 |
| Titel |
Environmental signals in tree-ring δ18O from a temperate catchment in Switzerland |
| VerfasserIn |
Annika Oertel, Kerstin Treydte, Dominik Michel, Elisabeth Tschumi, Ansgar Kahmen, David Frank, Sonia Isabelle Seneviratne |
| Konferenz |
EGU General Assembly 2017
|
| Medientyp |
Artikel
|
| Sprache |
en
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 19 (2017) |
| Datensatznummer |
250145201
|
| Publikation (Nr.) |
 EGU/EGU2017-9118.pdf |
|
|
|
|
|
| Zusammenfassung |
| Oxygen isotopes (δ18O) in tree rings are a valuable proxy for past environmental conditions.
Yet, the contribution of source water δ18O versus signals generated at the leaf level as well as
the influence of tree-physiological responses on tree-ring δ18O differences between
individual trees at a site remain uncertain.
To address this topic, we conducted a study at a catchment research site in northeastern
Switzerland. Its unique long-term sampling design allowed for bi-weekly δ18O measurements
of precipitation and creek water for the 2002 to 2014 period. Four ash trees (Fraxinus sp.)
situated at a creek and four on a nearby steep slope were selected for δ18O measurements of
tree-ring cellulose for the same 13 year period. δ18O of soil water as well as cryogenically
extracted stem and twig xylem water were determined for three days within the 2016
vegetation period for comparison of xylem and soil water δ18O between the slope and
the creek site. Gas exchange measurements with a LI-COR (Li-6400) allowed for
comparison of transpiration rate, stomatal conductance and productivity between
sites.
We calculated correlations to environmental variables and applied the mechanistic
Péclet-modified-Craig-Gordon (PMCG) model to simulate observed δ18O cellulose values
while varying the parameterization of physiological and environmental variables according to
the measured values.
Mean inter-series correlations between the tree-level δ18O time series are similarly high
at the slope and the creek locations, and both site-chronologies are tightly correlated (r=0.9)
although offset by 0.9 ‰ on average. Both chronologies contain a similarly strong
summer VPD/RH signal, but we find correlations to precipitation and creek discharge
δ18O are just as high. Our results suggest that i) both leaf-level and source water
signals are imprinted in cellulose δ18O, and ii) in addition to leaf-level evaporative
enrichment the VPD signal at least partly results from its correlation to precipitation δ18O
(r=0.57).
The cellulose δ18O values are on average 0.9 ‰ higher at the slope site compared to the
creek site despite similar stem xylem and soil water δ18O. Although it is commonly assumed
that no fractionation occurs during transport from the roots to the twigs, we find that twig
xylem δ18O is about 1 ‰ higher than stem xylem δ18O. Simulations with the PMCG
model (which captured the overall behaviour in the δ18O time series well) with
varying parameterization for stomatal conductance and twig xylem δ18O according to
measured values show that about 50 % of the 0.9 ‰ offset between creek and slope is
explained.
We conclude that i) it is unlikely that differences in source water δ18O cause the
0.9 ‰ offset between both sites alone, ii) despite the proximity of both sites the
transpiration rate and productivity of trees at the creek site is higher, and iii) according
to the PMCG model tree physiological responses are important for site internal
tree-ring δ18O variability although interannual variations are only slightly affected. |
|
|
|
|
|
|