The Biogeographic Significance of a Large, Deep Canyon: Grand Canyon of the Colorado River, Southwestern USA

Mountains and uplifted areas occupy more than 10% of the Earth’s surface, and their associated drainages rather commonly develop as constrained, canyon-bound channels. Therefore, large deep canyons (LDCs) are relatively frequently encountered, persistent landforms, occurring either as steeply dipping, V-shaped canyons that emerge from fold, fault-block, volcanic, and dome mountains, or occurring as drainages incised into uplifted plateaus. The latter type include the world’s deepest canyons, including: the 3.35 km-deep Cotahuasi Canyon of the Rio Cotahuasi, a tributary of the Rio Ocona in southwestern Peru; the 2.44 km-deep Hells Canyon of the Snake River in Idaho; the >2.1 km-deep Barranca de Cobre in Chihuahua, Mexico; and the world-renowned 2.48 km-deep Grand Canyon in northern Arizona. LDCs also occur in submarine environments at the mouths of large rivers: the Nile and the Rhône Rivers have large submarine canyons at their mouths created during repeated Pliocene desiccation of the Mediterranean basin; the Indus and Ganges Rivers form lengthy submarine deltaic canyons; the Hudson River in northeastern North America has a substantial submarine canyon; and large canyons form in other tectonically active submarine areas. Whether terrestrial or subaqueous, LDCs support or generate strong ecological gradients, and periodically or perennially provide cascading deliveries of flow, sediments, nutrients, and biota to lower elevations (e.g., Gurnell and Petts 1995; Butman et al. 2006; Canals et al. 2009), processes that may influence the distribution and evolution of life around them.

basin. The Verde River basin breaches the geologic province boundary, draining the central southern portion of the GCE. The Virgin River drains southwestern Utah and northwestern Arizona, also breaching the geologic province boundary. In addition, the GCE supports thousands of caves and springs (Stevens and Meretsky 2008), and thousands of km of escarpment edges.
Embedded wholly within the southwestern Colorado Plateau, GC is naturally separated into an eastern basin, which receives the flows of the Paria and LCR drainages, and a more open western basin that connects western Grand Canyon to the Mojave and Sonoran deserts to the west and south ( Fig. 3; Billingsley andHampton 1999, Stevens andHuber 2004). The two GC basins are separated by the steep, narrow 35 km-long Muav Gorge, which creates a formidable cliff-dominated barrier to upstream and downstream dispersal of numerous southwestern plant, invertebrate, and vertebrate taxa (Miller et al. 1982;Phillips et al. 1987;Schmidt and Graf 1990;Stevens and Polhemus 2008).  Billingsley and Hampton (1999), and prepared by J. D. Ledbetter.

Overview
LDCs create complex spatially and temporally intercorrelated environmental gradients, and ecological processes that often are stratified across elevation, strongly affecting the composition and structure of biotic assemblages. The major gradients in LDCs include: gravity, geology, regional climate, elevation, and aspect. Collectively, these affect, generate, or permit the solar radiation budget, micro-to synoptic climate, moisture availability, dip angle, pedogenesis, and natural disturbance regimes, all of which affect the development of canyon biotic assemblages over time. Here I briefly review the roles of these major physical gradients in the GCE.

Gravity
Gravity plays a profound role in species dispersal and ranges in LDCs. Passively dispersing taxa, including plants, non-flying invertebrates and vertebrates, and even some bird species more readily disperse and colonize downslope. Although the inner canyon corridor within 100 m of the river makes up less than 3 percent of the overall land area in the GC, that zone supports more than 760 plant species, more than 40% of the entire GC flora (Stevens andAyers 2002, Busco et al. 2011). Downslope dispersal accounts for the in-canyon occurrence of many rim-dwelling species of grasses, forbs, shrubs (e.g., Fallugia paradoxa, Parreyella filifolia), and trees. These upper elevation "waif" or "Gilligan" species may reach the canyon-floor desert, far from their normal habitats and, while able to become established, they may not be able to reproduce or form permanent colonies. Such species are particularly likely to occur in aspect refugia: boreal colonists often occur in the desert on north-facing slopes. Numeorus insects (e.g, Prionus heros, various scarabaeids, Pandora moths), wild turkey (Meleagris gallopavo), and other rim fauna also readily disperse downslope, and have been regularly detected along the river (Brown et al. 1987;LaRue et al. 2001;Stevens unpublished data).

Geology and geomorphology
Geologic context, parent rock, structure, and tectonic history play dominant roles governing the development, functioning, and characteristics of LDC ecosystems. Geology and geologic processes control: 1) elevation, a primary determinant of local climate; 2) aspect, in turn influencing solar radiation budget, microclimate, productivity, pedogenesis, habitat availability, and species distributions; 3) geomorphology, including width-depth relationships of the canyon stream channel; and 4) rock color, which may differentially affect heat loading. The subject of microtopographic ecological gradient impacts on assemblages is generally poorly studied, but topographic differences of less than 10 m can affect local microclimate and species distributions, particularly near water sources where surface heating affects humidity.
Although much debated, drainage evolution of the Colorado River in the American Southwest has occurred over the past 40 million yr through conclusion of the Laramide orogeny and the onset of Basin and Range orogenic uplift and stream capture (Hunt 1956;Dickinson 2002; review in Blakey and Ranney 2008). On-going debate over development of the Colorado River basin is divided between advocacy for an old canyon (Oligocene-Miocene origination) as opposed to a young canyon (late Miocene-Pleistocene origination), and top-down drainage integration versus integration of variably independent sub-basins. Basin and Range crustal extension caused the uplift of the Sierra Nevada Range and numerous other north-trending ranges, which block the on-shore movement of moist Pacific airflow. This geologic deformation and development of the regional rain-shadow over the past 15 million ys changed the GCE from a low-lying, mesic, savannah-dominated landscape to its present-day uplifted, arid character. Also of great biogeographic impact was the tectonic/volcanic connection between South and North America 2.7 million years ago (Wallace 1876;Webb 2006). The ensuing land bridge permitted movement of biota between continents, resulting in the Great American Biotic Interchange, resulting in the movement and extinction of numerous species. The interchange is largely responsible for the contemporary co-occurrence of 150 genera of plants and animals in South America and the American Southwest. Following integration of the Colorado River, western Grand Canyon was dammed by numerous lava flows over the past 750,000 yr, forming large Pleistocene lakes and perhaps large outburst floods (Hamblin 1994;Crow et al. 2008). Biogeographic analyses of aquatic nepomorph Hemiptera (Stevens and Polhemus 2008) demonstrate significant upstream attenuation in species richness, suggesting that such flooding reversed or retarded upriver colonization in Pleistocene time. Interestingly, the ranges of aquatic Heteroptera taxa in that study also reveal the "biological shadow" of an earlier Paleogene east-and north-flowing Nevada river system into western Grand Canyon, a river unrelated to the present-day Colorado River. www.intechopen.com The Biogeographic Significance of a Large, Deep Canyon: Grand Canyon of the Colorado River, Southwestern USA 175 Fluvial geomorphology has been thoroughly studied in GC (e.g., Howard and Dolan 1981;Schmidt and Graf 1990;Topping et al. 2005). Rapids form at the confluence of tributary canyons when rare debris flows deliver large boulders into the Colorado River, damming the mainstream (Webb et al. 1987). Rapids create recirculation zones (eddies) that, on descending flows, deposit fine sediments in characteristic locations both upstream and downstream of the debris fans. Thus, unlike alluvial rivers, the location of sandbars in the geomorphically constrained Colorado River are fixed within 13 geomorphic reaches (Table  2). More than 550 debris fan complexes generate a distinctive suite of fluvial microhabitats, each with discrete grain sizes, soils, stage elevation relationships, inundation frequencies, and stage-zoned riparian vegetation (Turner and Karpiscak 1980;Schmidt and Graf 1990;Melis et al. 1997). Analyses of flow regulation influences at a large suite of debris fan complexes revealed that flow regulation allowed extensive development of fluvial marshes, enhancing riverine plant species richness (Stevens et al. 1995;Waring et al. in press). Clearwater releases and flood control also greatly increased aquatic productivity within the river which, coupled with riparian vegetation expansion, has led to increased waterbird biodiversity, and peregrine falcon (Falco peregrinus anatum) population increases and the potential for formation of novel post-dam trophic cascades (Brown et al. 1992;Stevens et al. 1997a,b;).
Thus, flow regulation reduced flooding disturbance and increased productivity of this naturally flood-prone river ecosystem, changing the river ecosystem energetics from allochthonous to autochthonous sources (Carothers and Brown 1991), and increasing riverine species richness and trophic complexity, in accord with the predictions of both Connell's (1978) intermediate disturbance concept and Huston's (1979) dynamic equilibrium model (Stevens and Ayers 2002). But also of scientific and stewardship interest, flow regulation differentially influenced geomorphological influences on subaqueous versus subaerial biota, colonization, and production. Flow regulation has swamped geomorphic influences on biological organization in the aquatic domain of the river, reducing differences in benthic standing stock or species among various aquatic microhabitats (Stevens et al. 1997b). In contrast, flow regulation has enhanced variation in biotic development on the various geomorphic microhabitats in debris fans in the riparian domain (Stevens et al. 1995(Stevens et al. , 1997a. Differential responses of the river ecosystem to flow regulation on the aquatic versus the riparian domains vastly complicate environmental management of focal species and habitats, creating challenging administrative tradeoffs Lovich and Melis 2007).

Elevation
Elevation strongly affects local climate, productivity, species composition, and microsite ecology. Following the model of Von Humboldt and Bonpland (1805, Jackson 2009), C.H. Merriam was the first American naturalist to quantify the influences of elevation on biotic zonation across an Arizona transect from the floor of Grand Canyon to the top of the San Francisco Peaks and out into the Painted Desert (Merriam and Steineger 1890). Merriam attributed the discrete zonation of trees across elevation primarily to temperature and latitude; however, Holdridge (1947) and others subsequently recognized the importance of seasonal and total precipitation, evapotranspiration, and other factors controlling vegetation and biome development (Olson et al. 2001). Nonetheless, elevation remains an overwhelmingly important ecological state variable due to its strong negative relationship with air temperature and freeze-thaw cycle frequency, and its positive relationship to precipitation and relative humidity. The global adiabatic lapse rate is -6.49 °C/km. Analysis of paired daily minimum and maximum air temperature from 1941-2003 at Phantom Ranch (elevation 735 m) on the floor of GC with the South Rim (2100 m) produces a GC-specific lapse rate of -8.7 °C/km. The >1.3-fold steeper lapse rate in GC is likely a function not only of the dark red and black bedrock color of the inner canyon, but also to aspect. Steep, S-facing slopes in the GCE, particularly those with darker rock color, absorb and re-radiate more heat than do N-facing slopes, which often are shaded from direct sunlight, and are cooler and more humid than S-facing slopes across elevation. Overall, elevation strongly and broadly influences synoptic climate, while aspect exerts strong local control over microclimate and microsite potential evapotranspiration and therefore productivity.

Aspect and solar radiation
Solar radiation limitation is an important factor limiting ecosystem productivity of rivers and lakes (e.g., Yard et al. 2005;Karlsson et al. 2011). Depending on the latitude, depth, and cardinal orientation, LDCs also can be strongly limited by solar radiation, particularly eastor west-oriented canyons at higher latitudes ( Fig. 4; Stevens et al. 1997b). During pre-dam time, the Colorado River's naturally high suspended load prevented sunlight from reaching the floor of the river (Topping et al. 2005); however, sediment retention in Lake Powell and resulting clearwater releases now allow illumination of the river floor in the upper reaches, enormously increasing benthic productivity. Yard et al (2005) measured mid-day solar radiation in the river at 25 km intervals from Glen Canyon Dam to Diamond Creek and described variation in the availability of photosynthetically active radiation (PAR) of the Colorado River aquatic domain. They reported that river surface PAR varied strongly in relation to cardinal orientation of the canyon and in relation to tributary-contributed suspended inorganic sediment load, ranging from little limitation in upstream, wide, and east-west oriented reaches to significantly reduced PAR in downstream (more turbid), narrow, north-south-oriented reaches. These results help explain the low pre-dam productivity of the pre-dam sediment-laden Colorado River, postdam stairstep decreases of benthic standing stock at the confluences of the Paria River and the LCR, and spikes in mainstream productivity at the mouths of N-or S-flowing tributary (Stevens et al. 1997b). Yard et al. (2005) focused attention on interactions of physical solar radiation limitation on turbidity, canyon geomorphology, and aquatic productivity. To better understand potential solar radiation variation on the Colorado River floodplain in GC, I used a solar pathfinder (SPF; Solar Pathfinder 2008) to measure the percent of mean monthly potential solar radiation at riverside and at the 10-year return flood stage at each bend in the river, typically every 1-2 km between Glen Canyon Dam and Diamond Creek ( Table 2). The general model for this analysis includes latitude (sun angle) and canyon configuration (width, depth, cardinal orientation), but does not account for cloud cover or atmospheric aerosol obstruction of solar radiation intensity (Fig. 4). The use of a SPF is more accurate than landscape modeling of solar reception on the river banks because the floodplain is often narrow, and minor cliff projections and large rocks strongly influence shading patterns. Fig. 4. A conceptual schematic of solar insolation limitation in large deep canyons. A e -slope angle on the sun-exposed side; A s -slope angle on the shaded side; B e -angle of incidence on the sun-exposed side; d L -depth of canyon on sunlit side; d tot -total depth on the shaded side; I-angle of solar incidence; S L -sunlit slope, S D -unlit dark slope; S O -slope length on the shaded side; w-width of the canyon. The full solar budget of a canyon is the sum of solar radiation integrated across momentary to annual time scales. This model does not illustrate or take into account atmospheric interference, which reduces the canyon's total solar energy budget.
These data demonstrate that limitation of solar radiation is substantial on the Colorado River floodplain in GC, varying by the cardinal orientation of the canyon and both local and rim cliff structure (Fig. 5). The overall average percent of potential sunlight received by the Colorado River floodplain in GC is 69.5% of that available on the rims, varying from an average low of 54.8% with virtually no direct wintertime insolation in the steep, narrow, west-flowing Muav Gorge, to 75-88.1% in the relatively wide Permian, Furnace Flats, and Lower Granite Gorge reaches. The most extreme limitations of solar limitation occur on north-facing slopes in east-west flowing segments of the canyon in the deepest portions of the Canyon. Extremely steep, narrow, deep geomorphic reaches, such as the Muav Gorge, receive no direct insolation during winter months, regardless of aspect, allowing mesic rim grasses, such as galleta-grass (Pleuraphis spp) to grow even on S-facing slopes. In less steep but still narrow reaches (e.g., the Aisles) galleta-grass grows only on refugial N-facing slopes, and otherwise is rare to non-existent in the lower elevations of GC.
River corridor vegetation composition varies in relation to aspect and the solar radiation budget, with greatest differences on N-versus S-facing slopes. N-facing slopes, along the river, such as those at Rkm 63-65R, 193-198L, 228-232R, and 315-318L, typically support upland and boreal Great Basin Desert plant species whose ranges are normally 600 m higher www.intechopen.com The Biogeographic Significance of a Large, Deep Canyon: Grand Canyon of the Colorado River, Southwestern USA  Table 2. The Muav Gorge (MG; circled) is the most deeply canyon-bound reach, receives the least solar radiation, and is a significant barrier to upriver and downriver range extensions.
in elevation, including netleaf hackberry (Celtis laevigata var. reticulata), galleta grass (Pleuraphis spp.), juniper (Juniperus spp.), buffalo berry (Shepherdia rotundifolia), and other perennial grasses, shrubs, and woodland taxa. In contrast and with little compositional overlap, S-facing slopes in those segments support Sonoran Desert vegetation, including several cacti taxa, brittlebush (Encelia farinosa), and upper riparian zone western honey mesquite (Prosopis glandulosa var. torreyana), but little desert grass cover. Analysis of the aspect of all stands of western honey mesquite and netleaf hackberry reveal strong, opposite affinity for S-and N-facing slopes, respectively. E-and W-facing slopes support a mix of Nfacing slope boreal and S-facing slope desert taxa, and thus have roughly twice the species density as N-or S-facing slopes. Additional study is needed to distinguish plant compositional differences between E-facing slopes that receive early warming sunlight, from W-facing slopes, which receive hotter, late afternoon radiation. Aspect similarly influences small desert mammal composition, with N-facing slopes supporting a mixture of canyon deer mice (Peromyscus crinitus), woodrats (Neotoma spp.), and seed-feeding Chaetodipus spp. pocket mice, while S-facing slopes primarily support a lower diversity of weedy rodent species, dominated by western cactus deer mouse (Peromyscus eremicus) and less common white-tailed antelope squirrel (Ammospermophilus leucurus).
Elevation and aspect interactively affect solar radiation, which affects vegetation at at springs throughout the GCE. Cliff Spring near Cape Royal on the North Rim of Grand Canyon is a hanging garden (a contact spring on a cliff face; Fig. 6A). It is a high-elevation, E-facing site that receives direct insolation during winter, allowing it to thaw and warm quickly after cold winter nights. But due to the overhanging cliff, it is protected from direct insolation during the summer months. This configuration moderates the springs' summertime microclimate, allowing the highest elevation population of Primula specuicola wall plants to persist there. In comparison, Vaseys Paradise is a gushet spring along the Colorado River at Rkm 51R (Fig. 6B). Also E-facing, its depth in the canyon precludes direct insolation during the winter, but its relatively warm (16ºC) water and rapid warming in the morning hours, coupled with early shade, allows the endangered Kanab ambersnail (Oxyloma haydeni kanabensis) to persist there, one of only three naturally occurring Oxyloma populations in Arizona (Meretsky and Stevens 2000). More detailed modeling of elevationaspect relationships will be productive for predicting potential climate change impacts on springs and regional vegetation.
Fig. 6. Solar radiation budgets at two east-facing springs in GC: A -Cliff Spring is a hanging garden with direct insolation in winter, but is shaded by the overhanging cliff during summer, permitting persistence of a small population of Primula specuicola (inset); B -Vaseys Paradise is a gushet spring that received morning light throughout the year, but is shaded in early afternoon during the hot summer months, factors that allow endangered Kanab ambersnail (Oxyloma haydeni kanabenss; inset) to persist.
These observations, models, results, and studies demonstrate strong, pervasive, and controlling impacts of solar radiation limitation on LDC ecosystem ecology. The extent and patterns of solar radiation limitation created by cliffs in the GCE indicate that, depending on LDC width and depth, solar radiation limitations are most influential in deep E-W oriented canyons, impacts that increase with latitude and canyon depth. Solar limitation in LDCs also may exert impacts on canyon geomorphology. At the latitude of GC (35 o N), N-facing slopes freeze in winter and may remain frozen for prolonged periods, while S-facing slopes generally receive direct solar radiation every day and thus undergo daily freeze-thaw cycles. www.intechopen.com The Biogeographic Significance of a Large, Deep Canyon: Grand Canyon of the Colorado River, Southwestern USA

181
Can such aspect differences lead to faster erosion rates and cliff retreat of S-facing slopes as compared to N-facing slopes? This question has yet to be studied in detail, but may help account for the order-of-magnitude greater width of the canyon from river to rim of the Sfacing North Rim of GC, as compared to that of the N-facing South Rim (Fig. 3).

LDC biogeography 4.1 Biodiversity
GC likely supports more than 10,000 species of (non-microbial) macrobiota, while the GCE may support 15,000 or more taxa (Table 3; estimated from Carothers and Aitchison 1976;Suttkus et al. 1978;Hoffmeister 1986;Harper et al. 1994;Busco et al. 2011;Stevens unpublished data). Plants and vertebrates have garnered most inventory attention, and recent and on-going advances in invertebrate biodiversity research are expanding understanding of the biogeographic and evolutionary significance of GC as an LDC.  Table 3. Summary of selected GCE taxa, relative species richness, overall taxon vagility, biogeographical responses to GC as a LDC.

LDC landform effects
As a landform, GC influences species ranges and gene flow processes in four primary ways, each with complex subprocesses: 1) a partial or full corridor of low elevation riverine and desert habitats through the uplifted Colorado Plateau; 2) a partial or full barrier across the Plateau or in an upstream-downstream fashion; 3) a refuge, particularly for species requiring rare microhabitats, such as springs, caves, and rim edges; and 4) a null effect, not limiting gene flow across or within the landscape (Stevens and Polhemus 2008; Table 3). Below I elaborate on each of these types of landform influences using GCE biodiversity data and I discuss biogeographic anomalies in GC and the GCE.
Corridor Effects: Several biogeographic corridor functions exist in GC, including range, movement, and migration corridor effects.
Full downriver range corridor: the Colorado River also serves as a corridor for dispersal of Colorado Plateau taxa from Utah and the upper Colorado River basin downstream through GC into the lower Colorado River basin (e.g., Corispermum americanum nonnative goosefoot, Salix exigua coyote willow, and Baccharis salicina seep willow).
Partial downriver range corridor: Some Utah and upper Colorado River basin species ranges descend part way through GC, including: Pariella filifolia (Fabaceae), Falugia paradoxa Apache plume, Symphoricarpos oreophilus snowberry, Quercus turbinella scrub oak and Q. gambellii Gambel's oak, Boehmeria cylindrica at Rkm 55.5R Spring, Carex specuicola, Oxyloma ambersnails, and Rana pipiens northern leopard frog). The larger faults of Grand Canyon also provide access down from the rims, and are actively used by ungulates, terrestrial predators, humans, and passively wind-dispersing "aerial plankton".
Annual migratory corridor: Prior to impoundment the canyon served as a long-linear migratory corridor for several native fish (e.g., Xyrauchen texanus razorback sucker and Ptychochelius lucius pikeminnow; Minckley 1973Minckley , 1991 and still provides that function for numerous migratory waterbirds (Stevens et al. 1997a), and probably bats and monarch www.intechopen.com The Biogeographic Significance of a Large, Deep Canyon: Grand Canyon of the Colorado River, Southwestern USA 183 butterflies (Garth 1950). An autumn migration route along the east side of the East Kaibab Monocline brings high densities of raptors across Grand Canyon from the north, and those birds use rising thermal air currents to ascend out of Grand Canyon on the South Rim at Lipan and Yaqui Points (Hoffman et al. 2002;Smith et al. 2008). Another hawk flyway likely exists along the Grand Wash and Hurricane Cliffs, but has not been studied.
Short-term lateral migration/movement corridor: Several taxa undergo temporal movements into or out of GC. Rabe et al. (1998) documented summertime daytime roosting of spotted bat (Euderma maculatum) along the river in central GC, with nocturnal forays of more than 38 km/night to North Rim meadows, likely to forage on abundant coniferous forest meadow moths and beetles. On a seasonal time scale, numerous taxa move into GC from the rims in autumn and winter, including rim-and montane-dwelling taxa as diverse as Culiseta incidens mosquitoes, American crow and common raven (Corvus brachyrhynchos and C. corax, respectively), desert mule deer (Odocoileus hemionus) and mountain lion (Puma concolor kaibabensis; Stevens et al. 2008a;Brown et al. 1987;Stevens unpublished observations). Thus, the biogeographic corridor effect operates in numerous, complex ways, affecting a great array of taxa in this LDC.

Barrier/Filter Effects:
The evolutionary consequences of GC as a barrier are well known through studies of the divided distribution of Abert's tassel-eared squirrel (Sciurus aberti aberti) from Kaibab squirrels (S. a. kaibabensis; Lamb et al. 1997). These two taxa were divided by the Pleistocene-Holocene climate transition, which eliminated suitable habitat within GC, isolating S.a. kaibabensis on the North Rim. The color shift between these two taxa involves a minor genetic change, and black S.a. aberti individuals have been reported south of GC (e.g., Allred 1995). Numerous other, but less well known examples of taxa with ranges divided by GC exist, and cryptic speciation may be commonplace. Taxa occurring north, but not south of GC include the landsnail family Oreohelicidae with the large genus Oreohelix (Bequaert and Miller 1973), Satyrium behrii hairstreak butterfly (Garth 1950), Plestiodon skiltonianus skink and Pituophis catenifer deserticola Great Basin gopher snake (Miller et al. 1982), and Thomomys bottae planirostris and T. talpoides kaibabensis pocket gophers (Hoffmeister 1986). Examples of taxa occurring on the South Rim but not the North Rim include the landsnail family Helminthoglyptidae with the large genus Sonorella (Bequaert and Miller 1973), Coenomorpha tullia furcae ringlet butterfly (Garth 1950), Plestiodon gilberti in Peach Springs Wash (L.E. Stevens personal observation) and P. multivirgatus skinks, Pituophis catenifer affinis Sonoran gopher snake (Miller et al. 1982), the oddly disjunct population of Tantilla hobartsmithi lyre snake between Lees Ferry and the LCR (Brennon and Holycross 2006), (from NE to SW) Thomomys bottae alexandrae, T.b. aureus, T.b. fulvus, and T.b. desertorum pocket gophers, and extirpated Panthera onca jaguar (Hoffmeister 1986).
The Colorado River corridor itself presents a barrier to the downstream distribution of Phrynosoma platyrhinos horned lizard and other species from southern Utah, and similar upstream exclusion of common Sonoran and Mohave Desert plant species from Lake Mead, such as Yucca brevifolia Joshua tree and Y. mohavensis Mohave yucca, Parkinsonia spp., and Psorothamnus spinosus desert shrubs, which probably have been excluded by the lack of suitable low-gradient bajada habitat within the walls of GC.

Refugia:
The two basins of GC generally support discrete endemic assemblages, with relatively few endemic species occurring in both. However, Camissonia specuicola hesperia Kaibab suncup and Polistes kaibabensis Kaibab paper wasp occur relatively widely through GC, and Rosa stellata abyssa wildrose occurs on the rims, bridging both basins. Although GC caves are numerous, they appear to support highly individualistic assemblages, with relatively low similarity among caves, and with endemic taxa often restricted to one or a few closely-related caves.
Springs: Springs are well known as refugial habitats in arid regions (Stevens and Meretsky 2008). In the GCE springs habitat makes up <0.001 % of the overall landscape, but springs support numerous springs-specialist plant species: at least 9.5% of the regional flora are springs-specialist taxa. Several groups of these springs-specialist plants occur in the GCE: 1) facultative springs plants, such as cardinal monkeyflower (Scrophulariaceae: Mimulus cardinalis) also occur in protected habitats along perennial streams. 2) obligate springs species may either be widespread across the GCE, such as Epipactis gigantea heleborine orchid, Primula specuicola cave primrose, and Cirsium rydbergii thistle, or 3) species very narrowly restricted to one or a few springs (e.g., Asteraceae: Flaveria mcdougallii ragweed along the Colorado River between Rkm 218-285; Ranunculus uncinatus buttercup at North Rim springs). Some plant species may be relatively common nationally, but only have been detected in GCE at springs, such as Persicaria amphibia water knotweed at Rkm -14.4L and Boehmeria cylindrica false nettle at 55.5R.
GCE springs also support rare and some endemic invertebrates, and springs assemblages are often highly individualistic. Stevens and Polhemus (2008) reported that 53% of the 89 aquatic Hemiptera taxa in the GCE were restricted to 3 or fewer of 444 localities (primarily springs) at which aquatic Hemiptera were detected. For example: Ochterus rotundus (Ochteridae) only occurred at GC springs, far outside of its range in south-central Mexico (Polhemus and Polhemus 1976 Escarpment Edges: Canyon rim edges are remarkably important refugial and ecotone habitats. Rim edges are abundant throughout the GCE, with its many steep, long escarpments. Rim edges have been poorly studied, but appear to be subject to greater climatic stresses than are other habitats, sustaining more severe temperatures from subsidence of cold air during winter and higher temperatures from rising hot air from desert canyons during the summer. In addition, such habitats are subject to strong, erratic winds during all seasons. As a consequence, many of the plant species that are restricted to rims are mat-or clump-forming or otherwise sturdy, low-growing shrubs. GC species wholly restricted to these harsh rim habitats include Astragalus cremnophylax var. cremnophylax sentry milk-vetch, an endangered legume that occurs in a few small patches primarily on the South Rim (Maschinski et al. 1997;Busco et al. 2011 Biogeographic Anomalies: Several common southwestern desert taxa are conspicuously missing from the Colorado River corridor in GC, and their absence highlights otherwise difficult-to-discern LDC ecological processes. Among the missing taxa are: horned lizards (Phrynosoma spp.), kangaroo rats (Dipodomys spp.), and lagomorphs (i.e., Sylvilagus spp. and Lepus spp.). All of these taxa are common on the canyon rims and around Lakes Powell and Mead, and all but the kangaroo rats are found at Lees Ferry; however, none have been detected along the Colorado River in Grand Canyon except at the Lake Mead boundary. In the case of the lagomophs, rabbits occur on both rims, and down onto the Esplanade Platform (a high-elevation platform) on the North (but not the South) Rim, and I observed a single Sylvilagus on the Tonto Platform at Cottonwood Creek in November 2001, the only individual seen by me in more than 10,000 km of trekking through the inner Grand Canyon. Predation by wall-nesting and migrating raptors likely causes extinction probability to exceed colonization potential, thus preventing lagomorph populations from becoming established in the inner GC. In the cases of horned lizards and kangaroo rats, the steepness of the surrounding terrain likely limits habitat patch suitability, and despite relatively likely downriver/downslope colonization over millennia via rafting, extinction probability is apparently too high to permit establishment.
Another biogeographic anomaly is the much-reduced presence and ecological role of termites (Isoptera) in Grand Canyon. Termites are abundant desert southwestern decomposers and provide substantial food resources to a wide array of insectivores (Ueckert et al. 1976). Their relative scarcity and small colony sizes in Grand Canyon are surprising, and few species have been detected (Jones 1985). The generally steep terrain of the canyon means that fallen wood is unstable as habitat, and driftwood piles along the river were, in pre-dam time, too commonly wetted or moved by flooding to provide suitable, long-term termite habitat. Thus, like the missing vertebrate taxa, habitat size and stability probably limit termite colonization on the floor of GC. One consequence of the absence of termites is that driftwood piles in Grand Canyon contain logs that regularly exceed 1,000 yr in age (A. McCord, University of Arizona Tree Ring Laboratory, personal communication), providing a largely overlooked wealth of paleoecological information.
Lastly, many records of highly disjunct species distributions exist in GC. For example: a single, large, isolated stand of Canotia holocantha cruxifixion-thorn (Celastraceae) exists in a small canyon at Rkm 196L; a single specimen of Xantusia vigilis was collected in the middle reaches of Clear Creek (Rkm 135R; Miller et al. 1982); the damselfly Coenagrion resolutum occurs in a few ponds on the North Rim (Stevens and Bailowitz 2009); the termite Incisitermes minor (Isoptera: Kalotermitidae) was collected at Cardenas Creek (Rkm 114L) but otherwise is known only from the Pacific Coast (Jones 1985); the Surprise Canyon relict leopard frog occurs in a highly isolated population in western GC; and numerous erratic bird records exist in GC, including a scissortail flycatcher (Tyrannus forficatus), a painted bunting (Passerina ciris) and 2 records of magnificent frigatebird (Fregata magnificens; Brown et al. 1987;LaRue et al. 2001). A perplexing skeleton of a collared peccary (Pecari tajacu) was found in middle Spring Creek (Rkm 328R; Stevens unpublished data, skull housed at the National Park Service collection at Grand Canyon). Although this species is expanding its range northward from the southern deserts, it was not previously known to occur north of the Colorado River. Each of these records represents range extension of populations considerably external to GC, primarily but not exclusively from the south. Additional basic inventory is needed to better understand such apparently enigmatic range records.

LDC biogeographic hypotheses
Six questions illuminate the extent to which GC influences the distribution and evolution of biotic assemblages. Testing these questions required three critical analyses: a) a landscape analysis of land area by elevation of the GC within the larger GCE, which was accomplished using geographic information systems (GIS) analysis of 30 m digital terrain data; b) a compilation of species richness, elevation range, and rarity data for a broad array of organisms with differing dispersal capabilities within GC and the surrounding GCE, including plants, vertebrates, and selected groups of invertebrates, work that is on-going www.intechopen.com The Biogeographic Significance of a Large, Deep Canyon: Grand Canyon of the Colorado River, Southwestern USA 189 through the Museum of Northern Arizona in Flagstaff, Arizona; and c) biogeographic affinity information for the taxa under consideration. The latter data are now available for most vascular plants and vertebrates, but only for a few of the better-known arthropod goups.

1) Do landform corridor, barrier or refuge effects differentially influence species richness over refuge and null effects?
If GC is a biogeographically significant landform, then corridor, barrier/filter, and refugial effects on species ranges should greatly predominate over null effects. Among the plants, as mentioned above, nearly 46% of the flora is found in the river corridor, demonstrating a pronounced corridor effect. Instances of refuge effects and restricted distributions of unique genomes are becoming more widely recognized (e.g., Stevens and Polhemus 2008;Bryson et al. 2010). While partial corridor effects are numerous, relatively few non-endemic plant taxa are restricted to one or the other rims, suggesting that barrier effects among plants are relatively weak (Table 3). Thus, a total of 81.5% of GC aquatic Hemiptera demonstrated range patterns related to LDC landform configuration. Differences in the strength of landform impacts between Odonata and aquatic Hemiptera may be related to vagility, with many Odonata being more capable of long-distance dispersal. Evidence from these studies also points to a more concentrated landform impact on both of these taxa in GC as compared to the general ecotonal impacts of the Mogollon Rim (Polhemus and Polhemus 1976).
Among the GC herpetofauna, the role of partial and full corridor distributions also is dominant, as mentioned above ( At least 178 mammal taxa have existed in the GCE in historic times, with 128 species among 67 genera in 23 families in 7 orders (Durrant 1952, Hoffmeister 1986, Flinders et al. 2005. The GCE fauna is overwhelmingly dominated by Rodentia, with 94 taxa, followed by Carnivora with 28 taxa and Chiroptera with 24 taxa. At least 113 mammal taxa exist in or on the immediate rims of Grand Canyon, with 97 species among 59 genera in 22 families. Excluding humans, 10 mammal species are non-native. Hoffmeister (1986) lists 145 recent mammal species in Arizona, a number that has changed somewhat due to further collecting and improved taxonomy. Therefore, the GCE supports at least 88% of the state's fauna, and the GC supports two thirds of the State's fauna. Hoffmeister (1986) reported 55 widelydistributed species in the GCE, with at least 14 taxa restricted to upper elevations, five species restricted to the north of the Colorado River, and 10 species restricted to south of the river. Thus, neotropical influences on the GCE mammal fauna prevail over those of the nearctic region, a pattern similar to that of other taxa.
Thus, barrier, corridor, and refuge effects collectively dominate over null effects (X 2 df=1 = 9.846, P < 0.002), indicating that GC is a significant biogeographic feature for mammals. Although the landform impact on GC species ranges is overwhelmingly evident, the significance of GC as an LDC requires closer analysis of the distribution and evolution of endemic taxa (see question 6, below).

2) Does biogeographic affinity influence assemblage composition in GC?
Greater-thanexpected species richness of GC and the GCE in general is a partial result of topographic complexity, as well as the position of the GCE as a mixing zone during late Cenozoic time.
The GC and GCE support several suites of taxa. a) A distinctive Madrean biota is comprised of Central American, Mexican, and pan-tropical plant and animal taxa whose ranges likely extended into the region during warm interglacial phases, and these taxa generally occupy lower elevations (e.g., Phillips et al. 1987;Stevens and Polhemus 2008;Stevens et al. 2008a,b). Peculiar to this group are the "Guatemalan taxa": isolated GCE populations with ranges otherwise occurring in Central and northern South America. For example, masked clubskimmer dragonfly and Ochterus rotundus waterbug are found at GC springs and also in Central America (Stevens and Polhemus 2008;Stevens and Bailowitz 2009 resolutum damselfly, various gerrid waterstriders (Hemiptera), and other taxa. c) In addition to species with exogenous biogeographic affiliations, a large suite of broadly distributed species exists that are more-or-less centered in their ranges (e.g., Anax junius and Rhionaeschna multicolor dragonflies; Danaus plexippus monarch and Vanessa cardui painted lady butterflies; and others). d) Also, in the case of Odonata, as many as 18 (20.2%) of the 89 GCE taxa have ranges that include or are centered on the Pacific Coast of North America, an intriguing pattern suggesting that some dragonfly ranges may predate the Basin and Range orogeny. The back-and-forth and elevational adjustment of assemblages during glacial advances and retreats between climate oscillations, and the stability of refugial microhabitats within GC contribute to the relatively high biodiversity of this LDC.

3) Does elevation influence species richness in a fashion analogous to the effects of latitude?
Many of the GCE faunal taxa studied thus far show the well-known latitudinal diversity gradient, a pattern of declining species richness across latitude (reviewed in Lomolino et al. 2010). Exceptions to the pattern include species, such as conifers, Salicaceae and tenthridinid sawflies, and ichneumonid wasps that are derived from boreal regions. Species richness also generally declines across elevation in a fashion analogous to latitude; however, the "mid-domain effect" of elevation on species richness commonly results in a unimodal peak of species richness at intermediate elevations (Romdal et al. 2005; review in Lomolino et al. 2010). Our work on plants (Figs. 7A, B), and macroinvertebrates ( Fig. 8) demonstrate that although the mid-elevation richness peak is distinctive (e.g., among plants - Fig. 7A; aquatic Hemiptera and Odonata -Stevens and Polhemus 2008 and Stevens and Bailowitz 2009, respectively; and landsnails and non-melittid bees -Stevens unpublished data), the effect is largely accounted for by the species-area relationship: the ratio of the log 10 -transformed insect species richness to the log 10 area of the GCE within 100 m belts revealed a strong negative response across elevation (Figs. 7B, 8). The GCE flora and fauna both consist of broad mixtures of Maderan and Rocky Mountain taxa. The 1600 m elevation zone may approximate the division between those two assemblages, and the sensitivity of composition and vegetation structural responses to climate change may warrant more focused research on that zone. 4) Does vagility influence species richness (are highly mobile taxa relatively more species rich than less mobile taxa)? I used the literature and available information to conduct a qualitative analysis of landform effects for a wide array of taxa for which biogeographic data are available from within the GC and across the GCE (Table 3). This analysis indicates that taxa with low overall vagility show stronger evidence of barrier/filter effects than those with higher vagility (e.g., strongly dispersive taxa), and the data indicate that taxa with high vagility are relatively more species rich than taxa with low vagility. While somewhat intuitive, these patterns reinforce the view that GC is an isolated, relatively young geologic feature, one still undergoing colonization and assortative assemblage development by species colonizing from the surrounding region and affecting in-canyon refugia.

5) Does species richness attenuate upstream through the Colorado River corridor?
Plant distribution data in the Colorado River corridor indicate some support for this pattern. The ranges of numerous conspicuous species extend part-way up into Grand Canyon, including: Yucca whipplei yucca, Fouquieria splendens ocotillo, Prosopis glandulosa mesquite, Acacia greggii catclaw, Ferocactus cylindrica barrelcactus, several Cylindropuntia cholla cacti species, Canotia holocantha crucifixion-thorn, and Larrea tridentata creosote-bush, to name a few.   8. Log 10 species richness/log 10 area (km 2 ) within an elevation belt, as a function of belt elevation for Grand Canyon ecoregion non-melittid bees, terrestrial gastropods, Odonata, and aquatic Hemiptera. Redrawn in part from Stevens and Polhemus (2008) and Stevens and Bailowitz (2009). Among the insects, analysis of aquatic Hemiptera and Odonata ranges conclusively demonstrate upstream attenuation of species richness through the Colorado River corridor. Most Nepomorpha Hemiptera are found in the lower reaches of the Colorado River drainage, with only a few species of Corixidae, Notonectidae, one Gerridae, one Gelastocoridae, and two Veliidae common in the upper reaches of GC (Polhemus and Polhemus 1976;Stevens and Polhemus 2008). Even though they are generally far more vagile, Odonata diversity declines with distance upstream as well. Several provocative examples of exclusion exist: Brechmorhoga pertinax and Hetaerina vulnerata both replace widespread and common congeners in western basin of GC. These two species exist in the eastern basin of GC only along perennial springfed tributaries (Stevens andBailowitz 2005, 2009). Whether exclusion of the more widespread Odonata congeners is the result of competitive superiority or other factors remains to be determined.
Among the herpetofauna, upstream species attenuation also is apparent, as described above (Fig. 9). Species richness on the rims is equivocal from the upper to the lower Canyon, with greater species richness on the North Rim likely attributable to the greater range of elevations there. However, both in the inner canyon and on the canyon floor, herpetofaunal species richness attenuates upstream markedly, with at least 7 taxa (24%) in the lower Canyon missing from the upper Canyon. Species-area influences may account for some of this attenuation, as the upper Canyon is narrower; however, habitat area is yet to be determined for most of these herpetofaunal taxa. Range restrictions are not clear for other terrestrial vertebrates. Fig. 9. Grand Canyon herpetofaunal species richness on the South Rim, the south side Inner Canyon, the Colorado River corridor, the north side Inner Canyon, and the North Rim, from the upper (eastern) to the lower (western) Grand Canyon.

6) Are levels of endemism and rarity consistent with the geologic developmental history of GC?
Most endemic taxa in GC and the GCE are restricted to harsh, constant environments (e.g., high-conductivity limnocrene springs, south-facing desert slopes, rim edges, caves, alpine habitats). Based on aquatic Hemiptera data, levels of endemism and rarity previously were regarded as being low in GC, a phenomenon attributed to the youth of GC as a landform (Polhemus and Polhemus 1976), and supported by the greater frequency of subspecific or varietal-level endemism as compared to species-level endemism. Varietal level endemics in GC are numerous, including: plants, such as Aletes m. macdougalii, Arabis g. gracilipes (found elsewhere in the GCE), and an undescribed Arctomecon californica variety in western GC (Phillips et al. 1987;Brian 2000;Arizona Rare Plants Committee 2001).
Numerous endemic invertebrates exist in GC and the GCE (Table 4). Many butterfly taxa that occur at lower elevations in GC appear slightly different from other populations in the Southwest, and at least 4 endemic butterfly and skipper subspecies are known from GC: Papilio indra kaibabensis, Speyeria atlantis schellbachi, Coenomorpha tullia furcae, and Agathymus alliae paiute (Garth 1950;Stevens unpublished data). Three endemic tiger beetle subspecies are known from the GCE: Cicindela hemorrhagica arizonae in inner GC, C. hirticollis coloradulae in the LCR, and C. terricola kaibabensis in North Rim meadows (Stevens and Huber 2004).
Other varietal-level endemics include: the likely new subspecies of Vulcans Well waterbug (Belostoma near flumineum; Stevens and Polhemus 2008), Grand Canyon rattlesnake (Crotalus oregonus abyssus), North Rim Thomomys talpoides kaibabensis, and Kaibab squirrel (Miller et al. 1982;Hoffmeister 1986 Table 4. Percent of endemism and rarity among selected GCE invertebrate taxa. Rarity was evaluated as the percent of localities at which a species was detected in relation to the total number of localities at which members of that taxon were detected. While early focus on these taxa suggested GC was not an evolutionarily significant landform, more collecting and observation in recent decades has revealed many more www.intechopen.com  (Phillips et al. 1987;Brian 2000;Arizona Rare Plants Committee 2001;Hodgson 2001); the above-mentioned cave endemic invertebrates; Kaibab monkey grasshopper -Morsea kaibabensis; at least four chironomid midge species (Sublette et al. 1998); Tapeats robber fly -Efferia tapeats (Scarbrough et al. in press); three unique Nebria ground beetles (D. Kavanaugh, California Academy of Sciences, personal communication); Schinia immaculata (Pogue 2004); Kaibab paper wasp -Polistes kaibabensis (Snelling 1974); and other species. While no species-level vertebrates are endemic to GC, 19.2% of 775 invertebrate species that have been studied in detail are rare and 9.7% of those species are endemic (Table 4). With at least 30 endemic faunal species in and on the periphery of GC, and with more than 200 varietal-and species-level endemic taxa in the surrounding GCE, GC is emerging as a far more important evolutionary landscape than previously recognized.

Comparison of LDCs with mountains and islands
As a LDC, GC clearly exerts a profound biogeographical influence on the ranges of biota in and around it, and increasing evidence points to the evolutionary importance of this landform. But how do LDC biogeographical processes compare with those of mountain range and island landscapes, landforms that are well known to affect species distribution and evolution (MacArthur and Wilson 1967;Brown 1971-contested by Lawlor 1998Lomolino et al. 2010)?
A comparison of the characteristics of these three types of landforms indicates that LDCs differ in biogeographic function from the other two types, but are somewhat more similar to mountain ranges than they are to islands or archipelagos (Table 5). The focal feature of LDCs are dendritic drainage networks, with directional, gravity-facilitated flow, sediment, nutrient, and species transport, usually in a downstream direction, but also upstream through Aeolian transport for some components. LDCs are more strongly characterized by connectivity and gravity-facilitated movement, with productivity, growing season length and species richness increasing towards the focal feature (the canyon floor). In contrast, the peaks and ridges that characterize mountain ranges are harsh, somewhat to extremely unproductive, difficult to access, with short growing seasons, and are generally inhospitable habitats. Although montane slopes may be used by wide-ranging terrestrial taxa and by migrating birds, mountain ranges appear to be considerably less likely to facilitate gene flow than are LDCs. Dispersal among islands is restrictive, often limiting an island's impact on gene flow to passive processes or to active habitat searching by highly vagile taxa. Rockfall and flooding are more dominant forms of natural disturbance in LDCs than in the other two landforms, and all three landforms provide various refugial habitats. The longevity of LDCs and islands is usually geologically shorter than that of mountain ranges, and the several to tens of millions of years that LDCs exist may not be sufficient to create as much genetic isolation. Likely the largest difference between LDCs and mountain ranges is that the former are generally narrower than the latter, a difference that limits the extent of isolation on the rims and in refugia. The corridor functions of LDCs contribute to, and can facilitate, regional gene flow, and LDC aspect influences enhance in-canyon species retention and genetic diversity. In contrast, the corridor function of mountain ranges may be less significant, and vicariance effects stronger, filtering gene flow around the peidmont peripheries. The several Ensatina salamander taxa whose non-overlapping ranges encircle the Sierra Nevada Mountains in California is a well-known example (Pereira and Wake 2009). However, such processes also occur around LDCs. A ring clade of pocket gopher (Thomomys bottae) subspecies has been reported around GC (Hoffmeister 1986). Moving clockwise from Lees Ferry around GC are the non-overlapping ranges of T.b. alexandrae, T.b. aureus, T.b. fulvus and T.b. desertorum on the south side of GC, and T.b. planirostris and T.b. fulvus on the north side (Hoffmeister 1986). Whether or not such subspecific differentiation represents cryptic speciation or morphological noise (e.g., Rios and Álvarez-Casteñeda 2007) remains to be determined, but it does suggest a process similar to that of the Ensatina salamanders of California.

Characteristic or Process LDCs Mountain Ranges Islands
Overall, and depending on landform size and structure, LDCs can play a significant role in regional biogeography, affecting species dispersal and gene flow. LDCs play complex roles, broadly functioning as corridors, barriers, and refugia and affecting the majority of species in the landscape, and with potentially strong evolutionary consequences on regional diversity. Additional research on gene flow in and around LDCs is warranted to better understand biogeographic patterns and processes, and comparative studies are needed to compare species-area relationships among LDCs of different sizes, cardinal orientation, and latitude.

Landform development and biogeography
Badgley (2010) proposed that mammalian diversity is greatest in tectonically active landscapes, where ecotones are abundant, habitat diversity is greatest, and ecologically gradients are steepest. The biotic assemblages of the Mogollon Rim ecotone and, to some extent, GC support this hypothesis (e.g., Stevens and Polhemus 2008;Stevens and Bailowitz 2009), but not completely. In particular, her predictions 4 ("endemism…should reflect origination within the region rather than range reduction from larger areas") and 7 ("species originating in topographically complex regions should colonize adjacent lowlands more often than the reverse pattern") are not fully supported in GC, where relictual endemism prevails over adaptational radiation, and elevated species richness in refugia is more the result of colonization from surrounding biomes during favorable climate conditions. Nonetheless, the upriver attenuation of aquatic Heteroptera and Odonata taxa reported in our work suggests that species richness is related to proximity to geologic province boundaries and regional topographic diversity.
Landform evolution is evident not only in regional geology, but also through the distribution and genomes of present-day species. However, the role of past landscape change on contemporary biogeography is difficult to determine. Mitochondrial analysis of GC Hyla arenicolor treefrogs indicates a discrete episode of introgressive hybridization with H. eximia in the latest Miocene and recent or on-going hybridization with H. wrightorum (Bryson et al. 2010). The timing of those introgressions may reflect initial opening of the Colorado River to the Gulf of Mexico, and post-Pleistocene faunal mixing, respectively. But many other examples of ancient landform change and contemporary biogeography are likely to exist. Is the range restriction of most GC nepomorph Hemiptera to westernmost GC the shadow of the ancient (Oligocene?) river drainage from what is now southern Nevada into Arizona and north into Utah (Stevens and Polhemus 2008)? Does the Pacific Coast affinity of at least 3, and perhaps as many as 18 dragonfly species reflect pre-Basin and Range orogenic connectivity? Why do more than twice as any neotropical/Mexican mammals reach their northern range limits at the Colorado River, as compared to boreal species (Hoffmeister 1986). Why are Ochterus rotundus, 8 other aquatic GCE Hemiptera, and masked clubskimmer dragonflies neotropical isolates, with ranges otherwise centered in central and southern Mexico? Further research, including distributional, autecological, and genetics analyses of these generally poorly known taxa, is needed to determine whether and at at what scale environmental changes influence present-day LDC biogeography in the GCE.

Conservation biogeography
Human impacts on the GCE have profoundly altered ecosystem structure, composition, and biogeography through three processes: habitat alteration, extirpation, and the introduction of non-native species. The Colorado River is one of the most regulated rivers in North America, with a dozen large dams and thousands of small impoundments throughout its catchment (Hirsch et al. 1990). Dams and irrigation systems have fundamentally altered flow, flood dynamics, sediment transport, water temperature and chemistry, and the distribution of riverine biota. The loss of 4 of the 8 native fish species from GC due to habitat changes and interruption of migration has been thoroughly described (Minckley 1991), but less recognized has been the impact of greatly diminished connectivity on plant colonization processes, and interruption of range among tiger beetles, herpetofauna, southwestern river otter (Lontra canadensis Sonora), and other river corridor biota (Stevens 2011). So to, the rims of GC have sustained significant human impacts from fire suppression and alteration of forest structure (Fulé et al. 2002), the loss of large predators through federal extermination programs (Rasmussen 1941), and the degradation of more than 90% of rim springs and natural ponds (Grand Canyon Wildlands Council, 2002).
GC is one of the world's great landscape parks, a vast wilderness, and a United Nationsdesignated World Heritage Site, so it comes as a surprise that rather many biota have been lost during its protection by the National Park Service (  Table 6. Extirpated, endangered (*), extirpated-but-reintroduced (**), and at-risk taxa, including those for which insufficient information exists to evaluate conservation status.
The other important human impact on GC has been the introduction of numerous nonnative species. At present, at least 194 non-native vascular plant species are known from GC National Park, 10.7% of the entire flora, and the same percent of non-native plants as exists in the United Kingdom (Stevens and Ayers 2002;Busco et al. 2011;Stevens unpublished data). As in Stohlgren et al. (1999), Stevens and Ayers (2002) reported that hotspots of native plant species richness, such as springs and riparian habitats, also support higher numbers of non-native species. Improved understanding of the levels of habitat alteration and loss, the extent and on-going threats of extirpation, and the role and impacts of non-native species on the region's ecosystems are essential to protect the native species, natural resources, and biogeographic processes of GC and the GCE. Notable successes have been made in the restoration of native species and natural ecological processes in the GCE, including restoration of riparian habitats at Lees Ferry, and restoration of springs habitats in northwestern Arizona by Grand Canyon Wildlands Council, Inc (www.grandcanyonwildlands.org). Protection of endangered sentry milkvetch from trampling at the South Rim has enhanced population viability (Maschinski et al. 1997). Population reintroduction/restoration successes in the GCE include: the reintroduction of formerly endangered peregrine falcons (Brown et al. 1992) and California condors (Gymnogyps californianus; California Condor Recovery Team 2007); on-going attempts to reintroduce black-footed ferret (Mustela nigripes); and protection and translocation of endangered Kanab ambersnail into an off-river springs complex in middle Royal Arch Creek (Meretsky and Stevens 2000). A robust non-native plant control program has been implemented by the National Park Service: several common exotic plant species have been eliminated in the Park, and riparian habitat restoration is being attempted. Selective removal of non-native rainbow trout (Oncorhynchus mykiss) may have contributed to increased population size of endangered humpback chub (Gila cypha) near the LCR (Coggins and Yard 2011). These examples demonstrate that focused conservation actions can be effective for protection and restoration of native species and natural ecosystems in this internationally recognized LDC.

Summary
Large deep canyons (LDCs) are relatively common landforms on Earth, but their regional biogeographic roles and significance has received little scientific attention. Here I summarize information on ecological gradients, species richness, and ecosystem structure on the world's best known LDC, Grand Canyon (GC) of the Colorado River in the context of www.intechopen.com The Biogeographic Significance of a Large, Deep Canyon: Grand Canyon of the Colorado River, Southwestern USA 201 the surrounding GC ecoregion (GCE) on the southern Colorado Plateau. I first describe the extent and influences of major physical gradients in LDC biogeography affecting its biodiversity, including geomorphology, elevation, gravity, and climate. By virtue of its depth and narrowness, the inner canyon is naturally light-limited, receiving only 69.5% of ambient solar radiant energy. I then briefly review the ecology of the Colorado River ecosystem and the impacts of Glen Canyon Dam, reporting that nearly 50 yr of flow regulation has swamped geomorphic differences and limited assemblage composition in the aquatic domain, but had the opposite effect on the riparian domain. Next, I use regional biodiversity and range information on GC and GCE plants, invertebrates, and vertebrates to evaluate the biogeographic influences of GC on its biota. As a landform, GC influences species ranges and gene flow in four ways, as: 1) a partial or full range corridor of low elevation riverine and desert habitats through the uplifted Colorado Plateau; 2) a partial or full barrier across the Plateau; 3) a refuge, particularly in microhabitats like caves, springs, and escarpment rim edges; and 4) a null effect, not limiting gene flow across the landscape.
Available data indicate that GC functions primarily as a corridor and barrier/filter, and also supports refugial functions, and the ranges of relatively few taxa are unaffected by GC as a landform. GC has greater species richness than expected because it is a mixing zone of: a) Maderan (Mexican and neotropical) taxa occupying lower elevations and south-facing slopes; b) boreal and upland taxa occupying higher elevations and north-facing slopes; and c) range-centered taxa occupying middle elevations. Aspect refugia likely acquire taxa during climate extremes and support those populations well into climate transitions. Strongly vagile (e.g., flying taxa like butterflies, dragonflies, birds, and bats) tend to be relatively more species rich than low-vagility taxa (e.g., non-volant taxa, such as land Mollusca and non-flying beetles). Endemism is not as low in the region as previously reported, with 9.7% endemism among 745 invertebrate species in 10 orders studied thus far. Factors limiting development of endemism include the relatively young age of the landform (5-17 million years old), climate changes, and damming of the river by volcanic eruptions during the past half million years. At least 20 and perhaps as many as 29 vertebrate taxa, including nearly all large, wide-ranging predators have been extirpated from GC and the GCE in the past century, and more than 200 non-native plant and animal taxa have been introduced into GC, substantially altering the trophic structure of GC ecosystems. As an LDC, GC exerts a profound effect on the biota within and around it, functioning differently and in a more complex fashion than do other kinds of landforms. Due in part to this complexity, the assemblages and ecological functions of GC are susceptible to numerous human alterations, even when the best conservation practices are adopted.

Acknowledgements
I thank InTech, Inc., and particularly Ms. Dragana Manestar, for the opportunity to present this information. I received much-appreciated support from the Annenburg/Explore fund to prepare this manuscript. I warmly thank the Museum of Northern Arizona, Dr. Breunig its Director, and the MNA staff for administrative and office support. It has been my deepest pleasure and honor to work with numerous collaborators and governmental agencies on these topics over the past 4 decades, as reflected in the bibliography. Insights and stimulating discussions with these collaborators have allowed me to greatly deepen my understanding of regional biogeography; however, any errors herein are mine, not theirs. I