Cross Site Analysis

The Pearson product moment correlation coefficients for the period 1957-1990 between the SOI values for a given month and the standardized temperature or precipitation anomalies do not show very high values (tables 6.1 and 6.2) because the data are inherently noisy. However, most of the values shown are statistically significant, partly because of the large number of pairs of observations (397-408) in the analyses. Another reason that the correlation coefficients are low is that all SOI values during the period are used, and thus both El Niño (extreme negative SOI values) and La Niña (extreme positive SOI values) events occur in the series along with intermediate values. This approach differs from one where correlations are found between one or more climate variables of El Niño years only and some other ecosystem variable. Many studies of this kind tend to deal only with extreme ENSO values instead of all the data (e.g., Cayan et al. [1999] use the 90th percentile ENSO events). In addition, Sardeshmukh et al. (2000) have noted that away from the tropical Pacific Ocean, an ENSO event is associated with relatively minor changes of the probability distributions of atmospheric variables. Nonetheless, it is important to estimate the changes accurately for each ENSO event, because even small changes of means and variances can imply large changes in the likelihood of extreme values. Wolter et al. (1999) have also quantified similar relationships. The higher correlation coefficients in Greenland (1999) are used to indicate sites where the climatic variable displays an association to both El Niño and La Niña events.

Table 6.1 Correlation coefficients between monthly SOI values and LTER site temperatures

Lag

0

1

2

3

4

5

6

7

8

9

10

11

AND

-0.196

-0.226

-0.193

-0.187

-0.15

-0.089

-0.088

-0.085

-0.065

-0.041

-0.014

0.005

ARC

-0.082

-0.046

-0.07

-0.073

-0.077

-0.043

0.057

0.014

0.028

0.011

-0.028

-0.002

BNZ

-0.081

-0.098

-0.108

-0.105

-0.104

-0.096

-0.014

-0.042

-0.024

0.011

-0.025

-0.03

CDR

-0.09

-0.011

0.025

-0.028

-0.041

-0.023

-0.095

-0.064

-0.075

-0.066

-0.078

-0.064

CWT

0.019

0.093

0.142

0.037

0.003

-0.028

-0.005

-0.037

-0.062

-0.047

0.005

-0.062

HBR

0.004

-0.055

-0.006

-0.013

-0.022

-0.018

-0.073

-0.103

-0.093

-0.035

-0.07

-0.049

HRF

0.007

-0.03

-0.025

-0.027

-0.052

-0.056

-0.106

-0.137

-0.11

-0.05

-0.093

-0.081

JRN

-0.059

-0.089

-0.079

-0.093

-0.131

-0.136

-0.095

-0.088

-0.045

-0.061

-0.022

0.011

KBS

-0.127

-0.103

-0.063

-0.046

-0.017

-0.001

-0.041

-0.029

-0.049

-0.075

0.045

0.083

KNZ

-0.007

0.065

0.116

0.091

0.037

0.044

0.009

-0.08

-0.043

-0.017

-0.002

0.014

LUQ

-0.157

-0.199

-0.279

-0.238

-0.223

-0.267

-0.252

-0.234

-0.152

-0.157

-0.15

-0.141

NTL

-0.092

-0.087

-0.023

-0.049

-0.082

-0.128

-0.142

-0.178

-0.156

-0.069

-0.068

-0.066

NWT

0.005

0.054

0.093

0.023

0.032

0.083

0.019

-0.01

0.03

0.03

-0.015

0.036

PAL

0.166

0.208

0.202

0.217

0.185

0.169

0.156

0.164

0.121

0.154

0.135

0.133

SEV

-0.005

0.026

0.05

0.032

0.036

0.031

-0.002

-0.062

-0.085

-0.132

-0.165

-0.117

SGS

-0.096

-0.029

0.012

-0.01

-0.014

-0.004

-0.022

-0.072

-0.051

-0.03

-0.011

0

VCR

0.059

0.065

0.096

0.062

0.011

0.013

-0.022

-0.067

-0.1

-0.023

-0.008

-0.048

aTrailing numbers represent lags by month of correlation (i.e., 0 = no lag, 1 = SOI value correlated against the temperatures of the following month).

aTrailing numbers represent lags by month of correlation (i.e., 0 = no lag, 1 = SOI value correlated against the temperatures of the following month).

Table 6.2 Correlation coefficients between monthly SOI values and LTER site precipitation

Lag

0

1

2

3

4

5

6

7

8

9

10

11

AND

0.024

0.086

0.069

0.139

0.072

0.02

0.068

0.01

-0.024

-0.024

0.013

-0.027

ARC

0.02

0.019

0.019

0.004

-0.056

-0.027

0.006

-0.011

-0.006

0.003

-0.011

0.022

BNZ

0.097

0.1

0.1

0.051

0.013

0.029

0.013

0.063

-0.021

-0.027

-0.002

0.008

CDR

0.037

0.079

-0.006

-0.087

-0.034

-0.055

-0.051

-0.069

-0.027

-0.052

-0.058

-0.051

CWT

0.075

0.005

0.057

0.002

0.074

0.081

0.067

0.061

0.036

0.059

0.04

0.045

HBR

0.05

0.023

-0.01

0.017

-0.009

0.036

-0.015

0.033

0.032

0.034

0.018

-0.015

HRF

0.008

0.001

-0.007

0.021

0.036

0.081

0.046

0.016

0.014

0.051

0.01

-0.038

JRN

0.003

-0.031

0.058

0.081

0

-0.028

0.027

-0.011

-0.043

-0.091

-0.049

-0.111

KBS

0.094

0.069

0.07

0.032

-0.005

0.057

0.002

0.014

-0.024

-0.02

-0.032

0.001

KNZ

-0.069

-0.05

-0.037

-0.042

0.017

-0.007

-0.028

-0.015

-0.041

-0.038

0.001

-0.027

LUQ

0.012

0.022

-0.042

-0.037

-0.089

-0.025

-0.041

-0.079

-0.073

-0.019

-0.035

-0.048

NTL

0.034

0.036

-0.032

-0.012

-0.008

-0.059

-0.071

-0.108

-0.1

-0.111

-0.1

-0.086

NWT

-0.028

-0.019

-0.061

-0.008

-0.026

-0.033

-0.025

-0.039

-0.055

-0.06

0.012

-0.025

SEV

-0.106

-0.093

-0.159

-0.172

-0.124

-0.094

-0.088

-0.043

-0.068

-0.015

0.015

0.051

SGS

-0.082

-0.062

-0.084

-0.085

0.006

0.019

0.007

0.005

-0.04

-0.046

-0.042

-0.044

VCR

-0.024

-0.083

-0.105

-0.104

-0.05

-0.028

0.028

-0.005

0 -0.054

0.017

-0.017

aTrailing numbers represent lags by month of correlation (i.e., 0 = no lag, 1 = SOI value correlated against the precipitation of the following month).

aTrailing numbers represent lags by month of correlation (i.e., 0 = no lag, 1 = SOI value correlated against the precipitation of the following month).

Table 6.3 Classification of sites by signal strength and duration of signal

Very strong signal (r>0.20)

Site and climate variable

Signal duration in lag months

ANDta

1

LUQt

2,3,4,5,6,7

PALt

2,3,4

Signal (r=0.1 to 0.2)

Site and climate variable

Signal duration in lag months

ANDpb

3

BNZt

2,3,4

BNZp

1,2

CWTt

2

HBRt

7

HRFt

6.7

JRNt

4,5

JRNp

11

KBSt

0,1

NTLt

5,6,7,8

NTLp

7,8,910

SEVt

9,10,11

SEVp

0,2,3,4

VCRt

8

VCRp

2,3

No signal (r<0.1)

ARCt

ARCp

CDRt

CDRp

CWTp

HBRp

HRFp

KBSp

KNZt

KNZp

LUQp

SGSt

SGSp

NWTt

NWTp

a t represents temperature. b p represents precipitation.

Unless otherwise specified, I use the term ENSO in the subsequent discussion to refer to both of these extremes. Although there are a few exceptions, the results of the current study are consistent with the expected patterns of the geography of ENSO effects on the climate as illustrated in the Synthesis to this section.

Given these facts, it is appropriate to classify the LTER sites in terms of their ENSO responses into three categories (table 6.3). The first category consists of three data series (ANDt, LUQt, and PALt) that show a strong response (r > 0.2). Here, and in the following, I use the abbreviation letters of the LTER sites (chapter 1, table 1.1) and a lowercase "t" or "p" to indicate temperature and precipitation, respectively. The second category consists of 15 data series that display a detectable signal (r = 0.1 to 0.2). The third category is where there is no signal according to the definition (r < 0.1). Some of the data series in this category do come close to the cutoff r value and might well have an ENSO signal by other definitions. The duration in months of the ENSO signal is also shown in table 6.3. In some data series the signal lasts for up to 6 months, whereas in other series the signal occurs only in a single month. The general types of patterns in the data are represented by the data series for PALt, SEVp, and ARCp. The ENSO signal at the Palmer site in the Antarctic is strong and actually lasts in some form for the whole 12 months. Although the analysis does not extend to longer periods, the signal may be found before the beginning and after the end of the 12 months considered here. The SEVp ENSO signal is the strongest in the second and third month after the extreme SOI value and then gradually decreases in strength, becoming insignificant after the seventh month. The ARCp series is an example of a data series with no ENSO signal.

Some interesting details and implications arise from these results. At some sites, such as LUQ, higher or lower than average precipitation values are more ecologically important than higher or lower than average temperature values. The LUQ, Puerto Rico, site shows strong and long-lasting higher than normal temperatures associated with an El Niño occurrence and lower than normal temperatures with a La Niña. The site also displays drier than normal conditions for zero- and one-month lags between the SOI value and the rainfall. But, with a higher number of monthly lags, the LUQ precipitation is higher when the SOI has indicated El Niño conditions, or at least negative SOI values two or more months previously. This is generally consistent with Schaefer's findings (chapter 8). However, he deals only with true El Niño years when SOI values are negative and large, rather than with all months of SOI values. Specifically, Schaefer finds that although El Niño years have higher precipitation than average this effect occurs only in May. The wetter or drier than normal conditions actually have more effect on the ecosystem than do the warmer than average conditions. Temperatures are usually high at LUQ, and a little increase will not make much ecological difference. However, the precipitation at this site is the highest of all LTER sites (2470 mm annually), and the variability around this value can be very large. Between 1961 and 1990, the wettest year (1979) had 3955 mm of precipitation, whereas the driest year (1967 following the 1966 El Niño) had 1540 mm. The increase or decrease of precipitation, according to Schaefer, has a large ecological impact on streamflows and their sediment load and water chemistry. This can have an even greater ecological impact when the below-normal precipitation occurs in the dry season between January and March or during large storm events at this site. Gianinni et al. (2000) have demonstrated a complex geography of Caribbean climate in response to El Niño, with some parts of the region wetter and others drier than average.

At the Colorado alpine site (NWT), the ENSO precipitation response is also important. This site had one of its highest precipitation years (1581mm) in the super El Niño year of 1983. Net primary productivity was above average during this year. However, the El Niño-related precipitation signal at NWT is not strong in the correlation analysis described here. Most likely, the high precipitation of 1983 cannot be explained by the 1982-1983 El Niño alone.

SEV is another site where the importance of the ENSO-related precipitation appears more important than the temperature signal. The ENSO-related precipitation increase is manifested shortly after the occurrence of an extreme SOI value. The effect of the increase in winter precipitation in the case of El Niños has been well documented (Molles and Dahm 1990; Dahm and Molles 1992). After the forecast of the 1997-1998 El Niño event, workers at the SEV site issued a warning to New Mexico residents to be particularly careful not to allow the buildup of household and other waste. Such waste would add to the natural increased rainfall-derived accumulation of vegetative material on which rats and other small mammals feed. By issuing the forecast, scientists hoped to decrease the possibility of outbreaks of the ratborne hantavirus. Such outbreaks did not occur during the months after the 1997-1998 El Niño, but it is not possible to assess the direct effectiveness of the warning (Robert Parmenter, pers. comm., 2000).

At sites such as NTL, El Niño-related temperatures are more important than precipitation. El Niño occurrence is associated with higher than normal temperatures at the NTL site in Wisconsin, and La Niña corresponds to lower than average temperatures. However, the greater effect is found in the following summer for a wintertime maximum or minimum SOI value. Investigations have not yet been made to see whether this has effects on the ecosystem. The work of Robertson et al. (1994) focused on air temperatures during the spring melt of lake ice, and this had obvious ecosystem effects. The ENSO effect for El Niños also is manifested by increasing NTL precipitation values in summer and into fall. This may affect the at-mosphere/groundwater water input ratio to the lakes that, in turn, affects the water chemistry and has a cascading effect through the ecosystem.

At the Antarctic site, PAL, the ENSO-related signal in temperature is extremely strong in the context of the present study. Smith et al. (1996) suggested the lag may extend to 19 months. During El Niño events, temperatures at PAL tend to be colder than average. Smith et al. (1996) have noted that El Niño occurrence is associated with above-average ice extents in the Western Antarctic Peninsular area. Here the effect at the quasi-quintennial timescale somewhat offsets the strong warming trend that has been noted at this site during the last 40 to 50 years. An important ecological linkage is associated with penguins in this location. Optimum sea ice conditions no longer exist for Adélie penguins in the Western Antarctic Peninsular because of the lack of sea ice as the result of long-term warming. In contrast, Chin-strap penguin populations are increasing because they do better in open-water conditions (Fraser et al. 1992). Thus, at the longer timescale of five decades, the smaller timescale El Niños give a "momentary" respite to the Adélie penguins at the expense of the Chinstrap penguins, whereas La Niñas may have the opposite effect.

The result of the analysis of the climatic response to the 1982-1983 super El Niño compared to more normal-size warm events was not clear-cut. The LTER sites that had shown the highest response in the previous analysis to El Niños were examined. The 1982-1983 El Niño was certainly larger in terms of its SOI value than those of 1958, 1965, 1972, and 1987. However, the responses to these five ENSO events are not altogether consistent. At AND, in the Pacific Northwest, the 1982-1983 temperature anomaly was larger than for any of the other El Niño years. This was also true for the NTL, Wisconsin, temperature anomaly of the following summer. But with respect to temperature at LUQ, Puerto Rico, and precipitation at SEV, New Mexico, the 1982 to 1983 El Niño led to a smaller response than in some of the other El Niño years. The PAL temperatures in the Antarctic showed a larger positive anomaly for the 1982-1983 event than other years, but the pattern is confounded by the large negative anomaly for the 1958 event. Thus super El Niños might give rise to larger climatic responses than "normal" El Niños at some of the LTER El Niño-sensitive sites but not necessarily all of them.

Application of Framework for Investigating Climate Variability and Ecosystem Response

In this chapter, we have concentrated only on a periodic type of climatic variabil-ity—the ENSO. We have seen that any particular site may exhibit a variety of responses, ranging from long-lasting responses of several months, to short-lived responses of one month, to no response at all. At some sites, precipitation anomalies have the greatest effect on the ecosystem, whereas at others temperature anomalies are more important.

Although it is not possible to go into the details of the framework questions for all LTER sites, some examples are appropriate. Having identified the nature of ENSO as a climatic signal, we present the next part of the framework questions.

1. Are there any preexisting conditions that will affect the impact of the climatic event or episode?

One of the benefits of asking this question is that it will usually stimulate new research questions. An example relates to the effect of ENSO events that are themselves superimposed on a trend of longer warming and less ice at the Antarctic PAL site. The new question arises, At what stage does the environmental condition pass a threshold, for example, a change from pack ice to open ocean, that might lead to a fast decline or increase in the penguin populations? The issue of preexisting conditions further raises questions about the relationship between climate events at one timescale and those at another. It is possible, for example, that La Niña years set the stage in the Pacific Northwest for increased, short-term, rain-on-snow flood events at the Andrews rainforest. That the ENSO scale can be related to the individual storm scale has been shown in at least two cases. The first case is the increase of Atlantic hurricane frequency and damage during La Niña periods (Pielke and Land-sea 1999). The second case is the January 1998 ice storm in the northeastern United States that had a documented impact on the HBR LTER site and was attributed to the presence of an El Niño event (Barsugli et al. 1999).

2. Is the effect of climate direct or does it cascade? If it cascades, how many levels does it have, and is the interaction between each level linear or nonlinear?

A direct climatic effect on an ecosystem is exemplified by a windthrow event in which trees are severely broken; yet even this sets off a sequence of ecosystem responses at a relatively small spatial scale. Climate effects on ecosystems, however, most often go into cascades. So, for example, the increased El Niño-related precipitation at SEV increases the water in the aquatic systems and also sets the stage for increased primary productivity on the terrestrial systems. The latter, in turn, provides increased forage for small mammals, which provide transportation for the hantavirus. This example of a terrestrial ecosystem simplistically identifies a three-level system. Quantification of this system would help determine the degree to which the various stages were linear or not.

The subtle timing of the ecological response of the southern Wisconsin lakes to the El Niño signal illustrates that sometimes whether a cascade results depends on exactly how the signal and response are coupled. Much more work is needed to identify the coupling mechanisms and their temporal and spatial aspects in various ecosystems.

At this point, it is appropriate to address the question of whether correlation analysis should be used at all in attempts to relate ENSO to its climate signal and later to a potential ecological response. At the LTER All Scientist's Workshop on Climate Variability and Ecosystem Response (2000), this question was raised on the ground that meteorological, climatological, and, to a lesser extent, ecological data are inherently autocorrelated both in time and space and thus violate the assumptions of the statistical methods being used. This may be viewed as part of a larger question that has recently been discussed by Nicholls (2001), who points out many criticisms of null hypothesis significance testing in atmospheric science in general. As an alternative, Nicholls suggests we focus on the strength of the effect rather than on its significance. The correlation coefficients used in this chapter do just that. The correlation coefficients used here should be regarded as an index of the strength of the relationship between ENSO and temperature and precipitation rather than being viewed within the context of central tendency statistics.

3. Is the primary ecological effect completed by the time of the next climatic event or episode (or part thereof)? If the effect is complete, we may consider the next part of the cascade (if any). If the primary ecological effect is not complete (i.e., reaches a new constant level), is it still of sufficient magnitude to have an effect on the rest of the ecosystem? If so, we should pass the effect along the cascade.

The answer to the first question depends on the "characteristic time scale" of the ecosystem. On the one hand, at NTL the ecological effect of early ice melt during an El Niño is completed by the time of the next El Niño event. When an early melt occurs, primary productivity can have an early start, and, at least in hypothesis, there can be more productivity during the growing season at all higher trophic levels. The higher trophic levels represent the next part of the cascade. On the other hand, within the ENSO context, the ecological effect may not be complete for ecosystems, such as forests, acting at long timescales. In most of these kinds of cases, an individual El Niño will not have a measurable effect except possibly on the aquatic parts of the system, as in the case of LUQ. Apart from forest ecosystems, however, there will be few examples at the ENSO time scale where the ecological effect is not complete by the time of the next event.

4. Does the climatic event or episode have an identifiable upper or lower limit? If a limit exists, we can stop the consideration if necessary at the limit but keep the cascade going until it reaches limits that may exist in later parts of the cascade.

The ice melt at the NTL represents a more or less linear change. The change has a limit because there is always a finite amount of ice to melt. After the melt, the cascade of the energetics of the lake ecosystem will continue through the various trophic levels. Most changes will have limiting values. It will often be important to identify what these limiting values are for both the climatic system and the ecosystem. We might assume, in the case of the climate system, that the climatic effects of the super El Niño represent a set of limiting values. However, we showed that this does not necessarily lead to the identification of unequivocal limiting values. We also have to consider how limiting values might change as one moves across timescales and specifically how past and future climate change might affect the limiting values. In this context we must remember that the ENSO-intensity time series itself is not stationary (Torrence and Webster 1999).

5. Does the climatic event or episode reverse to some original state (i.e., is itpe-riodic, homeostatic, etc)? If so, what timescales are involved? Does the climate state go back to the original position or beyond? Do cascades reverse? Can we identify the timing of these events?

The first of these five questions has a relatively easy answer as far as ENSO climatic phenomena are concerned. The climatic variation is quasi periodic and returns more or less to its original position. Atmospheric cascades do not usually reverse. The timing of ENSO events is at a quintennial timescale in terms of the usual occurrence of the event, yet when an ENSO event occurs it does so at a seasonal and monthly timescale. These answers also apply to many El Niño-influenced ecosystems. An increase in Adélie penguin populations during a greater than average ice year associated with an El Niño will be reversed by the occurrence of a La Niña if we assume La Niña has the opposite climatic effect at the Palmer site. Energy flow through trophic levels is not reversed. The energy flow is always from the primary producers to the top carnivores. Therefore, in this sense, the flow of food along the food chain cannot be reversed. Regarding the final question in this series, the timing of many events in the Palmer, Antarctic, ecosystem is well established.

6. After the climatic event or episode, do the values of the climatic variables return along their outward path or is there hysteresis or some other trajectory in operation? If the latter, how does this affect the cascade?

Changes in the atmospheric part of the ENSO system tend to return along their outward path at least as far as the values of the climatic variables are concerned. This generally applies to atmospheric pressures in the Pacific Ocean source areas of the events and the values of temperature and precipitation in the affected climates of the world. The energy transfers related to the ENSO phenomenon in the Pacific Ocean do not return along the same path because of the operation of the second law of thermodynamics. El Niño-related ecosystem changes such as the increase of populations in the NTL lake ecosystem will often reverse themselves along the same or similar pathway after the El Niño event. El Niño-related changes such as the loss of aquatic species in the LUQ aquatic ecosystem conceivably may take some time to reverse, and a hysteresis effect might come into play. An extreme example of this is the episode in the late 1940s and mid-1950s of a series of dry La Niña events that led to significant dieback of pinyon pine and juniper at the SEV site.

All the above relates to a deterministic, nonchaotic system. A consideration of chaotic systems is beyond the scope of this chapter. Many, if not most, of our at mospheric systems and ecosystems display some degree of chaos, and it will be essential to address this topic in the future.

The application of this framework to the ENSO case of climate variability has been very effective in raising further research questions and providing a manner in which they can be posed. Although some of the answers to the framework questions yield nothing new and are sometimes even trivial, the realization that the climatic ENSO signal has to be specifically connected to some part of the ecosystem to be effective provides a great stimulus for further investigations. We have learned that the timing of, or trigger of a sensitive nonlinear mechanism by, the climate signal is critical for the effectiveness of the signal. Also, the particular climatic variable in which the ENSO signal is found has to be one with a direct link to the ecosystem. The existence of a coupling mechanism between the climate variability signal and an ecosystem-driving function is therefore critically important. The same would also be true for climate variability relations with human systems.

Conclusions

Clearly, the idea of a simple forcing event and its direct response must be extended when considering ecosystems. The example of an ENSO event has been a useful, and relatively simple, one for illustrating the utility of the framework questions of this book. The LTER sites that manifest strong, detectable, and weak or no climatic signals to ENSO events have been identified. We have learned that the timing of the ENSO and the identification of an ecosystem-coupling mechanism are critical for this particular form of climate variability to have an effect. A statistically significant climate signal at an LTER site does not necessarily mean there will be an ecologically significant response. ENSO signals in the temperature series at the AND, LUQ, and PAL sites are the strongest statistically. Of these, only the signal at the PAL site has an important direct ecological effect. Somewhat less statistically strong ENSO signals at NTL and SEV do have important ecological effects. The results of the analysis of the climatic response to the 1982-1983 super El Niño compared to more normal-size warm events were not clear-cut, although in some cases the effects of the super El Niño were more pronounced.

The framework questions about climate variability and ecosystem response have allowed us to at least begin a thorough consideration of ecosystem response to a climatic phenomenon. The framework must also be applied in a quantitative fashion. In other sections of this book, we apply the framework to climatic forcing functions at other timescales ranging from an individual storm to a major glacial period. Only after many such applications will we begin to see some of the important basic principles relating climate variability and ecosystem response.

Acknowledgments This study was supported by NSF Grant DEB 9416820 and the Crystal Harmony.

References

Barnston, A G., A. Leetmaa, V. E. Kousky, R. E. Livezy, E. A. O'Lenic, H. Van den Dool, A. J. Wagner, and D. A. Unger. 1999. Bulletin of the American Meteorological Society 80:1829-1852.

Barsugli, J. J., J. S. Whitaker, A. F. Lughe, P. D. Sardeshmukh, and Z. Toth. 1999. The effect of the 1997/98 El Niño on individual large-scale weather events. Bulletin of the American Meteorological Society 80:1399-1411.

Buizer, J. L., J. Foster, and D. Lund. 2000. Global Impacts and Regional Actions: Preparing for the 1997-98 El Niño. Bulletin of the American Meteorological Society 81:21212141.

Cayan, D. R., K. T. Redmond, and L. G. Riddle. 1999. ENSO and Hydrologic Extremes in the Western United States. Journal of Climatology 12:2881-2893.

Dahm, C. N., and M. C. Molles., Jr., 1992. Streams in semi-arid regions as sensitive indicators of global change. pp. 250-260. In P. Firth and S. Fisher, editors. Troubled Waters of the Greenhouse Earth. New York: Springer-Verlag.

Dahm, C. N., and D. I. Moore. 1994. The El Niño/Southern Oscillation Phenomenon and the Sevilleta Long-Term Ecological Research Site. Pages 12-21. In D. Greenland, editor. El Niño and Long-Term Ecological Research (LTER) Sites. LTER Publication No. 18. LTER Network Office. University of Washington. College of Forest Resources. AR-10. Seattle, Wash. 98195. 57 pp.

Fraser, W. R., W. Z. Trivelpiece, D. G. Ainley, and S. G. Trivelpiece. 1992. Increases in Antarctic penguin populations: Reduced competition with whales or a loss of sea ice due to environmental warming? Polar Biology 11:525-531.

Giannini, A., Y. Kushnir, and M. A. Cane. 2000. Interannual Variability of Caribbean Rainfall, ENSO, and the Atlantic Ocean. Journal of Climate 13:297-311.

Greenland, D., editor. 1994a. El Niño and Long-Term Ecological Research (LTER) Sites. LTER Network Office Publication No. 18. LTER Network Office, Univesity of Washington. College of Forest Resources. AR-10. Seattle, Wash. 98195. 57 pp.

Greenland, D. 1994b.The Pacific Northwest regional context of the climate of the H.J. Andrews Experimental Forest Long-Term Ecological Research Site. Northwest Science 69(2):81-96.

Greenland, D. 1999. ENSO-related phenomena at Long-Term Ecological Research sites. Physical Geography 20:491-507.

Kiladis, G. N., and H. E. Diaz. 1989. Global climatic anomalies associated with the extremes of the Southern Oscillation. Journal of Climate 2:1069-1090.

Kiladis, G. N., and H. E. Diaz. 1992. Atmospheric teleconnections associated with the extreme phase of the Southern Oscillation. Pp. 7-28 in El Niño: Historical andpaleocli-matic aspects of the Southern Oscillation. H. F. Diaz and V. Markgraf, editors. Cambridge: Cambridge University Press. 474 pp.

Leung, L. R., A. F. Hamlet, D. P. Lettenmaier, and A. Kumar. 1999. Simulations of the ENSO hydroclimate signals in the Pacific Northwest Columbia river basin. Bulletin of the American Meteorological Society 80:2313-2339.

Mason, S. J., L. Godard, N. E. Graham, E. Yulaeva, L. Sun, and P. A. Arkin. 1999. The IRI seasonal climate prediction system and the 1997/98 El Niño event. Bulletin of the American Meteorological Society 80:1853-1873.

Molles, M. C., Jr., and C. N. Dahm. 1990. A perspective on El Niño and La Niña: Global implications for stream ecology. Journal of the North American Benthological Society 9:68-76.

Nicholls, N. 2001. The insignificance of significance testing. Bulletin of the American Meteorological Society 82:981-986.

Pielke, R. A., Jr., and C. N. Landsea. 1999. La Niña, El Niño, and Atlantic Hurricane Damages in the United States. Bulletin of the American Meteorological Society 80:20272033.

Robertson, D. M., W. Anderson, and J. J. Magnuson. 1994. Relations between El Niño/Southern Oscillation events and the climate and ice cover of lakes in Wisconsin. Pages 48-57. in D. Greenland, editor. El Niño and Long-Term Ecological Research (LTER) Sites. LTER Publication No. 18. LTER Network Office. University of Washington. College of Forest Resources. AR-10. Seattle, Wash. 98195. 57 pp.

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

Sardeshmukh P. D., G. P. Compo, and C. Penland, 2000. Changes of Probability Associated with El Niño. Journal of Climate 13:4268-4286.

Smith, R. C. S. E. Stamerjohn, and K. S. Baker. 1996. Surface air termperature variations in the Western Antarctic Peninsula Region. Pages 105-121 in Foundations for Ecological Research West of the Antarctic Peninsular. Antarctic Research Series. Vol. 70. American Geophysical Union. Washington, D.C.

Torrence, C., and P. J. Webster. 1999. Interdecadal Changes in the ENSO-Monsoon System. Journal of Climate 12:2679-2690.

Trenberth, K. E. 1997. The Definition of El Niño. Bulletin of the American Meteorological Society 78:2771-2777.

Trenberth, K. E., and J. M. Caron. 2000. The Southern Oscillation Revisited: Sea Level Pressures, Surface Temperatures, and Precipitation. Journal of Climate 13:4358-4365.

Wolter, K., R. M. Dole, and C. A. Smith. 1999. Short-Term Climate Extremes over the Continental United States and ENSO. Part I: Seasonal Temperatures. Journal of Climate 12:3255 -3272.

Woodhouse, C. A. 1994. Tree-growth response to ENSO events near Niwot Ridge in the central Colorado Front Range: An extended abstract. Pages 22-28 in D. Greenland, editor. 1994. El Niño and Long-Term Ecological Research (LTER) Sites. LTER Network Office Publication No. 18. LTER Network Office, Univesity of Washington. College of Forest Resources. AR-10. Seattle, Wash. 98195. 57 pp.

Was this article helpful?

0 0

Post a comment