The Geography of Cities: Why They Form Where They Do

When Alexander the Great chose the site for Alexandria in 331 BCE, he reportedly walked the city’s future boundaries himself and marked them with barley flour because his surveyors had run out of chalk. He chose a narrow strip of land between the Mediterranean and Lake Mareotis on the Egyptian coast, and he chose it deliberately: a natural harbor on the sea, a freshwater lake serving as an inland port, and the proximity of the Nile Delta’s agricultural wealth. Alexandria became one of the ancient world’s greatest cities not because Alexander was lucky but because the geographical logic of that location was overwhelming. Any city founded there would have had advantages; Alexander’s founding simply accelerated an outcome that geography had been preparing for millennia.

The relationship between geography and urban formation is one of the most consistent patterns in human history. Cities do not appear randomly on the landscape. They cluster at points where geographical features reduce the cost of moving goods, people, and information. When transportation technology changes, some cities grow and some decline, but the underlying logic that governs where cities form has remained remarkably stable across ten thousand years of urban history.

The Confluence Principle

The most reliable predictor of city formation across world history is the river confluence. Cities form where rivers meet, where rivers meet the sea, and where rivers become navigable after a cascade or falls. The reason is straightforward: before roads, rivers were highways, and where two highways intersected, trade concentrated.

Pittsburgh sits at the confluence of the Allegheny and Monongahela rivers, which form the Ohio. The French built a fort there in 1754; the British seized and rebuilt it. When industrialization arrived, Pittsburgh’s river access made it the natural center of steel production: coal could come down one river system, iron ore down another, and finished steel could move by water to markets in every direction. The geography that made it a contested military site in the eighteenth century made it an industrial capital in the nineteenth. The city did not create its advantages; it inherited them from the landscape.

Khartoum sits at the confluence of the Blue and White Nile, which made it the natural administrative capital of Sudan regardless of which colonial power controlled the territory. Cairo sits at the apex of the Nile Delta, the point where the river fans into its distributaries, which made it the gateway between Upper and Lower Egypt and therefore between sub-Saharan Africa and the Mediterranean world. Wuhan sits at the confluence of the Han and Yangtze rivers, which made it the natural center of central China’s internal trade for centuries before industrialization arrived to confirm the choice.

The pattern repeats across every continent because the underlying physics does not change. Water transport is an order of magnitude cheaper than land transport using animal power. Before steam, before rail, before motorized vehicles, the city that sat astride river confluences had a structural cost advantage in trade that no amount of human ingenuity could overcome by founding a competing city in an inferior location. The geography was destiny.

Natural Harbors and the Sea-City Relationship

The cities that dominated world trade for most of recorded history were not river cities but harbor cities, and the most dominant harbor cities were those with natural features that made them simultaneously accessible to ocean shipping and defensible against attack. The ideal harbor was deep enough for large vessels, sheltered from storm winds, narrow enough to defend with shore fortifications, and located near productive hinterland that generated exports and demanded imports.

Carthage occupied a peninsula in the Bay of Tunis with a natural harbor that could be blocked at its entrance by a chain. The city built an additional artificial circular harbor for its warships behind the commercial port. The geography gave Carthaginian merchants and sailors advantages over competitors that the city exploited across three centuries of Mediterranean commercial dominance. When Rome destroyed the city in 146 BCE, it was acknowledging that the geographical advantage of the site was so great that leaving a competing power there indefinitely was strategically unacceptable.

Constantinople’s geographical advantages were so extreme that the city remained the most important urban center in Europe and western Asia for over a thousand years. The site sits where Europe meets Asia, where the Black Sea meets the Mediterranean, and where a triangular peninsula of land creates a naturally defensible position requiring fortification on only one side. The Golden Horn, the natural harbor on the city’s northern edge, was deep and sheltered and could be closed with a chain across its entrance. The city controlled the only sea passage between the Mediterranean and the Black Sea, which meant that every trader moving between those two worlds paid whatever toll Constantinople set.

The Byzantine Empire held this position for a millennium largely because the geography was so advantageous that even military incompetence and political dysfunction could not immediately dislodge it. When the Ottomans finally took the city in 1453, they did not relocate the capital; they kept Constantinople and renamed it Istanbul, because the geographical logic of the site was as compelling to them as it had been to Constantine eleven centuries earlier.

The Defensibility Constraint

Geographical advantages for trade often conflict with geographical advantages for defense, and the cities that survived long enough to become great were those that found sites offering both. The conflict between accessibility and defensibility is one of the fundamental tensions in urban geography, and how different civilizations resolved it explains much about their urban patterns.

Greek cities tended to cluster around an acropolis, a defensible high point that could serve as a refuge when the lower city was threatened. Athens’ Acropolis was not chosen as a temple site randomly; it was an existing fortress around which the city grew. The commercial port at Piraeus was five miles away, down a long fortified corridor, because the geography that made Piraeus a good harbor made it indefensible without the hinterland fortress. Athens was thus a city of two geographies stitched together by long walls.

Medieval European cities often chose sites that appear economically suboptimal from a pure trade perspective: hilltops, river bends that created natural moats, peninsulas with narrow necks. Carcassonne sits on a ridge above the Aude river plain in southern France not because the ridge is a convenient trading location but because it was defensible. The trade routes adjusted to pass through Carcassonne because the city existed there, rather than the city existing there because the trade routes had already chosen the site. Defense determined location; trade adapted.

The pattern shifted with the coming of gunpowder, which made traditional stone walls obsolete, and later with the growth of states powerful enough to provide external military protection to cities in inferior geographical positions. Once cities could rely on national armies rather than their own walls, the defensibility constraint relaxed, and pure trade logic began to dominate site selection. The result, visible in the nineteenth century, was explosive growth in ports and rail junctions that would have been militarily untenable in earlier centuries.

The Railway Correction

The industrial revolution represents the largest single disruption to city geography in recorded history, because railways changed the relative cost of land versus water transport. Before railways, moving a ton of goods one mile by road cost roughly thirty times more than moving it the same distance by water. Railways reduced land transport costs dramatically, but not to the level of water transport: early railways were still more expensive than river shipping per ton-mile.

What railways did change was the geographical flexibility of industry. Manufacturing no longer needed to cluster at water-accessible points. It could locate anywhere a railway line could reach, and railway lines could reach anywhere terrain permitted. The result was the growth of inland cities that had previously been limited by their lack of water access.

Birmingham in England is the paradigmatic case. The city had no river navigation of significance, no natural harbor, no geographical advantage visible to a pre-industrial analyst. Its growth into the world’s first major industrial city was a product of the canal network that preceded the railway and then the railway system itself. When transportation technology changed the cost structure of moving goods, Birmingham’s location at the center of England’s population suddenly looked like an advantage rather than a liability.

Chicago’s rise in the mid-nineteenth century reflects a similar dynamic. The city sat at the southern tip of Lake Michigan, which gave it water access, but the decisive factor in its growth was its position as the natural rail hub connecting eastern manufacturing to western agriculture. Every railway line from the east seeking to penetrate the agricultural interior hit Chicago as the logical transfer point. The city’s geographical position relative to the rail network was determined by terrain, not by any decision that Chicago’s founders made. They did not know they were founding a rail hub; they were founding a trading post on a lake. The railway system, following the logic of terrain, found them.

Why Geography Persists

Modern transportation technology has reduced the cost advantages of specific geographical sites enormously compared to pre-industrial periods. Containerization, air freight, interstate highways, and digital communication have all reduced the penalty of being in a geographically suboptimal location. Yet the largest cities in the world still cluster at the same types of sites they occupied three thousand years ago: natural harbors, river confluences, and coastal plains adjacent to productive agricultural hinterland.

The reason is agglomeration. Once a city forms at a geographically advantageous site, it accumulates human capital, institutions, infrastructure, and commercial networks that become advantages in themselves, independent of the original geographical logic. New York’s harbor advantage has been diminished by containerization and air freight, but New York remains the dominant financial center of the Western hemisphere because the concentration of financial expertise, legal infrastructure, regulatory relationships, and commercial networks that accumulated when the harbor mattered continues to compound.

Geographical advantages create first-mover advantages in urban formation, and those first-mover advantages create institutional accumulations that outlast the original geographical rationale by centuries. Alexandria’s library and scholarly culture persisted as an advantage for the city long after cheaper sea routes had reduced the importance of its specific harbor position. Constantinople’s role as a commercial entrepot continued under the Ottomans because the merchants who knew the trade routes and had the relationships were already there.

The implication is that the study of city geography is not merely antiquarian. Understanding why cities formed where they did illuminates why they remain where they are. The geographical logic is now partially embedded in institutions and networks rather than in raw physical advantages, but the initial geographical selection still explains why those institutions and networks formed in those specific places. The landscape shaped the city; the city then preserved itself against a landscape that had changed around it.

Every city is, in the end, a bet that its founders made on a particular reading of geography. The cities that prospered were those where the geographical reading proved correct, or where agglomeration forces were powerful enough to sustain the city after the original advantages faded. The cities that vanished were those where the geography turned out to matter less than expected, and where nothing else accumulated quickly enough to replace it.