Our framework question asks, Is the ecosystem effect or response completed by the time of the start of the next climate event or episode? This question can be asked in different ways such as in the three questions posed by Boose (chapter 2) in his hurricane study. The question can also be posed implicitly in different contexts. For example, Parmesan et al. (2000, p. 446) have stated "the initial resistance, trajectory of response, and extent to which a system returns to original conditions (resilience) after a disturbance depend on the frequency, intensity, duration, and extent of disturbance, as well as the inherent properties of the biological system, including evolutionary history. . . . " LTER studies confirm this in many cases of climatological or meteorological disturbance.
It is also likely that different parts of an ecosystem will have different recovery times. The example of the February 1996 flood at the Andrews LTER site is interesting because the range of recovery times for different parts of the ecosystem have been documented for this event (Swanson et al. 1998). The recovery times ranged from less than 3 months for aquatic algae, through 1-3 years for cutthroat trout, to more than 30 years for coniferous trees. Boose (chapter 2) recognizes an important caveat to our thinking on recovery times when he points out that some adaptive responses, such as the creation of new foliage and branches following a hurricane, cannot be repeated indefinitely at short intervals. In contrast to an individual flood or hurricane event, suggested ecosystem process and component response times to decade- to century-long temperature increases have a wide range. The range extends from one day for leaf photosynthesis and respiration, through a decade for litter mass change, to over 1000 years for soil organic matter development and plant migration and invasion processes (Shaver et al. 2000).
LTER studies reveal some cases where the ecological response to a drought is completed and other cases where the response is not completed by the time the next drought occurs. Gage (chapter 4) points out that the effects of a one-year drought, such as that of the 1988, on annual rotational agronomic systems are minimal. However, the same 1988 drought had long-lasting, documented effects on some of the natural vegetation species at the Cedar Creek LTER site in southern Minnesota. Citing Tilman and Downing's (1994) work, Gage notes that the effects of the 1988 drought were evident in the oak savanna complex 5 years later. About 30% of the pin oaks and 19% of bur oaks died. As a result of such episodes, some parts of the ecosystem may return to their original state, whereas other parts are affected for many years to come or even permanently. It is certainly true that the landscapes of most LTER sites exhibit long-term legacies after a severe, relatively short-term, climate episode. This raises the question, What is it that determines which parts of the ecosystem will be most negatively affected? Based on our small number of examples, it seems that vegetation with longer life spans such as trees, as opposed to grasses, is most vulnerable.
The Cedar Creek finding is reminiscent of the fact that dead junipers still reside on the landscape at the Sevilleta LTER site, the result of La Nina-related droughts of the 1950s (chapter 15). However, studies based on tree rings, which provide information for almost 400 years at the Sevilleta site, place this result in an even more surprising context. Sevilleta researchers hypothesize that the 1950s drought was one of a series of droughts that recur at an interval of 55-62 years. Partly due to the fact that it was accompanied by the introduction of cattle ranching, the ecosystem has not yet, and possibly never will, recover from the previous cyclic drought of the 1890s. At both Sevilleta and the Jornada LTER site in southern New Mexico, shrub-land took over from grassland, and there is no sign of a return before the beginning of the next drought period. Thus, although in some senses, these semiarid ecosystems may return to "normal" in terms of biomass productivity levels, for example, after an El Nino-related season or two of above average winter and spring precipitation, they still may not be returning to "normal" at the multicentury timescale.
Cross-timescale considerations are also important at the Palmer Antarctic site, where ENSO-scale events govern key biophysical interactions and many of the interactions are complete by the time of the next ENSO event. However, the circa 60-year warming trend at this location complicates matters such that Palmer researchers find it difficult to envisage an "end scenario" (chapter 9). Nevertheless, it is very interesting that the 600-year fossil record at Palmer shows the current presence of chinstrap and gentoo penguins to be unprecedented and that the site was dominated throughout most of the 600 years by Adelies. In some cases, long-term trends, or their operation in association with oscillatory climate phenomena, can set new "preexisting" conditions for each cycle of "cyclic" climate variability as in the case of variability in PNW salmon abundance at the decadal timescale.
Two of the studies in this book extend to the timescale of 25,000 years. Beetle assemblage evidence from Elias (chapter 18) suggests that certainly at this timescale the ecosystem has made its necessary adjustments by the time of the next ice advance or return to warmer climates. However, these "adjustments" in the insect assemblage are to a "lowest common denominator." As Elias (p. 381) expresses it, "the current group of species in the alpine ecosystem may not be the best fit for the en-vironment—they are simply the best fit among those species able to persist regionally through the last glacial cycle." In this case we have a climate filter acting at the millennial timescale. Furthermore, Elias (p. 381) believes ". . . that at the century to millennial timescale, the response of major components of the vegetation in high altitude ecosystems of the Colorado Front Range lags behind major temperature changes." The lag is in the order of 500-1000 years. Analogous, lagged responses are described by Monger (chapter 17) for the arid environments of the Southwest as they change between glacial and interglacial times. In the Antarctic Dry Valleys, however, the ecological response to warming conditions is still dependent on events that started at least 24,000 years ago, as described previously and in chapter 16. In this case, even at these large timescales, the ecological response to Holocene warming cannot be said to have been completed. This is because it remains to be seen what would happen to the current soil communities if the carbon source derived from relic benthic algal mats were ever to be completely depleted. Indeed, Fountain and Lyons (p. 334) suggest "given the extremely slow cycling of nutrients and the pace of geomorphic change, we suspect that ecosystem responses are overprinted on each other and are not completed before the next event occurs."
Was this article helpful?