UC Berkeley

Our ability to forecast future climate scenarios and manage our water resources requires an understanding of the fluxes of transpired water in forested areas. In addition to quantifying the amounts of water moving through trees, we also need to know where different species of trees get their water and how responsive species are in changing their source water to drought or other climatic factors. The critical zone (CZ), which on hillslopes can be defined as the area between the tree tops and the top of underlying fresh bedrock, has emerged as an important multidisciplinary framework by which to study the flux of water up into trees, or down into runoff generating pathways. A common approach to defining the water sources used by the vegetation employs the use of different stable isotope ratios of water (2H/H and 18O/16O) in both the various reservoirs and in the trees; such information can also be exploited as a hydrologic tracer for any given catchment or in the CZ. This study was conducted at “Rivendell”, an intensively instrumented site at the South Fork Eel Critical Zone Observatory. Rivendell spans a north-south divide. The North Slope is an unchanneled 32° hillslope draining to Elder Creek, upstream from the confluence with the South Fork Eel. The South Slope is 25°, and is a large head scarp of a deep-seated landslide draining directly to the South Fork Eel. Soils are thin (10-70 cm), and are underlain by saprolite (a friable, soil-like material of variable thickness with relict rock properties), vertically dipping weathered bedrock of argillite and sandstone above fresh bedrock. The depth of the transition from weathered bedrock to fresh bedrock varied from 6 m near Elder Creek to approximately 26 m at the hillslope divide. The site is within the Coastal Belt of the Franciscan Formation. The climate is Mediterranean, and the bulk of the 1535 mm of annual rainfall that occurred during the study period fell between October and May.

I used nearly 2000 samples of natural abundance water isotopes to quantify the variation in potential sources of water used by the vegetation and to trace precipitation inputs through the different materials of the critical zone: soil, saprolite, and weathered bedrock and to follow subsurface mobile waters through the weathered zone to a seasonally dynamic groundwater that then drained to Elder Creek. All runoff occurred as groundwater discharge. Most samples were collected bi-weekly from June 2011-January 2013. The isotopic composition of the shallow soil showed strong seasonally dynamics, shifting from evaporatively enriched (heavier than “average” rainfall) in the summer (detected by having more positive H and O stable isotope ratios) to depleted (lighter than “average” rainfall or the mobile water fraction collected with lysimeters) in the winter (detected by having more negative H and O stable isotope ratios). In contrast, the isotopic composition of the saprolite behaved oppositely: it was lightest in the summer, and shifted toward the average annual rainfall values during the winter, only to return to very light values progressively through the summer. Deep weathered bedrock was always distinct in being isotopically light (negative). Mobile water sampled from lysimeters passed through the soil in winter but due to mixing within the saprolite and weathered bedrock of the critical zone, by the time the water reached the groundwater it was isotopically invariant and appears to have become homogenized such that its value was equal to the average annual precipitation isotope values. As a consequence, the groundwater and resulting Elder Creek flow were also isotopcially nearly invariant. Surprisingly, water held more tightly within the interior of argillite (but not sandstone) is very negative isotopically, falling well to the left of the Local Meteoric Water Line (LMWL)–the precipitation inputs that fell on to the site during the study period. The persistently negative isotopic compositions of critical zone materials suggest the possibility of either inherited paleo-meteoric water or post-precipitation isotope effects could be influencing the isotope values of water at depth and that these post-precipitation isotope effects may be common, yet are largely unaccounted for, in eco-hydrologic applications. Despite a transient isotopic response due to precipitation inputs, the soil moisture, rock moisture, and mobile water have distinct isotopic compositions that lead to a structured heterogeneity that persist over multiple years. Thus, the material properties in the critical zone impart an isotopic ‘fingerprint’ on their waters that should be traceable into the xylem of vegetation growing on the site.

During the study period (2011-2013), I continually sampled tree xylem from 33 Douglas-fir (Pseudotsuga menziesii), 8 madrone (Arbutus menziesii), 16 live oak (Quercus wislizenii), and 5 tanoak (Notholithocarpus densiflora) trees to determine the source water of different species of vegetation growing across the north-south divide. The δD of the Douglas-fir matches that of the saprolite and weathered bedrock moisture, but not the soil.

In order to determine the source of Douglas-fir water, we conduced a targeted tracer experiment in which D2O enriched water was slowly added via 12 pipes to the saprolite in early June, 2014. I measured the isotope composition of xylem water of 17 trees and found the “spiked” isotope tracer in two Douglas–fir but not in the hardwoods. This tracer experiment showed that Douglas-firs do use rock moisture. However, the long term monitoring shows that the δ18O composition of Douglas-fir xylem water is commonly more positive than rock moisture and no mixing line of observed sources can explain this seemingly deviant heavy δ18O. Based on these combined results I hypothesize that the dominant water source used by Douglas-fir xylem may have experienced fractionation in the subsurface before being taken up by the trees. This fractionation may occur in the critical zone due to uptake of rock moisture through mycorrizhae (that are symbiotically connected to Douglas-fir roots), or (less likely) there may be mixing within the bole of the Douglas-fir that leads to a positive δ18O offset. If true, the isotope composition of Douglas fir xylem water would not be a reliable indicator of source water.

In contrast to the Douglas fir, hardwoods on this site appear to be opportunistic water users. Hardwoods use mobile moisture and bulk soil water in the winter, and the evaporatively- enriched bulk soil moisture, and, to a lesser extent, non-evaporatively enriched saprolite moisture, in the summer. An analysis of an adjacent Douglar fir and madrone tree showed that even the roots from the different trees that crossed each other in the subsurface and the adjacent boles of the trees both in summer and after winter rains held isotopically distinct water. The roots of an individual tree, especially in summer, are isotopically diverse and size dependent, indicating that the bole values record a mixture of sources. South Slope hardwoods may show signs of greater reliance on deeper, non-evaporatively enriched saprolite moisture in late summer. If this finding is supported with further observations it would suggest that these trees compete with Douglas-firs for rock moisture.

The use of mobile water by the hardwoods on our study site, the strong summer diurnal oscillations in stream flow in Elder Creek, and the documented rise in stream runoff (including summer baseflow) after clear cut logging in nearby experimental watersheds all demonstrate that forests use water that would otherwise contribute to streamflow. Subsurface isotopic dynamics of the critical zone moisture causes the retained moisture to be isotopically distinct from streamflow, but the mobile that drains to depth and mixes enroute results in a groundwater (and resulting streamflow) that is nearly invariant isotopically (and similar to annual rainfall values). Thus, forests do use water that contributes to runoff and yet can be isotopically distinct from streamflow. These observations challenge the recently proposed concept of only ‘two water worlds’: a hypothesis that argues that plant available moisture is hydrologically disconnected from stream flow.

The forest canopy at our study site and in much of the northern California landscape is mixed hardwoods and Douglas-fir. The opportunistic hardwoods can extract moisture to lower water potentials, giving them access to different water sources than conifers like Douglas-fir can use and thus, in severe drought, may effectively compete with Douglas-fir that prefer, and may require, using water sources that are at a higher water potential. Such a competition needs to be investigated as it matters to future forest composition under drier and warmer future climates in the Northern California Coast Ranges.

Collectively, the results of my investigations point toward the importance of material properties in influencing the isotopic composition of the waters they hold, and to vegetative water-source use that is species-specific. The observed isotope effects require future study to determine the mechanisms responsible for the chronically negative rock moisture of the soil and weathered bedrock, the cause of the invariant Douglas-fir xylem water that lacks similarity with any subsurface reservoir, and the possibility of hardwood use of rock moisture and subsequent competition with Douglas-fir trees at some places and at some times in the watershed.