Home       About us   Issues     Search     Submission Subscribe   Contact    Login 
Conservation and Society
An interdisciplinary journal exploring linkages between society, environment and development
Conservation and Society
Users Online: 839 Home Print this page Email this page Small font sizeDefault font sizeIncrease font size



 
Previous articleTable of Contents Next article

BOOK REVIEW
Year : 2004  |  Volume : 2  |  Issue : 1  |  Page : 208-210

Book Review 4


Fellow, Ashoka Trust for Research in Ecology and Environment, Bangalore, India

Correspondence Address:
Jagdish Krishnaswamy
Fellow, Ashoka Trust for Research in Ecology and Environment, Bangalore
India
Login to access the Email id

Source of Support: None, Conflict of Interest: None


Rights and Permissions
Date of Web Publication18-Jul-2009
 


How to cite this article:
Krishnaswamy J. Book Review 4. Conservat Soc 2004;2:208-10

How to cite this URL:
Krishnaswamy J. Book Review 4. Conservat Soc [serial online] 2004 [cited 2019 Sep 18];2:208-10. Available from: http://www.conservationandsociety.org/text.asp?2004/2/1/208/55833

A.K. Ghosh, J.K. Ghosh and Barun Mukhopadhyay (eds), Sustainable Environments: A Statistical Analysis. New Delhi: Oxford University Press, 2003, 208 pp., Rs 495. ISBN: 019565858-2.



There is growing recognition that humans are altering the earth's natural systems at all scales, from local to global, at an unprecedented rate. Global warming and ecological degradation are no longer processes to be neglected or ignored. Yet it is equally evident that separating anthropogenic effects from natural variability or dynamics is also very challenging. It is also clear that understanding and responding to these challenges requires a creative scientific framework. The reductionist approach that limited the ability to model and analyse natural and human-influenced systems is no longer in vogue. As an example, hydrologic and metrologic data are collected across many time scales, from seconds to annual, and span spatial scales from less than a metre to thousands of square kilometres. Variability of hydrologic, metrologic and ecological processes across space and time as well as seasonality and changes in statistical properties over time (nonstationary) imply the need for statistical and stochastic models that account for spatial dependence (a violation of the traditional independence assumption) as in geostatistical approaches and time-varying parameters as in dynamic linear models. The challenges of dealing with increasingly large and complex data sets with spatially-dependent and time-evolving structure will require intensive collaboration between hydrologists, ecologists, climatologists and statisticians.

The magnitude and complexity of environmental and ecological problems that confront us have led to efforts to develop a strong quantitative--analytical framework to accommodate specific types of data sets and questions that are sought to be answered. A major contribution of this field has been to provide a framework for decision making for policy makers in the face of uncertainty. The need to explicitly quantify uncertainty and provide decision makers with choices rather than one solution has led to a new area in environmental statistics called decision sciences. One other outcome of this is an increasing tendency to embrace a suite of non-linear, spatially dependent and dynamic, Bayesian, likelihood and parameter estimation methods in place of the traditional, frequentist and hypothesis-testing approaches, which are still favoured by many ecologists and environmental scientists elsewhere but especially in India. This statistical renaissance has flowered in the United States in the specific context of environmental and ecological applications, but has yet to make a mark in India beyond a few statisticians. The book under review is one of the first to be published in India that can claim to be part of this revolution.

This book is the outcome of an international workshop on .Statistical Science and Environmental Policy: Possible Interactions. organised by the Indian Statistical Institute and the Bernoulli Society in Calcutta in January 2000. (A very good example is The Ecological Detective by Ray Hilborn and Marc Mangel (1997)). The twenty selected contributors have come up with nine chapters, including a very comprehensive overview in the introduction. Eight specific chapters on methodological approaches cover a wide-ranging suite either in detail or cursorily of methods that range from logistic regression modelling to space-time modelling. These are applied to an equally wide-ranging set of environmental problems including air and water pollution, toxicity and landscape degradation.

The book has some major weaknesses and shortcomings. The papers are not representative of research done in India or by Indian scientists and applied statisticians. The paper on space-time modelling of water quality curiously ignores many seminal paper on dynamic linear models, especially from the Duke University School of Statistics. The Sunderbans paper is essentially a quantitative ecology application rather than in the area of ecological statistics and seems out of place in this book.

At least chapter 8 is able to communicate with a larger audience and has a clear message. The chapter on trends in Indian temperature and monsoon rainfall in India is an example where environmental statistics are able to help tease out natural or global sources of variability from those that can perhaps be attributed to anthropogenic effects. Sources of long-term variability in hydrologic time series include the contributions from both natural variability and human-induced changes of the landscape and climate. Separating these using innovative techniques will remain a challenging arena for applied statisticians and environmental scientists.

The book would have been strengthened by a chapter on exploratory data analyses techniques. I was also disappointed by the absence of good conceptual descriptions of how prior information and data should and can be incorporated or a critique of traditional hypothesis testing. Bayesian statistics provides the framework for using past information and expert insight, which is especially appropriate for ecological and environmental applications, more so when one has limited data.

The importance and strengths of this book lie primarily in its potential in kickstarting the field of environmental and ecological statistics in India through an easily accessible and inexpensive introduction to the state of the art to those without ready access to the expensive peer-reviewed journals and classical books. However, the style and content of most individual chapters, except for the excellent introduction, make it forbidding to those outside the field of statistics, namely, the ecologists and environmental scientists who could have gained from an exposure to the quantitative-analytical conceptual framework. The level of statistical and mathematical detail in most of the chapters would frighten away most except those who dabbled in some of the techniques. It is hoped that a follow-up volume or a similar initiative in the future can bridge the communication gap by adopting a different style of presentation where the concept is explained in as simple and straightforward a manner as possible, and the details are in an appendix attached to each chapter. The material in this book illustrates, although not explicitly, the view that statistical science is an integrative discipline that lies at the heart of the scientific method and not an afterthought. India has produced theoretical and applied statisticians of the highest calibre in the past and it is hoped that in the emerging field of environmental and ecological applications our statisticians will make a mark.[1]

 
   References Top

1.Hilborn, Ray and Marc Mangel (1997), The Ecological Detective: Confronting Models with Data: Princeton: Princeton University Press.  Back to cited text no. 1      




 

Top
 
Previous article Next article
 
  Search
 
    Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
    Access Statistics
    Email Alert *
    Add to My List *
* Registration required (free)  

 
    References

 Article Access Statistics
    Viewed1134    
    Printed80    
    Emailed0    
    PDF Downloaded222    
    Comments [Add]    

Recommend this journal