iiinova — where your digital self keeps evolving

Reposted by Andrés Baselga

Stephen Heard @stephenbheard.bsky.social · 21/09/2026

"But the trouble with all of these additions and applications piled atop the humble research article is that they’re new solutions to a comprehensively solved problem." Excellent from @jbyoder.org

1146 14

Andrés Baselga @andresbaselga.bsky.social · 10/09/2026

I didn’t say “not correctly”, but “not directly”. To go from the data to the scores you need at some point to compute variance across communities (or equivalently, a distance matrix). The way you do that is what affects the ecological meaning of scores (or ordination distances)

101 1

Andrés Baselga @andresbaselga.bsky.social · 10/09/2026

Anyway, thanks again for the interesting discussion!

100 1

Andrés Baselga @andresbaselga.bsky.social · 10/09/2026

If two communities have different scores, what does it mean? Is this affected by double zeros, by mean species richness, by richness differences? In what way are these communities different? I find this question relevant

200 2

Andrés Baselga @andresbaselga.bsky.social · 10/09/2026

I think the site scores cannot represent the data directly. The geometry of the scores is the same geometry of the distance matrix in the ordination space. I believe it is important to know what the distances (or the scores) mean in ecological terms.

200 2

Andrés Baselga @andresbaselga.bsky.social · 10/09/2026

And sorry for insisting, but model based ordinations do use a distance measure, even if implicitly. Knowing which facet of community variation is being captured would be relevant

200 2

Andrés Baselga @andresbaselga.bsky.social · 10/09/2026

Thanks for sharing your thoughts! I disagree with this view, though. Dissimilarities capture relevant information about how communities vary, and different dissimilarity measures capture different facets of that variation

200 2

Andrés Baselga @andresbaselga.bsky.social · 10/09/2026

Distances and variance are two sides of the same coin (as @comecology.bsky.social mentioned): doi.org/10.1111/ele.... It would be valuable to check how the ordination distances resulting from different models relate to dissimilarities with well-known properties (chi-sq, hellinger, sorensen...)

doi.org

100 1

Andrés Baselga @andresbaselga.bsky.social · 10/09/2026

So my question was (is) if it is possible to know how to specify the models in a way in which the ordination distances have a specific ecological meaning (analogous to different dissimilarity measures)

100 1

Andrés Baselga @andresbaselga.bsky.social · 10/09/2026

An NMDS reflects one type of dissimilarity measure, and its ecological meaning depends on the dissimilarity you have chosen. My intuition is that the ecological meaning of the ordination distances on a model-based ordination depends on how you specify the model

200 2

Andrés Baselga @andresbaselga.bsky.social · 10/09/2026

If so, what do these distances mean?

000 0

Andrés Baselga @andresbaselga.bsky.social · 10/09/2026

Thanks! I think I understand that, but my question is whether using different distributions results in ordinations in which the ordination space is proportional to a reasonable “dissimilarity”. Even if you “avoid” distance matrices, the ordination is an implicit distance matrix, right?

100 1

Andrés Baselga @andresbaselga.bsky.social · 09/09/2026

It is thus not trivial what type of dissimilarity is represented in the ordination space, as it would completely change the ecological meaning and any derived inference

200 2

Andrés Baselga @andresbaselga.bsky.social · 09/09/2026

My question is if is it possible that the ordination resulting from the multi-species model does accommodate a meaningful dissimilarity measure. If an ordination is produced, distance between two communities in the ordination space represents community dissimilarity

100 1

Andrés Baselga @andresbaselga.bsky.social · 09/09/2026

Thanks again. I would need to study and probably would need a lot of help to understand the maths of how the model can resolve these problems to make the distance in the ordination reflect a meaningful ecological dissimilarity

100 1

Andrés Baselga @andresbaselga.bsky.social · 09/09/2026

Yes, my question is how is the ordination calculated in a multi-species model? How do you go from the data to the ordination? Either explicitly or implicitly, you need to compute distances (=dissimilarities) between communities, and translate that into an ordination.

100 1

Andrés Baselga @andresbaselga.bsky.social · 09/09/2026

Indeed, I think some indices are more meaningful than others :-) Euclidean distances are probably the least meaningful. For example, calculating the euclidean distance to centroid to estimate uniqueness would produce measures with no ecological meaning (double zero problem, pairwise symmetry)

000 0

Andrés Baselga @andresbaselga.bsky.social · 09/09/2026

Thanks, I was aware of that. However, these models do build an ordination space, and the distances in the ordination space are in fact "dissimilarities" between communities. The problem with euclidean distance is that it is not ecologically meaningful (i.e. it is affected by double zeros).

100 1

Andrés Baselga @andresbaselga.bsky.social · 09/09/2026

I am curious about this. Can multi-species models handle different community dissimilarity indices, so the distances between sites in the ordination space reflect an ecologically meaningful community dissimilarity? @comecology.bsky.social @vdveenb.bsky.social

300 3

Andrés Baselga @andresbaselga.bsky.social · 08/09/2026

We made a video-abstract summarizing the major points of our paper on ecological uniqueness: youtu.be/Spft2aR8ZUI?...

youtu.be

Ecological uniqueness: New measures to reconcile method with concept

YouTube video by Biogeography Lab (BiBiCI-USC)

011 1

Andrés Baselga @andresbaselga.bsky.social · 07/09/2026

In sum, to quantify ecological uniqueness do use: 1. An asymmetric index of dissimilarity 2. Average dissimilarity instead LCBD (they are linearly related, but average dissimilarity can be interpreted in absolute terms, while LCBD cannot be compared across studies).

000 0

Andrés Baselga @andresbaselga.bsky.social · 07/09/2026

Average dissimilarity and LCBD are mathematically related (linear transformation). However, average dissimilarity does reflect community differentiation and is independent on the number of sites (N). LCBD is blind to community differentiation and depends on N, so it is not comprable across studies.

100 1

Andrés Baselga @andresbaselga.bsky.social · 07/09/2026

This has important consequences. Uniqueness based on symmetric indices identify as unique communities that are impoverished subsets (i.e. they are not communities with actually unique species). Also, note the perfect linear relationship between uniqueness measured as average dissimilarity and LCBD

100 1

Andrés Baselga @andresbaselga.bsky.social · 07/09/2026

Using symmetric indices results on measures of uniqueness that do not actually reflect the intended concept (i.e. unique composition of species). This occurs either for average dissimilarity (dots) or LCBD (squares). Only uniqueness based on an asymmetric index (red) truly quantifies the concept

100 1

Andrés Baselga @andresbaselga.bsky.social · 07/09/2026

Asymmetric dissimilarity measures have been scarcely used, and less so to quantify ecological uniqueness. However, ecological uniqueness is intrinsically asymmetric. Community C is more unique than community D, so if uniqueness relies is quantified with a dissimilarity index, it must be asymmetric

100 1

Andrés Baselga @andresbaselga.bsky.social · 07/09/2026

New paper by Carola Gómez-Rodríguez and me on ecological uniqueness. We show that: 1. Asymmetric dissimilarity is a must for quantifying the uniqueness of biological communities. 2. Average dissimilarity is an absolute measure of uniqueness, while LCBD is a relative indicator doi.org/10.1111/geb....

doi.org

Ecological Uniqueness: New Measures to Reconcile Method With Concept

Aim Ecological uniqueness quantifies how different a community is relative to all other communities within the study. This broad-sense definition of uniqueness makes it dependent on the choice of th.....

194 9

Andrés Baselga @andresbaselga.bsky.social · 22/07/2026

This is the relationship between betavolume and beta diversity for all combinations of shared and unique species between two sites (and 100 species in total). I was tempted to call this a "curlew-shaped relationship":

020 2

Andrés Baselga @andresbaselga.bsky.social · 21/07/2026

Song et al introduced a geometric measure intended to quantify beta diversity, called betavolume. I assessed its behaviour: 1) betavolume does not measure beta diversity or differentiation; 2) links to species interactions remain to be proven. esajournals.onlinelibrary.wiley.com/doi/10.1002/...

esajournals.onlinelibrary.wiley.com

A geometric approach to beta diversity: Comment

Click on the article title to read more.

273 7

Reposted by Andrés Baselga

Biogeography Lab (BiBiCI-USC) @biogeography-usc.bsky.social · 20/07/2026

📽️ Primer video de la serie #BeetleBits sobre métodos de muestreo de coleópteros, en ell marco del proyecto #MetaBeetle (AEI) Empezamos con unos consejos que nos da Mauro Quintanilla-Viana para evitar picaduras de garrapatas. -.- Tick bite prevention tips (in Spanish) youtu.be/kW15M_jV2cg?...

youtu.be

Beetle Bits: Prevención de picaduras de garrapata en muestreos de campo

YouTube video by Biogeography Lab (BiBiCI-USC)

021 2

Reposted by Andrés Baselga

gavin jones @ecologyofgavin.bsky.social · 14/12/2025

I’m really concerned about early career scientists using generative AI in writing, because writing is how we think and reflect. There’s almost nothing *more* human than that, and to give that gift away to a machine is almost unconscionable to me

713534 135

Reposted by Andrés Baselga

Dr Selena Wisnom @lswisnom.bsky.social · 03/11/2025

Academics in Assyria in the 7th c BC complain that admin is preventing them from doing research and teaching

5245181434 4.5K

Reposted by Andrés Baselga

Stephen Heard @stephenbheard.bsky.social · 21/10/2025

New post: All The Things I Was Never Trained To Do. Which turns out to be a sneakily disguised meditation on what university education is for, and thus what university is, and what it isn't - despite the persistence of idiots who don't understand that!

scientistseessquirrel.wordpress.com

All the things I was never trained to do

I’m reminded just about every day that the disconnect between my education, and the career I built from it, is enormous. Or at least, it seems so. I had a pretty unremarkable educational sequence f…

0185 18

Andrés Baselga @andresbaselga.bsky.social · 15/10/2025

Important thoughts about museums, extinct species, and all the biological diversity that is going extinct unnoticed (not even preserved as museum specimens)

110 1

Reposted by Andrés Baselga

Fernando Ojeda @ojedacopete.bsky.social · 04/10/2025

Álvaro Pérez Gómez defendió ayer su tesis doctoral en la UCA. Ha sido fácil y fascinante trabajar con él; entre otras cosas, por su formación naturalista, que traía ya de su mentor desde pequeño: Íñigo Sánchez. Álvaro ha realizado este breve documental de su tesis: m.youtube.com/watch?v=kjpp...

m.youtube.com

Vídeo presentación Tesis Doctoral Álvaro Pérez Gómez

YouTube video by Álvaro Pérez Gómez

0146 14

Reposted by Andrés Baselga

Biogeography Lab (BiBiCI-USC) @biogeography-usc.bsky.social · 23/09/2025

We have developed a new methodological framework to measure biodiversity. We focus on the uniqueness (or singularity) of biological communities and apply it to the GBF 30x30 target. If you only have 4:35 min, check the visual abstract to know a little bit more about it ☺️: youtu.be/dHMBhVJ8dns?...

youtu.be

Areas of high biodiversity value evidenced by the spatial scaling of phylogenetic uniqueness

YouTube video by Biogeography Lab (BiBiCI-USC)

054 5

Andrés Baselga @andresbaselga.bsky.social · 23/09/2025

Thanks Erik! Glad you found it interesting

010 1

Reposted by Andrés Baselga

Erik Simon Lledó @bokamero.bsky.social · 23/09/2025

Evolutionary "hills" or "islands"? Very interesting new approach to measure the uniqueness of a biological community, with great potential use as baseline to identify high-value biodiversity sites at multiple scales.

153 5

Andrés Baselga @andresbaselga.bsky.social · 23/09/2025

This video summarizes the new framework we introduce to quantify and model phylogenetic uniqueness, and our main results for terrestrial vertebrates: youtu.be/dHMBhVJ8dns?...

youtu.be

Areas of high biodiversity value evidenced by the spatial scaling of phylogenetic uniqueness

YouTube video by Biogeography Lab (BiBiCI-USC)

000 0

Andrés Baselga @andresbaselga.bsky.social · 23/09/2025

Phylogenetic endemism is best to maximise global phylogenetic diversity in the smallest area, and spatial scaling of phylogenetic uniqueness identifies globally distinct but regionally less unique sites (evolutionary hills), from highly irreplaceable sites even at small scales (evolutionary islands)

100 1

Reposted by Andrés Baselga

Graciela Gil-Romera @gilromera.bsky.social · 05/09/2025

Gracias al programa CICERON del @csic.es mi colega y amigo @oscar-godoy.bsky.social y yo coordinamos este itinerario que persigue acercar la conservación y restauración de la naturaleza a la sociedad. youtu.be/WlJCXlgp310?... gracias! @actisandres.bsky.social @clementealvarez.bsky.social

youtu.be

CSIC Cicerón 13.8 Coordinación Itinerario Cicerón Agresta

YouTube video by CSIC Comunicación

2156 15

Andrés Baselga @andresbaselga.bsky.social · 05/09/2025

I would always appreciate colourblind friendly palettes, please!

020 2

Reposted by Andrés Baselga

The International Biogeography Society @biogeography.bsky.social · 02/09/2025

Abstract submission for the meeting in Aarhus has been extended by one week! The new abstract submission deadline is 8 September. conferences.au.dk/tibs-aarhus-...

032 3

Reposted by Andrés Baselga

GCiencia @gciencia.bsky.social · 02/09/2025

🐝 A captura masiva da 'Vespa velutina' con trampas non selectivas mata tamén insectos polinizadores ⚠️ É urxente apostar por métodos de control sostibles Así o explican investigadores da Universidade da Coruña neste artigo de @es.theconversation.com

gciencia.com

Por que a captura masiva de velutinas con trampas debe parar?

O uso indiscriminado de trampas contra a avespa asiática provoca a morte de milleiros de especies de insectos beneficiosos

042 4

Reposted by Andrés Baselga

CRETUS @cretus.usc.gal · 01/09/2025

Esta semana os IPs @andresbaselga.bsky.social e #C.Gómez-Rodriguez imparten “Código para biólogos: introdución a R” Curso da #UniversidadeDeVerán no que guiarán aos asistentes na aprendizaxe de R desde as bases ata a súa aplicación no ámbito da bioloxía. @biogeography-usc.bsky.social @usc.gal

022 2