The Restless Caspian

A sea with no ocean tide

The Caspian is the world’s largest lake. Its surface now stands about 27–29 metres below ocean level, enclosed between the Caucasus, Iran, Central Asia and the East European Plain. Most of its water arrives through rivers, above all the Volga. Water leaves primarily by evaporation. If inflow overtakes evaporation, the lake expands; if evaporation wins, it contracts. An ocean can rise while the Caspian falls.

Around 21,000 years ago, enormous ice sheets held enough water to lower the global ocean by more than a hundred metres. Yet the Caspian’s late Pleistocene history includes the Khvalynian transgression, when lake water occupied the northern lowlands above modern ocean level. The lake was receiving and retaining continental runoff while distant shorelines across the world lay exposed.

Cold climate alone does not dictate the result. Lower temperature reduces evaporation, but glacial conditions can also reduce precipitation and river discharge. Ice-dammed drainage can redirect water between catchments. Meltwater can enlarge rivers without raising the global ocean immediately. Changes in the Volga basin, the Eurasian ice sheet, the Urals and possible eastern catchment connections all enter Caspian history. The lake records a balance assembled across millions of square kilometres.

The Black Sea and open ocean rose during deglaciation. The Caspian followed its own water balance, sometimes moving in the opposite direction. Blue bodies of water did not rise together.

The northern lowland magnified every metre

The southern Caspian occupies a deep tectonic basin. North of the Mangyshlak threshold, the geometry changes. The modern northern Caspian is extraordinarily shallow, much of it only a few metres deep. Beyond it lies the Caspian Depression, broad country below global ocean level. When lake level rises, the shore can travel hundreds of kilometres north across this low gradient. When it falls, an inland sea can disappear from the northern basin while deep water remains in the south.

Early Khvalynian deposits carry that former lake onto land. Marine and brackish molluscs occur in terraces and “chocolate clays” along the Lower Volga. At its large stages, the lake invaded river valleys and covered the Caspian lowland far north of the present delta. Traditional reconstructions divide the event into an Early Khvalynian highstand, reaching positive elevations above global ocean level, and lower Late Khvalynian stages.

The terrace numbers are spectacular. Shoreforms and deposits have been assigned to stages near +22 and +35 metres, while the highest classic reconstruction reaches +48–50 metres. That largest surface would flood the entire northern depression and push a long estuary up the Volga. But a terrace height is not automatically a precise date. Reworked shells, reservoir effects in radiocarbon, inherited sediment and correlations between distant outcrops have produced widely different chronologies.

A +35-metre late-glacial crest still creates an immense northern sea. The disputed +48–50-metre crest would push it farther again because the lowland magnifies every metre.

The Khvalynian water advanced

Optically stimulated luminescence has tightened the sequence. Instead of dating shell carbonate that may carry old carbon, OSL dates the last time mineral grains were exposed to light before burial. Redzhep Kurbanov and colleagues applied it to Lower Volga sections where marine chocolate clays sit in a clear sedimentary order. Their best-resolved late-glacial deposits place renewed Early Khvalynian marine sedimentation after the glacial maximum, broadly between about 17,000 and 13,100 years ago.

The sediments describe movement, not merely presence. Lower units record initial penetration of water; dense clay records a deeper bay; upper laminated material records regression and renewed river influence. The lake moved north, persisted over drowned lowland and then retreated. Newer work distinguishes stages in that motion while also demonstrating why a single age for the entire named transgression is misleading.

Alina Tudryn and colleagues compared the radiocarbon, uranium-series and luminescence ages across the basin. Firm chronological anchors coexist with stages dated largely by correlation between distant deposits. The disagreement changes the precise dates, but not the advance and retreat preserved in the Lower Volga sediments.

The advancing shore would have transformed ecology at continental scale. Steppe became shoal, river mouth and bay. Salinity shifted as freshwater expanded across the basin. Volga channels backed up and deposited fine sediment. Islands became peninsulas and then disappeared. Along the shallow north, each vertical metre redrew a long front between terrestrial and aquatic life.

The sea found an outlet

West of the Caspian, the Manych Depression cuts a low passage north of the Caucasus toward the Sea of Azov and the Black Sea basin. Today it contains rivers, saline lakes and reservoirs divided by subtle ridges. During high Caspian stages, it became the basin’s spillway. The evidence is material: sediments within the depression contain Caspian molluscs, carried beyond the enclosed lake’s modern watershed.

Daria Semikolennykh, Andrei Panin and Elya Zazovskaya dated shells from deposits tied to the final phase of flow. Their new AMS radiocarbon ages agree with OSL work and place the last discharge near 14,500 calibrated years ago. They reconstruct the Manych connection as functioning approximately 18,000–14,500 years ago. Older liquid-scintillation dates had supported a later end; the newer samples are single shells taken from sections with better geomorphic control.

The current-laid sands imply moving water at appreciable speed, measured in decimetres per second rather than a stagnant marine bay. Caspian water crossed the divide, ran west and entered the Azov–Black Sea system. In that interval the Caspian ceased to be endorheic. Its changing water balance found a threshold, and once the lake surface rose above that threshold, surplus water acquired a route out.

An outlet also changes the lake itself. Below the sill, inflow can accumulate across the northern lowland. At the sill, additional water escapes. The Manych passage therefore caps a transgression, turns stored water into discharge and connects two great basins whose other histories remain distinct.

Overflow and catastrophe are different claims

The Manych landscape contains ridges, scours and a cut channel that have inspired a “great flood” reconstruction. In that account, rapidly rising Khvalynian water broke across the divide and released an exceptional outburst toward the Black Sea. The volume available in the Caspian and the low corridor make violent flow physically possible. The dated molluscs and current-laid sand establish a real, energetic discharge.

They do not by themselves specify its hydrograph. The same deposits can belong to powerful flow sustained across years or centuries, to multiple episodes, or to a shorter outburst. The 2025 Manych study dates the end of the final discharge; it does not measure an instantaneous peak flow. Even the height of the threshold is reconstructed differently: some workers place the ancient divide near +45–50 metres and argue that overflow eroded it downward, while the modern runoff threshold lies below ocean level.

The lake may have risen across the northern plain over centuries while its first crossing of the sill produced a much faster local event. A catastrophic flood along Manych can coexist with gradual inundation elsewhere. Accepting the overflow does not require turning the entire Khvalynian transgression into one flood.

The distinction matters: a changing shoreline, a new connection and a violent discharge can unfold at different speeds. The Manych connection and its high-energy flow are secure. A short cataclysm remains one interpretation of their force and duration.

The floor became country again

After the late Khvalynian stages, the lake withdrew. The Mangyshlak regression of the early Holocene exposed the northern basin and cut valleys into earlier Caspian deposits. In deep-water cores and coastal sections, the event separates late Pleistocene Khvalynian conditions from the Neocaspian recovery. It reversed the geography: bays became isolated lakes, shoals became steppe, and the Volga ran farther south before reaching standing water.

The proposed minimum ranges widely, from a reference −50 metres relative to global ocean level down to −113 metres in the deepest reconstructions, with other workers favoring an intermediate −80 metres or lower still. The lowest contours contract the lake dramatically toward the middle and southern basins. Dating also varies. A short lowstand around 11,500–10,500 years ago appears in a south-basin core — the reconstruction mapped here — while coastal stratigraphies and regressional surfaces have supported later or longer spans.

The map here follows that shallower −50-metre reference; other reconstructions favor a deeper −80 metres, and some run the exposed floor down toward −113. Whichever figure is closest, every serious reconstruction exposes the northern floor on a vast scale and yields a Caspian much smaller than the Khvalynian lake.

As water returned during the Neocaspian transgression, the north flooded again. By roughly 6,500 years ago the lake had approached the range of later Holocene levels. This was not a simple postglacial drowning. The Caspian expanded, spilled, fell and expanded again.

A continental water balance

The lake’s movements convert climate into geography through a watershed much larger than the Caspian itself. The Volga supplies most modern inflow, collecting rain and snow from the East European Plain. Small changes spread across that area can add or remove enormous volumes at the lake. Evaporation responds to temperature, wind, humidity and surface area. As the lake expands shallowly northward, it creates more water surface from which to evaporate, introducing a feedback that can slow further rise.

Late-glacial meltwater offers one route to the Khvalynian highstand, but proposed sources differ. Modeling has tested enhanced runoff from the Volga basin, changing glacial margins and water reaching the Caspian from east of its modern catchment. Strontium and salinity records address which waters entered. Each mechanism must account for both the volume of the high lake and the timing recorded in sediment.

Compare the drowning of the wider Florida shelf: that ocean rose mainly because ice melted into a connected world sea. The Caspian rose or fell according to how water moved through its own basin. Both transformed hundreds of kilometres of shore, in opposite responses to the same deglacial world.

The inland shoreline in motion

The northern Caspian is a machine for turning vertical change into horizontal history. During high Khvalynian stages, water covered the depression and pushed into the Volga valley. At the Manych threshold, the enclosed lake opened west. During the Mangyshlak regression, its former floor returned to steppe. Neocaspian water then remade the shallow north.

None of these states is the lake’s timeless form. The present shoreline is one balance among inflow, evaporation and basin geometry. Terraces above it and drowned valleys below it are earlier balances written onto the same lowland. Disputed dates shift the precise crest or minimum; they do not change the lake’s fundamental restlessness.

The Caspian’s changing water balance repeatedly moved its shoreline across the northern lowlands.