Joint meeting of the Royal Statistical Society Environmental Statistics Section and the
European Region of The International Environmetrics Society

Thursday 2 June 2005, 10.00am to 5.00pm

The Royal Statistical Society, 12 Errol Street, London EC1Y 8LX, UK
(for directions see www.rss.org.uk )

STATISTICS AND ENVIRONMENTAL MONITORING

Design and analysis of environmental monitoring programs - an example from the Kattegat

Reporting on the state of the marine environment can be improved using statistical methods for design of monitoring programs and linked data analysis, illustrated using measured dissolved inorganic nitrogen for 1993-97 in the Kattegat. Proposed approaches are relatively simple but can handle missing data and utilize spatial and temporal correlation.

How best can geoscientists use detailed environmental data?

Geoscientists can now collect large volumes of data to describe systems in detail, and the incredible value of continuous detailed data sets, as with remediation of acidified water bodies, is recognised. Extracting temporal and spatial trends and assessing use of system descriptions in a predictive manner are important statistical issues.

What can we learn from routine environmental monitoring data?

Many millions of pounds are spent on routine collection of environmental data, subsequently used for many purposes including assessment of long-term change, ascertaining impacts of accidental pollution, and assessing compliance with objectives. I review some statistical methods used to analyse routine monitoring data with illustrations from air and water quality.

Combining deterministic and stochastic models for modeling, estimation and prediction of spatial-temporal environmental processes

Estimating spatial temporal trends of air pollution levels is vital for air quality management. Generally there are pollution measurements at only a sparse set of monitoring stations and there are outputs of regional scale air quality models. A hierarchical statistical model combines different sources of information for improved trend estimation.

Uncertainty analysis of boreal forest phytomass and Net Primary Productivity (NPP) in Central Siberia

This talk describes an uncertainty analysis of a greenhouse gas inventory of the Siberian terrestrial ecosystem, focussing on forest phytomass and forest NPP. The sizes of these components are highly uncertain, due to natural variability and systematic uncertainty. Methods and results from Monte Carlo modelling in S-Plus will be presented.

Objective analysis, sequential data assimilation and geostatistics: some interactions

Meteorologists and oceanographers use 'objective analysis', actually simple kriging for a spatial interpolation of differences between observations and forecasts from a numerical model. This is also the analysis step of extended and ensemble Kalman filters for sequential data assimilation. The potential for implementing geostatistical concepts in data assimilation is discussed.

Registration for this meeting is required using booking forms available from the RSS website (www.rss.org.uk) or by e-mail from meetings@rss.org.uk. The charge (including lunch, coffee and teas) will be

Posters

We are inviting participants to bring posters (max size A0 portrait) to this meeting. Please send abstracts of your proposed poster to ris@ceh.ac.uk by 18 May 2005.

Further details are available at www.jiscmail.ac.uk/envstat, through the RSS website www.rss.org.uk or through the TIES website (www.nrcse.washington.edu/ties/).