**The monograph
Statistical
Analysis in Climate Research
ISBN 0521450713**

co-authored by Hans von Storch and Francis W. Zwiers, by Cambridge University Press. £ 65 or US $ 110.

**About the book**

See also comments by Bob Livezey in *nature*, by Robert Lund (Journal of The American Statistical Association, 95, 1375-1376, 2000).by Manfred Mudelsee (in Computers and Geosciences and in J. Roy. Stat. Soc D (The Statistician), Vol. 49 (2000), Part 3, P. 450) and by others.

The tools of mathematical statistics find wide application in climatological research. Indeed, climatology is, to a large degree, the study of the statistics of our climate. Mathematical statistics provides powerful tools which are invaluable for this pursuit. Applications range from simple uses of sampling distributions to provide estimates of the uncertainty of a climatological mean to sophisticated statistical methodologies which form the basis of diagnostic calculations designed to reveal the dynamics of the climate system. However, even the simplest of statistical tools has limitations and pitfalls which may cause the climatologist to draw false conclusions from his or her data if the tools are used inappropriately and without a proper understanding of their conceptual foundations. The purpose of this book is to help the climatologist understand the basic precepts of the statistician's art and to provide some of the background which is needed to apply statistical methodology correctly and usefully.

So far, only very few errors have been found.

**The book features the following contents**:

Prologue, v

1. Introduction, 1

2. Probability Theory, 19

3. Distributions of Climate Variables, 53

4. Concepts in Statistical Inference, 73

5. Estimation, 83

6. The Statistical test of a Hypothesis, 105

7. Analysis of Atmospheric Circulation Problems, 139

8. Regression, 153

9. Analysis of Variance, 177

10. Time Series and Stochastic Processes, 205

11. Parameters of Univariate and Bivariate Time Series, 225

12. Estimation of Covariance Functions and Spectra, 261

13. Empirical Orthogonal Functions, 303

14. Canonical Correlation Analysis, 320

15. Principal Oscillation Pattern Analysis, 341

16. Complex Eigentechniques, 361

17. Specific Statistical Concepts in Climate Research,

18. Forecast Quality Evaluation, 418

Appendices, 421

References, 471

Index, 487

**Readers' Comments**

Here are some comments on earlier versions of the book:

**Dennis Shea (NCAR, Boulder) in
December 1996**

How do I view the book? I think it will be a book that many active researchers should have within reach. It covers MANY topics at varying degrees of sophistication. The examples are the real bonus because they are real world applications

Do I think it will be used in graduate classes? My guess is that atmos/ocean applied stat courses will probably use Wilks book as "required" with this book "recommended." I think the more leisurely pace of Wilks book makes it more amenable to a graduate course while the "information content" within your text may make it a bit formidable for graduate students. Note: if it is a two semester course I think that this text could become a "required" book for the second semester.

**Hans Wackernagel (Ecole
de Mines, Paris), 13 December 1996**

The overall impression is very positive: this is a book for climatologists which is not merely a compilation of material from standard texts in statistics. It presents many methods that have been developed specifically for analysing climatological problems. All techniques are permanently illustrated with applications, so that their physical meaning becomes apparent to the practitioner. The book by its homogeneous notation and improved nomenclature, is especially an excellent guide through the modern methods of eigenanalysis, from empirical orthogonal functions to principal oscillation patterns.

**Silvia A. Venegas
(Centre for Climate and Global Change Research, McGill University,
Canada) May 1997**

I wanted to tell you that the material included in that book is what I've been waiting for since long time. It is absolutely great! I've been unsuccessfully looking for a course in statistics applied to geophysics since long, and I've learned what I know only by reading other's papers. I think that it contains exactly what I was looking for: some basic statistics and some more complicated methods applied to climate data. I think it is something the community needed desparately!

**Mark A. Saunders (University College
London, UK) August 1997**

In recent years the use of statistical concepts and tools in climate research has become widespread . However, many of the powerful statistical techniques such as empirical orthogonal functions and canonical correlation analysis remain accessible only in the refereed literature. An influential up-to-date book describing these and other new developments has been long overdue and von Storch and Zwiers are to be congratulated on addressing this challenging task.

"Statistical Analysis in Climate Research" is written principally for postgraduate students and researchers. It is readable, well structured and packed with information. Indeed the range of material covered is so vast that the depth of description possible on individual topics is, perforce, often restricted. The book's main strength is its numerous examples which illustrate the tools presented through a wide range of important climate problems. I recommend the book to all climatologists as a useful and valuable new reference.

**Susan Chen (Canadian Institute for Climate Studies; Victoria) September 1997**

The chapter 18 has been very helpful to me. It not only gave me an overview of t he forecast verification system, but also pointed me to some detailed references. I found the chapter to be concise, very clear and easy to understand.

**Michel Piot (Institut für
mathematische Statistik, Universität Bern, Switzerland) February
1998**

The book "Statistical Analysis in Climate Research" provides an excellent overview on statistical methods - also a very helpful introduction to probability theory - which are used in climate research. The mathematician appreciates the numerous applications which are a great enrichment for studying and reading this book. Let's hope that "Statistical Analysis in Climate Research" becomes a reference book, so that the language of mathematicians and climate researchers can be unified.

**Daniel Herron, Ph.D Research Student
(Department of Geological Sciences
University College London, UK), 11 May 1998**

The book by von storch and zwiers provides a welcome authoratitive guide
to the statistical analysis of climate data. Each chapter is noted for it's
clear and concise description of all the relevant methods, as well as the
use of worked examples. The book covers all the traditional techniques, as
well as summarising modern methods such as pop's and eof's.

The chapters on probabilty theory and time series analysis are particularly
relevant to my work.

- Formula (6.26) in paragraph [6.6.8] on page 115. In the denominator a 2 is missing in front of the summation: 1 + 2 \sum ....
- The Mann Whitney test in incorrectly described. Correct is: "The way to do a two sided test is to do two one-sided tests at half the significance level. Compute the rank sum for the X sample and decide whether the sum is less than the critical value (at the (1-p)/2 level). Reject the null if it is. Otherwise, compute the rank sum for the Y sample and decide whether the sum is less than the critical value (make sure that you reverse nx and ny if they are not equal). Again, reject the null if it is. If neither test results in a reject decision, then you can state that there was insufficient evidence to reject the null in a two sided test at the (1-p) level."
- More errors are listed in Mudelsee's review.

*Hans von Storch, 16 November 2000 *