Kilian Eichenseer

Co-authored by @PalaeoSmith, @archaeometer, Matthias Sinnesael and myself (@EichenseerK)

📈New method for Stratigraphic Correlation and Age Modelling released today! Discuss the preprint, including a Cambrian case study, at egusphere.copernicus.org/preprints/2025… 💻Available as an R package at stratobayes.github.io #stratigraphy #correlation #agemodelling

The first session of our new #lecture series is in one week on Feb 29! Dr. Emily Mitchell will present "Using spatial point process analyses to quantify the behavioural ecology behind trace fossils" Sign up: bit.ly/palaeoverse-le… #computational #palaeoecology #palaeobiology

@wolexsegzy @foradamifera Thanks for your interest! This workshop is in-person only, but we may organise some online training further down the line. At this point I can't make any promises, though.

@EichenseerK @foradamifera Pls, is there a zoom link for those outside UK or a transcript of the training? I will really appreciate a positive answer. Thanks

Are you interested in stratigraphic correlation and age modelling? From 18-20 March in Durham, UK, we’re hosting a workshop on the automated correlation of chemostratigraphic and geophysical data, and on constructing age-depth models. Find out more at stratobayes.github.io

@fred_bowyer @foradamifera Thanks Fred, looking forward to seeing you in Durham!

Our new method integrates disparate proxy data and prior knowledge to produce quantitative palaeotemperature estimates. Bridging spatial gaps, this approach can increase the accuracy of proxy-based palaeoclimate reconstructions, especially in intervals with limited sampling. 📈🌎

Applying this model to geochemical and ecological sea surface temperature proxies from the early Eocene climatic optimum (EECO) suggests a flatter latitudinal temperature gradient than most earlier studies. 🌊🌡️

How can we reconstruct past climates from patchy and spatially biased records?🌍🌡️❓ @LewisAlanJones and I address this question in our new paper, published today in Climate of the Past: doi.org/10.5194/cp-20-… #palaeoclimate #isotopes #palaeoecology #Eocene #Bayesian #rstats

🚨ANNOUNCEMENT!🚨 We're excited to share our new lectures series which will bring you the latest on novel software tools and methodological approaches in #palaeobiology! The lectures will run on the last Thursday of every month at 14:00 UTC. Sign up: bit.ly/palaeoverse-le…

We’ve had a fantastic two days this week at UCL, London for the 1st Edition of the Palaeoverse Workshop and Hackathon! A big thank you to all attendees for bringing such enthusiasm (despite the warmth!) and being such a lovely bunch. #palaeoverse #palaeontology #palaeobiology

The image I created with @piginatutu introducing @ThePalaeoverse R package was chosen for this month's @MethodsEcolEvol issue! So proud🥹🥹🥹 📰 Check out our paper here: doi.org/10.1111/2041-2…

How hot was the early Eocene? 🔥 @LewisAlanJones and I have developed a Bayesian model to estimate latitudinal temperature gradients from spatially incomplete geochemical and ecological proxy records. Read and comment the preprint at egusphere.copernicus.org/preprints/2023… #palaeoclimate

@_jcken That means you also managed to install it in the first place, congrats :-D

Wrote a Stan program. Compiled first time.

@_jcken I got confused by this as well at some point. You can get the widening confidence intervals by specifying "ma" for moving average in the ci.interval argument: stackoverflow.com/questions/1426…

Does anyone know what the dashed line in the base R acf() means? I was always taught that they were upper/lower CI limits. But my intuition is that the width of these limits should be proportional to 1/sqrt(sample_size - lag). The given limits are constant #rstats

Our paper on the palaeoverse R package is out. Great for organising and plotting palaeobiological data. Some functions are useful for geological data more broadly. besjournals.onlinelibrary.wiley.com/doi/10.1111/20…

@_jcken That is glorious, congrats!

Dr Jack Kennedy (subject to minor corrections)

@_jcken @jumping_uk Agree, that message definitely came across

@EichenseerK @jumping_uk Good point! The key message I wanted to stress was simple models -> JAGS is probably fine. Complex models -> seriously consider Stan

You've heard about Stan and want to learn some more? Maybe you're about to step into the Bayesian paradigm and don't know where to start 🤔 In this week's blog post, we'll take a look what you can do with #Stan and draw comparisons to #JAGS! jumpingrivers.com/blog/why-stan/ #datascience