How landscape memory, hysteresis shape the way Indian cities flood

Next to the highway lies a lake but the boundary between water and land has blurred. What was once contained spreads across the wetland, dampening the mud path joggers run on, seeping into the road beneath the churning wheels of buses, cars, and motorcycles.

It is water that seems out of place, yet it moves with familiarity, following paths the land remembers long after they have been paved over.

What does it mean for a landscape to remember rain? 

In cities across India, streets remain waterlogged long after the downpour has passed.

These familiar scenes are often dismissed as failures of human-made drainage systems or excess rainfall. But hydrology offers an additional insight: landscapes don’t respond to rain instantly or forget it quickly. Instead they retain a memory of past rainfall, shaping how water moves through soils, wetlands, rivers, and cities.

This phenomenon is called hydrological hysteresis.

Memory of water

Hydrological hysteresis describes how a landscape’s response to rainfall depends on the current volume of rainfall as well as on past events. A catchment that has already absorbed weeks of monsoon rain will behave differently from a parched one, even if both receive the same amount of rain today.

Water is stored over time in soils, aquifers, wetlands, and floodplains but they release it at different rates. As a result, the relationship between rainfall and river flow is not linear. It shifts as the land moistens and dries.

When dry, a sponge will absorb water readily; but once it’s saturated, adding just a little more water will cause the sponge to leak. Similarly, dry soils and wetlands fill as the monsoon begins and excess water is stored in the soil and adjacent vegetation. As the rains continue, the soils and wetlands near saturation and infiltration drops. Rain water that might have been absorbed increasingly becomes runoff, leading to flooding even without rainfall intensifying.

When rivers outgrow their banks

Monsoon floods in India are often described as simple responses to heavy rain, but rivers don’t react to rain alone. They respond to how water reshapes and occupies the landscape over time. This evolving interaction between flow and land gives rise to hydrological hysteresis, and explains why rivers behave differently when floods rise or recede.

As monsoon rain intensifies, river channels fill rapidly. Water levels rise because more water enters the system and because flow accelerates and pressure builds within the channel. At this time, much of the surrounding floodplain is disconnected and the river is largely confined, with its energy directed downstream.

Once water exceeds the river’s banks, the system changes. It spreads laterally into the floodplains, wetlands, abandoned channels, and low-lying agricultural land. Large volumes shift from fast-moving channels into slow or near-stagnant floodplains. As sediment settles and flow slows, the local hydraulic gradients flatten.

These changes persist even as rainfall weakens. Floodplains don’t drain instantly back into rivers. The stored water seeps slowly through soils, reentering channels via backwaters or remaining ponded for weeks. As groundwater levels rise, drainage is further delayed.

When a river returns to a given water level on the falling limb, it is physically different from when it first reached that level.

A river thus acquires memory through altered storage and resistance across its landscape.

When lakes spill into the city

In October 2024, the Kogilu and Doddabommasandra lakes in Bengaluru’s Yelahanka area overflowed after days of sustained rainfall. Water breached nearby roads, including stretches of the Outer Ring Road. At first glance, the cause seemed straightforward: the lakes had filled up, leaving little capacity to absorb additional stormwater.

But what unfolded was not simply a matter of full lakes. It was a path-dependent response of the city’s drainage system.

As rain water accumulated, lake levels rose while remaining largely contained. Stormwater drains continued carrying runoff into the lakes. Once water crossed a critical elevation, however, the system shifted. Lakes spilled laterally into roads and open land, submerging drains that had previously acted as outlets. Water was now stored outside basins, on the streets and in saturated soils.

When rainfall intensity dropped, lake levels fell back toward earlier values, but flooding did not recede at the same pace. At the same lake levels that caused no flooding on the way up, roads remained inundated on the way down. Water trapped on urban surfaces drained slowly, constrained by saturated ground, flattened gradients, and submerged or clogged drains. The system no longer behaved as it had earlier in the event.

Bengaluru’s history helps explain this. The interconnected lakes established during Kempegowda’s rule in the 16th century were once linked by natural streams and wetlands that allowed water to spread and return gradually. But over time, these connections were straightened into concrete channels and the city built over the floodplains. The result was a system that filled quickly, spilled abruptly, and emptied slowly — leaving floods to linger even after a rain had eased.

Land’s memory

Hydrological hysteresis shows why rainfall totals alone are poor indicators of flood risk. Rivers and cities respond to how wet the landscape already is, which is why floods often arrive suddenly or persist long after the rain has stopped.

For policymakers, this goes beyond reactive flood control towards basin-scale planning. Urban lakes, wetlands, and floodplains are not redundant spaces but critical infrastructure that store water early in the monsoon and release it gradually. As climate change intensifies rainfall, recognising the land’s hydrological memory will matter more than engineering responses alone.

Priya Ranganathan is a doctoral student at the Ashoka Trust for Research in Ecology and the Environment who studies freshwater swamps in the Western Ghats.

Published – March 03, 2026 07:15 am IST