Statistics in Weather and Climate
Joint meeting of the Royal Statistical Society General Applications Section, Environmental Statistics Section and Reading Local Group
Monday 21st March 0930-1745
University of Reading, Palmer Building, Room 105
http://www.reading.ac.uk/maths-and-stats/news/Statistics-Weather-Climate.aspx
Invited speakers:
Dr Claudia Neves, University of Reading
Dr Andrew Parnell, University College Dublin
Prof. Jonty Rougier, University of Bristol
Dr David Walshaw, University of Newcastle
Schedule:
0930-1015 Jonty Rougier (Bristol)
1015-1045 coffee
1045-1115 Kate Fradley (Reading)
1115-1145 Rebecca Killick (Lancaster)
1145-1215 Emma Suckling (Reading)
1215-1315 lunch
1315-1400 Andrew Parnell (UCD)
1400-1445 David Walshaw (Newcastle)
1445-1515 Paul Sharkey (Lancaster)
1515-1545 coffee
1545-1615 Ross Towe (JBA Trust)
1615-1645 Sandra Chapman (Warwick)
1645-1745 Claudia Neves (Reading)
Description:
The meeting will focus on the use of probability and statistics for applications in weather and climate including, but not limited to, climate predication, paleoclimate and extreme events. We aim to bringing together specialists with varying backgrounds and view points: the background may be meteorology, physical modelling, geostatistics or of probabilistic/statistical nature.
This event will form part of the Mathematics for Planet Earth Centre for Doctoral Training Jamboree 2016.
Registration:
Places are very limited. Please register as soon as possible by emailing Richard Everitt ([log in to unmask]). There is no registration fee.
Directions:
The meeting is on the first floor of the Palmer Building, which is building number 26 on the map at http://goo.gl/AtV6rU. The university is easily accessed by bus from the railway station (see http://goo.gl/Ybe9AB for further details). Parking is limited on campus - please contact Richard Everitt ([log in to unmask]) if you require a parking permit.
Titles and abstracts:
Speaker: Dr Claudia Neves
Title: Extreme value statistics for non-stationary outcomes
Abstract: Extreme value theory provides a rigorous and prolific framework for analysing rare events with severe impact. The basic assumption is that the observations are independent and identically distributed. Although the celebrated extreme value theorem still holds under several forms of weak dependence, relaxing the stationarity assumption by considering a trend in extremes, for instance, leads to a changeling inference problem about the frequency of extreme events. Some studies dwell on the possible change in climate is not so much about the mean but rather in the frequency the extreme phenomena. The average rainfall may not change much, but heavy storms may become more or less frequent, meaning that observations have different underlying distributions. We present statistical tools for dealing with changes over time by inserting a trend on the probability that some high threshold is exceeded.
Speaker: Dr Andrew Parnell
Title: Statistical palaeoclimate reconstruction: recent results and opportunities for collaboration
Abstract: I will present the generic problem of reconstructing aspects of past climate based on proxy data from a statistical perspective. There are now numerous data sources to assist with such reconstructions, which all present their own advantages and issues. Whilst my group have developed one particular approach, based on Bayesian inversion of causal models of the climate-proxy relationship, we have currently only applied it to individual proxies (pollen) and individual sites. Our approach contrasts with much of traditional palaeoclimate reconstruction. I will point out the differences and explore some of the many possible extensions which require collaboration between climatologists, mathematicians, proxy specialists, and statisticians.
Speaker: Prof. Jonty Rougier
Title: A statistician’s viewpoint on weather, climate, and climate simulations
Abstract: There is plenty of agreement about what we mean by ‘weather’, some agreement about ‘climate’, and quite a lot of confusion about ‘climate simulations’ – not about what they are, but about what they mean. Statisticians are very familiar with the underlying issues, which centre on uncertainty and our attempts to define and quantify it. I propose that climate simulators represent expert opinions about future weather, and should be treated accordingly. This in turn requires that we expose the inherently probabilistic nature of a climate simulator, so that we can interpret it as offering a set of bets on future weather outcomes.
Speaker: Dr David Walshaw
Title: Improved Return Level Estimation through More Sophisticated Statistical Models
Abstract: The bottom line of an extreme value analysis is always the estimation of a return level, i.e. the value which will be exceeded on average once in any given fixed period of time. This would typically refer to extremes of rainfall, wind-speed, river height, sea level, temperature, pollution level, or sometimes something more unusual, such as earthquake magnitude or solar activity.
Many practitioners are familiar with the idea of fitting an extreme value distribution for block maxima or threshold exceedances to data, and using a fitted high quantile of the distribution as a return level estimate. In this talk we consider the limitations of this approach in terms of the huge uncertainties which often arise, and the ways in which this problem can be managed by more sophisticated modelling techniques.
Models considered generally involve supplementing the information available by incorporating other information. This may be other data on the same variable obtained at nearby locations, or it may be information on other variables which are strongly associated with the variable in question - covariate information in effect.
If time permits we will discuss the interpretation of return level estimates in the face of different estimation philosophies (Bayesian versus non-Bayesian) and confounding factors such as climate change, and the challenges posed by communicating the associated risks to non-specialists.
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|