Patterns and Processes Geomorphic Features of the Earth's Surface Regulate the Distribution of Organisms and Processes

Landform Effects on Ecosystem Patterns and Processes Geomorphic features of the earth's surface regulate the distribution of organisms and processes

F. J. Swanson, T. K. Kratz, N. Caine, and R. G. Woodmansee

nderstanding the form, be- types. The static and dynamic physi- havior, and historical context Landform effects provide cal characteristics of landscapes high- U of landscapes is crucial to un- lighted in the identification of these derstanding ecosystems on several temporal and spatial classes are important considerations temporal and spatial scales. Land- when analyzing individual ecosystems forms, such as floodplains and allu- perspectives for and making comparisons among eco- vial fans, and geomorphic processes, examining soil, systems. We will not consider here such as stream erosion and deposi- other classes of geomorphology-eco- tion, are important parts of the set- vegetation, and aquatic system interactions, such as effects of ting in which ecosystems develop and fauna and flora on rates of geomor- material and energy flows take place. ecosystems and phic processes. Over the long term, geomorphic pro- for interpreting This article is based on ongoing cesses create landforms; over a efforts to compare ecosystems among shorter term, landforms are boundary ecosystem processes the 15 diverse sites studied in the conditions controlling the spatial ar- Long -Te r m E co 1o gi c a 1 Re s e arch rangement and rates of geomorphic (LTER) program (Callahan 1984). processes. (Hack and Goodlett 1960). These sites include an old-growth co- Ecosystems respond to both land- Fascinating interactions among nifer forest in the Cascade Range of forms and geomorphic processes. The geomorphic processes, landforms, Oregon, Rocky Mountain alpine tun- history of geomorphic processes may and biota occur on various temporal dra in Colorado, lakes in forest land be expressed directly in the composi- and spatial scales. On a fine spatial of Wisconsin, and shortgrass steppe tion and structure of vegetation, scale, for example, parts of individual in Colorado. Within this diversity we where geomorphic events and vegeta- plants may retard soil erosion or may find valuable generalities. tion develop together. Geomorphic be damaged by earth movement. On a processes operating before the estab- much greater scale, the geographic Landscapes, landforms, and lishment of existing vegetation, or distribution and height of landmasses geomorphic processes those subtly coexisting with the vege- broadly control distributions of tation, may have their greatest influ- plants and animals through influences A discussion of effects of landforms ence on vegetation through control- on environmental gradients of tem- on ecosystems is hindered by a lack of ling patterns of soil properties across perature and moisture and on corri- concise, widely accepted definitions a landscape, as in toposequences dors of migration during environmen- of key terms and concepts. Naveh tal change. Rigorous examination of (1982) traces some of the history of F. J. Swanson is a research geologist at the interactions among geomorphology, usage and perception of landscape, a USDA Forest Service, Pacific Northwest ecosystems, and landscapes at the particularly crucial but poorly de- Research Station, Corvallis, OR 9733 1. spatial scales of hectares to thousands fined term. From roots in the Bible, T. K. Kratz is an assistant scientist at the of square kilometers is required for secular literature, and art emphasiz- Center for Limnology, University of Wis- further understanding of landscape ing a highly idealized view of nature, consin, Madison 53706. N. Caine is a ecology (Risser et al. 1984) and eco- landscape has gained increasing use in professor in the Department of Geogra- fields such as physical geography, phy, University of Colorado, Boulder system structure and function. 80309. R. G. Woodmansee is the director In this article, we identify and ex- ecology, and landscape architecture. of the Natural Resources Ecology Labora- plore four classes of effects of land- Landscape commonly refers to the tory, Colorado State University, Fort Col- forms on ecosystems, providing ex- form of the land surface and associat- lins 80523. amples from a range of ecosystem ed ecosystems at scales of hectares to

92 BioScience Vol. 38 No. 2 many square kilometers. Landform is usually used at a finer scale and more specifically, such as a landform carved out by a landslide or created by sediment deposition forming a gravel bar. Landscapes are composed of landforms and ecological units, such as patches (Forman and Godron 1986). By geomorphic processes we refer to mechanical transport of organic and inorganic material. In addition to surface erosion and mass movement, geomorphic processes include transport of material in solution in surface and subsurface water and biogenic soil movement by animals and root throw. A drawback of the term geomorphic process is the implication that the process is necessarily shaping landforms. Frequently, however, geomorphic process refers to transfer of material or disturbance of biota without regard to development of landforms or the time scale in which that Figure 1. Examples of the four classes of landform influence on ecosystem patterns and processes. a. Class 1. Topographic influences on rain and radiation (arrow) shadows. b. occurs. We use the term in this general Class 2. Topographic control of water input to lakes. Lakes high in the drainage system sense. may receive a higher proportion of water input by direct precipitation than lakes lower in the landscape where groundwater input (arrow)predominates. c. Class 3. Landform- Effects of landforms on constrained disturbance by wind (arrow) may be more common in upper-slope ecosystems locations. d. Class 4. The axes of steep concave landforms are most susceptible to disturbance by small-scale landslides (arrow). We consider four classes of landform effects on ecosystem patterns and processes (Figure 1): fluence other environmental and biot- ker and Niering 1965). Implicit in the 0 Class 1: Landforms-by their eleva- ic factors. Ecosystems are considered relations between patterns of vegetation, aspect (direction in which land static in class 1, but dynamic in class- tion and landforms are the influences surface faces), parent materials, and es 2 and 3. In class 4, we regard both of elevation and aspect on solar ener- steepness of slope-influence air landforms and ecosystems as gy and water regimes at patches with- and ground temperature and the dynamic. in complex ecosystems. As moist air quantities of moisture, nutrients, These four classes are valuable as a masses flow over hills and mountains, and other materials (e.g., pollut- framework for discussion rather than higher precipitation commonly falls ants) available at sites within a as a rigorous classification scheme. In at higher elevations (orographic ef- landscape. the following discussion, we give ex- fects), but such simple patterns may a Class 2: Landforms affect the flow amples of each class, address the ag- be confounded by the effects of rain of organisms, propagules (e.g., gregated effects landforms have in and shadow, fog and cloud belts, and seeds and sproutable root frag- determining the patterns of natural timing of snow accumulation and ments), energy, and material (water, landscapes, and consider implications melt. Neither the simple nor the com- dissolved material, and organic and for landscape ecology. Some effects of plex effects of landforms on environ- inorganic particulate matter) landforms on ecosystems at the spa- mental gradients, however, can com- through a landscape. tial scale of interest for this discussion monly be separated from the effects of 0 Class 3: Landforms may influence do not fall neatly into one of these landforms on movement of materials the frequency and spatial pattern of four classes. and energy. nongeomorphically induced disturbance by agents such as fire, wind, Landforms .and environmental gradi- Landforms and movement of materi- and grazing. ents. Elevation and aspect are envi- al, organisms, propagules, and ener- 0 Class 4: Landforms constrain the ronmental gradients that have been gy. Landforms regulate movement of spatial pattern and rate or frequen- widely recognized in mapping and in material, organisms, propagules, and cy of geomorphic processes that al- gradient analysis of patterns of vege- energy across a landscape by defining ter biotic features and processes. tation across landscapes ranging from gravitational gradients, influencing a few hectares to thousands of square flow paths of wind, and forming bar- In classes 1-3, we consider landforms kilometers (Billings 1973, Hack and riers and corridors for movement. to be unchanging boundaries that in- Goodlett 1960, Kessell 1979, Whitta- The role of landforms in controlling

February 1988 93 movement of material is implicit in forms strongly control the location Concentrations of silica, for exam- studies of nutrient cycling and sedi- and rate of processes that move both ple, are several orders of magnitude ment routing within drainage basins, soil and sediment. For example, steep, greater in groundwater than in pre- along transects crossing topographic concave hillslopes are landform units cipitation; biological productivity of features, and in other landform set- that are the predominant site of small, lakes high in the flow system can tings involving flow paths. The flow rapid landslides (Dietrich and Dunne become seasonally limited by silica, paths of material and energy move- 1978). whereas lakes low in the flow system ment across a landscape may vary Landforms also influence the tem- are likely to be limited by light or greatly depending on factors operat- poral and spatial patterns of fluxes of atmospherically supplied nutrients. In ing at several scales. Water follows material carried across landscapes by Crystal Lake, a lake high in the Wis- gravitational gradients. Dominant surface water. In lakes at LTER sites consin landscape, groundwater ac- wind direction or paths of animal in the Colorado alpine (Caine 1984) counts for less than 8% of the water migration, on the other hand, may and in the forest land of Wisconsin budget, yet it accounts for 50% of the produce other patterns of material (Magnuson et al. 1984), characteris- annual silica budget, the rest coming flux, controlled by landforms at tics of water quality vary with a lake's from regeneration from bottom sedi- broader scales. position in the landscape. At both ments (Hurley et al. 1985). Small To analyze transport through and sites, specific conductance-a general variations in the amount of ground- temporary storage of soil and sedi- measure of solute concentration and water entering the lake-caused, for ment in landscapes, scientists com- acid neutralizing capacity-is greater example, by differences in snowfall or monly use natural landform units to in lakes lower in the landscape, re- timing of snowmelt-can lead to rna- compartmentalize a landscape (Die- flecting the proportionate increase of jor differences in the spring algal trich and Dunne 1978, Swanson et al. surface or groundwater contributed blooms in the lake. In contrast, lakes 1982). Such landform units are addi- to lakes lower in a flow system (Fig- lower in the flow system are not tionally characterized by their pre- ure 2). More of the water entering limited by silica, so small fluctuations dominant process of transfer or such lakes has passed through the in the groundwater supply will not dynamics of storage, because land- vegetation and soil, entraining prod- affect algal dynamics in those lakes. ucts of rock weathering and decom- Landforms may also delimit the position and therefore giving higher ranges of some vertebrates (Forman 30 Green Lakes, Colorado solute concentrations. Lakes high in and Codron 1986). Gullies, streams, 1 the flow system, in contrast, receive a and cliffs may form physical barriers higher proportion of water by precip- to movement, or they may act as itation. which falls directly into the convenient, but passable, features to lakes or which drains rapidly from mark the boundaries between home adjacent steep, rocky slopes. ranges of neighboring animals. Cliff Not only can landforms influence faces and associated talus slopes form - Drainage area (km2) mean values of certain lake parame- habitats at several scales for a com- ters, but also their seasonal or annual munity of small mammals and birds 100 Northern Lakes, Wisconsin c variability. In the lakes of the Colora- in semiarid landscapes (Maser et al. 0 1 U3 do alpine, where spring snowmelt 1979). 5 80- 1 0 I causes large seasonal pulses in water The distribution of size and abun- 0 -.- flow, seasonal variation in specific dance of cavities in outcrops of bed- 60- "n conductance decreases as a function rock and in accumulations of boul- I of drainage area. This change reflects ders control the distribution of 40 i the buffering effects of larger lakes rodents. Raptors use the air space i lower in the landscape. Smaller lakes above, riding convective wind cur- to If at higher elevations appear be af- rents induced by the heating of cliff fected more by snowmelt and exhibit and talus surfaces. These birds prey a larger seasonal amplitude in specific on rodent populations occupying

High- Landscape position -Low conductance. fine-scale landform features below. In contrast, specific conductance in The grazing behavior of large ani- Figure 2. Specific conductance in the the Wisconsin lakes exhibits small mals may exhibit profound effects of Green Lakes, Colorado, and Northern variation, on the order of analytical landforms (Schimel et al. 1986, Senft Lakes, Wisconsin, study sites. The Green uncertainty, because of a much less et al. 1985). In semiarid grasslands, Lakes record shows conductance (K) in- pronounced effect of snowmelt and cattle preferentially graze lowland creasing with drainage area (1982: K = the relatively conservative behavior of swales, presumably because the influ- 8.1 + 1.8 area [km2] [r = 0.9961; 1985: K most of the important cations and ence of landforms on spatial patterns = 5.7 + 2.0 area [km2] [r = 0.992]). Values for Wisconsin lakes during spring anions. Although seasonal variation of soil moisture and nutrients results and fall mixes 1982-1985 demonstrate in specific conductance is low in the in better forage. This response of ani- higher conductances with lower landscape Wisconsin lakes, annual variation in mals to landforms may lead to redis- position. Specific conductance was mea- supply of limiting nutrients can be tribution of nutrients on the land- sured at 25" C. Values shown ? 1 stan- significant and can vary with a lake's scape (Schwartz and Ellis 1981, Senft dard deviation. position in the landscape. et al. 1985).

94 BioScience Vol. 38 No. 2 Landforms and nongeomorphically animals. In class 3, agents of distur- Creeks, lateral and vertical channel induced disturbances. The roles of bance operate through vegetation, changes appear to have occurred iandforms in controlling patterns of which may lead to secondary distur- much more catastrophically when in- ecosystem disturbance across land- bance by acceleration of geomorphic frequent, major floods move the scapes are not clearly understood ex- processes, such as increased surface coarse bed material. Chronic, over- cept in a few obvious cases, such as erosion after wildfire. In class 4, geo- bank deposition appears to be limited inundation by floodwaters in chan- morphic processes are the primary in part by a paucity of fine sediment, nels and on floodplains. Effects of disturbances. particularly in North Boulder Creek landforms on frequency and intensity Two dramatic examples of geomor- where upstream lakes trap sediment. of disturbance by fire, snow, and phic disturbances are landslides and Consequently, the composition of wind, for example, are recognized in lateral shifts of river channels. Por- floodplain vegetation along mountain anecdotes, but generally only qualita- tions of the landscape subject to these streams may be less restricted to spe- tively (e.g., Swanson 1981). disturbances are characterized by dis- cies that are best adapted to frequent, Landforms can protect certain ar- tinctive processes and frequencies of minor deposition of fine-grained eas from disturbance by providing change, and in each case landforms sediment. firebreaks and shelter from physical constrain the frequency and magni- These physical dynamics of fluvial damage by disturbances such as wind. tude of change. environments can make a substantial In studies of the forests of Mount RIVER CHANNELS. The geomorphic imprint on ecosystem patterns by Rainier National Park, Hemstrom dynamics of valley floors differ sub- controlling the distribution of sub- (1982) notes the importance of major stantially between mountain and low- strates on which plant and animal ridges and valley bottoms in con- land areas because of contrasts in the communities develop. Pastor et al. straining the spread of wildfire. Con- degree of landform control. Using age (1982) provide examples of how geo- sequently, boundaries between stands analysis of floodplain forests, Everitt morphic features, operating through of differing structure and age are (1968) determined that, during the soil moisture and chemistry, influence more likely to occur at these topo- preceding 1 00-year period, persistent patterns of vegetation composition graphic positions than at others. lateral migration of a reach of the and productivity. Landforms may also increase fre- meandering Little Missouri River LANDSLIDES. In many steep land- quency of disturbance by channeling (slope = 0.00085) in North Dakota scapes, small landslides are believed fire, wind, and other nongeomorphic eroded and reconstructed floodplain to occur repeatedly from depressions agents of disturbance into an area. surfaces in an area of valley floor on bedrock, variously termed bed- Topography, even subtle ridges and averaging 5.9 channel widths (0.54 rock hollows, swales, zero-order ba- depressions, controls patterns of km2 per km of valley length). Over sins, and headwalls (Dietrich and snow accumulation in alpine settings the same period, the steep (slope = Dorn 1984, Dietrich and Dune and the plains of Colorado and deter- 0.042), straight channel of French 1978). These hollows slowly collect mines the locations of snow sources Pete Creek in the Oregon Cascade soil from the surrounding hillslopes and sinks during periods of snow Range reset an area of only 1.0 chan- by surface erosion, root throw, soil redistribution by wind. Persistent nel width (0.015 km2 per km of valley creep, and other processes over peri- snow accumulation may suppress length) (Grant 1986). North Boulder ods of centuries and millenia. Eventu- vegetation, creating a site of bare soil Creek in the Colorado Front Range ally a landslide removes all or a por- that is a source of sediment associated has a similar gradient (0.030) and tion of the stored soil from the with a source of water-the snow- quasimeandering form, but it accom- hollow. Such a landslide commonly bank. Landforms thereby trigger a plished even less reworking of its val- occurs during periods of heavy rain- chain of events beginning with redis- ley floor (less than 0.0005 km2 per fall or rapid snowmelt. The potential tribution of water in the form of snow km of valley length) over a thousand for a landslide may increase as soil (a class 2 effect) that leads to localized years or more (Furbish 1985). The depth in hollows exceeds the rooting disturbances of vegetation by short- steep channel, coarse bed and bank depth of woody plants growing in ening the growing season and the sediment, and bedrock outcrops limit and around hollows, thereby limiting period of possible establishment (a lateral channel change in both the the effectiveness of soil mass anchor- class 3 effect). Disturbance of the Oregon and Colorado mountain ing by roots passing vertically and immediate site and adjacent areas by streams. horizontally into stable substrates. surface erosion and sediment trans- These mountain and lowland fluvi- Potential for occurrence of a landslide port (a class 4 effect) may follow. a1 environments also have contrasting also increases when the contribution regimes of chronic disturbance by of roots to soil strength is minimized Interactions among landform, geo- over b ank sediment a ti on. Ve r ti c a1 by decomposition after mortality morphic process, and ecosystem. The growth of floodplains beside the Little caused by wildfire, logging, or other distinction between landform effects Missouri River continues by over- disturbance of vegetation. of classes 3 and 4 is the emphasis on bank deposition for more than a cen- Landform units prone to landslides physical dynamics of the landscape in tury after establishment of initial veg- may experience dramatic change on class 4. For both classes, landforms etation on a freshly deposited ecologically relevant time scales. constrain the movement of agents floodplain surface (Everitt 1968). In From one geologic terrane to another, that disturb the ecosystem-fire, sur- the boulder-dominated channels of the frequency of a landslide reoccur- face and subsurface water, wind, and French Pete and North Boulder ring at a particular site can range

February 1988 95 from a few decades (Shimokawa geomorphic change relative to the An even more complex set of land- 1984) to 10,000 years or more (Die- rate of development of soil and bio- form effects on ecosystems are found trich and Dorn 1984, Kelsey 1982). logical features of interest. in studies of soii catenas in the short- At sites with a high frequency of In the Cascade Range of the Pacific grass steppe of Colorado. The Colo- landslides, landslides can truncate the Northwest, trees are long lived and rado State University LTER program succession of vegetation. Where land- changes in landform by fluvial pro- has analyzed catenas, the connected slides occur less frequently, soil devel- cesses and various types of mass series of soil and plant associations opment and climate change may alter movements are frequent, so the spa- extending down a topographic se- both the potential for landslides and tial distribution of age classes of vege- quence of sites from ridge to valley the character of vegetation developed tation on geomorphically active por- bottom (Jenny 1980). Schimel et al. on the site between landslides. Gradu- tions of the landscape reflect some of (1985) observed that, going down the al filling of landslide scars with soil the history of geomorphic distur- catena, organic carbon, nitrogen, and from adjacent slopes (Shimokawa bances. Other systems, such as much phosphorous and soil depth increase. 1984) may be viewed as a form of of the alpine environment of the These and similar patterns in other chronic disturbance of vegetation at southern Rocky Mountains and the soil constituents both result in and are the same time that it is contributing to lake district of northern Wisconsin, the result of higher productivity in soil and landform recovery, much like may be so stable geomorphically that lower slope positions. siltation on sites of overbank deposi- it is reasonable to consider the land- In the subtle topography of the tion on floodplains. scape as a static physical template for shortgrass steppe near Fort Collins, Catastrophic events are more im- ecosystem development, even on the Colorado, the effects of aspect and portant in the geomorphic distur- time scale of many millenia. In many relative elevation on the input of bance regimes of some landform units of these areas, however, there is abun- moisture and solar energy to sites than in others. Bedrock hollows sub- dant evidence that the landscape was (class 1 effects) are not pronounced, ject to periodic landslides, for exam- much more dynamic in the past. In but landform effects on movement of ple, operate on a pulse-reset basis. the southern Rockies this was espe- fluvial and aeolian material have a Other landform units, such as flood- cially true during and immediately variety of influences on landscape plains of meandering rivers, experi- after glacial stages, the “paraglacial” patterns. Higher biomass production ence progressive resetting by fre- conditions of Church and Ryder in lower slope positions occurs in quent, small, discrete increments of (1972). response to more fertiie soiis and change. In yet other landform units, The complex interactions among higher moisture availability as a result the soil, geomorphic processes, and landforms and spatial patterns of eco- of subsurface water flow from ups- landform are believed to have co- system characteristics are exemplified lope (a class 2 effect). Furthermore, evolved gradually through time, re- by vegetation in the steep, high-relief surface and subsurface flow transport sulting in conditions of a steady-state landscape of Mount Rainier National fine soil particles and organic matter dynamic equilibrium along a topo- Park (see cover). Long, smooth slopes to lower slope positions (a class 2 graphic sequence, as inferred for with more than a 1000-meter range in effect). Aeolian deposits also occur in some soil catenas. elevation result in distinctive zones of lower slope positions (Schimel et al. vegetation habitat types by altitude, 1985).Topographic effects on alluvial Complex landscapes clearly visible from the opposite val- and aeolian deposits, combined with ley wall (Hemstrom 1982, Hemstrom faster pedogenesis in the more moist Patterns of biota across natural land- and Franklin 1982). This pattern de- toeslopes, result in higher concentra- scapes typically reflect interactions velops in response to temperature and tions of silt and clay in toeslope loca- among most or all of the four classes moisture gradients (a class 1 effect). tions and, therefore, greater moisture- of landform effects on biota, soil, and Superimposed on this pattern is one holding capability. geomorphic disturbances. These in- of higher productivity in valley bot- This description of a catena is con- teractions are commonly so complex toms, where moisture and nutrients sistent with conventional wisdom, as and entwined in the history of a site are least limiting as a result of downs- described by Jenny (1980), and it is (embodied in its soils and other sys- lope movement (a class 2 effect). valid for sites in the Colorado short- tem components with long memory) Hemstrom (1982) documents two grass steppe, where toposequences that individual landform effects may class 3 effects. Landform features, es- are strongly developed. Field studies, be impossible to identify. pecially major ridges and valley bot- however, reveal that conventional ca- The usefulness and appropriateness toms, serve as barriers to the passage tena concepts are not sufficient to of examining each of the classes of of fire, a major disturbance in the explain soil and vegetation patterns in landform effects we have discussed forest around Mount Rainier. Snow much of this area. The pervasive ef- depends on the objectives of a study avalanches repeatedly move down fects of wind redistribution of soil and the characteristics of the ecosys- bedrock-defined paths, cutting dis- and snow impose other landform- tems and landscapes involved. In the tinctive swaths through forest vegeta- controlled patterns on ecosystem most dynamic systems, geomorphic tion. The floors of several valleys properties that are not explained by and ecological changes occur on simi- draining Mount Rainier also experi- toposequence landform elements. lar time scales. The potential impor- enced severe disturbance by mud- The importance of interactions be- tance of class 4 interactions, for ex- flows triggered high on the volcano (a tween wind and landform is indicated ample, depends in part on the rate of class 4 effect) (Figure 3). by soil stratigraphic studies, which

96 BioScience Vol. 38 No. 2 include observation of buried soils (Doehring et al. 1984, Schimel et al. 1985). Erosion and deposition events forming buried soil must have consti- tuted geomorphic disturbances of vegetation (a class 4 effect). Further- more, snow drifting to the lee side of hills creates localized sites with aug- mented water supply, which can lead to substantial differences in vegetative production (Woodmansee and Adam- sen 1983). These, and perhaps other, processes have resulted in a surpris- ingly uniform A-horizon (the upper soil layer) given the highly variable properties of the B-horizon (the un- derlying soil layer). Because the water-holding capacity of the A-horizon (1.5-2.0 cm) exceeds most single rainfalls and snowmelts, the water resource is relatively uniformly dis- tributed over the landscape. Conse- quently, the composition of the vegetation also tends to be rather constant spatially.

Conclusions The effects of landforms on ecosystem development and change have several important implications for the developing field of landscape ecology (For- man and Godron 1986, Naveh 1982, Risser et al. 1984). Landscape ecology as defined by Risser considers “the development and dynamics of spatial heterogeneity, spatial and temporal interactions and exchanges across heterogeneous landscapes, influences of spatial heterogeneity on biotic and abiotic processes, and management of Figure 3. The valley of White River draining the flanks of Mt. Rainier contains numerous examples of landform influences on vegetation and disturbance patterns. spatial heterogeneity.” In this article we have identified various ways in which patterns of ecosystem structure, composition, and function are trum of landform effects has received important factor in designing sam- controlled by landforms and geomor- little attention in the literature of pling programs for many aspects of phic processes. Knowledge of these landscape ecology. lake water chemistry and biology. geomorphic underpinnings of ecosys- Several points emerge: 0 Patterns of soils and vegetation tems is essential to interpreting eco- studied in the shortgrass steppe of systems in the context of their p In general, patterns of landscape Colorado indicate that landform- landscapes. and landforms are easier to observe ecosystem interactions may take The influence of landforms on pat- than ecosystem processes, so an un- multiple forms and that patterns terns of soil, vegetation, animals, and derstanding of how landforms affect imposed by one set of interactions aquatic ecosystems across a landscape processes yields some power in pre- may be overridden by another set. contribute to developing and main- dicting ecosystem behavior. The ef- Careful field studies to explore mul- taining a patchwork ecosystem. This fects of landscape position and soil tiple working hypotheses are essen- influence occurs through the effects of factors on lake processes, for exam- tial to elucidating the full range of landforms on environmental gradi- ple, imply that lakes high in the important interactions. ents (class 1 and 2 effects) and regula- landscape will be more sensitive to 0 The interactions of landform effects tion by landform of the patterns and acid deposition than lower lakes on ecosystem development during frequency of the disturbance (class 3 (Eilers et al. 1983). Therefore, posi- periods without major disturbance and 4 effects). To date, the full spec- tion within a drainage basin is an and on movement of disturbances

February 1988 97 across a landscape may influence Eilers, J. M., G. E. Glass, K. E. Webster, and Managed Forests-The Blue Mountains of strongly the persistence and geome- J. A. Rogalla. 1983. Hydrologic control of Oregon and Washington. USDA Handbook lake susceptibility to acidification. Can. J. 533, Portland, OR. try of the landscape mosaic defined Fish. Aquat. Sci. 40: 1896-1904. Naveh, Z. 1982. Landscape ecology as an by patches of vegetation, soils, or Everitt, B. L. 1968. Use of the cottonwood in emerging branch of ecosystem science. Adv. other ecosystem components. To an investigation of the recent history of a Ecol. Res. 12: 189-237. our knowledge, these important in- floodplain. Am. J. Sci. 266: 417-439. Pastor, J., J. D. Aber, C. A. McClaugherty, and teractions have not been examined Forman, R. T. T., andM. Godron. 1986. Land- J. M. Melillo. 1982. Geology, soils, and veg- scape Ecology. John Wiley & Sons, New etation of Blackhawk Island, Wisconsin. Am. quantitatively. York. . Midl. Nat. 108: 266-277. Furbish, D. J. 1985. The stochastic structure of Risser, P. G., J. R. Karr, and R. T. Forman. These concepts of landform effects on a high mountain stream. Ph.D. dissertation. 1984. Landscape ecology-directions and ecosystems provide broad temporal University of Colorado, Boulder. approaches. Illinois Natural History Survey and spatial perspectives for designing Grant, G. E. 1986. Downstream effects of tim- Special Publication No. 2, Champaign, IL. sampling of soil, vegetation, and ber harvest activities on the channel and Schimel, D. S., W. J. Parton, F. J. Adamsen, aquatic ecosystems and for interpret- valley floor morphology of western Cascade R. G. Woodmansee, R. L. Senft, and M.A. streams. Ph.D. dissertation. Johns Hopkins Stillwell. 1986. The role of cattle in nitrogen ing community and ecosystem pro- University, Baltimore, MD. budget of a shortgrass steppe. Biogeochemis- cesses in dynamic landscape. Hack, J. T., and J. C. Goodlett. 1960. Geomor- try 2: 39-52. phology and forest ecology of a mountain Schimel, D., M. A. Stillwell, and R. G. Wood- region in the central Appalachians. US Geo- mansee. 1985. Biogeochemistry of C, N, and Acknowledgments logical Survey Professional Paper 347, Res- P in a soil catena of the shortgrass steppe. ton, VA. Ecology 66: 276-282. This work was supported by grants Hemstrom, M. A. 1982. Fire in the forests of Schwartz, C. C., and J. E. Ellis. 1981. Feeding BSR 8315174, BSR 8514325 (F.J.S.), Mount Rainier National Park. Pages 121- ecology and niche separation in some ungu- BSR 8508356 (F.J.S.), BSR 8514330 126 in E. E. Starkey, J. F. Franklin, and J. W. lates on the shortgrass prairie. J.Appl. Ecol. (T.K.K.), BSR 8514329 (N.C.), and Mathews, eds. Ecological Research in Na- 18: 343-353. tional Parks of the Pacific Northwest. Pro- Senft, R. L., L. R. Rittenhouse, and R. G. BSR 8114822 (R.G.W.) to our respec- ceedings of the Second Conference on Scien- Woodmansee. 1985. Factors influencing pat- tive Long-Term Ecological Research tific Research in the National Parks. Oregon terns of cattle grazing behavior on shortgrass sites from the National Science State University Forest Research Laboratory, steppe. J. Range Manage. 38: 82-87. Foundation. Corvallis, OR. Shimokawa, E. 1984. A natural recovery proc- Hemstrom, M. A., and J. F. Franklin. 1982. ess of vegetation on landslide scars and land- Fire and other disturbances of the forests in slide periodicity in forested drainage basins. References cited Mount Rainier National Park. Quat. Res Pages 99-107 in. C. L. O’Loughlin, and A, J. 18: 32-51. Pierce, eds. Symposium on Effects of Forest Billings, W. D. 1973. Arctic and alpine vegeta- Hurley, J. P., D. E. Armstrong, G. J. Kenoyer, Land Use on Erosion and Slope Stability, tion: similarities, differences, and susceptibil- and C. J. Bowser. 1985. Groundwater as a East-West Center, University of Hawaii, ities to disturbance. BioScience 23: 697-704. source of silica for diatoms in a precipita- Honolulu. Caine, N. 1984. Elevational contrasts in con- tion-dc ainated lake. Science 227: 1576- Swanson, F. J. 1981. Fire and ge morphic temporary geomorphic activity in the Colo- 1579. processes. Pages 401-420 in Proceedings of rado Front Range. Studia Geomorphologica Jenny, H. 1980. The Soil Resource: Origin and the Conference on Fire Regimes and Ecosys- Carpatho-Balcanica 18: 5-31. Behavior. Springer-Verlag, New York. tem Properties. USDA Forest Service Gener- Callahan, J. T. 1984. Long-term ecological re- Kelsey, H. M. 1982. Hillslope evolution and al Technical Report WO-26, Washington, search. BioScience 34: 363-367. sediment movement in a forested headwater DC. Church, M., and J. M. Ryder. 1972. Paragla- basin, Van Duzen River, North Coastal Cali- Swanson, F. J., R. J. Janda, T. Dunne, and cia1 sedimentation: a consideration of fluvial fornia. Pages 86-96 in F. J. Swanson, R. J. D. N. Swanston. 1982. Sediment Budgets processes conditioned by glaciation. Bull. Janda, T. Dunne, and D. N. Swanston, eds. and Routing in Forested Drainage Basins. Geol. SOC.Am. 83: 3059-3072. Sediment Budgets and Routing in Forested USDA Forest Service General Technical Re- Dietrich, W. E., .and R. Dorn. 1984. Signifi- Drainage Basins. USDA Forest Service Gen- port PNW-141, Portland, OR. cance of thick deposits of colluvium on hill- eral Technical Report PNW-141, Portland, Whittaker, R. H., and W. A. Niering. 1965. slopes: a case study involving the use of OR. Vegetation of Santa Catalina Mountains, Ar- pollen analysis in the coastal mountains of Kessell, S. R. 1979. Gradient Modeling. izona: a gradient analysis of the south slope. northern California. Geology 92: 147-158. Springer-Verlag, New York. Ecology 46: 429-452. Dietrich, W. E., and T. Dunne. 1978. Sediment Magnuson, J. J., C. J. Bowser, and T. K. Kratz. Woodmansee, R. G., and F. J. Adamsen. 1983. budget for a small catchment in mountain- 1984. Long-term ecological research (LTER) Biogeochemical cycles and ecological hierar- ous terrain. Zeitschrift fur Geomorphologie on north temperate lakes of the United chies. Pages 497-516 in R. R. Lowrance, Suppl. Bd. 29: 191-206. States. Verb. Int. Verein. Limnol. 22: 533- R. L. Todd, L. Asmussen, and R. A. Leonard, Doehring, D. O., C. M. Yonker, and D. S. Schi- 535. eds. Nutrient Cycling in Agricultural Ecosys- mel. 1984. Geomorphic history and soil gen- Maser, C., J. E. Rodick, and J. W. Thomas. tems. Special Publication No. 23, University esis at the LTER shortgrass steppe site. Bull. 1979. Cliffs, talus, and caves. Pages 96-103 of Georgia, College of Agriculture Experi- Ecol. SOC. Am. 65: 153. in J. W. Thomas, ed. Wildlife Habitats in ment Station, Athens, GA.

98 BioScience Vol. 38 No. 2