
build and run a simple computational model that links emissions, the carbon cycle and temperature, and use it with global climate model results to show how human choices change Earth’s systems.
Phase 3: Make Sense
Do Now. A weather forecast is unreliable beyond about ten days. Write one reason a climate forecast for 2050 can still be useful. (Hint: think about averages.)
Climate models are run over the past with ONLY natural drivers (the Sun and volcanoes). What do they show for 1950–2020?
What a climate model is. A global climate model divides the atmosphere, ocean, land and ice into millions of boxes and uses the laws of physics (conservation of energy, mass and momentum) to calculate how energy, water and air move between them, step by step, decades ahead. Models are tested by running them over the past, a hindcast: started in the 1800s with the known changes in the Sun, volcanoes and greenhouse gases, they reproduce the observed warming; run with natural drivers alone, they do not. We cannot build a full model in class, but we can build a small one that captures the core relationships, a computational representation of how Earth’s systems connect and how people are changing them.
The class model. Each step is one year. Its boundary is the whole Earth treated as one box of air; its starting condition is the 2025 Mauna Loa mean; its input is emissions; its outputs are concentration, forcing and warming.
Worked example 1. Start at the 2025 Mauna Loa mean, 427.35 ppm. Scenario 1 keeps emissions constant at 11 GtC per year. Each year C rises by 0.45 × 11 / 2.13 ≈ 2.32 ppm, so after 25 years C ≈ 427.35 + 25 × 2.32 ≈ 485 ppm in 2050. Forcing: 5.35 ln(485/280) ≈ 2.9 W/m2. Warming from CO2: 1.8 × ln(485/280) / ln 2 ≈ 1.4 °C above pre-industrial. If instead emissions fall steadily to half by 2050, C reaches about 470 ppm; if they grow 1% a year, about 494 ppm.
Worked example 2: checking the model against reality. For 2024 (424.61 ppm) the model gives 1.8 × ln(1.516) / ln 2 ≈ 1.1 °C from carbon dioxide alone. Observed warming is about 1.2–1.3 °C over the last decade. The difference is expected: the model leaves out methane and other greenhouse gases (which warm), aerosol pollution (which cools), and natural swings. A good modeler states what the model omits.
What the full models project. The IPCC’s 2021 assessment reports, for 2081–2100 compared with 1850–1900, best estimates of about 1.8 °C of warming in a very low-emissions scenario (SSP1-2.6), 2.7 °C in an intermediate one (SSP2-4.5) and 4.4 °C in a very high one (SSP5-8.5). The spread between scenarios, set by human choices, is larger than the uncertainty within any one. For snow-fed regions like the valley in Unit 1, warming means a smaller share of precipitation falling as snow and earlier melt, and more days of extreme heat.
Systems and human activity. Your model shows the chain: human energy use → carbon in the atmosphere → energy balance → temperature → oceans and ice (hydrosphere and cryosphere) → back to people through water supply, heat and sea level. Two links the class model hides: extra CO2 helps plants grow and take up some carbon (photosynthetic biomass rises), and the ocean’s uptake lowers its pH. Changing the first link changes all the others.
Build and run the model with a partner.
Write up the model in your Forecast Notebook.
Draw the chain of your model as a system diagram, from human energy use to carbon in the air, forcing, temperature, snowpack and sea level, and back to people. Label the boundary, the starting condition, the input and the outputs. Add the two hidden links (plant growth and ocean pH) as dashed arrows.
Use your three scenario runs to predict the 2050 carbon dioxide level and CO₂-only warming in each, and explain in three sentences how a human choice changes at least two Earth systems. Name one thing the model leaves out that makes its warming too low.
A three-line model already shows the main point: the future depends on the input we choose. Next lesson we check that every gigaton is accounted for.
How confident are you that you can I can use a computational model to show how human choices change Earth's systems.
Exit ticket: give the 2050 CO₂ level from two of your scenarios and one reason the full IPCC models show more warming than the class model.
Timing: this workshop spans two 50-minute periods: day 1 build and run the model; day 2 scenarios, the system diagram and the IPCC comparison.
Materials: devices with a browser (JupyterLite runs offline once loaded) or calculators and the every-fifth-year worksheet.
Listen for: “the model predicts the weather in 2050”; treating model output as certain; forgetting the model counts CO2 only.
Support / stretch: give the code pre-typed with blanks for E. Stretch: add a line that lowers the airborne fraction by 0.002 each year and explain what real change that would represent.