Carbon Related Data Science - nsporillo/GlobalWarming GitHub Wiki

This project may one day entail interesting data science work. Designing a model that is representative of real life for use in game requires relationships between in-game concepts and real world concepts.

Excellent Fossil Fuel Data

Consumption

For instance, tracking coal consumption will be trivial since we're storing all the data we need.

Similarly, we might want to track coal production too. This is also trivial.

Reserves

Minecraft, unlike the real world, knows ahead of time the overall coal reserves in the explored world. Suppose the server administrator sets a world border and all of the chunks are explored, then we can analyze every chunk to determine the total number of coal blocks. However, 1 block of coal ore does not equal 1 piece of coal in Minecraft. So, like the real world, there is some fuzziness in our data.

Minecraft Ore Height Distribution

Source

How can we use this information to fine tune our models?

Suppose there is an arbitrary world border set for the map, we can detect this world border programmatically. Now we can estimate the exploitable world's fossil fuel reserves. Furthermore, we can apply this same logic to estimate the total available land that can be used to offset coal consumption via tree planting.

Since players can plant a sapling, wait for it to grow naturally, and then cut the tree back down. We can design our model to assume that if in a small world with lots of consumption, there ought to be a fair bit of tree farming. An important note is that we calculate the carbon reduction from tree growth in a non-conventional way. Instead of tracking the life of the tree (an expensive task) we calculate the carbon reduction based on the total size of the tree which grew and subtract the reduction immediately.

Per the Minecraft Wiki:

All trees in the active chunk radius around the player make attempts to grow at random intervals. For any given tree this can work out to about 3 growth attempts per minute. When a tree attempts to grow, it first checks that it has enough light, then randomly chooses which variant of that species of tree to become; for example, an oak sapling will choose to grow as either a small or large oak tree.

It's a reasonable trade-off between realism, performance, and fun. Players dont want to wait years for the effects of their behavior to be realized in-game.

What about bonemeal?

Configurable options regarding bonemeal. It can be eliminated as a fertilizer for saplings, or have it's successful fertilization rate slashed to balance gameplay.

What about items and decoration made from wood? wood fired furnaces?

Items and decoration are considered to be sequestered carbon. Unless players throw the wood materials into lava or fire, then it should not impact the atmospheric carbon dioxide.

Furnaces fired with wood or wood based items are tracked and have a carbon emission defined in our model.

Conservation of Mass

Minecraft basically ignores this, so we will have to consider the implications of our model. Players can store 54 stacks of 64 wood in a 2x1 chest (also made of wood) nearly endlessly. We don't want that to be the only option, but we certainly dont want the model to aggressively induce a sense of helplessness (regardless of how real life is).