Douglas G Goodin

Timescale is the organizing framework of this volume. In various sections, we consider the effects of climate variability on ecosystems at timescales ranging from weeks or months to centuries. In part III, we turn our attention to interdecadal-scale events. The timescales we consider are not absolutely defined, but for our purposes we define the interdecadal scale to encompass effects occurring with recurring cycles generally ranging from 10 to 50 years. A recurring theme in many of the chapters in this section is the effect on ecosystem response of teleconnection patterns associated with recognized quasi-periodic atmospheric circulation modes. These circulation modes include the well-known El Nino-Southern Oscillation (ENSO) phenomenon, which is generally thought to recur at shorter, interdecadal timescales but also includes some longer-term periodicities. Several other climate variability modes, including the Pacific North American index (PNA), North Atlantic Oscillation (NAO), Pacific Decadal Oscillation (PDO), and North Pacific index (NP) also show strong interdecadal scale signatures and figure prominently in the chapters of part III.

McHugh and Goodin begin the section by examining the climate record at several North American LTER sites for evidence of interdecadal-scale fluctuation. They note that interdecadal-scale contributions to climate variability can best be described in terms of two types of variation: (1) discontinuities in mean value, and (2) the presence of trends in the data. Evaluation of interdecadal periodicities in LTER data is complicated by the relatively short time series of observations available. McHugh and Goodin approach the problem mainly through the use of power spectrum analysis, a widely used tool for evaluating the periodicity in a time series of data. Principal components analysis is used to decompose the time series of growing-season climate data for each of the LTER sites into their principal modes of variability. These modes are then subjected to power spectrum analysis to evaluate the proportions of the variance in the data occurring at various timescales. McHugh and Goodin's results suggest that significant effects on precipitation and temperature at interdecadal timescales are uncommon in these data, although significant periodicities at both shorter and longer frequencies do emerge from the data (a finding of relevance to other sections of this volume). McHugh and Goodin further note that these results are for all LTER sites considered together and that significant interdecadal effects may emerge at specific sites, an observation relevant to the cross-timescale analysis appearing elsewhere in this volume. These results also suggest something about the geography of climate variation and how data from LTER sites may contribute to an understanding of the effects of climate variability in a number of ecosystem types.

In chapter 12, Glenn Juday and colleagues turn their attention toward a more direct observation of the interaction of climate and ecosystem response. Juday et al. use dendrochronological techniques to characterize 200 years of climate variability at the Bonanza Creek LTER, near Fairbanks, Alaska. Their recent reconstruction of nineteenth century temperatures in this area of Alaska indicates a pattern of quasi-decadal-scale variability, possibly associated with the PDO teleconnection pattern. Juday et al. compare this proxy temperature record to tree-ring width records for other parts of Alaska. They then use temperature records as part of a model of seed production in white spruce trees. Their results show a surprisingly warm climate over most of Alaska during the past two centuries. Temperature and precipitation patterns are coordinated in an inverse relationship, indicating a tendency for climatic conditions to remain either hot and dry or cool and wet. Juday and colleagues show that climate variability is closely tied to the reproductive cycle of the white spruce. They note that white spruce cone and seed production occurs episodically and in response to environmental cues. They suggest climate variability as an obvious candidate to provide these environmental cues. Examination of a 39-year record of white spruce reproduction at and around the BNZ LTER shows that seed drops are episodic, with major seed production events generally occurring at intervals of 12-17 years. Juday et al. note that major seed production events appear to be initiated during El NiƱo years. These results are an excellent example of the close relationship between climate variability and an important ecological response.

A somewhat different but equaling compelling link between climate and ecological response is made in Greenland's chapter (13) on climate variation and Coho salmon catch. Greenland uses the example of Coho salmon catches to illustrate how climate observations made at one LTER site (H. J. Andrews Forest, Oregon, AND) were generalized to a regional scale (the coastal Pacific Northwest) and then linked to a biological observation (Coho salmon population as represented by catch) that is important both ecologically and economically to the region of interest. In so doing, Greenland also provides an illustration of the fundamental framework questions underlying this volume, and a suggestion for further research guided by these framework questions. Greenland's investigation of salmon catch grew out of his analysis of climate at the AND LTER; he found salmon catch to be well related to temperature in western Oregon and to the broader region of the coastal Pacific Northwest. Comparisons of these temperatures with salmon catch data showed a strong inverse relationship, apparently at a cycle of approximately 20 years. Greenland links these two cycles through a five-level cascade model relating the PDO to nutrient availability, and thus to salmon population. In Greenland's cascade model, air temperature effects are a by-product of pressure changes (thus having no direct effect on salmon ecology), but are nevertheless strongly linked statistically to salmon catch. Greenland's results well illustrate the indirect and sometimes complex interaction between climate and ecosystem response.

In chapter 14 Hayden and Hayden examine potential changes in storm frequency at LTER sites in North America. They use data from the National Weather Service to map the presence and path of extratropical cyclonic storms during their mature (i.e., close isobar) stage. For their analysis, they selected 19 LTER sites, all located in the middle latitudes. They divided these sites into five geographically defined groups representing the interior west, west coast, Midwest, Appalachian region, and east coast. They examined storm data for these regions in terms of three questions: (1) Has storm frequency changed over the past century, and have these changes been geographically consistent? (2) Are changes in storminess consistent with model forecasts of climate change? (3) Is there a characteristic pattern associated with ENSO effects? Hayden and Hayden approach this problem using a grid-ded database of storms. They selected the grid cells associated with each LTER site and then extracted the index of storm counts from each LTER grid cell. Their results indicate that storm frequencies over the continental United States have changed during the past century, but the pattern of change can only be resolved at spatial scales below continental. Large changes in storm frequency are noted at most LTER sites; however, when averaged at a continental scale, no changes are apparent. The ENSO phenomenon showed little influence over storm frequency, and changes in storminess were not consistent with model predictions. As in McHugh and Goodin's chapter, Hayden and Hayden's results provide insight into the geographic effects of climate variability, as well as advancing important conclusions about the effects of scale in the observation of climate change.

In the final chapter of part III, Milne et al. use dendrochronology to analyze drought cycles at the Sevilleta LTER in New Mexico. They converted annual estimates of precipitation into anomalies by subtraction of the long-term mean, then used a probabilistic technique based on cumulated precipitation anomalies to evaluate the data time series. Their analysis revealed repeated, persistent wet and dry periods, most notably seven lengthy strings of dry years constituting persistent drought conditions. Milne et al.'s analysis tentatively suggests a 55- to 62-year recurrence interval for drought conditions, an interval coinciding with recent major droughts in the 1890s and 1950s. Based on this recurrence interval, they predict a prolonged period of lower precipitation in this region beginning in 2001 and persisting well into the second decade of the twenty-first century. Along with their climatic analysis, Milne et al. consider the effect that persistent drought might have on the transitional biome found at the SEV LTER. They suggest that long-term drought conditions could lead to changes in species composition, particularly between shrubland and woodland. Based on past interaction between two dominant species Larrea tridentata (a desert shrub) and Juniper monosperma (a woodland species) during the drought of the 1950s, Milne et al. hypothesize that drought conditions should favor establishment of L. tridentata, but thus far a simple model of species interaction under drought conditions remains elusive. Results of Milne et al. suggest the complex links between climate and ecosystem response and the variety of mechanisms by which these links may be expressed.

The five chapters in this section represent a cross section of climate and ecosystem effects at interdecadal scales. The variety of climatic processes considered and the frequent indication of periodicities not strictly defined at an interdecadal scale clearly indicate the difficulty of isolating one time period or one climatic process for intensive study. These chapters reinforce the across-the-scale complexity of the climate-biosphere system.

0 0

Post a comment