qpt_mixsiar

Scripts for plant wax mixing model (MixSIAR) experiments and results processing for Lake QPT, Baffin Island, Nunavut, Canada (Lindberg et al., in prep).

Manuscript title: Bayesian modeling plant wax sources to an Eastern Canadian Arctic lake sediment record reveals stable plant wax vegetation sources following postglacial shrub colonization

Authors: Kurt R. Lindberg, Elizabeth K. Thomas, Martha K. Raynolds, Helga Bultmann

Study site: Lake Qaupat (QPT)

Lake Qaupat (QPT) is located in southern Baffin Island near the city of Iqaluit and Frobisher Bay (63.68°N, 68.20°W). The modern catchment vegetation is predominantly shrubs on land, including Betula nana and Salix arctica, and submerged aquatic mosses within the lake, such as Calliergon richardsonii (Hollister et al., 2022).

Sedimentary ancient DNA (sedaDNA) evidence revealed that Betula sp. first colonized the catchment at 6.1 ka (Crump et al., 2019), following local Laurentide Ice Sheet deglaciation around ~9-8 ka (Miller, 1980). However, sedaDNA can only be interpreted on a presence/absence basis, leaving questions about how the overall catchment vegetation community and biomass may have changed following Betula sp. colonization.

To address these questions, we use the Bayesian mixing model framework, MixSIAR (Stock et al., 2018), to quantitatively estimate the sources of plant waxes post-Betula sp. colonization to represent changes in vegetation/biomass.

[INSERT FIGURE HERE]

Data: Plant wax n-alkanoic acids

We use plant wax n-alkanoic acid carbon chain-length distributions and $\delta$¹³C as tracers to model the contributions of different endmember priors. $\delta$¹³C values are used as the primary tracers and are weighted by the concentration of each carbon chain-length (Upadhayay et al., 2018).

MixSIAR Workflow

1. Generate model input files

• Mix: Mixture files are the $\delta$¹³C values of specified n-alkanoic acid carbon chain-lengths in Lake QPT sediments. Downcore sediment measurements (Age $\geq$ -39 cal yr BP) are from this study, while other samples are from Hollister et al. (2022). These files are currently not produced by qpt_mixsiar_makeinputs.ipynb.

• Source: Means and standard deviations of user-defined vegetation endmembers in qpt_mixsiar_makeinputs.ipynb. Endmember statistics are created from individual samples in "qpt_plantwax_source.xlsx".

  • Note (02/11/2025): Published Lake QPT plant wax endmember data (Hollister et al., 2022; DOI: https://doi.org/10.1029/2022JG006903) currently available. Eastern Canadian Arctic plant wax data not yet published as part of Lindberg et al. (in prep): Ecological and environmental controls on plant wax production and stable isotope fractionation in modern terrestrial Arctic vegetation
  • Lake QPT plant wax endmember data from Hollister et al. (2022) is publically available on figshare: https://figshare.com/s/96c3bd6c11bae87c1e28?file=34381742

• Discr: Discrimination factors that account for potential alteration/fractionation between the endmember and mixture. These factors are not well-defined in sedimentary mixing studies (Menges et al., 2020), so the Python input script sets them all to "0".

2. Run model experiments

• MixSIAR experiements are run using qpt_mixsiar.R (Jupyter Notebook version in progress). The script is set to output two files upon model completion:

  1. output_summary: Lists model-calculated proportional (out of 1) contributions for each endmember as a mean, standard deviation, and percentile values.

  2. diagnostics: Determines if the model run reached sufficient convergence as dictated by the Gelman and Geweke diagnostic criteria described in each file.

3. Generate plots from output summaries (In progress)

References

Crump, S. E., Miller, G. H., Power, M., Sepúlveda, J., Dildar, N., Coghlan, M., & Bunce, M. (2019). Arctic shrub colonization lagged peak postglacial warmth: Molecular evidence in lake sediment from Arctic Canada. Global Change Biology, 25(12), 4244-4256.

Hollister, K. V., Thomas, E. K., Raynolds, M. K., Bültmann, H., Raberg, J. H., Miller, G. H., & Sepúlveda, J. (2022). Aquatic and terrestrial plant contributions to sedimentary plant waxes in a modern Arctic lake setting. Journal of Geophysical Research: Biogeosciences, 127(8), e2022JG006903.

Lindberg, K.R., Thomas, E.K., Raynolds, M.K., Bultmann, H. 2024 (talk). Evaluating plant waxes as tracer for reconstructing mid-late Holocene vegetation change in a sourthern Baffin Island lake catchment. The 52nd International Arctic Workshop, Amherst, MA, Mar. 13-16.

Menges, J., Hovius, N., Andermann, C., Lupker, M., Haghipour, N., Märki, L., & Sachse, D. (2020). Variations in organic carbon sourcing along a trans-Himalayan river determined by a Bayesian mixing approach. Geochimica et Cosmochimica Acta, 286, 159-176.

Miller, G. H. (1980). Late foxe glaciation of southern Baffin Island, NWT, Canada. Geological Society of America Bulletin, 91(7), 399-405.

Stock, Brian C., Andrew L. Jackson, Eric J. Ward, Andrew C. Parnell, Donald L. Phillips, and Brice X. Semmens. "Analyzing mixing systems using a new generation of Bayesian tracer mixing models." PeerJ 6 (2018): e5096.

Upadhayay, H. R., Bodé, S., Griepentrog, M., Bajracharya, R. M., Blake, W., Cornelis, W., & Boeckx, P. (2018). Isotope mixing models require individual isotopic tracer content for correct quantification of sediment source contributions. Hydrological Processes, 32(7), 981-989.