Optimal Traits of Plant Hydraulic Architecture for Rock-Dominated Landscapes

Optimality models can only be as good as assumptions about the relevant constraints on plant function. To date, Dynamic Global Vegetation Models (DGVMs) have utilized relatively simple representations of the rhizosphere, chiefly assuming uniform, thick soil without restrictions to root development. In reality, many terrestrial landforms have features that severely impede root growth. These include habitats with shallow or skeletal soils over bedrock, karst or caliche. Experiments have shown that plants in these habitats are not limited to using soil water, but use a variety of strategies to extract water from rocky substrates, e.g., growing extended structural roots along rock crevices into soil pockets or perched water tables, developing flattened root mats inside planar fissures or associating with mycorrhizae to extract water directly from the rock matrix. While these strategies expand plant-available water sources beyond soil, the added pools are expected to have extraction and recharge characteristics quite different from soil. Here I ask how the dynamical differences in non-soil water pools should influence plant hydraulic traits.

I built upon earlier work to determine how predictions of optimal plant function types change when model details are adjusted to reflect water uptake from non-soil sources. The model is a hydraulic continuum model based on Darcy’s law with optimization parameters representing biomass allocation between leaves, stems and roots, variable stem water storage capacity, and sensitivity of leaf and root conductivity to water potential. The rhizosphere is represented by two dynamically distinct water pools, the first representing a component with quick recharge and depletion (remnant soil), the second a non-soil component with restricted root density, potentially high storage capacity but possibly low hydraulic conductivity. The prediction of optimal plant functional types was significantly altered for non-soil compared to soil substrates, indicating a shift towards more extreme shallow-rootedness or stem water storage. Implemented within the context of a DGVM, these modifications produce lower productivity but higher drought resilience for vegetation in rock-dominated landforms and vegetation dynamics unlike that produced by merely depth-restricted soil.