Lenia Orbium artificial life visualization by Raven Carrigg. QGIS digital elevation model visualization of the Boring Lava Fields in Oregon by Raven Carrigg.

Complex systems · Computational art · Visual modeling

Systems and Simulations

Computational art, artificial life, cellular automata, terrain visualization, remote sensing, environmental data, and experiments with language, number, and form.

Systems and Simulations gathers Raven Carrigg’s computational art, complex systems experiments, remote sensing visualizations, mathematical modeling, and interdisciplinary research work.

These projects grow from a background in applied mathematics, computational modeling, environmental data analysis, creative coding, research, and documentation. They explore how form emerges from rules, data, feedback, spatial structure, nonlinear interaction, and time.

Across artificial life, cellular automata, LANDSAT time series, social dynamics, digital elevation models, biomechanics, and computational poetics, these projects treat language and number as neighboring abstractions for noticing pattern, tracing transformation, and giving shape to complexity without making it simple.

Artificial life

Lenia and Orbium

Lenia is a continuous cellular automaton for artificial life, developed by Bert Wang-Chak Chan. Orbium is one of its most recognizable self-organizing lifeforms: a small computational organism whose motion and persistence emerge from local update rules.

Lenia Orbium. Artificial life visualization inspired by Bert Wang-Chak Chan’s Lenia system, where local update rules produce motion, persistence, and organism-like behavior.
Still image from a Lenia Orbium artificial life visualization by Raven Carrigg.
Emergence as motion. A still from the Orbium simulation, where continuous local interactions create form, apparent agency, and a small luminous creature of computation.

Reference: Bert Wang-Chak Chan, Lenia: Biology of Artificial Life, 2019.

Read the Lenia paper on arXiv.

Remote sensing

LANDSAT time series

These time series are processed remote-sensing visualizations created with LANDSAT imagery in SentinelHub. They use false-color, natural-color, and NDVI treatments to show environmental and spatial change across decades: glacial retreat, vegetation recovery, urban expansion, agricultural mosaics, volcanic landscape recovery, and changing water boundaries.

Bear Glacier, Alaska. False-color LANDSAT time series showing glacial retreat, ice movement, coastal change, and shifting boundaries between ice, water, and land.
Bozeman, Montana. NDVI LANDSAT time series showing vegetation change, agricultural mosaics, urban expansion, and mountain-valley spatial structure.
Mount St. Helens, Washington. False-color LANDSAT time series showing eruption-scar contrast, vegetation recovery, long-term regrowth, and landscape reorganization.
Las Vegas and Lake Mead, Nevada. Natural-color LANDSAT time series showing urban growth, desert development, shifting water boundaries, and long-term Lake Mead change.

Computational poetics

Language, recombination, and nonlinear memory

This project treats personal archival writing as a living text system: a field of fragments, motifs, images, repetitions, and emotional signals that can be filtered, recombined, and re-arranged into new poetic structures. In this work, computational poetics became a way to listen for patterns across time: recurring images, emotional echoes, ruptures, contradictions, and fragments of language that began to speak differently when placed in new relation to one another.

Cover of Liminal Afterhood: Fragments from the Edge of Survival by Raven Indigo Carrigg.

Archival fragments, Markov-inspired recombination, and GNOEMs

The method began with more than 111,000 words of personal writing spanning over a decade. From that archive, metaphor-dense fragments were selected and processed through phrase-level recombination inspired by Markov chains and n-grams, producing cross-era constellations of language.

The final poems were manually selected, arranged, edited, and shaped into GNOEMs for the poetry collection Liminal Afterhood, where computational recombination became a way of listening for nonlinear memory, trauma crosstalk, recurrence, rupture, and emergent meaning.

The term GNOEM is used here as a personal adaptation of generative poem, inspired by the human-computer poetry experiments documented at Gnoetry Daily.

Social systems

Integrate-and-fire network model of group dynamics

This project explores human group dynamics as an integrate-and-fire network driven by LLM-based sentiment classification of text data. Granger causality analysis identifies significant directional relationships, lags, and influence patterns between sentiment dyads, which are then used to shape the simulated network.

Text data was classified using Claude 3.5 Haiku to generate sentiment signals for the simulation. The model is exploratory rather than predictive: it uses time-series analysis, sentiment dynamics, and firing thresholds to investigate how local affective signals can produce emergent group-level patterns, phase shifts, and critical transitions.

Integrate-and-fire group dynamics model. A simulation of activation, inhibition, directional influence, and threshold-driven firing in a network of sentiment-derived signals.
Granger causality analysis plot showing significant directional sentiment dyads and lag structure used to inform an exploratory group dynamics model.
Granger causality analysis. Significant sentiment dyads, lag structure, and relative influence patterns were used to inform network connections, timing, and firing thresholds.

Biomechanics

Squat biomechanics study

This independent biomechanics study modeled attainable back squat depth as a function of body-segment length ratios, ankle dorsiflexion range of motion, and joint constraints, then tested the model against Notch® accelerometer and video-frame data from 80 healthy adults.

Squat biomechanics study preview showing Notch sensor placement and video-frame movement analysis.

Attainable squat depth, anatomy, and movement constraints

This project combines geometric modeling, constrained optimization, anthropometric measurement, wearable-sensor data, and video analysis to examine why squat depth varies across bodies.

The study found that ankle dorsiflexion and lower-body-to-torso segment ratios were meaningfully associated with attainable squat depth, challenging one-size-fits-all standards of squat technique.

Terrain visualization

Digital elevation models

These QGIS-based digital elevation model visualizations translate terrain into color, relief, drainage, and structure. They show landform as pattern: volcanic fields, river islands, channels, ridges, and subtle surface memory.

QGIS digital elevation model visualization of the Boring Lava Fields in Oregon by Raven Carrigg.
Boring Lava Fields, Oregon. QGIS digital elevation model visualization of volcanic terrain, drainage traces, surface relief, and small landforms made visible through color.
QGIS digital elevation model visualization of Kiger Island near Corvallis, Oregon by Raven Carrigg.
Kiger Island, Corvallis, Oregon. QGIS digital elevation model visualization of river-shaped land, channel memory, floodplain structure, and subtle elevation change.

Undergraduate thesis

Advection, forest turbulence, and carbon fluxes

Raven’s undergraduate thesis developed a theoretical model of sensible heat transport in heterogeneous forest terrain, using micrometeorological data from a mature Douglas-fir site in the Oregon Coast Range to examine advection, turbulence, storage, and energy balance closure.

Undergraduate thesis preview showing a forest sensor network in heterogeneous terrain.

Is Advection Important?

This thesis evaluates whether common simplifying assumptions, including horizontal homogeneity, uniform gradients, and zero advection, hold in complex forest terrain.

The work connects fluid dynamics, boundary-layer meteorology, ecological measurement, and carbon-cycle questions by exploring whether sensible heat advection can help explain subcanopy carbon loss.

Cellular automata

Cyclic cellular automata

Cyclic cellular automata use simple local rules to produce traveling waves, spirals, domains, collisions, and other emergent spatial patterns. These images explore rule-based pattern formation through color, repetition, and nonlinear interaction.

Cyclic cellular automata visualization by Raven Carrigg using rule 1/1/16/M.
CCA rule 1/1/16/M. A maze-like cyclic cellular automata pattern with nested color bands, sharp wavefronts, and rule-driven spatial structure.
Cyclic cellular automata visualization by Raven Carrigg using rule 1/3/3/M.
CCA rule 1/3/3/M. A warm spiral field where local update rules generate curling fronts, flame-like forms, and repeating spatial turbulence.
Cyclic cellular automata visualization by Raven Carrigg using rule 2/5/3/N.
CCA rule 2/5/3/N. Blocky domains and color territories emerge through neighborhood interactions, local thresholds, and simple competition.
Cyclic cellular automata visualization by Raven Carrigg using rule 2/11/3/M.
CCA rule 2/11/3/M. Dense spatial turbulence appears as flecks, fronts, and competing patches chase one another through cyclic transitions.
Cyclic cellular automata visualization by Raven Carrigg using rule 3/5/8/M.
CCA rule 3/5/8/M. Smooth spiral waves and soft gradients unfold as cyclic states propagate, collide, and reorganize across the grid.
Cyclic cellular automata visualization by Raven Carrigg using rule 3/10/2/N.
CCA rule 3/10/2/N. Large blue-green domains form through simple local competition, producing slow boundaries and broad emergent territories.

Cyclic cellular automata were developed by David Griffeath. Rule notation and examples here follow the cyclic cellular automata framing described by Softology.

Read Softology’s cyclic cellular automata overview.