Geoscience Reference
In-Depth Information
for an extended calibration and veri
cation period, we present the results of a
modeled precipitation reconstruction and
flood dating based on subfossil-oak, tree-
ring data from 3250 BC to 2250 BC. The researched time period was chosen due to
the fact that the HOC displays a noticeable drop in replication around 2750 BC and
a pronounced rise in raw total ring width (TRW) since 2731 BC. As a result of the
origin of the wood samples, the reconstruction emphasizes the situation of the Main
region (MR) in Germany (49
E). In addition to TRW data, which had
been used for the seasonal precipitation reconstruction, wood anatomical anomalies
(WAA) had been used as a proxy for strong and long-lasting
°
45
N, 9
°
30
′
′
ood events.
2 Materials and Methods
Chronologies from TRW from living (1876 to 2004 AD) (Land
2013
) and subfossil
(3250 to 2250 BC) (e.g. Friedrich et al.
2004
) oak trees from the MR were used in
this study. To preserve high- to mid-frequency variability, tree-ring series were
individually detrended using 67 % cubic smoothing splines with 50 % frequency
response cutoff (Cook and Peters
1981
). Annual indices were calculated as ratios
from the
fitted growth curves. Variance adjusted chronologies (living and subfossil)
were generated using a bi-weight robust mean (Cook
1985
). Signal strength of the
two chronologies was assessed using a moving window of 50 years (with 25 years
overlap) of the inter-series correlation (RBar) and expressed population signal
(EPS) (Wigley et al.
1984
).
Additionally all living and about 20 % of the subfossil oak samples had been
analyzed for characteristic changes in the wood anatomy which can be regarded as a
reliable proxy for long-lasting
flood events (e.g. Astrade and Begin
1997
; Land
2013
;
Tardif and Bergeron
1997
). WAA in oak trees are formed in the submerged parts of a
stem. The characteristics of WAA allow for differentiation between summer and
winter
floods (Astrade and Begin
1997
; Land
2013
; Tardif and Bergeron
1997
). In
order to identify the exact timing of the growth response of TRW indices to the
climatological trigger an algorithm programmed in MATLAB (Schoenbein
2011
)
was used. Based on climate data in daily resolution (e.g. temperature, precipitation,
hours of sunshine, drought index) the script aggregated the data for each year altering
(a) the length of the data sample used for correlation between 21 and 181 days and (b)
the starting date of the sample derived from (a) between January 1st of the previous
year and December 15th of the current year (Schoenbein
2011
). A Pearson correlation
between each generated data set and the TRW index for the MR was calculated.
Highest correlations were selected and response functions were generated. A sig-
ni
cant relation was found between TRW index and the sum of the daily precipitation
from April 1st to July 10th. The linear climate-growth model derived from the cor-
relation [Precipitation (mm) = 56.97 + 127.81
×
TRW] had been applied to the TRW
index from subfossil oak trees.
Search WWH ::
Custom Search