- There is Only One CloudHow a visual illusion became a scientific assumption
- cloudyview↗3D cloud field visualization toolkit — three tiers from fast optical depth to photorealistic Monte Carlo path tracing in Mitsuba 3.
- Global sonde datasets do not support a mesoscale transition in the turbulent energy cascade↗Structure functions of global radiosonde wind data find no spectral transition at the mesoscale, supporting an anisotropic-cascade theory in which large flat circulations continuously deform into more circular ones at smaller scales. First independent test of the Lovejoy–Schertzer cascade.
- The ignored lesson of cloud shapeClouds are scale invariant. That means we are modeling climate wrong.
- On form and pattern in fractal cloudsFractal geometry and its application to the natural world
- The new AI for science is differentAI systems that output explanations or algorithms should not be conflated with prior systems that guess a solution to a problem.
- Multifractal ExplorerGenerate scale-invariant fields in-browser. Three parameters (α, C₁, H) plus anisotropy and outer scale.
- scaleinvariance↗Python package for multifractal field simulation and exponent analysis. FIF, fBm, fractional integration, anisotropy, CPU/GPU support.
- objscale↗Fractal dimensions and size distributions for 2D binary arrays. Useful for cloud masks and other 2D object analysis.
- Too many exponentsEmpirical and analytical relationships for the many exponents defining scale invariant fields
- scaleinvariance-wasm↗Rust/WebAssembly port of the scaleinvariance simulation API. Powers the in-browser Multifractal Explorer.
- Experiments in automationReflections on a year of coding agents in science
- Toward less subjective metrics for quantifying the shape and organization of clouds↗Two distinct fractal-dimension metrics with clear physical interpretations are superior to subjective categories, but there are common pitfalls in estimating their values.
- How to visualize scale invarianceA walk-through of simulating multifractal fields using the Fractionally Integrated Flux algorithm.
- Climate, visualized, from space5 great visualizations of the difference between weather and climate
- Beware your emotional firewall
- The other superintelligenceHuman collectives as a kind of superintelligence
- Business DensityExplore business density heatmaps across different cities.
- Anisotropic TurbulenceParticle flow through stratified atmospheric turbulence.
- Photon PathInteractive ray tracing through simulated clouds.
- Finite domains cause bias in measured and modeled distributions of cloud sizes↗Previous disagreement in cloud-size scaling exponents trace primarily to finite-domain truncation artifacts. A correction procedure based on filtering bins where truncated clouds dominate is proposed.
- Scaling ExplorerVisualize how correlation dimension and Haar fluctuations are computed on images.
- Climatologically invariant scale invariance seen in distributions of cloud horizontal sizes↗Cloud size distributions follow power laws across more than five orders of magnitude with remarkably stable scaling exponents across season, latitude, ocean vs. land, and satellite platform.
- Updates to Thought CloudWhy Thought Cloud is becoming an interactive, non-linear map — and moving from Substack to my own site.
- Eddy Mixing3D circulation advecting a passive scalar — vortex tubes, rings, and chaotic flows. Hill's vortex and ABC flow.
- Fields, objects, and a rabbit hole to get lost inIsn't a field simply an object in a space with one higher dimension?
- Clouds, their edges, and how to define themAn overview of our recent study on the theory of cloud edge
- How do you know I am a human? And how do I?