BOOK REVIEWS

Rethinking the Fifth Discipline: Learning Within the Unknowable by Robert Louis Flood. Sami Potatoes: Living with Reindeer and Perestroika by Michael Robinson and Karim-Aly Kassam. The Self Managing Organization: How Leading Companies Are Transforming the Work of Teams for Real Impact by Ronald E. Purser and Steven Cabana.     

Forecasting time series with increasing seasonal variation

Four options for modeling and forecasting time series data containing increasing seasonal variation are discussed, including data transformations, double seasonal difference models and two kinds of transfer function-type ARIMA models employing seasonal dummy variables. An explanation is given for the typical ARIMA model identification analysis failing to identify double seasonal difference models for this kind of data. A logical process of selecting one option for a particular case is outlined, focusing on issues of linear versus non-linear increasing seasonal variation, and the level of stochastic versus deterministic behavior in a time series. Example models for the various options are presented for six time series, with point forecast and interval forecast comparisons. Interval forecasts from data-transformation models are found to generally be too wide and sometimes illogical in the dependence of their width on the point forecast level. Suspicion that maximum likelihood estimation of ARIMA models leads to excessive indications of unit roots in seasonal moving-average operators is reported.     

Prediction intervals for exponential smoothing using two new classes of state space models

Three general classes of state space models are presented, using the single source of error formulation. The first class is the standard linear model with homoscedastic errors, the second retains the linear structure but incorporates a dynamic form of heteroscedasticity, and the third allows for non‐linear structure in the observation equation as well as heteroscedasticity. These three classes provide stochastic models for a wide variety of exponential smoothing methods. We use these classes to provide exact analytic (matrix) expressions for forecast error variances that can be used to construct prediction intervals one or multiple steps ahead. These formulas are reduced to non‐matrix expressions for 15 state space models that underlie the most common exponential smoothing methods. We discuss relationships between our expressions and previous suggestions for finding forecast error variances and prediction intervals for exponential smoothing methods. Simpler approximations are developed for the more complex schemes and their validity examined. The paper concludes with a numerical example using a non‐linear model. Copyright © 2005 John Wiley & Sons, Ltd.     

Full-glacial shoreline vegetation during the maximum highstand at Owens Lake, California

Owens Valley, California, was markedly different during the Wisconsin glacial stage from what it is today. Alpine glaciers bounded the Sierra Nevada, and pluvial Owens Lake reached highstands and overflowed its natural basin. We analyzed three layers from two packrat middens, dated to ca 23,000-14,500 yr BP, obtained from Haystack Mountain (1155 m) only 10 m above and Juniperus osteosperma ) and single-needle pinyon pine ( Pinus monophylla ) woodland existed at the site. In the layers dated to ca 17,500 and 16,000 yr BP, macrofossils document the presence of Rocky Mountain juniper ( Juniperus scopulorum ), a species that no longer occurs in California. It is suggested that meltwater from the retreating glacial ice inundated the Owens River Lake chain causing pluvial Owens Lake to reach its highstand. This caused an increase in effective moisture, due to high groundwater, allowing the mesophytic Rocky Mountain juniper to exist at the site.     

Habitat use and food selection of small mammals near a sagebrush/crested wheatgrass interface in southeastern Idaho

Habitat use and food selection data were collected for deer mice ( Peromyscus maniculatus ), montane voles ( Microtus montanus ), Ord's kangaroo rats ( Dipodomys ordii ), and Townsend's ground squirrels ( Spermophilus townsendii ) near a sagebrush ( Artemisia tridentata )/crested wheatgrass ( Agropyron cristatum ) interface in southeastem Idaho. Significantly more captures occurred in the native sagebrush habitat than in areas planted in crested wheatgrass or in disturbed sites. Crested wheatgrass, a prolific seed producer, still accounted for over 30% of the total captures. Montane voles and Townsend's ground squirrels (during periods of aboveground activity) used the crested wheatgrass habitat throughout the summer, while deer mice and, Ord's kangaroo rats exhibited heavy use after seed set.     

Fine-scale recombination patterns differ between chimpanzees and humans

Recombination rates seem to vary extensively along the human genome. Pedigree analysis suggests that rates vary by an order of magnitude when measured at the megabase scale, and at a finer scale, sperm typing studies point to the existence of recombination hotspots. These are short regions (1-2 kb) in which recombination rates are 10-1,000 times higher than the background rate. Less is known about how recombination rates change over time. Here we determined to what degree recombination rates are conserved among closely related species by estimating recombination rates from 14 Mb of linkage disequilibrium data in central chimpanzee and human populations. The results suggest that recombination hotspots are not conserved between the two species and that recombination rates in larger (50 kb) genomic regions are only weakly conserved. Therefore, the recombination landscape has changed markedly between the two species.