Abstract Multivariate statistical approaches were applied on geochemical data of lake sediments and the catchment basin materials (Al-enriched crusts and soils) from four upland lakes (namely LB1, LB2, LB3, and LB4) in the Serra da Bocaina, Southeast Amazon, to recognize the geochemical signatures of lake sediments and their link with source rocks. This is also compared with Fe-enriched crusts and soils, which collected from Serra Sul de Carajas. The major and trace elements were analyzed by Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES) and X-ray Fluorescence (XRF), and Inductively Coupled Plasma Mass Spectrometry (ICP-MS), respectively. The data closure effect was reduced by centred log ratio (clr) transformation. The upper continental crust (UCC) normalized patterns show that sediments are mainly enriched with Ti, Fe, P, Se, V and Cr, similar to catchment basin crusts and soils. The distribution of elements varied between lakes, which is mainly influenced by the morphology and lithology of the basin. The chemical index of alternation (CIA) values were very high (>90) for both lake sediments and catchment materials, which indicates intense source area weathering. Index of lateritization (IOL) shows that the catchment crusts are strongly lateralized and it identifies samples between extremely weathered and reworked samples that do not meet the criteria for strict in situ formation. The principal components analysis (PCA) and Spearman correlation matrices based on clr-transformed data revealed four geochemical groups in sediments. The major detritic geochemical association of sediments (Group 1: Al–Ti-Ga-Cr-V-Ti-Nb-Sc, which indicative of mafic/metavolcanic rock; and Group 2: Zr-Hf-HREEs, which refers to the heavy mineral fractions that remain stable during lateritization) had a close relationship with catchment lateritic crust, except a few alterations, which possibly influenced by weathering and transport processes. REE pattern normalized to chondrite and the geochemical ratios show the analogies between the lake sediments and the catchment lateritic crusts. The linear discriminant analysis (LD) and PCA distinguished Al-rich laterites from Fe-rich laterites and indicates that the former (which likely derived from mafic/metavolcanic rocks) is the main source of lake sediments.

中文翻译：

---

理解源汇关系的多元统计和地球化学方法——以亚马逊东南部近海湖泊为例

摘要 运用多元统计方法对亚马逊东南部 Serra da Bocaina 四个高地湖泊（LB1、LB2、LB3 和 LB4）的湖泊沉积物和集水盆地材料（富铝结壳和土壤）的地球化学数据进行了研究，认识湖泊沉积物的地球化学特征及其与烃源岩的联系。这也与从 Serra Sul de Carajas 收集的富含铁的地壳和土壤进行了比较。主要和痕量元素分别通过电感耦合等离子体发射光谱法 (ICP-OES) 和 X 射线荧光 (XRF) 以及电感耦合等离子体质谱法 (ICP-MS) 进行分析。中心对数比 (clr) 变换降低了数据闭合效应。上大陆地壳 (UCC) 归一化模式表明沉积物主要富含 Ti、Fe、P、Se，V 和 Cr，类似于集水盆地结皮和土壤。湖泊之间元素分布不同，主要受盆地形态和岩性影响。湖泊沉积物和集水材料的化学交替指数 (CIA) 值都非常高 (>90)，这表明源区风化强烈。红土化指数 (IOL) 显示集水结皮强烈偏侧化，它可以识别极端风化和不符合严格原位形成标准的返工样品之间的样品。基于 clr 转换数据的主成分分析 (PCA) 和 Spearman 相关矩阵揭示了沉积物中的四个地球化学组。沉积物的主要碎屑地球化学组合（第 1 组：Al-Ti-Ga-Cr-V-Ti-Nb-Sc，指示镁铁质/变质火山岩；和第 2 组：Zr-Hf-HREEs，指在红土化过程中保持稳定的重矿物部分）与集水区红土地壳密切相关，除了少数可能受风化和运输过程影响的蚀变。归一化为球粒陨石的 REE 模式和地球化学比率显示了湖泊沉积物和集水红土结壳之间的类比。线性判别分析 (LD) 和 PCA 将富铝红土与富铁红土区分开来，并表明前者（可能来自基性/变质火山岩）是湖泊沉积物的主要来源。可能受风化和运输过程的影响。归一化为球粒陨石的 REE 模式和地球化学比率显示了湖泊沉积物和集水红土地壳之间的类比。线性判别分析 (LD) 和 PCA 将富铝红土与富铁红土区分开来，并表明前者（可能来自基性/变质火山岩）是湖泊沉积物的主要来源。可能受风化和运输过程的影响。归一化为球粒陨石的 REE 模式和地球化学比率显示了湖泊沉积物和集水红土结壳之间的类比。线性判别分析 (LD) 和 PCA 将富铝红土与富铁红土区分开来，并表明前者（可能来自基性/变质火山岩）是湖泊沉积物的主要来源。