U–Pb ages and Hf isotopic composition of zircons in Austrian last glacial loess: constraints on heavy mineral sources and sediment transport pathways

Abstract

Loess sediments in Austria deposited ca. 30–20 ka ago yield different zircon age signatures for samples collected around Krems (SE Bohemian Massif; samples K23 and S1) and Wels (halfway between the Bohemian Massif and the Eastern Alps; sample A16). Cathodoluminescence (CL) imaging reveals both old, multistage zircons with complex growth histories and inherited cores, and young, first-cycle magmatic zircons. Paleoproterozoic ages between 2,200 and 1,800 Ma (K23 and S1), an age gap of 1,800–1,000 Ma for S1 and abundant Cadomian grains, indicate NW African/North Gondwanan derivation of these zircons. Also, A16 yields ages between 630 and 600 Ma that can be attributed to “Pan-African” orogenic processes. Significant differences are seen for the <500 Ma part of the age spectra with major age peaks at 493–494 and 344–335 Ma (K23 and S1), and 477 and 287 Ma (A16). All three samples show negative initial ɛHf signatures (−25 to −10, except one grain with +9.4) implying zircon crystallization from magmas derived by recycling of older continental crust. Hf isotopic compositions of 330- to 320-Ma-old zircons from S1 and K23 preclude a derivation from Bavarian Forest granites and intermediate granitoids. Rather, all the data suggest strong contributions of eroded local rocks (South Bohemian pluton, Gföhl unit) to loess material at the SE edge of the Bohemian Massif (K23 and S1) and sourcing of zircons from sediment donor regions in the Eastern Alps for loess at Wels (A16). We tentatively infer primary fluvial transport and secondary eolian reworking and re-deposition of detritus from western/southwestern directions. Finally, our data highlight that loess zircon ages are fundamentally influenced by fluvial transport, its directions, the interplay of sediment donor regions through the mixing of detritus and zircon fertility of rocks, rather than Paleowind directions.

Appendix

Uncertainty propagation for ¹⁷⁶Hf/¹⁷⁷Hf_{t, zircon} (or initial ¹⁷⁶Hf/¹⁷⁷Hf_zircon)

The ¹⁷⁶Hf/¹⁷⁷Hf composition of zircon at the time of crystallization (i.e., initial ¹⁷⁶Hf/¹⁷⁷Hf_zircon or ¹⁷⁶Hf/¹⁷⁷Hf_t,zircon) is calculated as

$$D_{t} = D_{\text{m}} - P_{\text{m}} ({\text{e}}^{\lambda t} - 1),$$

(1)

where D _t  = ¹⁷⁶Hf/¹⁷⁷Hf _t,zircon, D _m = ¹⁷⁶Hf/¹⁷⁷Hf_{zircon-measured}, P _m = ¹⁷⁶Lu/¹⁷⁷Hf_{zircon-measured}, λ = λ _176Lu = 1.867 ± 0.008 × 10⁻¹¹ a⁻¹ (Söderlund et al. 2004) and t is the crystallization age of zircon.

Using the law of propagation of uncertainty, the combined standard uncertainty for D _t is given by

$$\sigma_{{D_{t} }}^{2} = \left( {\frac{{\partial D_{t} }}{{\partial D_{\text{m}} }}\sigma_{{D_{\text{m}} }} } \right)^{2} + \left( {\frac{{\partial D_{t} }}{{\partial P_{\text{m}} }}\sigma_{{P_{\text{m}} }} } \right)^{2} + \left( {\frac{{\partial D_{t} }}{\partial \lambda }\sigma_{\lambda } } \right)^{2} + \left( {\frac{{\partial D_{t} }}{\partial t}\sigma_{t} } \right)^{2}$$

(2)

Since

$$\frac{{\partial D_{t} }}{{\partial D_{\text{m}} }} = 1$$

(3)

$$\frac{{\partial D_{t} }}{{\partial P_{\text{m}} }} = {\text{e}}^{\lambda t} - 1$$

(4)

$$\frac{{\partial D_{t} }}{\partial \lambda } = ({\text{e}}^{\lambda t} - 1)P_{\text{m}} t$$

(5)

$$\frac{{\partial D_{t} }}{\partial t} = ({\text{e}}^{\lambda t} - 1)P_{\text{m}} \lambda,$$

(6)

the combined uncertainty of D _t  = ¹⁷⁶Hf/¹⁷⁷Hf _t,zircon is

$$\sigma_{{D_{t} }} = \sqrt {\sigma_{{D_{\text{m}} }}^{2} + \left( {({\text{e}}^{\lambda t} - 1)\sigma_{{P_{\text{m}} }} } \right)^{2} + \left( {({\text{e}}^{\lambda t} - 1)P_{\text{m}} t\sigma_{\lambda } } \right)^{2} + \left( {({\text{e}}^{\lambda t} - 1)P_{\text{m}} \lambda \sigma_{t} } \right)^{2} }$$

(7)