Why Disturbances Can Be Predictable: a Perspective on the Definition of Disturbance in Streams Author(S): N

Home , Disturbance (ecology)

Why Disturbances Can Be Predictable: A Perspective on the Definition of Disturbance in Streams Author(s): N. LeRoy Poff Source: Journal of the North American Benthological Society, Vol. 11, No. 1 (Mar., 1992), pp. 86-92 Published by: The University of Chicago Press on behalf of the Society for Freshwater Science Stable URL: https://www.jstor.org/stable/1467885 Accessed: 26-08-2019 01:31 UTC

REFERENCES Linked references are available on JSTOR for this article: https://www.jstor.org/stable/1467885?seq=1&cid=pdf-reference#references_tab_contents You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profitSociety for service Freshwater that helps Science, scholars, The researchers, University and of students Chicago discover, Press are use, collaborating and build upon with a wide JSTOR to digitize, preserve and extend access to Journal of the North American range of content Benthologicalin a trusted digital Society archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at https://about.jstor.org/terms

This content downloaded from 24.9.112.12 on Mon, 26 Aug 2019 01:31:04 UTC All use subject to https://about.jstor.org/terms J. N. Am. Benthol. Soc., 1992, 11(1):86-92 ? 1992 by The North American Benthological Society

Why disturbances can be predictable: a perspective on the definition of disturbance in streams

N. LEROY POFF

Department of Zoology, University of Maryland, College Park, Maryland 20742-4415 USA

Abstract. Resh et al. (1988, J. N. Am. Benthol. Soc. 7:433) assert that lotic ecosystems are unique in that disturbances are necessarily "unpredictable" events. Acceptance of this "predictability clause" in the definition of disturbance introduces problems that influence our ability to identify, describe, and discuss important aspects of disturbance in streams. I focus on four specific aspects of Resh et al.'s treatment of predictability, particularly as it relates to hydrologic disturbance (e.g., spates). First, Resh et al. offer only purely statistical criteria for identifying hydrologic disturbance, an approach that has important limitations for characterizing predictability of the disturbance regime. I suggest adoption of more physically based measures of disturbance (e.g., flow at which movement of bed is initiated) to which specific ecological responses can be determined. Second, lotic species are assumed to be "adapted" to predictable hydrological events, despite the general absence of supporting empirical evidence. I suggest that this assertion be reframed as a testable hypothesis. Third, defining disturbances as necessarily unpredictable can result in a tautology, in which biological response to disturbance and predictability of disturbance are expressed in terms of one another. Fourth, a more explicit distinction between ecological and evolutionary time scales of response to disturbance allows predictability to be properly considered as a separate (and non- tautological) component of the disturbance regime. In sum, disturbances (including predictable ones) always have ecological effects; however, the magnitude of ecological response to a particular disturbance may be constrained by evolutionary (historical) adjustments of the biota if the disturbance regime is characterized by high predictability.

Key words: disturbance, predictability, streams, hydrology, geomorphology.

Communication among scientists often de-group consisting of 10 participants, the paper pends on careful definition. This is certainly by Resh et al. attempted to distil a "consensus" true in ecology, where many terms or phrases view on a complex issue. However, the authors attempt to generalize across diverse systems expressed and alternative views throughout their levels of organization. Differences in definition paper, and they invited further discussion of may arise from discernably different philo-these views. In this paper, I wish to address a sophical starting points, or they may arise componentmore of Resh et al.'s definition of distur- subtly from ambiguities associated with com-bance, viz. the assertion that lotic systems are plex phenomena and frame of reference. Some- unique in that disturbances must be unpre- times, ambiguities arise because terms or dictable.con- I critically examine this assertion and cepts implicit in the definition are themselves explore its implications to argue that acceptance subject to variable interpretation (e.g., discus- of this "predictability clause" unnecessarily re- sion of "niche" often relies on some common stricts our ability to identify, describe, and dis- notion of "resource"). Regardless of the cuss under- important aspects of disturbance in streams, lying reasons, we may speak at cross-purposes primarily because it blurs the distinction be- because definitional assumptions have nottween been the ecological and evolutionary time adequately specified. scales of biotic response to disturbance (both of The word "disturbance" is becoming whichincreas- are important in disturbance studies). ingly important in the ecologist's lexicon. This Theis an important issue because stream ecol- recent seminal paper by Resh et al. (1988) ogists has need to explicitly evaluate the assump- played a significant role in drawing attention tions, ambiguities, and implications that un- to this concept as applied to lotic ecosystems. derlie our use of the word "disturbance", a Furthermore, it has greatly helped stream relativistic ecol- concept that changes with scale of ogists to focus on many basic issues relevant observation to (Rykiel 1985). Hereafter, distur- the role of disturbance in structuring these bance sys- will refer to hydrologic disturbance un- tems. Produced during a workshop discussion less otherwise indicated. I will focus primarily

This content downloaded from 24.9.112.12 on Mon, 26 Aug 2019 01:31:04 UTC All use subject to https://about.jstor.org/terms 1992] DEFINITION OF DISTURBANCE 87 on spates; however, similar arguments could be proach appropriately considers the impact of made relative to other types of disturbance as disturbance in the context of normal environ- well (e.g., "droughts", chemical pollution, etc.). mental variation, I wish to emphasize some spe- Two widely cited general definitions of dis- cific limitations associated with such a purely turbance exist. Sousa (1984) defines disturbance statistical approach to defining predictability of in a population-dynamics context as "a discrete, disturbance. Resh et al. indicate that any purely punctuated killing, displacement, or damaging statistical definition of disturbance is arbitrary of one or more individuals (or colonies) that because, in their example, all months would directly or indirectly creates an opportunity for have the same expected disturbance frequency new individuals (or colonies) to become estab- (i.e., 5% assuming a normal distribution) given lished". White and Pickett (1985) adopt a broad- a sufficiently long period of record. However, er definition of disturbance as "any relatively because all months have equal disturbance fre- discrete event in time that disrupts ecosystem, quencies, disturbances would also be equally community, or population structure, and likely for each month, leading to the unex- changes resources, substrate availability, or the pected conclusion that disturbances (as defined) physical environment". Note that both defini- have no seasonal predictability (i.e., they are tions consider disturbances as "discrete" events, uniformly distributed throughout the year). the impact of which can be measured in terms Thus, using this (or other) arbitrary statistical of ecological responses. Further, according both criterion to define "disturbance", one could ar- to Sousa and to White and Pickett, a given dis- gue that all streams are similar in having no turbance regime is characterized by the statistical seasonal predictability of disturbance. distribution (based on historical record) of in- Resh et al. partially address this issue in their dividual events in terms of such attributes as second approach, where they apply a formal frequency, intensity (magnitude), duration, measure and of predictability (Colwell 1974) to predictability. monthly maximum and minimum daily dis- Resh et al. (1988:434) modify the White charges. and They show convincingly that temporal Pickett definition for streams to include events distribution of these extreme flows varies among that are only "... outside a predictable rangestreams (fig. 3, tables 2, 3). However, maximum ... [because] ... organisms are adapted to or pre- minimum monthly extremes are arbitrary in- dictable seasonal fluctuations of discharge, sofar tem- as they do not necessarily represent actual perature, dissolved oxygen, etc. ...". Further, hydrologic disturbances, so calculated predict- they state that "... when measured by the ability im- scores do not necessarily describe pre- pact on a community, we feel that it is the dictability un- of disturbance per se (a point not ex- predictable event that typically constitutes plicitly a addressed by the authors). Resh et al. disturbance". Although intuitively appealing appropriately conclude that such a quantitative on its face, this "predictability clause" intro- analysis is useful both for considering temporal duces significant methodological and concep- hydrologic constraints on the biota and for site- tual problems that deserve discussion. matching in comparative studies. However, similar arguments can be made about formal How predictability of disturbance can measures be of predictability for other arbitrarily non-arbitrarily determined selected levels of flow (e.g., mean daily flow- Poff and Ward 1989). Thus, I would argue that Resh et al. link disturbance and predictability arbitrarily chosen hydrographic statistics (e.g., using two statistical approaches. First, mean they or SD of flow defined over any arbitrary present graphical arguments (figs. 1, 3) that time hy- interval) can be used to characterize pre- drological events may be considered distur- dictability of streamflow generally (and thus fa- bances if they are extreme relative to the cilitate ex- among-stream hydrographic compari- pected long-term mean value for some specified sons); however, they cannot necessarily be used time interval. In their example, monthly tomeans characterize the predictability of disturbance + 2 standard deviations (SD) are used to specifically. represent the "predictable range" included in their Non-arbitrary criteria for disturbance are definition (though they caution against general needed to characterize the statistical predict- use of this specific criterion). While their ability ap- of a disturbance regime. Objective mea-

This content downloaded from 24.9.112.12 on Mon, 26 Aug 2019 01:31:04 UTC All use subject to https://about.jstor.org/terms 88 N. L. POFF [Volume 11 sures are needed to identify events that satisfy kiel 1985, also cf. Resh et al. 1988). However, the criteria provided in the Sousa (1984) or White the specification of a disturbance is scale-de- and Pickett (1985) definitions (viz., resource pendentloss, and is guided by the ecological ques- mortality, etc.). In streams, flows that move sub-tion(s) of interest. For example, less flow is re- strata (spates) are events that clearly satisfy thesequired to remove surface benthos than to scour conditions. Different approaches can be taken, deep hyporheos, leading to fundamentally dif- depending on the frame of reference. For ferentex- perceptions of what intensity of physical ample, bankfull discharge could be considered disruption constitutes a disturbance for these as a channel-wide measure of disturbance, be- two different ecosystem compartments. That a cause this level of flow is often correlated with physical event may constitute a disturbance at active channel formation and significant sedi- one level but not another indicates the hierar- ment transport (e.g., Richards 1982) and there- chical nature of disturbances (cf. Allen and Starr fore directly and indirectly modifies many 1982,lev- Frissell et al. 1986). Therefore, it is critical els of ecological organization. Poff and Ward that a researcher clearly specify the scale of ob- (1989) used this approach and showed that servation,sea- both with respect to the character- sonal predictability of an index of bankfull isticsflow of the physical disturbance (i.e., magni- varies greatly among streams across the USA. tude, duration) and to the ecological level(s) of Some streams experience bankfull discharge response. In principle, once a particular thresh- with a relatively fixed periodicity (e.g., in spring old disturbance flow is established, the statis- during snowmelt), while in others bankfull ticalflow properties of a disturbance regime (e.g., pre- occurs without seasonal pattern (i.e., at any dictability,time frequency) can be determined, if of the year). Importantly, some streams do hydrologic not data are available. This approach have a clear pattern of predictability using forcesthis us to link specific hydrologic events ex- criterion, indicating a gradient in disturbance plicitly to specific system characteristics (e.g., predictability across lotic ecosystems. substratum movement) and to specific ecolog- A finer-grain approach for identifying dis-ical responses, thereby facilitating non-arbi- turbance in a particular stream might be tai-trary descriptions of disturbance regimes. In- lored by determining at what level of flow patch- formation provided by this approach would specific disruptions occur. Indeed, Resh et allowal. us to contrast spatial and temporal dif- emphasize differential susceptibility of substra- ferences in ecological responses to specific hy- tum types to disturbance as an important drologicalme- events both within and among diator of biotic responses to hydrologic distur- streams. It would also encourage us to examine bances. Biota inhabiting a particular type differences of and similarities of response from substratum (e.g., sand) may be "disturbed" both by ecological and evolutionary perspectives a given hydrologic event, whereas biota on (see an- below). other patch type (e.g., gravel) may not be. Sed- iment transport processes are sufficiently well Why organisms are not necessarily "adapted" understood for us to estimate when substratum to predictable disturbances movement will occur (given data on median particle size, channel geometry, gradient, and Let us assume that disturbance predictability discharge). This habitat-specific approach hascan, in fact, be measured for a system. We are been adopted by researchers working on a stillva- left with Resh et al.'s assertion that, because riety of organisms, from meiofauna (Palmer organisms et are "adapted" to predictable envi- al. 1992) to macroinvertebrates (McElravy et ronmental al. fluctuations, only unpredictable 1989, Rader and Ward 1989, cf. Feminella and events count as disturbances. This argument in- Resh 1990) to fish (Erman et al. 1988) and cave troduces important issues and problems. Ini- fauna (Poulson and Culver 1969). tially, we ought to ask ourselves under what Both channel-wide and patch-specific ap- conditions adaptation to environmental varia- proaches attempt to identify threshold flows that tion can be expected. Thiery (1982) identified potentially constitute ecologically meaningful three general properties of a given set of en- physical events. The disturbance is thus defined vironmental conditions that collectively justify in terms of the physical event (a cause), not in the cost of adaptation. (This argument presup- terms of the biotic response (an effect) (see Ry- poses that species have the potential to adapt

This content downloaded from 24.9.112.12 on Mon, 26 Aug 2019 01:31:04 UTC All use subject to https://about.jstor.org/terms 1992] DEFINITION OF DISTURBANCE 89 to the set of conditions to begin with.) If the all if the environment simply "filters out" spe- environmental selection regime consists of ex- cies that lack appropriate mechanisms for per- treme events that are frequent in their occur- sisting under a given disturbance regime (see rence (relative to organism life span), not too Poff and Ward 1990). Even if lotic organisms deviant (relative to normal conditions), and con- are globally adapted as "weedy" (fugitive) spe- tingent (i.e., predictable), then organisms may cies inhabiting relatively variable, disturbed adapt, e.g., by modifying life cycles to minimize systems (see Townsend 1989), different degrees exposure to extremes. An important point, how- of "weediness" may be favored under different ever, is that different combinations of these environmental regimes (e.g., stable vs. flashy properties lead to different expectations about streams) (cf. Minshall 1988, Resh et al. 1988, whether adaptation would be favored and what Townsend 1989). The main point is that we just form it might take (Slobodkin 1968, Slobodkin do not know generally whether "adaptation" and Sanders 1969, Thiery 1982). For example, or "filtering" (or both, or neither) occurs, and adaptation may be possible where a set of con- it probably does not really matter for many eco- ditions is simultaneously deviant, infrequent, logical questions. So, rather than asserting a yet predictable, or where it is deviant, frequent, priori that organisms are "adapted" to the "pre- yet unpredictable. But if extreme events are in- dictable" portion of the environmental spec- frequent and unpredictable, or if they are sim- trum, this thought can be recast in terms of ply very rare (regardless of predictability), then testable hypotheses such as the following: or- appropriate selective pressures for maintaining ganismal attributes (e.g., life history tactics) or adaptation would presumably be absent. ecosystem attributes (e.g., primary or secondary These arguments indicate that just because an production) will vary among streams depend- environment is predictable does not mean that ing on the predictability or various interactive organisms are adapted to it; potential adapta- components of the disturbance regime (cf. Resh tion to disturbance predictability is itself con- et al. 1988). Data are needed to test these hy- tingent upon other attributes of the disturbance pothesis. regime. In north temperate streams, highly periodic seasonal changes in temperature or re- Linking disturbance and predictability source inputs (which are extreme and frequent can result in a tautology relative to population turnover times for many lotic species) may arguably result in evolution- Accepting the assertion that disturbances must ary adjustments of biota (e.g., Vannote et al.be unpredictable also involves the risk of re- 1980). However, the frequency and seasonal lying on the magnitude of ecological response predictability of hydrologic disturbances can to define disturbances-something Resh et al. vary widely in many streams across a range ofstate they wish to avoid. As an extreme and values (Poff and Ward 1989), leading to pre- hypothetical example, suppose a stream is sam- sumably noisy environmental selective forces pled only twice: once prior to and once shortly that may not promote local adaptation. following a "major spate". Was the spate a dis- In a more pragmatic context, the imbroglio turbance? If resources and biota were severely of the last two decades concerning the "adap- depleted by this spate, it would indeed be dis- tationist programme" and the "just-so stories" turbance according to Sousa (1984) or White and (Gould and Lewontin 1979) should caution us Pickett (1985). But the stream definition (Resh from over-interpreting the limited data sug- et al.) apparently requires some knowledge of gestive of local adaptation to hydrologic re- "predictability", and, unfortunately, hydrolog- gimes. While there is some evidence that pop- ic details for this hypothetical stream are lack- ulations inhabiting separate locales (where ing (often the case in the real world). According environmental selective pressures differ) may to the Resh et al. definition, the biota respond be characterized by different life history strat- only to unpredictable events; therefore, we egies (Leggett and Carscadden 1978, Crowl 1990) might deduce, tautologically, that this was a or behaviors (Flecker 1992), we cannot yet say (unpredictable) disturbance because the biota whether any of these patterns result from ge- responded. Had the ecological response been netic isolation or phenotypic flexibility. Indeed, less dramatic (e.g., 50% or 10% of the observed it is not necessary to invoke local adaptation atvalue), we might be tempted to conclude that

This content downloaded from 24.9.112.12 on Mon, 26 Aug 2019 01:31:04 UTC All use subject to https://about.jstor.org/terms 90 N. L. POFF [Volume 11 it was not a disturbance and thus "predictable". for example, no population can be perfectly This example is intentionally exaggerated "adapted" to to a hydrological event that disrupts make the point that strict adherence to the habitat"un- or exports resources. By contrast, in evo- predictability clause" in the definition of dis-lutionary terms, the loss of some individuals from turbance introduces ambiguity that can place a population during an extreme event may be us in the uncomfortable position of document- irrelevant because it is a normal component of ing disturbance depending on the magnitude species population dynamics. This evolutionary of biological response(s). As argued above, spec- perspective is related to the views of Lewontin ifying a physical threshold for disturbance with(1969) and Holling (1973) that higher-level sys- the potential to alter specified ecological or temeco- relationships (e.g., species persistence as system properties allows one to define predict- opposed to maintenance of population densi- ability of disturbance independent of ecologi- ties) may resist displacement by disturbance cal response to the event. Admittedly, ecological from a "basin of attraction" due to high system responses to such threshold flows may vary stabilityspa- (for recent discussions see Connell and tially or temporally, but this variation itself Sousa is [1983] and Sutherland [1990]). In this view, of prime ecological and evolutionary interest. a system incorporates its disturbance history, such that only unexpected events (in magni- Why disturbances can be predictable: tude or timing) have the potential to "push" ecological vs. evolutionary the system beyond its "stable" boundaries. This perspectives appears to be what Resh et al. have in mind when they state that disturbances can only be With the above arguments, I have attempted events occurring "outside a predictable range". to illustrate what I view as methodological and This is certainly a valid and important per- conceptual ambiguities associated with includ- spective, and one that is particularly applicable ing predictability in a definition of disturbance. to perturbation-dependent ecosystems (e.g., fire- Yet I agree with the authors of the Resh et al. driven forest and grassland systems), where paper that "unpredictable" disturbances typi- organisms and communities have become de- cally have substantially greater impact on lotic pendent on, and actually contribute to, the dis- ecosystems than "predictable" ones of similar turbance factors that are inherent to the system magnitude. This seeming contradiction can be (see Vogl 1980). resolved, I believe, by making explicit what in An evolutionary perspective on disturbance the Resh et al. paper is a largely implicit yet is inherent in statements such as only unpre- very important distinction about the time scale dictable events are disturbance (Resh et al. 1988, of biotic response to disturbance. Resh et al. Reice et al. 1990), or the absence of hydrologic state (p. 443) that the "effects of extreme flows disturbance following river impoundment is it- often have been observed as loss of numbers or self a disturbance (cf. Ward and Stanford 1983). biomass of certain taxa through flood scour But, or use of Sousa's (1984) or White and Pickett's desiccation". In considering the importance (1985) of definitions compels us to view distur- predictability of hydrologic events, they say,bances as "discrete" events that always have "We assume, therefore, that predictability ecologicalof impacts, though the ecological re- hydrologic regime is important with respect sponses to may well be constrained by evolution- evolution of behavior, life history strategies, ary (historical) events. Therefore, as suggested and competitive interactions". The authors byare Resh et al., populations may be adjusted to distinguishing between two very different scales a particular disturbance regime such that high of biotic responses to disturbance: ecological discharge occurring at an unusual time of the and evolutionary. In ecological terms, the con-year (in evolutionary or historical terms) indeed sequences of disturbance are clear: resources would are have a much greater impact (in ecolog- exported, population densities are reduced, etc.ical terms) on community structure than that These are, in fact, the criteria by which sameone spate occurring at the usual time of year identifies disturbance according to definitions (as illustrated in Resh et al.'s fig. 1). The pos- of Sousa (1984), White and Pickett (1985) and sibility of evolutionary adjustments to hydro- Resh et al. (1988). Therefore, in ecological time, logic disturbance itself generates important in- disturbances can clearly be predictable because, sights and hypotheses. For example, we could

This content downloaded from 24.9.112.12 on Mon, 26 Aug 2019 01:31:04 UTC All use subject to https://about.jstor.org/terms 1992] DEFINITION OF DISTURBANCE 91 hypothesize that under a particular disturbance disturbance contributes to stream ecosystem regime, biological mechanisms or attributes (e.g., structure and function (Resh et al. 1988). We mobility, behavior, physiology, life history, want to know what ecological and evolutionary morphology) that confer differential survival mechanisms contribute to species success and and persistence under that environmental se- persistence in lotic systems characterized by hy- lection (or filtering) regime should be favored. drologic events of variable intensity, frequency, Across a gradient of disturbance regimes within duration and predictability. Clearly, these issues or among streams, we would expect to find sta-go far beyond the narrow concern over defi- tistical associations between patterns of physi- nition of terms, and differences in perspective cal environmental variability and appropriate will persist and should be encouraged. But it is biological attributes (e.g., Lake et al. 1985, Min- worthwhile to explore the assumptions that un- shall 1988, Resh et al. 1988, Poff and Ward 1989, derlie our concepts and perspectives, so we do 1990, Townsend 1989, Reice et al. 1990, Wallace not communicate at cross-purposes while at- 1990). tempting to discern nature's web.

Concluding remarks Acknowledgements

In summary, I see four limitations in Resh et I would like to thank Margaret Palmer, Doug- al.'s assertion that disturbances are necessarily las Gill, and Rose Lew for commenting on an unpredictable. First, use of arbitrary hydrolog- early version of the manuscript and for spirited ical statistics may be appropriate for character- and stimulating discussions on the general top- izing predictability of streamflow generally but ic of predictability and disturbance. Two anon- not of disturbance specifically. Physically based ymous reviewers also provided helpful com- criteria for disturbance (e.g., flows initiating bed ments that improved the final version of the movement) with clear ecological consequences paper. (e.g., faunal or resource displacement) are needed to quantify differences in predictability of disturbance per se. Second, the assertion that Literature Cited lotic biota are "adapted" to "predictable" hy- ALLEN, T. F. H., AND T. B. STARR. 1982. Hierarchy: drologic disturbances lacks a firm empirical ba- perspective for ecological complexity. University sis and should be regarded as a hypothesis to of Chicago Press, Chicago. be tested. Third, defining disturbances as nec- COLWELL, R. K. 1974. Predictability, constancy, and essarily unpredictable can result in a tautology, contingency of periodic phenomena. Ecology 55: wherein response to disturbance and predict- 1148-1153. ability of disturbance are expressed in terms of CONNELL, J. H., AND W. P. SOUSA. 1983. On the ev- each other. Fourth, the important distinction idence needed to judge ecological stability or persistence. American Naturalist 121:789-824. between ecological and evolutionary time scales CROWL, T. A. 1990. Life-history strategies of a fresh- of response to disturbance, though implicit water snail in response to stream permanence throughout Resh et al., should be made explicit. and predation: balancing conflicting demands. Doing so allows us to consider predictability as Oecologia 84:238-243. a separate component of the disturbance re- ERMAN, D. C., E. D. ANDREWS, AND M. Yo- gime. Disturbances are, by definition, ecologi- DER-WILLIAMS. 1988. Effects of winter floods on cal events, but if their temporal distribution is fishes in the Sierra Nevada. Canadian Journal of "predictable enough", ecological responses may Fisheries and Aquatic Sciences 45:2195-2200. be small because organisms and communities FEMINELLA, J. W., AND V. H. RESH. 1990. Hydrologic have adjusted to them in evolutionary (histor- influences, disturbance, and intraspecific com- petition in a stream caddisfly population. Ecology ical) terms. Differences within or among streams 71:2083-2094. with respect to ecological responses to specified FLECKER, A. S. 1992. Fish predation and the evolu- disturbance intensities can be hypothesized to re- tion of invertebrate drift periodicity: evidence flect differences in disturbance regimes (pre- from neotropical streams. Ecology (in press). dictability, frequency) among sites, but data are FRISSELL, C. A., W. J. LISS, C. E. WARREN, AND M. D. needed to test such hypotheses. HURLEY. 1986. A hierarchical framework for Ecologists are ultimately concerned with how stream habitat classification: viewing streams in

This content downloaded from 24.9.112.12 on Mon, 26 Aug 2019 01:31:04 UTC All use subject to https://about.jstor.org/terms 92 N. L. POFF [Volume 11

a watershed context. Environmental Manage- SHELDON, J. B. WALLACE, AND R. WISSMAR. 1988. ment 10:199-214. The role of disturbance in stream ecology. Jour- GOULD, S. J., AND R. C. LEWONTIN. 1979. The span- nal of the North American Benthological Society drels of San Marco and the Panglossian para- 7:433-455. digm: a critique of the adaptationist programme. RICHARDS, K. 1982. Rivers: form and process in al- Proceedings of the Royal Society of London 205:luvial channels. Methuen, London. 581-598. RYKIEL, E. J. 1985. Toward a definition of ecological HOLLING, C. S. 1973. Resilience and stability of eco- disturbance. Australian Journal of Ecology 10:361- logical systems. Annual Review of Ecology and 365. Systematics 4:1-23. SLOBODKIN, L. B. 1968. Toward a predictive theory LAKE, P. S., L. A. BARMUTA, A. J. BOULTON, I. C. CAMP- of evolution. Pages 187-205 in R. C. Lewontin BELL, AND R. M. ST. CLAIR. 1985. Australian (editor). Population biology and evolution. Syr- streams and Northern Hemisphere ecology: com- acuse University Press, Syracuse, New York. parisons and problems. Proceedings of the Eco-SLOBODKIN, L. B., AND H. L. SANDERS. 1969. On the logical Society of Australia 14:61-82. contribution of environmental predictability to LEGGETT, W. C., AND J. E. CARSCADDEN. 1978. Lati- species diversity. Brookhaven Symposium on Bi- tudinal variation in reproductive characteristics ology 22:82-95. of American shad (Alosa sapidissima): evidence SOUSA, for W. P. 1984. The role of disturbance in natural population specific life history strategies in fish. communities. Annual Review of Ecology and Journal of the Fisheries Research Board of Canada Systematics 15:353-391. 35:1469-1478. SUTHERLAND, J. P. 1990. Perturbations, resistance, LEWONTIN, R. C. 1969. The meaning of stability. and alternative views of the existence of multiple Brookhaven Symposium on Biology 22:13-24. stable points in nature. American Naturalist 136: MCELRAVY, E. P., G. A. LAMBERTI, AND V. H. RESH. 270-275. 1989. Year-to-year variation in the aquatic mac- THIERY, R. G. 1982. Environmental instability and roinvertebrate fauna of a northern California community diversity. Biological Review 57:691- stream. Journal of the North American Bentho- 710. logical Society 8:51-63. TOWNSEND, C. R. 1989. The patch dynamics concept MINSHALL, G. W. 1988. Stream ecosystem theory: a of stream community ecology. Journal of the global perspective. Journal of the North Ameri- North American Benthological Society 8:36-50. can Benthological Society 7:263-288. VANNOTE, R. L., G. W. MINSHALL, K. W. CUMMINS, J. PALMER, M. A., A. E. BELY, AND K. E. BERG. 1992. R. SEDELL, AND C. E. CUSHING. 1980. The river Response of invertebrates to lotic disturbances: continuum concept. Canadian Journal of Fish- a test of the hyporheic refuge hypothesis. Oecolo- eries and Aquatic Sciences 37:130-137. gia (in press). VOGL, R. J. 1980. The ecological factors that produce POFF, N. L., AND J. V. WARD. 1989. Implications of perturbation-dependent ecosystems. Pages 63-94 streamflow variability and predictability for lotic in J. Cairns (editor). The recovery process in dam- community structure: a regional analysis of aged ecosystems. Ann Arbor Science, Ann Arbor, streamflow patterns. Canadian Journal of Fish- Michigan. eries and Aquatic Sciences 46:1805-1818. WALLACE, J. B. 1990. Recovery of lotic macroinver- POFF, N. L., AND J. V. WARD. 1990. Physical habitat tebrate communities from disturbance. Environ- template of lotic systems: recovery in the context mental Management 14:605-620. of spatiotemporal heterogeneity. Environmental WARD, J. V., AND J. A. STANFORD. 1983. The inter- Management 14:629-645. mediate disturbance hypothesis: an explanation POULSON, T. L., AND D. C. CULVER. 1969. Diversity for biotic diversity patterns in lotic ecosystems. in terrestrial cave communities. Ecology 50:153- Pages 347-356 in T. D. Fontaine and S. M. Bartell 158. (editors). Dynamics of lotic ecosystems. Ann Ar- RADER, R. B., AND J. V. WARD. 1989. The influence bor Science, Ann Arbor, Michigan. of environmental predictability/disturbance WHITE, P. S., AND S. T. A. PICKETT. 1985. Natural characteristics on the structure of a guild of disturbance and patch dynamics: an introduc- mountain stream insects. Oikos 54:107-116. tion. Pages 3-9 in S. T. A. Pickett and P. S. White REICE, S. R., R. C. WISSMAR, AND R. J. NAIMAN. 1990. (editors). The ecology of natural disturbance and Disturbance regimes, resilience, and recovery patchof dynamics. Academic Press, New York. animal communities and habitats in lotic ecosystems. Environmental Management 14:647-660. Received: 18 July 1991 RESH, V. H., A. V. BROWN, A. P. COVICH, M. E. GURTZ, Accepted: 19 November 1991 H. W. LI, G. W. MINSHALL, S. R. REICE, A. L.

This content downloaded from 24.9.112.12 on Mon, 26 Aug 2019 01:31:04 UTC All use subject to https://about.jstor.org/terms