
Yossi (Yosef) Ashkenazy
Edward and Bertha Rose Professor of Environmental Physics
Department of Solar Energy & Environmental Physics
The Jacob Blaustein Institutes for Desert Research
BenGurion University of the Negev
Sede Boqer campus
84990 Midreshet BenGurion
Israel
Email: ashkena@bgu.ac.il
Tel: 97286596858 (work)
Tel: 97286529857 (home)
Fax: 97286596921 
Research Interests
 Climate dynamics and modeling
 Sand dunes dynamics and modeling
 Paleoclimate
 Statistical analysis (linear and nonlinear) of climate
records
Recent Projects
Stochastic
nonlinearity of paleoclimate time series:
The climate of the last
800,000 years (800kyr) is characterized by pronounced
glacialinterglacial oscillations with approximate period of 100kyr.
Most researchers link the glacial dynamics to changes in the
distribution of the solar radiation (insolation) received by Earth.
Other researchers showed that glacialinterglacial records have
rednoise spectrum that usually characterize stochastic processes. In
our studies we show that these “erratic” time series have in fact
unique pattern for the stochastic fluctuations: Climate fluctuations
(e.g., of temperature or icevolume) appear in clusters of big and
small fluctuations where these clusters are self similar and obey
scaling laws. The clustering phenomenon is related to the
nonlinearity of paleoclimate time series and to the multifractal
nature of these time series. We developed several stochastic models
that account for this stochastic nonlinearity and showed that some
other
climate stochastic models do not reproduce this observed stochastic
nonlinearity.
The midPleistocene
transition of iceages:
Glacialinterglacial oscillations were
very different prior to ~800kyr ago. Before 800kyr ago the
oscillations in icevolume were of smaller amplitude (about 2/3)
compare to icevolume oscillations of the last 800kyr, and with
approximate period of 40kyr compare to the ~100kyr period of the
glacial oscillations of the last ~800kyr. We developed a simple
piecewise linear model to explain this midPleistocene transition and
showed that the glacial dynamics before and after the transition is
essentially the same; only the maximum icevolume threshold was
different. We obtain a good fit with paleoclimate records when adding
insolation forcing.
Lead and lag of the
tropics in glacial dynamics:
Recent proxy records from the
tropical region indicated the lead of tropical temperature over
global icevolume by several thousands years. We raised the
possibility that such tropical lead is due to seasonal bias of the
foraminifera from which the local temperature and global icevolume
are constructed; i.e., insolation of different seasons have relative
lag/lead between them and thus foraminifera that tend to grow at
certain season reflect this lead/lag. In this case the observed
tropical lead has nothing to do with glacial dynamics.
Box modeling of the
Eastern Mediterranean Sea:
The deep water formation in the
Eastern Mediterranean changed drastically in the last 10 years or so:
a new additional source of deep water formation was found in the
southern part of the Aegean Sea and is 3 times larger than the old
source of deep water formation in the Adriatic Sea. We develop a
simple box model to explain this Eastern Mediterranean transient and
show that such transition is consistent with multiple equilibrium
states of the fluxes between the different basins of the Eastern
Mediterranean Sea. Such transition may be caused by altering the
moisture and heat fluxes of the different boxes.
Statistical properties
of river flow fluctuations:
We study the statistical properties
of 30 world rivers. We found that river flow fluctuations obey
scaling laws and have nonlinear properties including volatility
correlations. We developed statistical descriptive model to
understand the possible sources of these nonlinear properties. We
also develop filtering procedure that efficiently excludes the
seasonal periodicity of climate records. [In some cases inadequate
filtering of the annual periodicity of climate records may lead to
erroneous conclusions regarding the underlying processes.]
