The research, led by Wanting Pan of Peking University’s State Key Laboratory of Vegetation Structure, Function and Construction, together with Chongyang Xu of Inner Mongolia University, Jingyu Dai of the University of Georgia, and senior author Hongyan Liu of Peking University, examined native broad-leaved woody plants across natural gradients of soil thickness and bedrock fracture development in Guizhou Province. The team’s central insight was that the geochemistry of the underlying bedrock, particularly its calcium and magnesium concentrations, offers a surrogate for the physical architecture of the subsurface. Limestones rich in calcium weather differently from magnesium-rich dolomites, producing distinctive combinations of soil depth, soil porosity, and fracture networks. By sampling along these geochemically defined gradients, the researchers could disentangle the influence of soil water storage from that of fracture-stored water.

The technical approach relied on a suite of water-use-related functional traits measured directly on plants. Leaf-level traits, including measures associated with conservative water use such as water-use efficiency and stomatal regulation characteristics, were combined with stem-level hydraulic traits that reflect investment in water transport infrastructure, including xylem anatomy linked to cavitation resistance and conductive capacity. In parallel, the team quantified three indicators of subsurface water-holding capacity at each site: soil thickness, soil porosity, and the thickness of the fractured bedrock layer. Soil thickness and porosity determine how much water the fine-earth reservoir can retain against gravity, while fracture thickness governs the volume of epikarst, the weathered upper zone of rock, where water can be stored in cracks and fissures and tapped by deep-rooting plants.

The results revealed a clear hierarchy of control. Bedrock differences, operating through their mediation of soil thickness and fractured layer thickness, explained between 10.4 percent and 25.5 percent of the spatial variation in plant water-use strategies. That is a substantial fraction for a field where aboveground climate variables have traditionally dominated explanations of vegetation behavior. More striking was the asymmetry between the two subsurface compartments. Soil thickness exerted consistently stronger effects than bedrock fractures on plant water use, while fractures played a weaker, though still positive, role. In other words, a thick, porous soil profile does more to relax plant water stress than an equivalent volume of fractured rock beneath it.

The trait patterns told a coherent story of adaptation. Plants growing in true karst habitats, defined by higher bedrock calcium concentrations, thinner soils, and more developed fracture networks, displayed more conservative leaf water-use strategies. Conservative strategies, in plant physiological terms, mean tighter stomatal control, lower stomatal conductance, and typically greater intrinsic water-use efficiency, the ratio of carbon gained through photosynthesis to water lost through transpiration. Such plants effectively trade rapid growth for survival, closing their stomatal apertures early as soil moisture declines to avoid the dangerous drop in leaf water potential that can trigger hydraulic failure. This confirms that where soils are shallowest, trees behave as careful water savers rather than aggressive spenders.

Meanwhile, thinner soils correlated with higher bedrock magnesium concentrations and greater investment in hydraulic structure. This pattern suggests that plants on dolomitic substrates, where soils tend to be thin and rocky outcrops abundant, compensate by building costlier xylem systems, with features that protect against cavitation, the formation of air embolisms in water-conducting vessels that can shut down water transport during drought. The connection between magnesium-rich parent material and hydraulic investment adds a novel geochemical dimension to the well-documented global pattern in which woody plants exposed to chronic water limitation evolve denser, safer wood, sometimes at the expense of conductive efficiency.

The study’s framing situates it within a rapidly evolving body of research on rock water use by plants worldwide. Recent work, including a landmark 2021 survey in Nature, established that woody plants across diverse ecosystems routinely tap water stored in bedrock, and subsequent studies in the American Southwest and Mediterranean climates have shown that rock moisture can sustain transpiration through months-long dry seasons. Southwest China’s karst, covering hundreds of thousands of square kilometers and supporting some of the world’s most extensive carbonate terrains, represents an extreme case because the bedrock is not merely a passive water tank but an actively weathering, geochemically dynamic medium. Previous work by some of the same authors had shown that bedrock geochemistry influences vegetation growth by regulating regolith water-holding capacity, and that rock crevices help determine woody plant cover in the karst critical zone. The new study extends this line of inquiry from ecosystem-scale productivity down to the level of individual functional traits.

The significance of distinguishing soil from fracture contributions goes beyond academic taxonomy of plant strategies. Karst ecosystems in Southwest China have undergone dramatic transformations over recent decades, including large-scale afforestation programs that contributed substantially to China’s greening trend and carbon sequestration, alongside persistent challenges of rocky desertification, a process in which soil erosion exposes barren bedrock and degrades ecosystem services. Because these landscapes are projected to face increasingly variable rainfall and more intense compound droughts under climate change, understanding which subsurface reservoir buffers vegetation against dry spells directly informs predictions of forest resilience. The study’s results suggest that sites with thicker soils confer greater drought buffering to their vegetation than sites whose storage capacity lies mainly in fractured rock, partly because soil water is more readily accessible to fine roots and less subject to rapid drainage through macro-pores into the deep karst aquifer.

Methodologically, the use of bedrock calcium and magnesium concentrations as proxies for subsurface physical structure is an elegant solution to a stubborn measurement problem. Directly mapping soil thickness and fracture networks across rugged, boulder-strewn terrain is labor-intensive and spatially limited. Because limestone and dolomite weather along predictable pathways, the concentration of these alkaline earth metals in the parent rock encodes information about how much regolith accumulates and how extensively the epikarst fractures. This opens the possibility of scaling trait-based predictions of drought vulnerability across entire karst regions using geochemical maps, an approach that could be integrated with remote sensing of vegetation condition and with earth system models that increasingly seek to represent water storage in bedrock.

The authors emphasize that their findings provide a structured framework for evaluating belowground ecological processes that have long been overlooked in shallow-soil landscapes globally, not only in karst. Thin-soil ecosystems on volcanic, granitic, and sedimentary substrates all feature some combination of soil and fractured bedrock storage, and the relative balance between the two likely shapes plant strategies in each. The quantified contribution of bedrock-mediated variation, roughly a tenth to a quarter of the observed trait variation, establishes that lithology is not a background variable but an active determinant of plant ecology. At the same time, the clear dominance of soil thickness delivers a practical message for restoration and management: protecting and rebuilding soil profiles in degraded karst may do more to secure the future of these forests than any intervention targeting the rock itself.

The work, supported by the National Natural Science Foundation of China under grant 41571130044 and conducted in collaboration with the Puding Karst Ecosystem Observation and Research Station, marks a step toward a more complete accounting of the hidden water economy of forests. As the researchers conclude, the pivotal role of soil thickness and the weaker yet positive role of bedrock fractures together refine our ability to assess how heterogeneous karst ecosystems will weather the climatic variability ahead. In a region where millions of people depend on karst forests for water regulation, carbon storage, and soil retention, knowing that the thinnest soils host the most frugal trees is knowledge with immediate consequence.