In geological and hydrological systems, the relationship between a trigger—such as rainfall or groundwater extraction—and its physical effect is rarely instantaneous. Instead, these systems exhibit hydraulic hysteresis, a phenomenon where the state of a system depends not only on its current environment but also on its past history. This "memory" manifests as a temporal lag between external hydrological changes and internal mechanical responses, complicating our ability to predict disasters like landslides and land subsidence.
The Mechanism of Pore Pressure Lag
The primary driver of this delay is the time required for water to move through porous media, a process known as seepage. When rainfall hits a slope or river levels rise, the water does not immediately weaken the entire geological mass. Instead, it infiltrates the soil and gradually alters the internal pore water pressure.
As Lan et al. explain, hydraulic hysteresis models must account for this temporal gap because the reduction in soil shear strength—the force keeping a slope from sliding—occurs only as water seeps deep into the mass. This creates a measurable time lag between changes in river discharge and the resulting movement of adjacent landslides. Clements frames this mechanical response as a diffusion process, where the rate of response is governed by the "hydraulic diffusivity" of the local geology. In this framework, mountainous regions with high diffusivity react rapidly to water input, while sedimentary basins exhibit much slower changes due to prolonged water retention.
Antecedent Rainfall and System Memory
Because of this lag, a single day of heavy rain is often less dangerous than the cumulative effect of weeks of moderate precipitation. This is known as effective antecedent rainfall.
Qiu et al. demonstrate that landslide displacement is often driven by a weighted summation of past rainfall events, where the influence of older rain gradually attenuates over time. This cumulative soil moisture reduces shear strength far more effectively than sudden, isolated storms. Similarly, Ge et al. found that the cumulative rainfall over the two days preceding an event is a critical threshold for triggering disasters.
This "memory" effect is not limited to soil. Zhang et al. note that groundwater systems respond to precipitation with delays influenced by the thickness of the vadose zone (the unsaturated soil layer) and aquifer permeability. Even large-scale features like lakes show this lag; Wang et al. observed that the lowest water levels in Lake Qinghai can occur a full year after the lowest annual precipitation.
Hysteresis in Land Subsidence
Hysteresis also plays a critical role in land subsidence, particularly in urban areas where groundwater is extracted. When water is pumped out, the surrounding soil layers compact. However, this deformation does not stop the moment pumping ceases.
Yong et al. highlight a significant challenge in subsidence prevention: soil deformation often continues even after groundwater levels have stabilized. This delayed mechanical response means that surface sinking can persist for years after mitigation efforts begin. Similarly, Du et al. found that surface deformation in mining regions persists long after industrial activities end, driven by the slow compaction of underground cavities and ongoing fluctuations in groundwater.
Implications for Early Warning Systems
Understanding the temporal distribution of triggers is essential for safety. For example, Wu et al. found that "pre-peak" rainfall patterns—where intensity is highest at the start—are particularly destabilizing because they saturate the soil early, leading to faster failure.
Furthermore, different types of precipitation introduce different lag scales. Liu et al. point out that snow acts as a reservoir, creating a dual-phase lag: first, an immediate increase in gravitational load from the snowpack, followed by a much later infiltration of meltwater into the soil.
Current early-warning systems often struggle because they lack integration of these hydraulic lags. As Lan et al. argue, incorporating time-dependent seepage effects is necessary to provide precise evacuation windows. Without accounting for the delayed response of the Earth, our models will continue to be surprised by disasters that were set in motion weeks or months before the first crack appeared.
Key Foundational Work
Landslide Identification and Monitoring along the Jinsha River Catchment (Wudongde Reservoir Area), China, Using the InSAR Method (2018) — contributed 1 concept elements to this direction.
A global analysis of soil moisture derived from satellite observations and a land surface model (2012) — contributed 1 concept elements to this direction.