My
personal (non-work) homepage
is
mostly full of photos of Japan, and can be found here.
I have a blog
too.
Research
in probabilistic climate prediction
I
am a member of the Global
Change Projection Research Programme at RIGC
(formerly known as FRCGC, and before that FRSCG, as you might guess
from the URL).
I am currently (2010-present) mostly working on ways to use the CMIP3
multi-model ensemble. This seems to be a hot topic these days, the IPCC
held an Expert
Meeting early in 2010:
"The expert meeting will provide tentative best practices
in selecting
and combining results from multiple models for IPCC AR5; in short the
beginning of a quantitative framework for analysis and assessment of
the models. Specific aims of the meeting will be to maximize the
robustness and policy relevance of the projections provided in the
presence of model error, projection uncertainty, observational
uncertainties and a heterogeneous set of models."
Unfortunately, it seems to me that much of what has been written about
the multi-model ensemble is misleading, inadequate or plain wrong. I'm
writing a series of papers through which I hope to correct some
misunderstandings. The first of these was this
one published in GRL at
the start of 2010 (see also this follow-up),
this one
eventually got into J Climate and there is another GRL paper here.
I think there's still a bit of work
to
do on this, especially on the topic of model independence (what do we
even mean to call models independent?) and how/why to discard or
downweight models from the ensemble.
Another major interest of mine is the use of paleoclimate
simulations to assess
and evaluate climate models. According to the climate models, we can
expect our current climate to change substantially through the current
century (and beyond, for as long as we keep on emitting large
quantities of CO2). Paleoclimates provide the only opportunity to
actually evaluate the models' ability to simulate substantially
different
climates to today's, through the comparison of these simulations with
proxy data. The
proxy data are limited and imprecise, and the most recent paleoclimate
eras - for which data are most plentiful and reliable - were
generally colder rather than warmer than the present, but such testing
is surely better than nothing. For the most part, the models seem to do a mostly reasonable job,
as far
as we can tell - but with substantial regional problems, inasmuch as
the proxy data can be trusted. I've recently analysed proxy data and
models to generate a new reconstruction of temperature at the Last
Glacial Maximum and used this
result for another estimate of climate sensitivity here.
Previously, I mostly worked on the problem
of parameter estimation in single-model ensembles and
its specific relevance to climate prediction (but now I'm not so convinced that single model ensembles
are a good idea). Some of my work has been
in collaboration with the UK-based GENIE
project,
and the rest is mostly in collaboration with other researchers here
at RIGC, also with NIES and CCSR. There are now several
researchers
here working
on related topics forming the JUMP
group.
Betting on
Climate Change
I'm
also interested in the use of prediction markets and ideas futures. Here's a page I wrote, based
on a poster
presentation I gave at the EGU in 2005. I've also had an article
published on
realclimate.org
on the same topic. It's not clear yet where,
if
anywhere, it is going, but it's been interesting and instructive so
far. I've co-authored various articles with some financial people - to
be honest, this is almost all Daniel Bloch's work - which can be found
on the SSRN site through this link. The main problem with getting anything meaningful
established is the long time scale associated with
climate-change related damage, especially sea level rise.
I
obviously shouldn't omit to mention the loss
of one bet
with David Whitehouse from the pressure group Global Warming Policy
Foundation. I was expecting the 1998 record (in the Hadley Centre
analysis) to be beaten in the 4 years 2008-2011 inclusive, and
according to HadCRUT3, this did not occur. Interestingly, but unknown
to me at the time, they had been working on an improved version of
their temperature analysis HadCRUT4, with improved data coverage.
According to this new analysis, 2010 (and also 2005) actually was
warmer than 1998.
Publications
2013
G.
A. Schmidt, J. D. Annan, P. J. Bartlein, B. I. Cook, E. Guilyardi, J.
C. Hargreaves, S. P. Harrison, M. Kageyama, A. N. LeGrande, B. Konecky,
S. Lovejoy, M. E. Mann, V. Masson-Delmotte, C. Risi, D. Thompson, A.
Timmermann, L.-B. Tremblay, and P. Yiou. Using paleo-climate comparisons to constrain future projections in CMIP5. Climate of the Past Discussions (under review).
T. Lenton, Y. Aksenov, J.D. Annan, T. Cooper-Chadwick, S. Cox, N.
Edwards, S. Goswami, J.C. Hargreaves, P. Harris, Z. Jiao, V. Livina, D.
Lunt, R. Marsh, T. Payne, A. Price, A. Ridgwell, I. Rutt, J.G.
Shepherd, P. Valdes, G. Williams, M. Williamson, A. Yool., A modular,
scalable, Grid ENabled Integrated Earth system modelling (GENIE)
framework: Effects of dynamical
atmosphere and ocean resolution on
bi-stability of the thermohaline circulation, Climate
Dynamics DOI 10.1007/s00382-007-0254-9 (their
version).
J. K.
Hargreaves, A. Ranta, J.
D.
Annan and J. C. Hargreaves, The temporal fine structure of
night-time spike events in auroral radio absorption, studied by a
wavelet method. Journal of Geophysical Research, 106(A11):24621--24636.
J. D. Annan. Modelling under uncertainty: Monte Carlo methods for
temporally varying parameters. Ecological Modelling, 136(2--3):297--302.
2000
F.
Chen and J. D. Annan. The
influence of different turbulence closure schemes on modelling primary
production in a 1D coupled physical-biological model. Journal of Marine
Systems 26:259--288.
J. C. Hargreaves and J. D. Annan. Comments on ``Improvement of the
Short-Fetch Behaviour in the Wave Ocean Model (WAM)''. Journal of
Atmospheric and Oceanic Technology, 18(4):711--715.
1999
J. D.
Annan. Reply to
``Comments on the paper: On repeated parameter sampling in Monte Carlo
simulations''. Ecological Modelling 124:255--257.
J. C. Hargreaves and J. D. Annan. The impact of fierce weather on lazy
modelling. In The wind-driven air-sea interface, M. Banner (Editor),
ADFA Document Production Centre, Canberra, Australia, 125--132.
J. D. Annan. Numerical methods for the solution of the turbulence
energy equations in shelf seas. International Journal for
Numerical Methods in Fluids, 29:193--206.
1997
J. D.
Annan. On
repeated parameter sampling in Monte Carlo simulations. Ecological
Modelling 97(1--2):111--115.
1996
J. C.
Hargreaves, G. Gilmore
and J. D. Annan. The influence of binary stars on Dwarf Spheroidal
Galaxy kinematics. Monthly Notices of the Royal Astronomical Society
279(1):108--120.
1995
J. D.
Annan. The complexities
of the coefficients of the Tutte polynomial. Discrete Applied
Maths 57:93-103.
1994
J. D. Annan. An approximation algorithm for counting the number of
forests
in dense graphs. Combinatorics, Probability and Computing 3:273-283.
J. D. Annan. The complexity of counting problems. D. Phil. Thesis,
University of Oxford.
