBM for Dataset01 and Dataset05 had the same majority prediction. It seems that RF and GBM produce consistent results in map outputs according to this methodology and pedologist expertise. To evaluate the uncertainty and the consistency of soil unit prediction, we used the majority maps process. Random Forest, similar to GBM, presented the best results. The increase in the number of covariates was not a guarantee of improvement in the OA or in the quality of the map output. Geographic position and distance raster did not improve the map output according to expert evaluation. Because the variance between the ROICs, when the training and validation datasets were split based on it, the subsets are quite different in relation to the covariates, and this is the reason for the worse results of this model, comparing with the Dataset05. On the other hand, when considering one complete dataset not based on ROICs, the variance of training and validation subsets is lower and produced more accurate parameters of quality.

ABSTRACT Soil maps are important to evaluate soil functions and support decision-making process, particularly for soil properties such as pH, carbon content (C), and cation exchange capacity (CEC), but the spatial resolution and soil depth should meet the needs of users. On another hand, the efficiency of statistical models to create soil maps, with an acceptable level of accuracy, often require a large number of samples with an appropriate distribution across the area of interest. However, accessibility for sampling can be a trouble in remote areas, such as the Itatiaia National Park (INP). The hypothesis of this work is that it is possible to obtain a viable result in soil mapping of areas with limited access by using DSM tools. The general objective of this paper was to create 2- and 3-D maps of the soil properties pH, carbon content, and CEC, with the correspondent spatial uncertainty, in the INP plateau. The sampling strategy was designed using conditioned Latin Hypercube Sample (cLHS), and different methods were tested to produce the soil properties maps. For calibration of the models: linear (MLR, multiple linear regression) and nonlinear (GAM, Generalised Additive Models). The results showed differences in predictive performance for all statistical methods and covariate selection approaches. The GAM, with covariates selection based on soil formation factors, was the best method for the limited number of soil samples. The greatest uncertainty was associated with areas with the lowest accessibility and, consequently, with low sampling density and/or noises in covariates. Even though the 2- and 3-D maps of soil properties, each associated with explicit uncertainty, can contribute to the INP decision makers/managers by providing information not available before.

ABSTRACT: Users of soil survey products are mostly interested in understanding how soil properties vary in space and time. The aim of digital soil mapping (DSM) is to represent the spatial variability of soil properties quantitatively to support decision-making. The goal of this study is to evaluate DSM techniques (Regression Trees - RT and Multiple Linear Regressions - MLR) and the ability of these tools to predict mineral fraction content under a wide variability of landscapes. The study site was the entire Guapi-Macacu watershed (1,250.78 km2) in the state of Rio de Janeiro in the Southeast region of Brazil. Terrain attributes and remote sensing data (with 30 m of spatial resolution) were used to represent landscape co-variables selected as an input in predictive models in order to develop the explanatory variables. The selection of sampling sites was based on the Latin Hypercube algorithm. A representative set of one hundred points with feasible field access was chosen. Different input databases were tested for prediction of mineral fraction content (harmonized and original data). The Spline algorithm was used to harmonize data according to the GlobalSoil. Net consortium standards. The results showed better performance from the RT models, using input from an average of six covariates; the simplest MLR model used twice as many input variables, creating more complex models without gaining precision. Furthermore, better R2 values were obtained using RT models, irrespective of harmonization of soil data. The harmonized dataset from the 0.00-0.05 and 0.05-0.15 m layers, in general, presented better results for the clay and silt, with R2 values of 0.52 (0.00-0.05 m) and 0.69 (0.05-0.15 m), respectively. Prediction of sand content showed better results when the original depth data was used as an input, although all regression tree models had R2 values greater than 0.52. The RT models provided a better statistical index than MLR for all predicted properties; however, the variance between models suggests similarity of performance. Regarding harmonization of soil data, both input databases (harmonized or not) can be used to predict soil properties, since the variance of model performance was low and generalization of the soil maps showed similar trends. The products obtained from the digital soil mapping approach make it possible to integrate the factor of uncertainties, providing easier interpretation for soil management and land use decisions.

ABSTRACT: Planning sustainable use of land resources requires reliable information about spatial distribution of soil physical and chemical properties related to environmental processes and ecosystemic functions. In this context, cation exchange capacity (CEC) is a fundamental soil quality indicator; however, it takes money and time to obtain this data. Although many studies have been conducted to spatially quantify soil properties on various scales and in different environments, not much is known about interactions between soil properties and environmental covariates in the Brazilian semiarid region. The goal of this study was to evaluate the efficiency of random forest and cokriging models applied to predict CEC in the Brazilian semiarid region. The covariates used to predict CEC consist of images from Landsat 5 TM and a legacy soil map (scale 1:10,000). The sample set comprises 499 samples from the topsoil layer (0.00-0.20 m), where 375 samples were used in training processes and 124 as validation samples. The cokriging model (R2 = 0.57 and RMSE = 7.22 cmolc kg-1) performed better in predicting CEC than the random forest model (R2 = 0.47 and RMSE = 7.89 cmolc kg-1). The approach used showed potential for estimating CEC content in the Brazilian semiarid region by using covariates obtained from orbital remote sensing and the legacy soil map.

Resumo O objetivo deste trabalho foi avaliar ferramentas pedométricas para caracterização da composição granulométrica (areia, silte e argila) e comparação de perfis de solos hidromórficos. O estudo foi realizado na bacia hidrográfica do rio Guapi-Macacu, RJ. Foram aplicados o algoritmo "slice-wise" para o fatiamento do perfil em intervalos de 1 cm, e a função "spline" para harmonizar o conjunto de dados segundo seis intervalos de profundidade predefinidos. A análise do conjunto de perfis, por meio das funções de profundidade (slice-wise e spline) e de dendrograma de dissimilaridade, revelou que Planossolos e Gleissolos são relativamente rasos e apresentam camadas arenosas nos horizontes superficiais e teores de argila e silte irregulares em subsuperficie. Essas duas ordens ocorrem na área da bacia hidrográfica geralmente em planícies aluviais, em vales e em sopés de elevação. Os Neossolos Flúvicos apresentaram maiores quantidades de argila e silte em superfície, as quais decrescem em profundidade, e são mais profundos e ocorrem nos terraços aluviais. A abordagem permitiu descrever a variabilidade da textura dos solos hidromórficos em perfil e realizar a análise de similaridade entre as ordens, o que pode apoiar a distinção de fases em unidades de mapeamento e a identificação de critérios quantitativos para a classificação dos solos.

Abstract The objective of this work was to evaluate pedometric tools to characterize the particle size distribution (sand, silt, and clay) and to compare the profiles of hydromorphic soils. The study was carried out in the Guapi-Macacu watershed, in the state of Rio de Janeiro, Brazil. The slice-wise algorithm was applied to slice the profile data at 1-cm intervals, and the equal-area spline function for database harmonization, according to six predefined depth intervals. The analysis of the soil profile collection, using the soil-depth functions (slice-wise and spline) and the similarity analysis, revealed that Alfisols and Entisols are relatively shallow and have coarse texture in the topsoil layer and a more irregular distribution of clay and silt along the soil profile. These two orders usually occur in the watershed area related to floodplains, valleys, and footslopes. Fluvisols had higher amounts of clay and silt in the topsoil layer, which decrease with soil depth, and they are deeper and occur in alluvial terraces. The approach allowed to characterize the variability of texture along hydromorphical soil profiles and to perform a similarity analysis between soil orders, which can support the differentiation of soil mapping units and the identification of quantitative criteria for soil classification.

Resumo O objetivo deste trabalho foi avaliar a influência da resolução espacial do modelo digital de elevação e da eficiência de modelos Random Forest sobre a predição dos teores de areia, argila e carbono orgânico, com uso de número reduzido de amostras. O trabalho foi realizado em área de Cerrado com diversidade litológica, no Estado do Mato Grosso do Sul, tendo-se utilizado atributos morfométricos, dados do sensor TM Landsat 5 e litologia como covariáveis preditoras. Dados da camada superficial (0,0-0,2 m) de 175 perfis de solos (0,009 perfis km-2) e de 26 covariáveis preditoras foram utilizados com resolução espacial de 30 (conjunto 1) e 90 m (conjunto 2). A análise realizada pelo Random Forest mostrou que as covariáveis de nível de base do canal de drenagem, da elevação e da litologia foram as mais importantes para explicar a variabilidade. A validação dos modelos apresentou similaridade entre os conjuntos quanto à predição de areia, argila e carbono orgânico, o que explica os seguintes valores de variabilidade espacial, respectivamente: 44, 40 e 33%, para a resolução de 30 m; e de 45, 46 e 33%, para a resolução de 90 m. A resolução espacial das covariáveis preditoras tem pouca influência sobre a predição dos atributos, e a abordagem por Random Forest apresenta potencial de utilização para estimar atributos do solo.

Abstract The objective of this work was to evaluate the effect of the digital elevation model spatial resolution and of the efficiency of Random Forest models on the prediction of sand, clay, and organic carbon contents, using few soil samples. The study was carried out in a Cerrado area with lithological diversity, in the state of Mato Grosso do Sul, Brazil, using morphometric attributes, TM Landsat 5 sensor data, and lithology as predictive covariates. The surface layer data (0.0-0.2 m) of 175 soil profiles (0,009 profiles km-2) and of 26 predictor covariates were used with 30 (set 1) and 90-m (set 2) spatial resolutions. The performed analysis by Random Forest models showed that channel base level, elevation, and lithology were the most important ones to explain the variability. The validation of the models showed similarity among sets for the prediction of sand, clay, and organic carbon contents, which explains the following values of spatial variability, respectively: 44, 40, and 33%, for the spatial resolution of 30 m; and 45, 46, and 33%, for the spatial resolution of 90 m. The spatial resolution of the predictive covariates has little effect on attribute predictions, and the Random Forest approach has potential use for estimating soil properties.

Resumo O objetivo deste trabalho foi o desenvolvimento de modelos com diferentes conjuntos de dados, para estimar a densidade de solos de regiões tropicais montanhosas, a partir de atributos de solos comumente encontrados nas análises de perfis de solos descritos nos levantamentos regionais. O conjunto total de dados compõe-se de 163 amostras e foi dividido em seis grupamentos, dos quais três com 73 amostras, com o máximo de 32 covariáveis, e três com 163 amostras, com o máximo de 18 covariáveis. Testaram-se modelos de regressão linear múltipla (RLM) e randomForest (RF). A menor incerteza entre os modelos foi alcançada pelo RLM2, com R2 de 0,56, 13 covariáveis e 73 amostras. Nos grupamentos com 163 amostras, os melhores modelos foram os RF, com R2 médio de 0,48. A raiz quadrada da média do erro ao quadrado variou entre 0,09 e 0,14. As covariáveis mais importantes no modelo RF foram: carbono orgânico, hidrogênio, areia fina e grossa, saturação por bases e capacidade de troca catiônica. Pelo método "stepwise regression", as variáveis mais importantes foram: a relação silte/argila; areia grossa e fina; carbono orgânico; saturação por bases; e potássio.

Abstract The objective of this work was the development of models with different sets of data for estimating soil bulk density in tropical mountainous regions, from soil attributes commonly found in the analyses of soil profiles described in regional surveys. The complete dataset is composed of 163 samples and it was divided into six groups, of which three groups have 73 samples and the maximum of 32 covariables, and three have 163 samples and the maximum of 18 covariables. The linear regression (RLM) and randomForest (RF) models were tested. The lowest uncertainty between the models was achieved by RLM2, with R2 of 0.56, 13 covariables, and 73 samples. Considering the groups with 163 samples, the best models were the RFs with mean R2 of 0.48. The root mean squared error ranged between 0.09 and 0.14. The most important covariables in the RF model were: organic carbon, hydrogen, fine and coarse sand, base saturation, and cation exchange capacity. By the stepwise backward regression, the main covariables were: silt and clay relation; fine and coarse sand; organic carbon; base saturation; and potassium.

RESUMO Este trabalho apresenta um modelo para determinar a fragilidade ambiental em bacias hidrográficas. O estudo foi realizado na Bacia do Rio Aldeia Velha, RJ, localizada na zona de contato e transição entre a baixada litorânea e o relevo montanhoso da Serra do Mar. Fatores que influenciam a ocorrência de processos erosivos foram integrados por algoritmos em um SIG para construção de classes de fragilidade. A análise multicriterial considerou o modelo numérico de terreno, dados oficiais sobre variáveis ambientais, imagem orbital de alta resolução e a opinião de especialistas. Através de informações secundárias sobre pedologia, intensidade das chuvas e declividade do terreno gerou-se o Mapa de Fragilidade Potencial (MFP). Através da combinação desse mapa com informações sobre uso e cobertura da terra obteve-se o Mapa da Fragilidade Emergente (MFE). Os resultados mostram que mais de 70% da área da bacia possui fragilidade ambiental considerada alta ou muito alta, tanto potencial como emergente. Os Processos Erosivos Aparentes (PEA) relacionaram-se positivamente com as áreas de alta fragilidade nos produtos cartográficos finais, destacando regiões mais propensas à intensificação de movimentos de massa e prioritárias para prevenção contra perda de solo. Os modelos geraram informações importantes para o planejamento territorial, possibilitando um zoneamento acessível e de fácil atualização para as prefeituras municipais e organizações da sociedade civil, inclusive para o monitoramento das áreas de alta fragilidade ambiental.

ABSTRACT This paper presents an approach to mapping the environmental fragility of river basins. The study was performed at the Aldeia Velha river basin, a rainforest watershed located between Rio de Janeiro’s coastal plains and the Serra do Mar highlands. A multiple-criteria analysis was performed involving factors that affect the risk of erosion; these variables were analyze using GIS tools and were integrated by algorithm in order to form a description of the different classes of environmental fragility in the basin. The multicriteria analysis considered the use of a numerical land model, official data, orbital imagery and the opinion of subject-matter experts. A Map of Potential Fragility (MPF) was generated through the collection of secondary information such as soil types, rainfall intensity and terrain slope. This initial map was later combined with a land-use projection to produce a Map of Emerging Fragility (MEF). The mapping results pointed to a highly fragile environment, where more than 70 percent of the basin’s area was classified with high or very high degree of fragility, in both the potential and emerging context. The Apparent Erosion Processes (AEP) were positively correlated in the areas with a higher level of fragility in the final map products, as well as highlighted the locations most prone to the intensification of erosion processes. The models generated important information for land-use planning, enabling affordable and easy upgrade for municipal governments and civil society organizations to monitor high fragility areas.

A amostragem é uma das etapas mais importantes dos levantamentos de solos. No entanto, os esquemas de amostragem utilizados nos levantamentos convencionais têm se evidenciado inadequados para o mapeamento digital de solos, pois podem comprometer os resultados e, além disso, não possibilitam a realização de análises estatísticas. Este estudo teve por objetivo avaliar o método de amostragem do hipercubo latino condicionado (cLHS, sigla em inglês), na presença de covariáveis ambientais (elevação, declividade, curvatura e mapa de uso e cobertura do solo), em comparação com a amostragem aleatória, na alocação de 100 pontos amostrais, buscando maior representatividade das características ambientais da bacia do rio Guapi-Macacu. O desempenho dos métodos foi avaliado pela análise qualitativa dos histogramas de frequência e das análises estatísticas pelos testes F, T de Student e Kolmogorov-Smirnov (K-S), para cada covariável. Os resultados apresentaram que os pontos selecionados pelo método cLHS possuíam distribuição geográfica mais adequada do que aqueles obtidos pela amostragem aleatória. Além disso, o método cLHS preservou mais a distribuição de frequência das covariáveis contínuas do que a amostragem aleatória; para covariável categórica uso e cobertura do solo os métodos foram equivalentes. Os testes estatísticos confirmaram o melhor desempenho do método cLHS, cujas amostras não diferiram estatisticamente da bacia. Entretanto, a amostragem aleatória apresentou diferença estatística para com a bacia, para todas as covariáveis contínuas para pelo menos um dos testes utilizados. Assim, o método cLHS pode ser considerado como um método satisfatório para seleção de locais de amostragem em áreas heterogêneas similares as deste estudo, visando a utilização no mapeamento digital de solos.

Soil sampling is one of the most important stages in soil surveys. However, in conventional soil surveys, sampling schemes have proven to be inadequate for digital soil mapping because they may affect the results obtained and, in addition, they do not allow statistical estimations to be performed. This study attempts to evaluate the conditioned Latin Hypercube Sampling (cLHS) method with four environmental covariates (elevation, slope, curvature and land cover/land use map), in comparison to random sampling, in allocating 100 sample points, seeking a more representative description of the environmental characteristics of the Macacu watershed area. The performance of the methods was evaluated by qualitative assessment of histograms of distribution, and statistical analyses by the F test, Student T Test and Kolmogorov-Smirnov test (K-S) for each covariable. The sampling points selected by the cLHS method showed better geographic distribution than those obtained by random sampling. Furthermore, the cLHS method better maintained the frequency distribution of the continuous covariables than random sampling did, and the categorical covariable of land cover/land use showed no significance differences between the cLHS and random methods. Statistical analysis confirmed the better performance of the cLHS method compared to the random method, and the cLHS samples showed no significance differences from the watershed data. In contrast, the random samples showed significance differences from the watershed data for all the continuous covariables for at least one of the statistical tests used. Thus, the cLHS method is a satisfactory method for selecting points for soil sampling in heterogeneous areas, like those of the present study, for the purpose of use in digital soil mapping.

O objetivo deste trabalho foi avaliar modelos digitais de elevação (MDE), obtidos por diferentes fontes de dados, e selecionar um deles para derivar variáveis morfométricas utilizadas em mapeamento digital de solos. O trabalho foi realizado na Bacia Guapi‑Macacu, RJ. Os dados primários utilizados nos modelos gerados por interpolação (MDE‑carta e MDE‑híbrido) foram: curvas de nível, drenagem, pontos cotados e dados de sensor remoto transformados em pontos. Utilizaram-se, na comparação, modelos obtidos por sensor remoto e por aerorrestituição (MDE SRTM e MDE IBGE). Todos os modelos apresentaram resolução espacial de 30 m. A avaliação dos modelos de elevação foi baseada na análise de: atributos derivados (declividade, aspecto e curvatura); depressões espúrias; comparação entre feições derivadas a partir dos modelos e as originais, oriundas de cartas planialtimétricas; e análise das bacias de contribuição derivadas. O modelo digital de elevação híbrido apresenta qualidade superior à dos demais modelos.

The objective of this work was to evaluate digital elevation models (DEM) obtained by different data sources and to select one of them for deriving morphometric variables used in digital soil mapping. The work was performed in the Guapi‑Macacu river basin, RJ, Brazil. The primary data used in the models generated by interpolation (DEM map and DEM hybrid) were: contour lines, drainage, elevation points, and remote sensor data transformed into points. The obtained models by remote sensing and aero‑restitution (DEM SRTM and DEM IBGE) were used in the comparison. All models showed spatial resolution of 30 m. The elevation model evaluations were based on: the terrain derived attribute analysis (slope, aspect, and curvature); spurious depressions (sink); comparison between features derived from the models and the original ones originated from planialtimetric maps; and the analysis of derived watersheds. The DEM hybrid showed a superior quality than the other models.