165-169 Yan.p65

Across western Europe and eastern North America, rates of acid deposition are falling in response to reductions in atmospheric S emissions (1, 2), and the water-quality of acidified surface waters is beginning to improve (1, 3). To aid in the management of S emissions, biogeochemical models designed to predict socalled “critical acid loads” (4), are now being used to predict the degree of water-quality improvements which should accompany particular reductions in emissions (5, 6). For their part, ecologists are actively seeking to quantify the rates and trajectories of biotic recovery from historical acidification (7–11). The initial results are promising. Acid-sensitive species either lost to acidification or missing from historically acidified waters are appearing. This suggests that we should be able to combine biotic recovery models with critical load models to produce a more complete “picture” of the recovery process, and better predictions of its future. Ultimately, the ecological, i.e. the biotic, components of these models should have 2 functions: i) to identify when, where and for which biota recovery should occur; and ii) to identify where recovery might not occur without active management including, but not limited to additional S emissions reductions. The construction of such ecological recovery models logically proceeds from simple, first generation empirical models to more complex process-oriented models, assuming the latter will be more accurate. A few first generation models have already been developed. They couple biogeochemical models, which predict future time trends of lake alkalinity or pH, with regression models that employ these time trends to predict changes in the occurrence or abundance of indicator taxa. The regression models are developed from synoptic surveys of biota in lakes that vary in acidity (5, 12). These first generation models operate under one key assumption—that the historical spatial patterns of biota in acidified landscapes provide sufficient information to predict future patterns of recovery. To the extent that this assumption is true, the models’ predictions should be good; however, to the extent that it is false the predictions of recovery will be flawed. Recent reviews (13, 14) indicate that rates and trajectories of recovery may be influenced by many different factors which might not be well-captured in the patterns observed in historical surveys. It is quite possible, for example, that some damaged ecological communities may be resistant to change despite waterquality improvements (14, 15). If we wish to build accurate models to predict the recovery of acidified lakes, we must begin by understanding the key processes that influence the rates and trajectories of recovery. Hence, our first objective is to construct a conceptual framework which summarizes our current understanding of the ecological processes that influence the recovery of acid-sensitive species in historically acidified lakes. While we believe the processes in the framework are generally applicable, we employ crustacean zooplankton as model species, given our wealth of acidification and recovery knowledge on these organisms (10, 13, 14, 16, 17). Given that understanding should both spur and improve action, we also provide a second framework with more of a management orientation. It embeds our understanding of key ecological recovery processes into a larger framework which identifies needed ecological models and recovery bottlenecks which can be countered by appropriate management. In summary, we have 3 objectives: i) to construct a conceptual framework of the ecological recovery sequence from historical acidification; ii) to identify points in the process where management interventions might be required to overcome recovery bottlenecks; and iii) to identify the models which will be needed to operationalize the framework. The lake-liming literature provides a starting point for the identification of the processes in ecological recovery. One key observation that emerges from this literature is that the pace and extent of recovery is critically influenced by the severity and duration of antecedent damage—the more severe the damage, the slower and less complete the recovery. The probable explanation is that colonist availability is reduced when damage is severe and long-lasting (8, 18). With this proviso, the liming literature provides quite positive evidence that periphyton (19), phytoplankton (20, 21), zooplankton (8, 22), littoral micro(16) and macrobenthos (9) and fish (23, 24), might eventually recover following reductions in habitat acidity. We must, however, be cautious in our use of the liming literature to build a conceptual recovery framework, because there are geochemical and ecological reasons why liming might not perfectly simulate the natural recovery process. For example, Ca concentrations influence zooplankton species composition in soft-waters (25), and Ca levels increase after liming while they fall in response to reduced S deposition (1, 26). For Ca-rich taxa, falling Ca levels may partially offset the benefits of rising pH as S deposition decreases, Developing Conceptual Frameworks for the Recovery of Aquatic Biota from Acidification Norman D. Yan, Brian Leung, Wendel (Bill) Keller, Shelley E. Arnott, John M. Gunn and Gunnar G. Raddum