--- id: about/faq about: label: Questions people ask description: >- Plain answers to what a digital twin is for, from people who have asked: rehearsal, a shared picture, the hidden landscape, and who holds the model. pinned: 2 provenance: origin: machine sources: - title: 'Lewis Carroll, Sylvie and Bruno Concluded (1893), Project Gutenberg' url: 'https://www.gutenberg.org/ebooks/48795' - title: 'Elinor Ostrom (Wikipedia): design principles from Governing the Commons' url: 'https://en.wikipedia.org/wiki/Elinor_Ostrom' - title: 'Tim Harford, The lessons that flow from Bali''s water temples (2012)' url: >- https://timharford.com/2012/07/the-lessons-that-flow-from-balis-water-temples/ - title: Ozone depletion (Wikipedia) url: 'https://en.wikipedia.org/wiki/Ozone_depletion' reviewed: '2026-10-24' reviewed_by: >- machine draft by Claude from a conversation with Anselm (devlog/20261024-what-is-it-for.md); not yet rewritten by him status: proposed --- ## What is a digital twin for? A digital twin is a model of a place that can be checked against the place itself. People build one to answer a question they cannot answer by standing in the street. Four uses come up again and again. - **Rehearsal.** You can flood a model a hundred times, or move a road, before anyone does it for real. [eReefs](/twins/ereefs/) simulates the water, sediment and chemistry of the Great Barrier Reef from the river catchments out to the open ocean. - **One picture for everyone.** Arguments about a place usually happen with each person holding a different map in their head. A shared model puts the same map and the same numbers in front of all of them. [Virtual Singapore](/twins/virtual-singapore/) was built so that agencies could run their analyses on one model of the city. - **Seeing what is hidden.** Most of what shapes a place cannot be seen from the street: the groundwater, the floodplain, next summer's heat, and the [rules](/terms/rules-layer/) about what may be built where. [Will this property flood?](/everyday/will-this-property-flood/) is the simplest version: one house, and the maps, rain, rivers and ground around it. - **Knowing where it breaks.** Most systems bend for a long time and then change all at once. A model can show how far a place is from such a [tipping point](/terms/tipping-points/). The [thresholds chart](/vitals/thresholds/) shows this for the whole planet. ## Isn't it just a copy of the place? A perfect copy would be no use. In Lewis Carroll's *Sylvie and Bruno Concluded*, a visitor boasts that his country made a map "on the scale of *a mile to the mile*".[^1] It has never been spread out: "the farmers objected: they said it would cover the whole country, and shut out the sunlight!"[^1] A model is worth having for what it leaves out. The first step in building one is to [name the question](/guides/name-the-question/), and that decides what goes in. ## Don't people already know what is needed? Often they do, and the best example is a model that proved it. Bali's rice farmers have long shared their water through subaks, irrigation groups whose water temples set when each one plants. In the 1970s farmers were pushed to plant new rice as often as they could, outside the temple schedule, and the new system was overwhelmed by pests.[^3] The anthropologist J. Stephen Lansing and the ecologist James Kremer built a [computer model of 172 subaks](/articles/lansing-kremer-balinese-water-temple-networks/), and it found the temple schedules gave the best harvests of any it tried.[^3] Lansing writes that consultants evaluating the Asian Development Bank's irrigation project later acknowledged its conclusions. The farmers already knew. The model made what they knew visible to the people who had power over them. The [Bali scenario](/scenarios/bali-subak-water-temples/) asks what a model would show now, as terraces are sold for villas. ## Isn't more data the problem? We need to connect with each other. Both are needed. Elinor Ostrom studied commons that lasted, among them irrigation systems, fisheries and pastures, and set out eight principles they shared. One is monitoring: the people who check the state of the resource and what its users do are answerable to those users, or are the users themselves.[^2] Read that way, a lasting commons keeps a shared account of its water, fish or grass, which is a small model of a place kept by the community. A neighbourhood that is well connected but cannot see the aquifer under it can still agree on the wrong thing. A model nobody talks about is just a record. [Talk to the people affected](/guides/talk-to-the-people-affected/) starts there. ## Who holds the model? It matters more than anything else on this page. Many detailed models of places belong to the companies and agencies that work in them, and the people who live there rarely see them. A model is a public good when the people it concerns can see it, question it and run it themselves. That is why every engine, twin and data source here carries a note on what in it is open: the source, the data, and who can take part. ## Can a model show what a forest is worth? It can show what an ordinary price leaves out. A forest cleared for cattle shows up as income; the water it held, the soil it built and the fire it kept down usually do not, and someone else pays for their loss later. Some people are trying to count them: [regenerative GDP](/terms/regenerative-gdp-regdp/) treats living ecosystems as productive infrastructure on a national balance sheet. The [Flying D Ranch](/scenarios/flying-d-ranch-cattle-to-bison/) moved from cattle to bison in 1989 and is a place to compare the two. ## Has a model ever changed anything? The clearest case is the ozone layer. In 1974 Mario Molina and Sherwood Rowland proposed that CFCs could destroy ozone high in the atmosphere.[^4] In 1985 Joseph Farman's team reported the ozone over Antarctica falling each spring.[^4] Controls on the chemicals took effect in 1989, ozone levels levelled off by the mid-1990s, and a 2023 United Nations assessment expects the layer over Antarctica to be back at 1980 levels around 2066.[^4] A model and a measurement together made an invisible harm undeniable and its cause nameable. Harms whose costs are spread thin and come late, like carbon, soil and species loss, have been much harder. ## Can my town or neighbourhood have one? Yes, and it can start small. [Build the smallest version](/guides/build-the-smallest-version/): one place, one scenario, one output, shown to the people who asked. Many [everyday questions](/everyday/), such as how hot a street gets, whether a roof suits solar or what will grow in a garden, already have tools that answer them for one address. [Check it against what happened](/guides/check-it-against-what-happened/) before anyone relies on it. ## Do I need to be technical? No. The [guides](/guides/) start with naming the question and talking to people, and the [communication](/communication/) shelf collects ways of putting a model in front of people who will never open one. Someone has to do the technical part, but the question and the judgement about the answer belong to the people who live there.