Nuanced Soil Types

[skol]

Common "dirt" could be nuanced during world generation with respect to the geological profile (rock layers and erosion/sedimention) into clay soil (e.g. luvisol), sandy soil (e.g. podsol), chalky soil (e.g. calcisol), loamy soil (e.g. cambisol), silty soil (e.g. fluvisol), peaty soil (e.g. histosol), well-weathered soil (e.g. oxisol), sodic soil (e.g. solonchak) and ashy soil (e.g. andosol). Sand, clay and chalk could be the "pure" soil variants, as well as loam, silt, peat, laterite, salt and volcanic ash.

In addition the hydrological profile affects the fertility of the basic soil layer: beach, tideland, marsh, water meadow, ground water, rubble (raw or eroded, thus minimal water retention), climate (watered, dry, permafrost). Last but not least the basic biome vegetation itself influences the fertility of the soil by the amount of humus formation: desert, shrubland, grassland, forest, rain forest.

This is of course a simplification of soil typology (see e.g. https://de.wikipedia.org/wiki/Bodenart and http://www.ahabc.de/typen-co/ )

The distribution of different soil types could influence the gameplay with respect to

- vegetation diversity (consistent stratification and succession of plant communities)

- farming abilities of tilled soil (workability, specific growth speed and yield per crop species)

- soil improvement (from fertilization to crafting of garden soil)

- erosion abilities (drainage, desertification)

- indication of ores (bean ores, bog iron ore)

 

 

[Tyron]

I would love realism based soil layers and I studied that topic many hours already. Unfortunately this requires an incredible amounts of calculations while running the terrain generator, as i would have to recalculate many values (and have many other dependencies, such as the geologic province layer) for each horizontal coordinate - so 32x32 = 1024 calculations per chunk. Furthermore it is very complex to actually put this into one big formula that outputs the desired values, so for now I have a simplified system and I will focus on this again when more important components are done. It would be awesome if  the heathlands, peatlands, bog ore and the like would merely be emerging features from an big soil-determination-formula.

But! Here's what we have already: 

• Gelogic Provinces that determine the kinds of rocks and amount of rock layers. Example map (same color coding as on https://en.wikipedia.org/wiki/Geologic_province)
  
   
• Some very simple formulas for calculating the temperature, rainfall, fertility and soil thickness of a given horizontal coordinate. Which use these heuristics:
  • Higher elevation => more rain fall
  • Higher elevation => lower temperature
  • More rain and higher temperature => more fertile soil
  • Higher elevation => less fertile soil (erosion)
  • High rock weathing factor => thicker soil layer
  • Higher elevation => thinner soil layer
     
• Some other input maps are forest (white = heavy forest, black = no forest) and climate (red component = temperature, green component = rainfall):
   

[skol]

Cool. I love the way you tackle the matter. That's already a great foundation. This weekend I will play with the existing input parameters, and maybe I will be hit upon an idea of a somehow smart solution.