Rock Water as a Key Resource for Patchy Ecosystems on Shallow Soils: Digging Deep Tree Clumps Subsidize Surrounding Surficial Grass

Abstract

Mediterranean mountainous areas of shallow soil often display a mosaic of tree clumps surrounded by grass. The combined role and dynamics of water extracted from the underlying rock, and the competition between adjacent patches of trees and grass, has not been investigated. We quantified the role rock water plays in the seasonal dynamics of evapotranspiration (ET), over a patchy landscape in the context of current and past seasonal climate changes, and land‐cover change strategies. Soil water budget suggests deep water uptake by roots of trees (0.8–0.9 mm/d), penetrating into the fractured basalt, subsidized grass transpiration in spring through hydraulic redistribution. However, in summer trees used all the rock water absorbed (0.79 mm/d). A 15‐year data set shows that, with increasing seasonal drought‐severity (potential ET/precipitation) to >1.04, the vertical water flux through the bottom of the thin soil layer transitions from drainage to uptake in support of ET. A hypothetical grass‐covered landscape, with no access to deep water, would require 0.68–0.85 mm/d more than is available, forcing shortened growing season and/or reduced leaf area. Long‐term decreasing winter precipitation and increasing spring potential ET suggest drying climate, so far with stable vegetation mosaic but progressively earlier peak of grass leaf area. Intervention policies to increase water yield by reducing tree cover will curtail grass access to rock moisture, while attempting to increase tree‐related products (including carbon sequestration) by increasing forest cover will limit water availability per tree leaf area. Both changes may further reduce ecosystem stability.

Multiscale Assimilation of Sentinel and Landsat Data for Soil Moisture and Leaf Area Index Predictions Using an Ensemble-Kalman-Filter-Based Assimilation Approach in a Heterogeneous Ecosystem

“…Using the EnKF, we assimilated observations of 1) NDVI, which is related to LAI through Equation ( 1) in the VDM that describes the evolution of the leaf biomass through (8) that is related to LAI through (12); and 2) the radar-derived dielectric constant of the ground, which is related to θ through (4), in the LSM that describes the evolution of θ using (5). Hence, in the proposed approach, the EnKF is applied distinctly to the LSM and the VDM of both grass and tree components (Figure 2).…”

“…An ensemble of saturated hydraulic conductivity values (k ζ s ) is generated as being log 10 normally distributed with a mean of log( ks ) (indicating with ks the base (i.e., best guess) value of the k ζ s ensemble) and the standard deviation of SD logks . In this way, an ensemble of θ ζ , which includes model errors, is generated and evolves in time according to (5).…”

“…Recent improvements in satellite remote sensing techniques obtain land information for ecohydrological studies at unprecedented fine spatial and time resolutions [1][2][3], which are also suitable for heterogeneous ecosystems, typical of semiarid and arid climates [4,5]. In these ecosystems, trees, grass, and bare soil components coexist, and grass and tree covers vary in time and almost randomly in space, at scales of the size of tree clumps (usually less than 50 m [6,7]).…”

“…Following Montaldo et al [38], we have chosen these parameters because the saturated hydraulic conductivity is a main soil-water balance parameter, which largely affects soil moisture predictions [54,55], and the vegetation growth predictions are highly sensitive to the maintenance respiration coefficients [56]. The proposed approach was tested in a Sardinian field site, largely employed in ecohydrological studies (e.g., [5,6,57,58]), where a grassland coexists with wild olive trees under waterlimited conditions, and the data of a micrometeorological eddy-covariance-based tower are available. In the data assimilation system, we used Sentinel 1 radar data for backscatter retrieval, Landsat 8 and Sentinel 2 data for NDVI estimates, and the LSM-VDM of Montaldo et al [14].…”

Climate Change Impacts on the Water Resources and Vegetation Dynamics of a Forested Sardinian Basin through a Distributed Ecohydrological Model

“…Typically, to survive dry conditions, trees optimize the use of water by extending their roots downward, where soils are wet for longer [39][40][41][42]. In Mediterranean basins, especially under water-restricted conditions, upstream hillsides are often characterized by trees growing on shallow soils above fractured bedrocks or cemented horizons, where tree roots can still enter through cracks, taking up water from the rock storage (up to 70-90% of transpiration) through the hydraulic lift process [40,[43][44][45][46][47][48][49]. Drier conditions can affect tree development, making past survival and water-use optimization strategies insufficient, and these need to be investigated further.…”