This was a very fun project to work on, with some neat stats & maths from fractals to fluid mechanics. We were surprised that for such a high profile dataset like sewage spills there had been such limited public statistical analysis until now! 💩 🧮 🧑‍💻

eartharxiv.org/repository/v... 8/8

This matters because water companies now need to create hydraulic models of thousands of CSOs to reduce their spilling… but unless they can model infiltration (hard!) they are likely to underestimate the frequency of the most damaging, long duration spills! 7/
www.gov.uk/government/p...

What causes the heavy tail? When trying to replicate it using hydraulic models we can only do so by including infiltration of groundwater into pipes through cracks… this suggests that these long duration spills are *not* only occurring after heavy rain…

eartharxiv.org/repository/v...

6/

We think this is driven by daily fluctuations in water usage (🚿 💩) which modulates the length of sewage spills…

eartharxiv.org/repository/v...

5/

A total (and cool 🤓) surprise for us, was the discovery of a periodic variation of the misfit to this modelled trend, fluctuating around spills that are multiples of 24 hours long… what’s causing this?

eartharxiv.org/repository/v...

4/

Statistically, this tail is well described by a stretched exponential distribution. It’s not quite as heavy as a power law, but much heavier than a vanilla exponential…

eartharxiv.org/repository/v...

3/

We find that sewage spills follow a very heavy-tailed distribution. 10% of spills account for 85% of all spilling *time*. So, it’s not the infrequent, short spills we should care about, it’s the infrequent long ones…

eartharxiv.org/repository/v...

2/

Is this all subjects? Undergrad & postgrad?

I mean c’mon you’re the CEH! You famously study things that are naturally messy and green and blue! Why opt for…. black with very hard lines?!

so pleased everyone is choosing such unique and distinctive typefaces

Finally, the scientific models used by agencies to model nutrient pollution are _remarkably_ opaque and under scrutinised given their ubiquitous importance (as I have written about previously: alexlipp.github.io/posts/2024/0...)

also the population change assumes (iirc) its all *new* population into catchment, rather than say residents moving *within* catchment (=no net increase). so, population density increasing due to limited supply would be a “win” for nutrient neutrality…

you’re right on the gist (nutrient neutrality is mad) but actually population growth *is* linear with nutrient outflow (ie it’s not a capacity thing). Problem is builders can’t pay to increase treatment efficiency at existing plants (!?) even though that can work…

The perfect conditions continue 😎 🌞!

… and blessed with some Aurora Borealis too! #studygeologyandseethenorthernlights

Focus on "universities" raising fees misses the point that the real injustice is the interest rates & repayment thresholds on student debt which I suspect plays greater role in determining how much is paid over a lifetime than the initial "up-front" fees...

...