Multi-isotope approach (δ^44/40Ca, δ^88/86Sr and ⁸⁷Sr/⁸⁶Sr) provides insights into rhizolith formation mechanisms in terrestrial sediments of Nussloch (Germany)

Highlights

• Different δ^44/40Ca, δ^88/86Sr and ⁸⁷Sr/⁸⁶Sr isotope values for the rhizolith than for the rhizosphere and loess from the same stratigraphic level

• Lower δ^44/40Ca and δ^88/86Sr values result from preferential precipitation of light isotopes.

• Higher ⁸⁷Sr/⁸⁶Sr values result from a mixing between carbonate and non-carbonate minerals digested around the roots or from downward transfer.

Abstract

Carbonates in loess-paleosol sequences are considered as a valuable paleoenvironmental proxy for climate change during the Quaternary period. In this study we propose a new dataset of δ^44/40Ca, δ^88/86Sr and ⁸⁷Sr/⁸⁶Sr to understand the formation of rhizoliths (roots encrusted with CaCO₃) within these kinds of sequences from Nussloch region (Germany). Our results indicate that the leached fractions of rhizoliths (sampled at 3.25 m, 6.65 m, and 8.45 m depths and leached with HCl) have lower δ^44/40Ca and δ^88/86Sr, and higher ⁸⁷Sr/⁸⁶Sr isotopic values than the leached fractions of rhizosphere and root-free loess of the same stratigraphic levels. This suggests different mechanisms of carbonate formation and different sources of Sr and Ca. A three steps model of formation of rhizoliths from Nussloch has been proposed. First of all, the roots grow in an existent stratigraphic level and the biological activity associated with roots allows for the dissolution of carbonate and non-carbonate minerals. A part of the available Ca and Sr is taken up by roots, which absorb preferentially the light isotopes (e.g., ⁴⁰Ca and ⁸⁶Sr). A higher rate of water absorption by roots than nutrients, in association with CO₂ released by biological respiration, allows to reach supersaturation with respect to CaCO₃. The precipitation of rhizolith carbonate selects preferentially the light isotopes (e.g., ⁴⁰Ca and ⁸⁶Sr) leading to lower δ^44/40Ca and δ^88/86Sr isotope values than the ones recorded on the closest rhizosphere and loess. The higher ⁸⁷Sr/⁸⁶Sr signatures of rhizolith carbonates, which is not modified by root uptake or by precipitation processes, is the result of a mixture between ⁸⁷Sr/⁸⁶Sr signatures of carbonate and non-carbonate minerals digested around the roots. Nevertheless, grey zones remain, in this main hypothesis, about the very high contribution of residue materials to the isotopic signatures of rhizoliths. In this light, other mechanisms of downward transfer of Ca and Sr from upper stratigraphic level, that could imprint the leached fractions of rhizoliths, have been discussed.

Introduction

The loess-paleosol alternations are valuable records to study paleo-environmental changes occurring during the Quaternary period (−2.58 Myr to present day) around the world (e.g., Zhisheng et al., 1990; Gallet et al., 1996; Wang et al., 2000; Muhs and Budahn, 2006; Gocke et al., 2014b). Within these sedimentary sequences, pedogenic carbonates are excellent recorders of climatic processes affecting the carbon (C), the oxygen (O) and the calcium (Ca) cycles, and serve as reference materials for paleoenvironmental reconstructions (Cerling, 1984). The latter are mostly based on isotope or molecular records such as δ¹⁸O (palaeo-temperature and hydric conditions), δ¹³C (organic and inorganic sources of C), ¹⁴C (dating) or lipid markers such as n-alkanes (sources of organic matter) (e.g., Pustovoytov et al., 2007; Barta, 2011; Gocke et al., 2011, Gocke et al., 2013; Gocke and Kuzyakov, 2011; Huguet et al., 2012; Prud'Homme et al., 2016).

Among the numerous morphologies of pedogenic carbonates (Zamanian et al., 2016), rhizoliths (i.e., fossil roots encrusted by CaCO₃; Klappa, 1980) are frequent features of pedogenic CaCO₃ in soils and loess-paleosol sequences (Becze-Deák et al., 1997; Gocke et al., 2014c). They have been widely described in the literature and have been used as paleo-environmental tracers of C₃-C₄ plant abundances and alternation of dry-wet and cold-warm seasons (e.g., Wang et al., 2004; Wang and Greenberg, 2007). The rhizolith formation remains debatable and is currently explained by two main mechanisms: (1) a higher uptake of water than Ca²⁺ by the roots, associated with a high amount of CO₂ from rhizomicrobial respiration, lead to a supersaturated surrounding solution with respect of CaCO₃, (2) a release of HCO₃⁻ by the roots (to compensate the uptake of anions), which generates a pH increase in soil solution and thus allows for the precipitation of CaCO₃ (Zamanian et al., 2016 and reference therein). In every case, the root encrustation by a tubular layer of CaCO₃ is considered as rapid (years to decades) and occurred during the lifetime of the plant or just after its death (Gocke et al., 2010, Gocke et al., 2011; Gocke and Kuzyakov, 2011).

The use of rhizoliths as paleo-environmental proxies is affected by various types of limitations: (1) not all rhizoliths have the same degree of preservation (heterogeneous impact of degradation); (2) their abundances and the recorded signals differ according to their size. In fact, the micro-rhizoliths are often much more abundant than large rhizoliths (Gocke et al., 2010) but the small plant roots have negligible lifespans compared to large ones (Strand et al., 2008). (3) the spatial and temporal distribution of the roots, which can form the rhizoliths, is heterogeneous (Silk, 1984); (4) rhizolith formation is restricted to specific physico-chemical conditions (e.g., high Ca²⁺ and CO₂ concentrations around the root networks), which could generate occurrence gaps along the sedimentary sequences if climatic conditions change considerably over time; (5) it is very difficult to determine to which plant species rhizoliths belong to if they are not connected to the aboveground part of the source vegetation (Gocke et al., 2014a); (6) the different maximum rooting depth capacity of each plant species can generate a mixture of rhizoliths from different ages in the same stratigraphic level. Such post-sedimentary rhizolith formation has already been confirmed in the literature through the ¹⁴C isotopic composition measurements (Gocke et al., 2011). Nevertheless, the radiocarbon dating remains subject to strong degradation effects as C inputs from bedrock alteration or atmosphere, which may overprint the original rhizolith signal (Chen and Polach, 1986; Amundson et al., 1994; Budd et al., 2002).

An improved understanding of the rhizolith formation mechanisms is thus required to apprehend an important part of carbonate's dynamics within the terrestrial ecosystems and to enhance paleo-environmental or chronological interpretations based on rhizoliths. Calcium isotopes (δ^44/40Ca) are considered as good tracers of bio-geochemical mechanisms occurring during nutrient uptake by vegetation, and during carbonate precipitation at the water-soil-plant interface (Cobert et al., 2011; Schmitt et al., 2012, Schmitt et al., 2013, Schmitt et al., 2017, Schmitt et al., 2018; Tipper et al., 2016 and references therein). Radiogenic strontium isotope ratio (⁸⁷Sr/⁸⁶Sr) is traditionally used as Ca source tracers due to physical similarities of Ca and Sr allowing considering them as analogous in biogeochemical cycles (Clow et al., 1997; Probst et al., 2000; Drouet et al., 2005; Bullen and Bailey, 2005; Bagard et al., 2013). In addition, over the past decade, the stable isotope ratio of strontium (δ^88/86Sr) was proven to have high potential as tracer of alteration, biological cycling and secondary precipitation processes within soils (e.g., Shalev et al., 2013; Wei et al., 2013; Andrews et al., 2016).

The aim of this study is to characterize the rhizolith formation mechanisms within the Quaternary loess-paleosol sequence of Nussloch (SW Germany) and to identify the Ca and Sr sources of the rhizolith carbonates. For this, we adopted an innovative multi-isotope approach coupling δ^44/40Ca, ⁸⁷Sr/⁸⁶Sr and δ^88/86Sr isotopic signatures of rhizoliths, soils, paleosols and loess samples from different depths of the sequence to access formation mechanisms along a vertical and a horizontal transect within the sequence.