Australia-wide predictions of soil properties using decision trees

Henderson, Brent; Bui, Elisabeth N.; Moran, Chris; Simon, D.

doi:10.1016/j.geoderma.2004.06.007

Cited by 217 publications

(120 citation statements)

References 37 publications

Supporting

Mentioning

101

Contrasting

Unclassified

Order By: Relevance

“…3). Em relação à variabilidade geral dos valores de densidade do solo verificou-se que para ambos os horizontes diagnóstico o ambiente geomórfico de sopé de transporte foi aquele que apresentou maiores intervalos de confiança, indicando relações intimas entre o comportamento deste variável e as condições da paisagem (HENDERSON et al, 2005).…”

Section: Resultsunclassified

See 1 more Smart Citation

Segmentos de vertente e atributos do solo de uma toposseqüência na região de Manicoré, AM

Campos¹,

Ribeiro²,

Júnior³

et al. 2010

Rev. Ciênc. Agron.

View full text Add to dashboard Cite

Resumo -As variações dos atributos do solo ocorrem em função de vários fatores tais como: clima, material de origem, vegetação e, especialmente do relevo, que regula o escoamento superficial e a drenagem, e tem influência no tempo de exposição dos materiais à ação do intemperismo. O objetivo deste trabalho foi estudar a influência dos diferentes segmentos de vertente nos atributos do solo em uma toposseqüência na região de Manicoré, AM. Foi estabelecido um caminhamento de 3.000 m a partir do "espigão" da vertente no sentido do caimento mais suave do relevo e os segmentos da vertente foram determinados com base na "quebra" de declive do terreno. Foram coletadas 20 amostras nos diferentes segmentos de vertentes identificados: topo alto: 0,0-0,19 e 0,37-0,60 m; meia encosta: 0,0-0,28 e 0,60-0,80 m; sopé de transporte: 0,0-0,15 e 0,42-0,71 m e topo baixo: 0,0-0,15 e 0,20-0,40 m. O critério de escolha das profundidades foi à coincidência com os horizontes diagnósticos superficiais e subsuperficiais determinados na descrição morfológica dos perfis. Foram realizadas análises físicas de textura, densidades do solo e das partículas, porosidade total e condutividade hidráulica, Nas análises químicas foram determinados pH, em água e em KCl, Ca, Mg, K, e Al trocáveis, P disponível, H+Al e C orgânico. Com o uso das técnicas estatísticas multivariadas foi possível distinguir diferentes ambientes geomórficos independentes dos segmentos de vertente identificados. As variações do relevo favoreceram a presença de solos distróficos no topo alto e solos eutróficos no sopé de transporte. Palavras-chave -Formas do relevo. Manicoré. Pedogênese.Abstract -Variations of soil properties are consequence of several factors such as: climate, parent material, vegetation and, particularly relief, which controls runoff and drainage, and has influence in the exposure of parent materials to the weathering action. The objective of this work was to study the influence of the slope segments on soil attributes of a topossenquence in the Manicoré region, AM. Twenty samples were collected in each of the several slope segments identified along a 3 km transect, from the top and toward the lower part of the topography. The slope segments and the sampling depths were: high top (0.0-0.19 and 0.37-0.60 m), medium slope (0.0-0.28 and 0.60-0.80 m); lower slope (0.0-0.15 and 0.42-0.71 m) and low top (0.0-0.15 and 0.20-0.40 m). The sampling depths were coincident with the surface and subsurface diagnostic horizons defined in the profile morphological description. Physical analysis involved particle size distribution, soil and particle density, total porosity and saturated hydraulic conductivity. In the chemical analysis pH in water and KCl, exchangeable cations, exchangeable Al, available P, H+Al, organic carbon were determined. The use of multivariate statistic techniques enabled the distinction of different soil geomorphic environments independent of the slope segments identifications. The relief variations contributed to the presence of dystrophic soils in th...

show abstract

Section: Resultsunclassified

“…Nesse sentido, vários estudos foram desenvolvidos buscando esclarecer e constatar as relações entre os atributos do solo e as condições da paisagem (BARTHOLD et al, 2008;HENDERSON et al, 2005).…”

Section: Introductionunclassified

Segmentos de vertente e atributos do solo de uma toposseqüência na região de Manicoré, AM

Campos¹,

Ribeiro²,

Júnior³

et al. 2010

Rev. Ciênc. Agron.

View full text Add to dashboard Cite

show abstract

“…Ongoing research in digital soil mapping has demonstrated that reasonably accurate soil maps can be produced using quantitative predictive models. Digital soil mapping may also expedite soil survey (Lagacherie and Holmes, 1997;Dobos et al, 2000;Zhu, 2000;Moran and Bui, 2002;McBratney et al, 2003;Scull et al, 2003;Cole, 2004;Shi et al, 2004;Henderson et al, 2005;Saunders, 2005;Scull et al, 2005;Cole and Boettinger, 2007;Saunders and Boettinger, 2007;Brungard, 2009).…”

Section: Introductionmentioning

confidence: 99%

Random Forests Applied as a Soil Spatial Predictive Model in Arid Utah

Stum

2010

Digital Soil Mapping

View full text Add to dashboard Cite

Initial soil surveys are incomplete for large tracts of public land in the western USA. Digital soil mapping offers a quantitative approach as an alternative to traditional soil mapping. I sought to predict soil classes across an arid to semiarid watershed of western Utah by applying random forests (RF) and using environmental covariates derived from Landsat 7 Enhanced Thematic Mapper Plus (ETM+) and digital elevation models (DEM). Random forests are similar to classification and regression trees (CART).However, RF is doubly random. Many (e.g., 500) weak trees are grown (trained) independently because each tree is trained with a new randomly selected bootstrap sample, and a random subset of variables is used to split each node. To train and validate the RF trees, 561 soil descriptions were made in the field. An additional 111 points were added by case-based reasoning using aerial photo interpretation. As RF makes classification decisions from the mode of many independently grown trees, model iii uncertainty can be derived. The overall out of the bag (OOB) error was lower without weighting of classes; weighting increased the overall OOB error and the resulting output did not reflect soil-landscape relationships observed in the field. The final RF model had an OOB error of 55.2% and predicted soils on landforms consistent with soil-landscape relationships. The OOB error for individual classes typically decreased with increasing class size. In addition to the final classification, I determined the second and third most likely classification, model confidence, and the hypothetical extent of individual classes.Pixels that had high possibility of belonging to multiple soil classes were aggregated using a minimum confidence value based on limiting soil features, which is an effective and objective method of determining membership in soil map unit associations and complexes mapped at the 1:24,000 scale. Variables derived from both DEM and Landsat 7 ETM+ sources were important for predicting soil classes based on Gini and standard measures of variable importance and OOB errors from groves grown with exclusively DEM-or Landsat-derived data. Random forests was a powerful predictor of soil classes and produced outputs that facilitated further understanding of soil-landscape relationships.(147 pages) iv ACKNOWLEDGMENTS

show abstract

“…Each of the resulting new nodes is taken and the process is repeated continuing the recursion until reaching certain predefi ned stopping criterion. CART is easy to interpret and discuss, when a mix of continuous and categorical type environmental parameters are used as predictors, furthermore, they have excellent predictive capabilities (Breiman, L. 2001;Lawrence, R. et al 2004;Henderson, B.L. et al 2005).…”

Section: The Applied Geostatistical and Data Mining Toolsmentioning

confidence: 99%

Compilation of novel and renewed, goal oriented digital soil maps using geostatistical and data mining tools

et al. 2015

View full text Add to dashboard Cite

Due to former soil surveys and mapping activities signifi cant amount of soil information has accumulated in Hungary. Present soil data requirements are mainly fulfi lled with these available datasets either by their direct usage or aft er certain specifi c and generally fortuitous, thematic and/or spatial inference. Due to the more and more frequently emerging discrepancies between the available and the expected data, there might be notable imperfection as for the accuracy and reliability of the delivered products. With a recently started project we would like to signifi cantly extend the potential, how soil information requirements could be satisfi ed in Hungary. We started to compile digital soil maps, which fulfi l optimally the national and international demands from points of view of thematic, spatial and temporal accuracy. In addition to the auxiliary, spatial data themes related to soil forming factors and/or to indicative environmental elements we heavily lean on the various national soil databases. The set of the applied digital soil mapping techniques is gradually broadened incorporating and eventually integrating geostatistical, data mining and GIS tools. Regression kriging has been used for the spatial inference of certain quantitative data, like particle size distribution components, rootable depth and organic matt er content. Classifi cation and regression trees were applied for the understanding of the soil-landscape models involved in existing soil maps, and for the post-formalization of survey/compilation rules. The relationships identifi ed and expressed in decision rules made the compilation of spatially refi ned category-type soil maps (like genetic soil type and soil productivity maps) possible with the aid of high resolution environmental auxiliary variables. In our paper, we give a short introduction to soil mapping and information management concentrating on the driving forces for the renewal of soil spatial data infrastructure provided by the framework of Digital Soil Mapping. The fi rst results of DOSoReMI.hu (Digital, Optimized, Soil Related Maps and Information in Hungary) project are presented in the form of brand new national and regional soil maps.

show abstract

Australia-wide predictions of soil properties using decision trees

Cited by 217 publications

References 37 publications

Segmentos de vertente e atributos do solo de uma toposseqüência na região de Manicoré, AM

Segmentos de vertente e atributos do solo de uma toposseqüência na região de Manicoré, AM

Random Forests Applied as a Soil Spatial Predictive Model in Arid Utah

Compilation of novel and renewed, goal oriented digital soil maps using geostatistical and data mining tools

Contact Info

Product

Resources

About