Douglas G Goodin Maurice J McHugh

The five chapters of part III provide a broad overview of decadal-scale climate processes and their ecological effect in a variety of ecosystems. Written by authors with disciplinary backgrounds that encompass climatology, biometeorology, and ecology, the chapters range from cross-site climate analysis with little direct attention to ecosystem effects (e.g., McHugh and Goodin, chapter 11; Hayden and Hayden, chapter 14) to more intensive studies of direct climate/ecological interaction at single sites or over more defined geographical areas (e.g., Greenland, chapter 13; Juday et al., chapter 12; Milne et al., chapter 15). Separately, each of these chapters contributes to understanding some aspect of the interaction of climate and ecology. As an integrated whole, they encapsulate many of the cross-disciplinary problems confronted by LTER scientists as they explore the interaction of climate and ecology. Despite the widely varying topics addressed and the disparate backgrounds of the contributors, similar themes emerge in each of the chapters. Here, we elucidate these themes and place them within the framework questions that have guided this volume.

Climatic Themes in Decadal-Scale Ecosystem Variability

Climatologists have long recognized the existence of cyclical or quasi-cyclical modes or patterns in the global circulation system. Typically, these patterns are characterized by variation in the strength or position of semipermanent pressure centers within the global circulation system. These variations occur at timescales ranging from seasonal to decadal, and such variability is frequently invoked as a causal mechanism for climatic trends or fluctuation at these various timescales. A

variety of indexes have been constructed to characterize these pressure patterns and the teleconnections that result from them (see van Loon and Rogers 1978, Rogers 1984, and Trenberth and Hurrell 1994 for in-depth discussion of the derivation and interrelationships of atmospheric circulation indices). Evidence of some of these patterns recurs throughout each of the chapters, suggesting their importance in decadal-scale climate/ecology interactions at LTER sites.

Although the chapters in this section concentrate on interdecadal variability, climate variability is a multiscale phenomenon in both space and time. Several authors acknowledge this, notably Milne et al. (chapter 15), McHugh and Goodin (chapter 11), and Greenland (chapter 13). Each of these chapters notes the importance of nondecadal variations, particularly the El Nino-Southern Oscillation (ENSO) phenomenon. While acknowledging the presence and interaction of circulation patterns at multiple timescales, in this Synthesis we will concentrate on those that operate at the decadal timescale (i.e., cycles of 10-30 years) on which this section is organized.

Two circulation indices, the North Atlantic Oscillation (NAO) and Pacific Decadal Oscillation (PDO) emerge as principal modes of interdecadal-scale climate variability. Each of the indexes is defined in terms of the seasonal behavior of large Northern Hemisphere semipermanent pressure centers. The NAO indexes pressure changes associated with the Icelandic high-pressure system, whereas the PDO indexes the Aleutian low-pressure system. Both the NAO and PDO affect atmospheric and ocean circulation in the Northern Hemisphere, manifest as trends in dominance of meridional or zonal atmospheric flow with associated tendencies for variation in moisture and energy advection. Thus NAO and PDO phenomena affect temperature and precipitation, often at great distances from the centers of pressure themselves.

Effects of the Pacific Decadal Oscillation are apparent in nearly all of the inter-decadal patterns discussed in these chapters. Greenland (chapter 13) notes that over a 60-year period from 1925 to 1985, temperature at the HJA LTER and salmon catch in the ocean off the Pacific Northwest are inversely correlated. He also notes that the time series of both of these variables suggests a cycle of approximately 20 years (noted by the author with caution, since the relative brevity of the data time series precludes strong conclusions in the absence of true cycles). Both of these cycles appear related to the PDO, although the links among PDO, temperature, and salmon harvest are indirect. Greenland attributes this close link between climate and ecosystem response to a five-level cascade coupling ocean and atmosphere variability to salmon harvest via nutrient availability.

Juday et al. (chapter 12) also note a strong relationship between PDO effects and ecosystem response, particularly the response of white spruce in the Alaskan interior. Using dendrochronological surrogates for temperature and precipitation, Juday et al. showed that quasi-decadal cycles in tree growth occurred throughout the nineteenth and twentieth centuries. The timing and periodicity of these fluctuations strongly suggest a link to the PDO, although the authors point out that, like the PDO, the North Atlantic Oscillation affects boreal zone climate through alteration of storm tracks and advection patterns. An interesting aspect of Juday et al.'s analysis is that the use of biological data (tree rings, in this case) as surrogates for climatic conditions implicitly assumes the closeness of the climate and ecology link. Juday et al. use this relationship to link reconstructed temperature at one site (Fairbanks, Alaska) to tree-ring pattern at the BNZ LTER. The coherency of results between the two sites allow these authors to integrate the effect of known features of the Alaskan interior climate system with dynamics of the upland white spruce ecosystem.

Juday et al.'s results in chapter 12 also illustrate the interaction of climate processes at multiple timescales. They note that favorable years for seed cone drop often correspond to poor radial growth years and that these key events are frequently associated with ENSO effects. ENSO, a shorter term (quasi-quintennial) phenomenon, is linked to PDO, such that ENSO represents a deepening and expansion of the Aleutian low-pressure center whose behavior is indexed by the PDO (Ropelewski and Halpert 1986).

Milne et al. (chapter 15) also used dendrochronological records to infer climate variability at the Sevilleta LTER site in the southwestern United States. Using a 393-year (1598-1991) record, they note an interdecadal periodicity of about 59-62 years, corresponding to drought cycles in the U.S. Southwest. As is the case in the Alaskan interior, climatic and ecological response in the arid southwest is linked to ENSO effect, although Milne et al. note that this regional linkage has a strong seasonal component—summer precipitation is not strongly influenced by ENSO, but fall, winter, and spring rainfall are. Although not explicitly examined by the authors of chapter 15, it is quite likely that other Pacific teleconnection processes influence the Sevilleta site. They note that fall, winter, and spring storms originate in the Pacific and are guided by zonal jet stream flows. Thus, processes influencing the balance of zonal/meridional flow, such as the strength and position of north Pacific pressure centers are likely to influence precipitation at this site. A link to indexes related to Pacific flow such as PDO and the Pacific North American index (PNA, an index of zonal to meridional flow over North America) can be hypothesized.

McHugh and Goodin's spectral analysis of temperature and precipitation at several LTER sites (chapter 11) also shows the influence of several well-known circulation patterns. These authors used Principal Components Analysis to decompose time series of mean growing season maximum/minimum temperature and precipitation into their principal modes of variability, then evaluated the proportion of variance occurring at interdecadal timescales using power spectrum analysis. Therefore, McHugh and Goodin's results reflect the temporal behavior across the entire LTER network, allowing individual sites to be placed within the context of the network. This approach also permits analysis of the geography of the network.

As in other chapters in this section, McHugh and Goodin found evidence of periodic behavior at timescales other than decadal. Quasi-quintennial variability resembling the ENSO signal occur in two of the mean temperature principal components, as well as in the maximum temperature component. Significant spectral power is also found at periods of 50 years (resembling the timescale of precipitation variability observed by Milne et al. (chapter 15), but there was little evidence of significant cycles at decadal timescales. This contrasts with results from individual sites (e.g., chapters 12, 13, 15), where decadal-scale cycles are present. Examination of periodic patterns at individual sites in the context of the whole net work could reveal important conclusions about the geography of climate variability across the LTER network. The analysis of storm frequency by Hayden and Hayden (chapter 14) also suggests the importance of network geography in cross-site analysis. In addition, Hayden and Hayden's analysis showed little evidence of a link between storm frequency and ENSO across the network as a whole, nor did significant links with decadal-scale teleconnections emerge. Their results did show geographic patterns of storm frequency change, with the greatest change occurring in the western and central LTER sites. Like the findings of McHugh and Goodin, these results provide a geographic framework in which to consider the spatial aspects of climate variability and climatic change at individual LTER sites.

Discussion of the Framework Questions

The chapters in part III provide some insight into the framework questions identified at the outset of the book. The concluding chapter of this volume (chapter 21) will provide a comprehensive summary of each chapter in terms of the framework questions; in this section we will consider a few observations related to the chapters dealing with the interdecadal scale.

Most of the chapters were successful in identifying some type of climate variability, usually of a periodic or quasi-periodic type. Greenland (chapter 13) found that Coho salmon catch and temperature in the Pacific Northwest are related and that their variation seems to be linked to periodic pressure changes in the Pacific Ocean. Juday et al. (chapter 12) found evidence of similar climate variation in their analysis of white spruce in the Alaskan interior. They also attributed this variation to periodic change in Pacific pressure systems. Milne et al. (chapter 15) and McHugh and Goodin (chapter 11) found longer term variability in climatic signals: Milne et al. at a single site (Sevilleta LTER), and McHugh and Goodin across the entire network. Although all report some type of variability, of the five chapters in part III, only Hayden and Hayden (chapter 14) report any evidence of a trend.

Scale dependence is a theme that emerges from these chapters. Although nominally devoted to decadal-scale climate dynamics, each chapter includes some discussion of variability at scales other than decadal. Given the complexity of the climate system and the number of variables influencing climate, it is perhaps not surprising that strict adherence to a given time framework is not practical. In addition to this temporal-scale dependence, it is also noteworthy that in those chapters in which multiple sites were considered (i.e., chapters 11 and 14), a spatial-scale dependence could be noted. Hayden and Hayden (chapter 14) were able to geographically stratify their sites based on storm frequency parameters. McHugh and Goodin (chapter 11) noted the strength of association between individual sites and climate dynamics across the entire network, thus establishing geographic patterns of site variability.

In the chapters where the interactions between climate and ecosystems are directly considered, questions concerning preexisting conditions and cascade effects prove relevant. In chapter 13, Greenland provides a clear example of system cascade, outlining a five-step model linking atmospheric circulation pressure features in the northern Pacific to salmon harvest off the coast of Washington and Oregon. Greenland's cascade model clearly links an important circulation pattern (the PDO) to an ecosystem response (salmon harvest). His model also shows that the observed inverse correlation between temperature and salmon harvest, which motivated the investigation, is not direct, but rather both are responses to a remote driving force.

Juday et al.'s analysis of spruce tree dynamics in Alaska (chapter 12) shows the importance of both system cascades and preexisting conditions. Juday et al. note that warm, dry years are necessary to "condition" the white spruce reproductive system for extensive cone production in years of favorable climate—a dependence on short-term preexisting conditions. Juday et al. also present a type of cascade model in which a series of key events or "gateways" must be passed before the reproduction process can begin. The thresholds implied by these gateways represent a type of system cascade in which flows of material or energy are controlled by "stocks" and "regulators." These systems are noted for their ability to often produce unexpected transient behavior (Chorley and Kennedy 1971).

Milne et al. (chapter 15) also demonstrate a system cascade in their analysis of drought and its ecosystem effects in New Mexico. This cascade may be distinguished by the presence of significant local effects along with larger scale climate variability—an example of the scale effect in climate/ecosystem interaction. The local feedback mechanisms are in the form of terpenes released by desert vegetation, which establishes localized heating via a small-scale "greenhouse effect." This localized heating provides a positive feedback mechanism within the cascade and reinforces ecosystems changes associated with climate variability.

Conclusion

The chapters in part III provide an overview of climate research issues at individual LTER sites and over larger geographic areas. Several themes emerge, including scale dependence, nested effects, geographical effect, complexity of interaction, cascade effects, and persistence effects. These topics lie within the theme of this book, and they illustrate the various modes of research needed to evaluate climate/ecosystem interaction. These chapters also suggest avenues of further research.

Clearly, more investigation is needed to refine our understanding of the mechanisms by which climate and ecosystems interact. Greenland (chapter 13) notes that a full understanding of the system cascade by which climate and salmon harvest are related will require a more comprehensive model than the five-level process outline in chapter 13. Milne et al.'s analysis of shrub dynamics in central New Mexico (chapter 15) and Juday et al.'s analysis of spruce reproduction (chapter 12) also provide examples of system cascades where further understanding of climate/ ecosystem interaction will refine explanatory models.

Another recurring theme in this synthesis is the interaction among climate processes at multiple time and space scales. Investigations of the interaction of one climate process (e.g., PDO, ENSO) with one or a few ecosystem parameters have been conducted, but systematic evaluation of how the processes quantified by the various indexes (e.g., NAO, PDO, PNA, etc.) interact is less commonly undertaken. Systematic investigations of the effect of multiple, nested climate processes on ecosystem response are needed, perhaps using indexes derived from existing time series of climate indicators.

References

Chorley, R.J., and Kennedy, B.A. 1971. Physical Geography: A Systems Approach. London: Prentice-Hall.

Rogers, J.C. 1984. The association between the North Atlantic Oscillation and the Southern Oscillation in the Northern Hemisphere. Monthly Weather Review 112:1999-2015.

Ropelewski, C.F., and Halpert, M.S. 1986. North American precipitation and temperature patterns associated with the El Niño/Southern Oscillation (ENSO). Monthly Weather Review 114:2352-2362.

Trenberth, K.E., and Hurrell, J.W. 1994. Decadal atmosphere-ocean variability in the Pacific. Climate Dynamics 9:303-309.

van Loon, H., and Rogers, J.C. 1978. The seesaw in winter temperatures between Greenland and Northwestern Europe. I, General description. Monthly Weather Review 106:296310.

This page intentionally left blank

Part IV

Century to Millennial Timescale

This page intentionally left blank

0 0

Post a comment