Multivariate Analysis and Machine Learning in Properties of Ultisols (Argissolos) of Brazilian Amazon

Abstract

Ultisols are the most common soil order in the Brazilian Amazon. The Legal Amazon (LA) has an area of 5 × 10 6 km 2 , with few accessible areas, which restricts studies of soils at a detailed level. The pedological properties can be estimated more efficiently using statistical procedures and machine learning techniques, tools which are capable of recognizing patterns in a large soil database. We analyzed the main chemical and physical properties of the B horizons of the Ultisols of the Brazilian Amazon, as well as the spatial variability of the most explanatory properties of these horizons. Physical and chemical data of 1,068 profiles of the RadamBrasil Project were used. A principal component analysis (PCA) was applied and the most explanatory variables were separated by morphostructural units and climate zones. The technique of machine learning was used for spatialization of the explanatory variables based on predictive covariates. In general, the horizons are thick, clay, with a predominance of negative charges, and low levels of exchangeable cations. The variables retained in the PCA were: sum of bases (SB), Al 3+ , degree of flocculation (Floc), ∆pH, and organic carbon content (C). Areas of greater precipitation have low SB, with higher values in the basement complex (BC) and in areas under the Andean influence. Higher levels of Al 3+ and degrees of flocculation were also associated with greater precipitation. However, the soils are predominantly electronegative, showing a kaolinitic mineralogy. The C contents in general were low, with an increase in more humid zones due to the process of mineralization and illuviation (podzolization), and in the BC due to the protection of C by the aggregation of clay. The use of multivariate analysis allowed a better understanding of the Ultisols' main properties in different morphostructural and climatic domains, and its spatialization facilitated the interpretation of properties and their relationships with environmental characteristics in the Legal Amazon.

Multicriteria Analysis and Machine Learning Algorithm for Definition of Areas for Micro-Dam, Southeastern Brazil

“…Therefore, the Random Forest (RF) selected seven covariates to determine potential areas for micro-dam and generated an R 2 of 0.43 in training and validation (Table 3; Table 4). Although there are no specific studies with the prediction of micro-dam by ML, R 2 values are consistent with other prediction studies (YESILNACAR E TOPAL, 2005;SOUZA et al, 2018;GOMES et al, 2019;ASSIS et al, 2021). Generally, the lower performance is due to the lack of covariates that can better explain the distribution of a variable (KUHN E JOHNSON, 2013).…”

“…This entire process was repeated 100 times, randomly changing the samples present in the training and validation at each repetition. The advantage of this excessive repetition is to avoid potentially biased predictions, as it evaluates with different sample groups (GRANITTO et al, 2006; KUHN E JOHNSON, 2013;SOUZA et al, 2018;GOMES et al, 2019). Besides, at each step (100 times), this methodological framework provides statistical data indicating the accuracy and error of the prediction: R-square (R 2…”

“…In addition, the findcorrelation function previously discarded several high-correlation covariates. Therefore, using a smaller set of covariates generates more fluid modeling without overestimated predictions and meets the principle of parsimony (GRANITTO et al, 2006;KUHN E JOHNSON, 2013;SOUZA et al, 2018;GOMES et al, 2019). Therefore, the Random Forest (RF) selected seven covariates to determine potential areas for micro-dam and generated an R 2 of 0.43 in training and validation (Table 3; Table 4).…”

“…In general, combining the micro-dam potential map by the AHP+RF method generates better spatial results, reducing the effect of underestimating areas for micro-dam allocation in restricted areas when applying the methods in isolation. Notably, the advantage of ML is the inclusion of covariates in prediction studies (KUHN E JOHNSON, 2013;CHAGAS et al, 2018;SOUZA et al, 2018;GOMES et al, 2019). The insertion of covariates opens up vast possibilities for future studies on areas for micro-dam allocation, especially by inserting covariates of hydrological and climatic factors, which are important factors in micro-dam allocation (SCHIETTECATTE et al, 2005;HIPÓLITO et al, 2019;MESHRAM et al, 2020).…”

Modelagem Espaço-Temporal De Ilhas De Calor Urbana E Correlação Com Variáveis- Montes Claros, Mg, Brasil

“…Nos dados, testes de correlação Pearson que mensura o grau de associação ou distanciamento linear entre variáveis quantitativas foi aplicada. Além disso, analise estatística multivariada (Análise de componente principal -ACP), que reduz sobreposições e seleciona as formas mais representativas a partir de combinações lineares foram aplicadas (Souza et al, 2018). Ao final, o critério de Kaiser (Kaiser, 1960) Fonte: Elaborado pelos autores.…”