Earth System Lab

Cinevva's analyzer read this game's design doc and its source, then wrote the brief below. "Make your own" hands it to the game creator as your opening message, so you start from a real spec rather than a one-line idea.

Create a browser-based interactive educational exploration called “Earth System Lab” / “The Changing Earth System,” focused on a West Kalimantan, Indonesia Earth-system case study. Use a dark futuristic Three.js-style 3D background scene with a fixed perspective camera looking diagonally across a stylized low-poly landscape: deep teal terrain, a bright blue Kapuas-like river, a brown/olive farmland plot, tropical palm trees, scattered rock/cliff props, a large raised dark-blue futuristic research-lab platform with a cyan glowing rim, pale translucent cloud puffs overhead, and many small orange satellite-fire hotspot markers surrounded by faint gray smoke spheres. Keep the scene atmospheric with navy fog, cyan/teal rim lighting, soft shadows, and a dark scientific UI aesthetic. There is no walkable player character or combat; the lab platform is the scene’s central object and the experience is controlled through the HUD.

Overlay the scene with a responsive HUD. At the top left show the small label “EARTH SYSTEMS LAB / WEST KALIMANTAN NODE” and the large title “The Changing Earth System.” At top right show a rounded pill with the current progress, “STAGE 1 / 8” through “STAGE 8 / 8.” Put a large translucent navy, cyan-bordered rounded information panel on the left/center, a smaller “RESEARCH NOTE” panel on the right reading “West Kalimantan node” plus the current stage title and “Correlation is not automatically causation,” and a horizontally scrollable bottom navigation bar with eight buttons: “1 · Conflict,” “2 · Cause & Effect,” “3 · Cloud Lab,” “4 · LUX Lab,” “5 · Timeline,” “6 · Graphs,” “7 · Systems Map,” and “8 · Generalization.” Make all navigation and controls clickable, with the active tab highlighted teal/cyan.

Implement these eight stages exactly as an interactive learning sequence:
1. Conflict: heading “Healthy Earth System → stressed Earth system.” Explain that the West Kalimantan landscape includes tropical forest, peatland, rivers, farmland, and communities, and that hotspot markers show possible satellite fire detections rather than an exact count of every fire. Show KPI cards for “4 Earth-system spheres,” “10 yr future scenario,” and “PM2.5 atmospheric focus.” Show the conflict statement that increasing wildfire activity can alter interactions between atmosphere, biosphere, hydrosphere, and human systems; El Niño is one climate influence, while human land use and local weather can interact with it. Display badges for ATMOSPHERE, BIOSPHERE, HYDROSPHERE, GEOSPHERE, and HUMAN SYSTEM.
2. Cause & Effect: present a clickable vertical chain of eight evidence statements: “El Niño / warmer & drier conditions,” “Reduced rainfall / drier vegetation,” “Higher fire risk,” “More wildfire hotspots,” “Smoke & atmospheric particles,” “Cloud / water-cycle processes,” “Reduced ecosystem health,” and “Earth-system change over time.” Clicking a chain item reveals its explanatory text, emphasizing that each arrow is a relationship to investigate and not proof of a single cause. Also show four explanatory cards for atmosphere, biosphere, hydrosphere, and geosphere.
3. Cloud Lab: show a virtual experiment chamber UI with a stylized bottle, lamp, cloud, sensor, and drifting smoke. The independent variable is aerosol/particle condition; dependent variables are LUX, cloud duration, and estimated cloud density. Provide three mutually selectable buttons: “A · Without aerosol,” “B · With aerosol,” and “C · Higher aerosol.” Visually change cloud opacity/density for the selected condition. Include a disclaimer that the cloud is a condensation visualization and not a direct PM2.5 measurement. Include controlled-variable text, a “Run trial” button, and a trial table recording up to three LUX trials per condition with averages. Use the model conditions: clear cloud density 18%, aerosol 46%, high aerosol 76%; trial duration values 11, 14, and 17 respectively.
4. LUX Lab: show the same lamp → bottle cloud → sensor diagram, with a range slider that moves the sensor horizontally. Explain that denser cloud means less light reaches the sensor, so LUX falls. Display live KPI values for LUX at sensor, particle condition, and model cloud density. Use the model formula equivalent to 620 × (1 − cloudDensity/125) × (1 − sensorPosition/120), rounded, so clear/aerosol/high conditions begin around 531/391/243 at the default sensor position and decrease as the sensor moves. Include the warning that LUX is light intensity and must not be converted into atmospheric PM2.5 concentration. Include two numeric fields for user-entered trial values and a “Save entered values” button that reports they are saved to the browser session.
5. Timeline: title it “2015–2026 timeline.” Provide clickable year buttons from 2015 through 2026 and an evidence card table with rows for Hotspot/fire, Rainfall/drought, ENSO/El Niño, PM2.5, and Haze/ecosystem note. Preserve missing data as “Data unavailable from selected source” rather than inventing numbers. Include the qualitative entries visible in the source: 2015 major drought associated with El Niño and reported El Niño conditions; 2019 drought-year association and reported El Niño; 2023 strengthening El Niño and prolonged meteorological drought in parts of Indonesia in Q3; 2024 regional haze reviews including Kalimantan observations; 2025 hotspot and smoke-haze activity affecting parts of Borneo; and 2026 hotspot data unavailable at build time. Include source links/labels for NASA FIRMS, ASMC annual VIIRS hotspots, BMKG climate data, and the WHO PM2.5 guideline.
6. Graphs: title the panel “Data graph dashboard” and state that evidence must not mix units. Show a dark canvas chart with buttons for “Hotspot availability,” “Experiment LUX,” and “Cloud duration.” Hotspot availability should use bars across 2015–2026, with available versus unavailable entries shown differently; do not fabricate hotspot counts. The local model graphs should show the three labels A clear, B aerosol, C high with LUX values 620, 390, 220 and cloud duration values 11, 14, 17; use a distinct green experiment color and orange for hotspot availability. Explain that hotspot products differ by satellite and detection method and that cloud cover, canopy, fire size, and satellite passes affect detection.
7. Systems Map: show a clickable list of ten variables: Human land-use activities, El Niño / ENSO, Temperature / dryness, Fire activity, Vegetation dryness, Wildfire hotspots, Smoke / aerosols, LUX in the experiment, Rainfall / water-cycle indicators, and Ecosystem condition. Clicking each reveals a short explanation. Make clear the bottle is a simplified light scattering/absorption model, not a direct atmosphere replica; satellite hotspots are thermal anomalies with detection limits; land use can include clearing, drainage, and burning choices; and rainfall/soil moisture connect atmosphere, hydrosphere, and land.
8. Generalization: show selectable evidence cards including denser model cloud reducing LUX, NASA FIRMS/ASMC hotspots being detections with limitations, BMKG’s El Niño/drier-conditions evidence, WHO PM2.5 guideline values of 5 µg/m³ annual and 15 µg/m³ 24-hour, smoke reducing visibility and stressing ecosystems, and the limitation that LUX is not a PM2.5 sensor. Let users toggle cards as selected. Display the evidence-based generalization that changes in one Earth-system component can affect interconnected components; dry climate conditions and human activities can increase wildfire risk, fire can produce smoke and particles affecting atmosphere and ecosystems, and drought/reduced water availability can increase vulnerability. Add a “2026 → 2036” scenario with eight clickable action buttons: Fire prevention, Peatland protection, Sustainable land management, Reforestation, Water management, Community education, Satellite monitoring, and Air-quality monitoring. Start the projected fire-risk index at 50 and ecosystem resilience at 50, and increase resilience by 5 per selected action up to 95; show projected fire risk as 100 minus resilience, ecosystem resilience, and community readiness as resilience plus 20 capped at 100. Label these as scenario projections, not guaranteed predictions. Finish with an improvement plan explaining that LUX does not directly measure PM2.5, why that is so, the improvement of using a particulate-matter monitor, and the expected effect of obtaining direct PM2.5 measurements.

Use a clean sans-serif UI, very dark navy translucent panels, cyan borders and headings, mint-green active states, amber evidence/disclaimer notes, muted blue-gray secondary text, rounded buttons, compact tables, and subtle glow effects. The design should feel like a polished interactive science exhibit rather than an arcade game. Ensure it works responsively: on narrow screens hide the right research-note panel, make the main panel fit the viewport, and keep bottom tabs horizontally scrollable.