Environmental Data Analysis

According to the author’s preface, the goal of this book is to present a “comprehensive range of topics in data analysis in space, time and spectral domains which are necessary knowledge for environmental research.” He notes that environmental data contain huge amounts of information, varied in type and dynamic in nature, that require a variety of advanced techniques for exploration, identification and analysis.

The book largely focuses on mathematical methods that might be of use in the analysis of environmental data. The first four chapters only discuss mathematical methods and they do so without benefit of examples. First the author summarizes techniques for time series analysis, and then offers similar surveys of topics in dynamical systems and chaos, approximation methods, and interpolation techniques. The treatment is very succinct: the whole subject of state space models, for example, gets a total of four pages.

A chapter on statistical methods follows a pattern similar to the earlier four chapters except that it provides one brief case study — tree-ring based climate reconstructions. The author describes a variety of numerical methods in the following chapter, but it again has no examples.

Optimization methods are treated similarly, except that there is a short summary of case studies. Each case study (building design, supply chain planning, coal mining, iron and steel industry, electrical energy generating systems) is described in one short paragraph.

The remaining three chapters have a different character. One describes Data Envelopment Analysis as “a very powerful service management technique to evaluate organizations”. This technique uses linear optimization to measure the relative performance of “decision making units” (such as industries or groups of industries) where the presence of multiple inputs and multiple outputs makes comparisons difficult. The development here is more extensive and there are several examples and a few case studies with very brief summary descriptions.

Risk assessments are treated next. The object here is to discuss systematic procedures for estimating the probability of harm to or from the environment, the severity of the harm, and the uncertainty. The development is again more extensive and there are several examples. The final chapter is a brief summary of issues in life cycle assessment.

The book has are no exercises, but each chapter includes a bibliography with papers that appear to apply the techniques of the chapter to environmental problems.

This is an odd book. Most of the material directly related to environmental questions is in the final four chapters, and the connections to the previous seven chapters are tenuous. Certainly the earlier chapters would have benefited from examples cited from the references that the author provides.

Bill Satzer (bsatzer@gmail.com) was a senior intellectual property scientist at 3M Company. His training is in dynamical systems and particularly celestial mechanics; his current interests are broadly in applied mathematics and the teaching of mathematics.