ABSTRACT: Little is known about the groundwater recharge potential of weathered tropical catchments, where increasing water uptake is widespread to meet various water demands. This study aimed to estimate the volume of groundwater recharge of the Doce River Basin, Minas Gerais, Brazil. The BALSEQ model was applied to calculate the water balance over a period of two years (2007-2009). Evapotranspiration, runoff, and potential groundwater recharge (PGR) were calculated, using daily data on rainfall, potential evapotranspiration, and plant-available water. A soil survey was undertaken for all major soils occurring in the basin. Soils samples were used to determine hydraulic conductivity, bulk density, and water content at field capacity and at the permanent wilting point. Vegetation data were obtained from the literature and used to determine the following parameters: canopy interception, crop coefficient, and root depth. The estimated groundwater was spatially predicted using the Random Forests model with digital elevation, vegetation index, pedological, lithological, and climate maps. During the two years, an average of 32 % of rainfall was converted to groundwater. Annually, the percent of rainwater converted to groundwater varied between 27 and 48 % for all soil classes, highlighting the great temporal variability. The spatial prediction showed a volume of approximately 17,484 and 35,410 m3 of rainfall being converted to groundwater for the first and second year, respectively. The BALSEQ model showed a feasibility for the water balance calculation and can be reapplied for updating the groundwater maps of the Doce River Basin. These maps could then be used to guide land use planning programs, with the aim to protect water resources.

ABSTRACT: Increasingly, applications of machine learning techniques for digital soil mapping (DSM) are being used for different soil mapping purposes. Considering the variety of models available, it is important to know their performance in relation to soil data and environmental variables involved in soil mapping. This paper investigated the performance of eight machine learning algorithms for soil mapping in a tropical mountainous area of an official rural settlement in the Zona da Mata region in Brazil. Morphometric maps generated from a digital elevation model, together with Landsat-8 satellite imagery, and climatic maps, were among the set of covariates to be selected by the Recursive Feature Elimination algorithm to predict soil types using machine learning algorithms. Mapping performance was assessed using the confusion matrix, and the Z-test among the Kappa indexes of the matrices. In a conventional soil survey, the soils described and classified in the Brazilian System of Soil Classification [Argissolos Vermelho-Amarelos Distróficos – PVAd (Acrisols), Cambissolos Háplicos Tb Distróficos - CXbd (Cambisols), Gleissolos Háplicos Háplicos Tb Distróficos - GXbd (Gleysols), Latossolos Amarelos Distróficos - LAd (Xanthic Ferralsos), Latossolos Vermelho-Amarelos Distróficos - LVAd (Rhodic Ferralsols), and Neossolos Litólicos Distróficos - RLd (Neossols)] were grouped into composite mapping units (MU) using the conventional method. The eight algorithms showed similar performance without statistical difference (Kappa 0.42-0.48). The mapping of soils with varying slopes (LAd, LVAd, CXbd) showed lower accuracy, whereas soils on hydromorphic lowlands (GXbd) were classified more accurately. In map algebra, the result was rather satisfactory, with 63-67 % agreement between the conventional soil map and maps produced by machine learning. The areas with the largest disagreement in the DSM occurred in the LAd unit due to subtle color variation in the Latossolos mantle without a clear relation to any environmental variable, highlighting difficulties in DSM regarding hill slope landforms. Model performance was satisfactory, and good agreement with the conventional soil map demonstrates the importance of the DSM as a potential complementary tool for assisting soil mapping in mountainous areas in Brazil for the purpose of land use planning.

ABSTRACT Digital soil mapping (DSM) has been increasingly used to provide quick and accurate spatial information to support decision-makers in agricultural and environmental planning programs. In this study, we used a DSM approach to map soils in western Haiti and compare the performance of the Multinomial Logistic Regression (MLR) with Random Forest (RF) to classify the soils. The study area of 4,300 km2 is mostly composed of diverse limestone rocks, alluvial deposits, and, to a lesser extent, basalt. A soil survey was conducted whereby soils were described and classified at 258 sites. Soil samples were collected and subjected to physical and chemical analyses. Recursive Feature Elimination (RFE) was used to select the most important covariates from auxiliary data, such as climate, lithology, and morphometric properties to describe the soil-landscape relationship. Mapping performance was assessed by the Kappa index and overall accuracy derived from a confusion matrix generated using a 5-fold cross validation process. In addition, an external mapping validation was carried out using an independent soil dataset. Accordingly, the soil dataset was split into 80 % and 20 % for training and validation of the models, respectively. No significant statistical difference (Z = 0.56< |1.96|) was found between maps generated with both classifiers (Kappa index 0.45 for MLR and 0.42 for RF). Based on the Kappa values, the classification performance can be characterized as moderate for both algorithms. Surprisingly, the RF classifier outperformed MLR in the validation process (Kappa values of 0.55 and 0.33, respectively). These results suggest a higher generalization ability of RF. However, no significant statistical difference (Z = 1.83< |1.96|) was observed. The soil map derived from RF indicated the occurrence of Leptosols (48.5 %), Gleysols (19.6 %), Chernozems (8 %), and Fluvisols (6.6 %) in most of the study area. The DSM approaches proved suitable for mapping soils in western Haiti and could be used in other parts of the country, thereby closing information gaps with regard to Haitian soils.

ABSTRACT Soil bulk density (ρb) data are needed for a wide range of environmental studies. However, ρb is rarely reported in soil surveys. An alternative to obtain ρb for data-scarce regions, such as the Rio Doce basin in southeastern Brazil, is indirect estimation from less costly covariates using pedotransfer functions (PTF). This study primarily aims to develop region-specific PTFs for ρb using multiple linear regressions (MLR) and random forests (RF). Secondly, it assessed the accuracy of PTFs for data grouped into soil horizons and soil classes. For that purpose, we compared the performance of PTFs compiled from the literature with those developed here. Two groups of data were evaluated as covariates: 1) readily available soil properties and 2) maps derived from a digital elevation model and MODIS satellite imagery, jointly with lithological and pedological maps. The MLR model was applied step-wise to select significant predictors and its accuracy assessed by means of cross-validation. The PTFs developed using all data estimated ρb from soil properties by MLR and RF, with R2 of 0.41 and 0.51, respectively. Alternatively, using environmental covariates, RF predicted ρb with R2 of 0.41. Grouping criteria did not lead to a significant increase in the estimates of ρb. The accuracy of the ‘regional’ PTFs developed for this study was greater than that found with the ‘compiled’ PTFs. The best PTF will be firstly used to assess soil carbon stocks and changes in the Rio Doce basin.

RESUMO Objetivou-se com este estudo usar geotecnologias para mapear a cobertura do solo e diagnosticar conflitos de uso da terra em áreas de preservação permanente (APPs) de uma microbacia de 17.759 ha, localizada no município de Alta Floresta - MT. A delimitação das APPs teve como base a Lei Complementar nº 412 do estado de Mato Grosso, de 13 de Dezembro de 2010. Para tanto, utilizou-se mapa de hidrografia e modelo digital de elevação, para a delimitação da bacia e nessa as áreas de preservação. O mapeamento da cobertura do solo foi realizado por classificação supervisionada pelo algoritmo MaxVer no softwareENVI 4.2, em uma imagem orbital de alta resolução espacial, do sensor SPOT 5, adquirida no ano de 2008. O processamento dos dados, ajuste das projeções, criação dos arquivos vetoriais e cruzamento entre os mapas de cobertura do solo e áreas de preservação foram realizados no software ArcGIS 9.2. As classes de cobertura do solo mapeadas foram: floresta, pastagem, água, solo exposto e ocupação urbana. A área total de APPs (1.966,51 ha) representa 11,1 % da área da microbacia, dos quais a maior parte de uso indevido ocorre com pastagem (47,7 %); áreas de floresta representam 39,5 % do total das APPs. As classes solo exposto e ocupação urbana representam respectivamente, 26,5 e 110,1 ha, equivalendo 6,9 % da área. Concluiu-se que a imagem utilizada mostrou-se satisfatória para o mapeamento das classes de cobertura do solo na área estudada. O mapeamento e cruzamento das informações obtidas mostraram que um total 1.074,92 ha está com uso conflitante com a legislação ambiental vigente, pois apresenta alguma forma de ação antrópica. As áreas com uso indevido cobre 54,6 % das APPs e são representadas pelas fisionomias de pastagem, solo exposto e ocupação urbana.

ABSTRACT The aim of this study is the use of geotechnology in mapping the soil cover and to assess conflicts of land use in permanent preservation areas (PPA) of a microbasin with an area of 17,759 ha located in the municipality of Alta Floresta, in Mato Grosso State. The demarcation of PPA was based on the Complementary Law No. 412 of the state of Mato Grosso, December 13, 2010. To carry out the work, it were used hydrographic map and digital elevation model to delimit the microbasin and the preservation areas. The mapping of land cover was carried out by means of supervised classification with the algorithm MaxVer using Envi 4.2 software, in a satellite image of high spatial resolution from SPOT 5 sensor, acquired in 2008. The data processing, adjustment of the projections, file creation and crossing of vector-maps of land cover and preservations areas were done in ArcGIS 9.2. The land cover classes mapped are the following: forest, pasture, water, bare soil and urban settlement. The total area of PPA (1,966.51 ha) represents 11.1 % of the microbasin area, of which most of misuse occurs with pasture (47.7 %), forest areas represent 39.5 % of total PPA. The class of exposed soil and urban occupation represent respectively 26.5 and 110.1 ha, equivalent 6.9 % of the preservation area. It could be concluded that, the image used, was found satisfactory for mapping the land cover classes on the study area. The mapping and crossing of information showed that an area of 1,074.92 ha has conflict of use according to the current environmental legislation, as there are any forms of human activity. The areas with improper use account for 54.6 % of the total PPA and are represented by pasture, bare soil and 54.6 % of the total PPA and are represented by pasture, bare soil and urban settlement.

O objetivo deste trabalho foi avaliar a cokrigagem ordinária, com dados de altitude e distância do mar, em comparação à krigagem ordinária, na espacialização da precipitação pluvial dos períodos anual, seco e úmido, no Estado do Espírito Santo. Os dados de altitude e distância do mar foram obtidos de pontos de amostragem em grades regulares e irregulares. Foram utilizados dados de 108 postos pluviométricos. A avaliação dos métodos e das variáveis foi realizada com base na validação cruzada, tendo-se considerado os erros dos valores preditos e o ajuste dos modelos de regressão linear para valores observados e preditos. A grade regular para amostragem das covariáveis apresentou melhor acurácia de predição em comparação à grade irregular. A cokrigagem produziu resultados mais acurados do que a krigagem, verificados por pequenas diferenças nos erros médios absolutos, capazes de produzir mapas estatisticamente diferentes. A interpolação por cokrigagem e o uso de grades regulares para amostragem são preferíveis, principalmente se as covariáveis são de fácil obtenção e de baixo custo.

The objective of this work was to evaluate the ordinary co-kriging algorithm, using data from elevation and distance from the sea, compared to ordinary kriging, in the rainfall spatialization associated to wet, dry and annual periods, for the state of Espírito Santo, Brazil. Data of altitude and distance from the sea were obtained in sampling sites on regular and irregular grids. Data from 108 rain gauge stations were used. The evaluation of methods and variables based on cross-validation was performed, considering the errors of the predicted values and the fit of linear regression models for observed and predicted values. The regular grid sampling of the covariates showed the best prediction accuracy compared to the irregular grid. Cokriging produced more accurate results than kriging, checked by small differences in mean absolute errors which were able to produce statistically different maps. Cokriging interpolation and use of regular grids for sampling are preferable, mainly if the covariates are easy to obtain and inexpensive.

Neste trabalho, determinaram-se as características morfométricas do relevo e da rede de drenagem da Bacia Hidrográfica do Rio Preto, com a finalidade de prever do comportamento hidrológico da área. A partir da delimitação da área da bacia, da digitalização das curvas de nível e do sistema de drenagem, foram obtidas características geométricas do relevo e da hidrografia da BHRP e das suas seis sub-bacias. Concluiu-se que a área apresentava forma alongada e estreita, aliada a uma alta densidade de drenagem, evidenciando menor possibilidade de ocorrência de cheias. Porém, foram observadas enchentes justificadas pelos solos rasos, pelos intensos processos de desmatamento e uso da terra e pelo relevo acidentado, sendo 73,1% da área caracterizada como de relevo forte-ondulado a montanhoso. Quanto às sub-bacias, verificou-se que em todas essas unidades menos de 28% da área era compreendida de relevo plano, facilitando a erosão hídrica.

It was determined in this work the morphometric characteristics of relief and drainage system of Rio Preto Watershed (RPW) to predict the area hydrological behavior. From delimitation of the watershed area, digitalization of the contour lines and from the drainage system, it was obtained the geometrical traits of the relief and RPW hydrography its six main sub-watersheds. It was concluded that the area presented elongated and narrow shape together with a high drainage density, evidencing lower risk of flood occurrence. On the other hand, it was observed floods justified by shallow soils caused by the intense process of deforestation and land use and by the rugged relief, with 73.1% of the area characterized from rugged relief to mountainous. Regarding the sub-watersheds, it was observed that less than 28% of the area was comprised by flat relief, facilitating water erosion.

A atividade da avifauna nos ecossistemas terrestres imprime profundas alterações químicas e mineralógicas nos substratos minerais. O objetivo do presente trabalho é investigar os impactos da atividade da avifauna na evolução da paisagem do arquipélago de Abrolhos através do estudo da distribuição espacial dos solos e suas características químicas, com ênfase na fosfatização. Foram estudados dez perfis de solos representativos das principais unidades de paisagem do arquipélago. Com base em observação de campo e interpretação de carta náutica da Marinha do Brasil (n. 1311), em escala 1:100.000, foi produzido o mapa de solos das ilhas. A presença generalizada de solos fortemente ornitogênicos, em áreas sem nidificação atual ou com fraca nidificação, indica uma ocupação pretérita por aves bem mais ampla que a atual. Alguns solos estudados em Abrolhos são de difícil classificação e refletem influência marcante da avifauna, atual ou pretérita.

The activity of the avifauna on terrestrial ecosystems results in the intense chemical and mineralogical alteration of mineral substrates. The objective of the present work is to investigate the impact of the avifauna on landscape evolution at the Abrolhos Archipelago, based on soils chemical characteristics and spatial distribution. A total of 10 soil profiles are described and analyzed, representing the main landscape units present in the Archipelago. Based on careful field observation and interpretation made on the existing nautical map (n.1311), produced by the Brazilian Navy (scale 1:100.000), detailed soil maps were produced. The widespread occurrence of strongly ornithogenic soils in areas without current nesting activity indicates a much broader occupation by birds in the past. Some of the studied soils are of difficult classification, and reflect the marked influence of present-day and past avifauna activity.