Zero to Kriging - An introductory course for geostatistics
|
|
Zero to Kriging |
||
|
As the name suggests, this course takes interested students from no knowledge of statistics or geostatistics through to the mysteries of ordinary kriging and its variants in 30 hours (or less). This is a classroom course and includes manual exercises to reinforce understanding of the techniques. WHO SHOULD ATTEND? This course is aimed at any professionals dealing with the
estimation from or interpolation between samples collected on a spatial
basis. This is not a course for specialist geostatisticians but intended
rather for those who wish to use geostatistics to enhance their practical
applications or research. All lectures will be augmented by desktop exercises covering many different applications and candidates should come equipped with a calculator capable of taking natural logarithms. Questions and discussions are actively encouraged and, indeed, form the basis of a successful course. |
Day 1: The basic assumptions of spatial
estimation methods; Inverse Distance interpolation methods; Classical
statistics: the Normal distribution. Day 2: Statistical methods: the lognormal
distribution; confidence levels; standard errors; grade/tonnage curves. Day 3: Geostatistical methods: the
experimental semi-variogram; modelling semi-variograms; trends; anisotropy;
geological features. Day 4: Geostatistical estimation: weighted
averages; estimation errors; standard errors; confidence levels; ordinary and
simple kriging; cross validation. Day 5: Variations on kriging; lognormal methods; indicator methods; non-linear geostatistics; regression problems (conditional bias); simulation. |
|
COURSE LEADER
Isobel Clark has taught, researched and consulted in the
field of geostatistics for almost 30 years. Possibly best known as the author
of the introductory text "Practical Geostatistics" (1979), she is
now co-author of a more complete textbook, Practical Geostatistics 2000
which is available as hypertext on CD and as a hardcopy
book. Software and data sets are available to all.
Short courses and seminars are offered on a regular
basis and, to date, have been hosted by companies and educational
institutions on four continents. Dr. Clark lectured for 11 years at the Royal
School of Mines, Imperial College, London, at the University of the Witwatersrand
in Johannesburg for 9 years and was Visiting
Professor at Camborne School of Mines for 2 years.
In between these academic engagements, she acts as Managing Director and senior partner of Geostokos (Ecosse) Limited, an international consultancy company based in Central Scotland.
Her recent consultancy assignments range from the evaluation of tantalite deposits in Mozambique to the study of protected sea-birds in the UK.
Geostatistics is the name given to a particular group of techniques which model spatial processes and allow estimation of values at unsampled locations. Geostatistical estimation is a two stage process:
i. studying the gathered data to establish the predictability of values from place to place in the study area;
ii. values at those locations which have not been sampled. This process "is known as 'kriging'.
In mining, geostatistics "is extensively used in the field of reserve valuation - the estimation of grades and other parameters from a relatively small set of borehole or other samples.
Geostatistics is now widely used in many other fields. Obviously there are geological and geographical applications. However, the techniques are also used in such diverse fields as hydrology, ground water and air pollution, soil science and agriculture, forestry, epidemiology, management of wildlife and weather prediction.