Elvio Giasson

Decision trees for digital soil mapping on subtropical basaltic steeplands

When soil surveys are not available for land use planning activities, digital soil mapping techniques can be of assistance. Soil surveyors can process spatial information faster, to assist in the execution of traditional soil survey or predict the occurrence of soil classes across landscapes. Decision tree techniques were evaluated as tools for predicting the ocurrence of soil classes in basaltic steeplands in South Brazil. Several combinations of types of decicion tree algorithms and number of elements on terminal nodes of trees were compared using soil maps with both original and simplified legends. In general, decision tree analysis was useful for predicting occurrence of soil mapping units. Decision trees with fewer elements on terminal nodes yield higher accuracies, and legend simplification (aggregation) reduced the precision of predictions. Algorithm J48 had better performance than BF Tree, RepTree, Random Tree, and Simple Chart.

Digital soil mapping using multiple logistic regression on terrain parameters in southern Brazil

Soil surveys are necessary sources of information for land use planning, but they are not always available. This study proposes the use of multiple logistic regressions on the prediction of occurrence of soil types based on reference areas. From a digitalized soil map and terrain parameters derived from the digital elevation model in ArcView environment, several sets of multiple logistic regressions were defined using statistical software Minitab, establishing relationship between explanatory terrain variables and soil types, using either the original legend or a simplified legend, and using or not stratification of the study area by drainage classes. Terrain parameters, such as elevation, distance to stream, flow accumulation, and topographic wetness index, were the variables that best explained soil distribution. Stratification by drainage classes did not have significant effect. Simplification of the original legend increased the accuracy of the method on predicting soil distribution.

Tillage Effects on Particulate and Mineral‐Associated Organic Matter in Two Tropical Brazilian Soils

Abstract Two Ferralsols (350 and 600 g kg−1 clay) from the Brazilian Cerrado Region were evaluated for long‐term effects (5 and 8 years) of no tillage on carbon (C) stocks in particulate (>53 µm) and mineral‐associated (<53 µm) soil organic matter (SOM) fractions. Carbon stocks in particulate SOM increased under no tillage compared with conventional tillage, and the rate was higher in the clayey soil (0.62 Mg C ha−1 yr−1) than in the sandy clay loam soil (0.31 Mg C ha−1 yr−1). In contrast, the mineral‐associated SOM in the top soil layer (0–20 cm) was not affected by tillage system. Sequestration of atmospheric C in tropical no‐tillage soils seems to be due to accumulation of C in labile SOM fractions, with highest rates in clayey soils probably due to physical protection.

Uso de regressões logísticas múltiplas para mapeamento digital de solos no Planalto Médio do RS

Logistic nominal regressions establish mathematical relations between continuous or discrete independent variables and discrete dependent variables. The prediction potential of the occurrence and distribution of soil classes in the region Ibiruba and Quinze de Novembro, RS, Brazil was evaluated. Using a digital elevation model (DEM) with 90 m resolution, were calculated several topographic characteristics (elevation, slope, and curvature) and hydrographic variables (distance to rivers, flow length, topographical wetness index, and stream power index). Multiple logistic regressions were established between the soil classes mapped on the basis of a traditional survey at a scale of 1:80.000 and the land variables calculated using the DEM. The regressions were used to calculate the probability of occurrence of each soil class. The final estimated soil map was drawn by assigning the soil class with highest probability of occurrence to each cell. The general accuracy was evaluated at 58 % and the Kappa coefficient at 38 % in a comparison of the original soil map with the map estimated at the original scale. A legend simplification had little effect to increase the general accuracy of the map (general accuracy of 61 % and Kappa coefficient of 39 %). It was concluded that multiple logistic regressions have a predictive potential as tool of supervised soil mapping.

Comparação de métodos para mapeamento digital de solos com utilização de sistema de informação geográfica

Soil maps are sources of important information for land planning and management, but are expensive to produce. This paper proposes testing and comparing single stage classification methods (Multiple Multinomial Logistic Regression and Bayes) and multiple stage classification methods (Classification and Regression Trees (CART), J48 and Logistic Model Trees (LMT)) using geographic information system and terrain parameters for producing soil maps with both original and simplified legend. The database was managed in ArcGis computer application in which the variables and the original map were related through training of the algorithms. The results from statistical software Weka were implemented in ArcGis environment to generate digital soil maps. The terrain parameters that best explained soil distribution were slope, profile and planar curvature, elevation, and topographic wetness index. The multiple stage classification methods showed small improvements in overall accuracies and large improvements in the Kappa index. Simplification of the original legend significantly increased the producer and user accuracies, however produced small improvements in overall accuracies and Kappa index.

Digital Soil Mapping Using Logistic Regression on Terrain Parameters for Several Ecological Regions in Southern Brazil

As the relationship between soils and landscape within the context of soil formation is well known, predictive relationships between soils and soil formation factors can be established by regression techniques, relating soil and terrain attributes to occurrence of soil classes. This study proposes the production of maps using logistic regression on soil and terrain information from a pilot area to reproduce the original map and predict soil distribution in other similar landscapes in three study areas (Ibibuba Municipality, Sentinela do Sul Municipality, and Arroio Portao Watershed) in map scales from 1:30,000 to 1:50,000 and located in three ecological regions in Southern Brazil (Planalto, Encosta da Serra do Sudeste, and Depressao Central, respectively). By using logistic regressions for digital soil mapping, the method predicts the occurrence of soil units based on reference soil maps (produced by conventional methods), and on several parameters derived from a USGS SDTS-SRTM DEM, namely slope gradient, profile curvature, planar curvature, curvature, flow direction, flow accumulation, flow length, Stream Power Index (SPI), and Topographic Wetness Index (TWI). Results show that parameters such as elevation, curvature, SPI, TWI, and distance to streams are more frequently selected as parameters for predicting the occurrence of soil classes, with overall percent correct from 61% to 71%, and Kappa Index from 36% to 54% when the maps produced are compared with the original soil maps with a simplified legend (which simulate the production of soil maps with smaller scales that the original soil map). The prediction of soil map units using logistic regressions generated reliable soil maps, and the method appears to deserve more research effort, given the reliability and low cost of the resulting information.

Utilização de P-Index em uma bacia hidrográfica através de técnicas de geoprocessamento

Due to problems caused by eutrophication of lakes and reservoirs, the identification of phosphorus source areas is important for planning the control of agriculture-related water pollution. This study applied a method designed to identify these phosphorus source areas, called P-Index, which is based on combining data layers related to soil phosphorus availability and phosphorus transport processes. The method was used for studying the Salto Reservoir waterbasin, in Sao Francisco de Paula, in the State of Rio Grande do Sul, Brazil. Data related to phosphorus availability, intensity of transport processes, and distance to waterways were combined to yield a map of phosphorus contribution classes, which showed that almost the entire study area was considered to have low phosphorus contribution potential. A few small areas, corresponding to potato and garlic fields that received heavy chemical fertilization, were classified as having high phosphorus contribution potentials.

Prediction of soil orders with high spatial resolution: response of different classifiers to sampling density

O objetivo deste trabalho foi avaliar a densidade de amostragem na acuracia de predicao de ordens de solos, com alta resolucao espacial, em area viticola da Serra Gaucha. Para isso, utilizou-se modelo digital de elevacao (MDE) do terreno, base cartografica, mapa convencional de solos e o programa Idrisi. Sete variaveis preditoras foram calculadas e lidas junto com as classes de solo, em pontos aleatoriamente distribuidos, nas densidades de 0,5, 1, 1,5, 2 e 4 pontos por hectare. Os dados foram usados para treinar uma arvore de decisao (Gini) e tres redes neurais artificiais: teoria da ressonância adaptativa, fuzzy ARTMap; mapa auto‑organizavel, SOM; e perceptron de multiplas camadas, MLP. Os mapas estimados foram comparados com o mapa de solos convencional para calcular erros de omissao e de inclusao, exatidao geral, e erros de quantidade e de alocacao. A arvore de decisao foi menos sensivel a densidade de amostragem e apresentou maior acuracia e consistencia. O SOM foi a rede neural com menor sensibilidade e maior consistencia. O MLP apresentou minimo critico e maior inconsistencia, enquanto fuzzy ARTMap apresentou maior sensibilidade e menor acuracia. Os resultados indicam que densidades de amostragem usadas em levantamentos convencionais podem servir de referencia para estimar ordens de solos na Serra Gaucha.

Avaliação de cinco algoritmos de árvores de decisão e três tipos de modelos digitais de elevação para mapeamento digital de solos a nível semidetalhado na Bacia do Lageado Grande, RS, Brasil

O mapeamento digital de solos (MDS) tem como base a geracao de sistemas de informacoes que permitem estabelecer relacoes matematicas entre variaveis ambientais e solos e, dessa forma, predizer a distribuicao espacial das classes ou propriedades dos solos. Dentre as abordagens mais utilizadas, as arvores de decisao tem se destacado por apresentar bons resultados no MDS. Por outro lado, dada a disponibilidade de novas fontes de informacao sobre a elevacao, torna-se necessario o teste e avaliacao de modelos digitais de elevacao (MDE) quanto ao seu uso para o MDS. Este estudo testa cinco algoritmos de arvores de decisao (Simple Chart, Random Tree, REP Tree, BF Tree e J48) e tres MDE (Aster GDEM, SRTM e SRTM V3) para o MDS a nivel semidetalhado, em situacoes em que o principal fator diferenciador entre os tipos de solo e o relevo. O uso do MDE Aster GDEM e arvore de decisao com algoritmo J48, Simple Tree e BF Tree foram os que produziram modelos de arvore de decisao capazes de produzir mapas de solo com maior similaridade ao mapa de referencia.

Ação dos térmitas no solo

ABSTRACT The order Isoptera is well known by its potential asa plague, although the number of species that are plagues issmall within the group. Termites are the dominant invertebratesin tropical terrestrial environments and are spread from tropicalrainforests to the savannahs, being found even in arid regions,in various habitats. These insects have a major role and arestill little studied in tropical ecosystems. Nutrient cycling,aeration, water infiltration of soil, bioturbation, aggregatesformation and organic material decomposition, are processesinfluenced by the action of termites, which , directly or indirectly,affect soil and landscape formation wherever they are. Wesuggest that a better approach must be addressed in futureresearches about these insects influence in the soil underspecified conditions of use and management, in sustainablefood production and climate changes . Key words : Isoptera, termite activity, soil biology, bioturbation,plague-termites . INTRODUCAO A importância ecologica dos isopteros emecossistemas tropicais e grande, principalmente quandoconsideradas as modificacoes que podem causar noambiente, desde alteracoes no aspecto visual dapaisagem, pela construcao de seus ninhos, alem dealteracoes da topografia, como dos microrrelevos demurundus, ate modificacoes nas propriedades fisicase quimicas do solo, efeitos no processo dedecomposicao, ciclagem de nutrientes, entre outros(HOLT & LEPAGE, 2000). Os cupins (termitas) sao amais importante fauna do solo nos tropicos quentessazonalmente secos (LOBRY DE BRUYN &CONACHER, 1990), sendo fundamentais para ofuncionamento do ecossistema, pois ocupam niveistroficos na cadeia alimentar do solo (SILVA et al., 2007).Os termitas sao reconhecidos como “engenheiros doecossistema” (DANGERFIELD et al., 1998), devido ahabilidade que possuem de modificar a estrutura do