Agricultural Productivity Potential

Variable APP of ten · scored for every 1 km site

---

The supply of farmland is essentially fixed. The demand for its output is not.

The world will need to feed roughly ten billion people by 2050, on roughly the same area of cultivable land, with less water and a more volatile climate. The long-run structural tailwind for high-quality agricultural land is as clear as any macroeconomic forecast can be.

Yet agricultural land remains systematically underpriced in many markets, valued on current yields rather than potential yields under improved management and infrastructure. Arcanium assesses what the ground can actually support, reading the land cover of each tile at fine resolution and aggregating it to the one-kilometer site.

The purpose is to separate genuinely high-potential farmland from land that merely occupies an agricultural classification.

What it measures

Three things, in one number: what covers the ground, what the soil beneath could grow, and for how much of the year the climate lets anything grow. The first two say what the site could carry; the third says for how long. A perfect 100 is rich, fertile soil under productive cover in a climate that grows all year. Any one of the three missing pulls the figure down; the climate missing altogether takes it to zero.

How the number is built

1. Cover. Every fine-resolution pixel in the site is read and mapped to how productive that kind of cover is — cropland high, pasture and grassland in between, forest, scrub, wetland and built ground low — and the site takes the area-weighted mean.
2. Soil, where it has been surveyed. Cover says what is growing, not what could. Where a soil-quality rating exists, it is blended in, in proportion to how much of the site’s cover is farmable at all: pasture on prime soil scores on the soil; a town on prime soil stays a town.
3. Season. The number of days a year the site is warm enough to grow scales the whole figure. A season of eight months earns full credit — long enough for any annual crop with margin. Beyond that, up to 15% more for ground that can grow all year, the double-cropping premium. Below it, proportionally less: a four-month season carries about half the potential of the same ground further south.

Why the season is warmth alone

A growing season could also be defined by rain, and then the most productive irrigated farmland on Earth would read as a quarter-season, because the rain does not fall there — the water arrives in canals. Whether a site has water is the Water Resource Index’s whole question. Asking it here as well would charge irrigated land twice for the same fact. So this variable asks only whether it is warm enough for long enough; whether it is wet enough is asked next door.

What scores zero

Ground that is wholly built, with nothing farmable in it; and ground whose climate has no growing season at all. A site can be excellent soil under productive cover and still score zero if nothing can grow there.

What is not measured

• Yield. This is potential, not production: it does not read harvests, prices or management.
• Irrigation and rainfall. Deliberately — see above.
• Soil outside the surveyed areas. Where no soil rating exists, cover and season stand on their own, and the site’s own detail says so.
• A changing climate. The season is a thirty-year average. It is what agronomy plans on, and it moves slowly; it is not a forecast.

How to read the score

~90	Productive crop cover on good soil in a year-round climate. The practical ceiling; nothing reaches 100.
~70	The same ground in an eight-month season — the corn belt’s best, scaled for its winter.
~35	Forest, scrub or wetland in a long season; or farmable ground in a short one.
0	Wholly built ground, or no growing season.

Every site’s own farmland line — its dominant cover, its crop share, its soil rating where surveyed, and its growing season in days — is shown in its dossier on the map and in the ground report.

← All ten variables