[From the U.S. Government Printing Office, www.gpo.gov]
----------------- U.S. DEPARTAENT OF THE INTERIOR NATIONAL. BIOLOGICAL SERVICE BIOL 0 GICA L SCIENCE REPORT 3 Or HABITAT SUITABILITY IND'EX MODELS: -NON-MIG,RATO,RY FRESHWATER. LfFE""'" STAGES,OF ATLANTIC SALMON ilP, MR mghh 29 E@ 9F g, QL638 52 4 S771 i995 National Biological Editorial Staff Service., Information Transfer Center 1201 Oak Ridge Drive, Suite 100 Technical, Report Series Fort Collins, CO 80525-5589 Phone: (970) 223-9709 The NationalBiological Service publishes four technical Fax: (970) 226-9455 report series. Manuscripts are accepted fro De' CCMail: [email protected] m partment of the Interior employees or contractors, students and faculty associated with cooperative research units, and MANAGING EDITOR Paul A. Opler other persons whose work is sponsored by the Department. Manuscripts are.received with the ASSISTANT BRANCH Paul A. Vohs understanding that they are unpublished. Manuscripts LEADER receive anonymous peer review. The final decision to publish fies with the editor. SCIENCE EDITORS Eliiabeth D. Rockwell James R. Zuboy Series Desc i tion!@ rip TE ciimm EDITORS Jerry D. Cox Technical Series Deborah K. Harris Biological Science Report ISSN 1081-292X VISUAL INfORMATION 'Constance M. Lemos Ifighest quality of original science, review, and inven- SPECIALIST ton'es, which include North American Fauna and periodic status and,trends reports.,Each publication will be available on the electronic library server of the EDITORIAL ASSISTANT Martl, a W. Nichols Information Transfer Center (ITC). Selected publica- tions will be converted to CD-ROM and may be EDITORIAL CLERK Donna D. Tait obtained from the Publications Unit or.t1fe National Technical Inf6rmation'Service". 1P CLEm-TypisT Linda K. Bogard Informatibn-and Technology Report ISSN 1081-2911 Serie's contains conference proceedings, syntheses, dnnotated bibliographies, new techniques, and other products. Other Reports National Biological Servicejnformation Bulletin (formerly Research Information Bulletin) Contains interim results of scie -ntific studies. Hard cop- ies of each bulletin are distributed to appropriate offices and stations throughout the.-department. Elec- tronic copies are available from the ITC library server. Copies of this publication may be obtained National Biological Service Open File Reports from the Publications Unit, U.S. Fish and Wildlife An all-electroiiic series of reports is available as part of Service, 1849 C Street, N.W., 'Mail Stop 136, the Other Reports series and may include pre-publica- Webb Building, Washington, D.C. 20240 (611 tion info -rmation or other material (data sets, 703-358-1711; FAX 703-358-2314), or may be geographic Information System [GIS] maps not appro- purchased from the National Technical priate for hard-copy publication). *Provisions are Information Service (NT.IS),,5285 Port Royal underway for hatd-copy access and registration with Road, Spongfield, Virginia 22161 .(call toll free the National Technical Information Service. 1-800-553-6847). 01'rinted on recycled paper /0001 U.S. DEPARTMENT OF THE INTERIOR NATIONAL BIOLOGICAL SERVICE WASHINGTON, D.C. 20240 BIOLOGICAL SCIENCE REPORT 3 MAY 1995 HABITAT SUITABILITY INDEX MODELS: NONMIGRATORY FRESHWATER LIFE STAGES OF ATLANTIC SALMON By Jon G. Stanley and Joan G. Trial LIBRARY NOAA/CCEH 1990 HOBSON AVE. CliAS. SC 29408-2623 Preface Information in this report is for impact assessment and habitat management. This Habitat Suitability Index (HSI) model for nonmigratory freshwater stages of Atlantic salmon is the third generation of a model that was developed originally from a review and synthesis of existing information on Atlantic salmon (Trial and Stanley 1984). We define a juvenile as either the fry or parr stage up to the time of transformation to the smolt. We also include model variables for the embryo stage. We report on how the model was modified based on field testing in Maine in 1984 and further evaluated by comparison of alternative model outputs with a long-term data base from Canada and habitat selection data gathered in Maine. Despite the testing that went into developing this HSI model, it is nevertheless a hypothesis of species-habitat interactions, not a statement of proven cause and effect. These interactions are presented as an index on a scale from 0 (unsuitable habitat) to I (optimally suitable habitat). Through further use of this HSI model in assessing habitat in relation to Atlantic salmon populations, this index can be further refined. The National Biological Service encourages model users to convey comments and suggestions that may help increase the utility and effectiveness of the model. A form is provided in the appendix for this purpose. Contents Page Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I Habitat Use Information For Atlantic Salmon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I Age, Growth, and Food . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Reproduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Specific Habitat Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Habitat Suitability Index (HSI) Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Applicability of the Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Model Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Suitability Index Graphs for Model Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Rationale and Assumptions for Suitability Indices (SI's) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Field Application of the Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Interpreting Model Outputs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Sources of Additional Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . 15 Cited References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Appendix. Model Evaluation Form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Habitat Suitability Index Models: Nonmigratory Freshwater Life Stages of Atlantic Salmon by Jon G. Stanley National Biological Service Great Lakes Science Center 1451 Green Road Ann Arbor, Michigan 48105 and Joan G. Trial Maine Department of Inland Fisheries and Wildlife Fisheries Division 650 State Street Bangor, Maine 04401 Abstract. A Habitat Suitability Index model was developed by evaluating individual suitability indices of 17 environmental variables that have been shown to affect productivity or survival of nonmigratory freshwater life history stages of Atlantic salmon (SaInto salar L.). These stages included egg, embryo, fry, and parr but not smolt. During summer base flows, the most suitable habitats had temperatures of 16-191 C, oxygen percent saturation exceeding 60%, and pH between 5.5 and 6.8. The most suitable current velocity was 10-30 cm/s for fry and 10-40 cm/s for parr. The most suitable depth was 10-40 cm for fry and 20-50 cm for parr. The Habitat Suitability Index model is useful for evaluating stream habitat for production and survival of juvenile Atlantic salmon when these variables cannot practically be measured directly. Keywords: Atlantic salmon, ecology, habitat, water quality, substrate, streams, parr, spawning, habitat suitability index. Habitat Use Information For Europe it ranges from Iceland to Portugal, including the Atlantic Salmon Baltic Sea (Netboy 1974). Anadromous populations once migrated into most New England streams and the St. General Lawrence River tributaries, including Lake Ontario and Lake Champlain. Dams, pollution, and overfishing have The Atlantic salmon, Salmo salar L., inhabits the North eliminated spawning runs over much of the Atlantic Atlantic Ocean basin from Greenland to the Connecticut salmon's range in North America and Europe (Danie et al. River of New England (Scott and Crossman 1973). In 1984; Mills 1989; Thompson 1993). Landlocked popula- fions in North America, on the other hand, were endemic to 'Present address: School of Natural Resources and Environment, only a limited number of large lakes and watersheds but now University of Michigan, Ann Arbor, Michigan 48109-1115. occur in numerous lakes, especially in Maine, because of I 2 BIOLOGICAL SCIENCE REPORT 3 stocking (Warner and Havey 1985). The juveniles of 30 cm/s, corresponding to the preferred habitat selected by anadromous parents can be distinguished from juveniles of parr (Heggenes and Borgstrom 1991). landlocked populations, but the difference is not great Juvenile Atlantic salmon occupy stations in streams enough to categorize any particular individual (Riley et a]. and feed on invertebrates driffing on the surface and in the 1989). Their habitat use in small streams is similar (Sayers water column (Bley 1987). The diet is variable, generally 1990), and the Habitat Suitability Index (HSI) model in this consisting of the larvae of mayflies, chironomids, caddis- report applies to both. flies, blackflies, and stoneflies; annelids; and mollusks The Atlantic salmon has high social and economic (Scott and Crossman 1973). Larger juveniles also eat the value. The adults are harvested commercially on their adult fortris of aquatic insects and terrestrial insects. Food feeding grounds off Greenland and southern Labrador, size varies in direct proportion to the size of the fish Canada, and they are caught in the recreational fishery (Sosiak et at. 1979). Atlantic salmon are opportunistic during their migration as they reenter fresh water. The feeders, readily changing their diet to the most abundant annual return of anadromous Atlantic salmon to U.S. prey available. streams is about 5,000 individuals; the number returning After juveniles reach a total length of 125-150 mm, from year-to-year varies about five-fold (Rideout 1989). environmental stimuli trigger transformation into a smolt In Canada, the annual production potential is about 1.5 ready to migrate to sea (Danie et al. 1984). At the smolt million large salmon and an equal number of grilse, which stage, landlocked salmon migrate from streams into lakes are young salmon that are returning to their native rivers (Warner and Havey 1985). Migration is keyed to environ- to spawn after one winter at sea (Lear 1993). mental stimuli of rising water temperature, freshets, and Worldwide there are more than 350 recognized stocks photoperiod (Bley 1987). A few individuals in one popu- of Atlantic salmon (Chadwick 1985). In general, each lation of landlocked Atlantic salmon migrated to lakes in major river system has its own stock uniquely adapted to autumn (Warner and Havey 1985). the local conditions (Thorpe 1988). In North America, the Reproduction populations of Atlantic salmon are bolstered by stocking of hatchery fish (Rideout 1989). In some runs, more than Atlantic salmon spawn in fresh water during October 90% of the fish are from hatcheries. The juveniles are and November when water temperatures reach 4.4-5.6* C generally stocked when they are 1-year-old smolts ready (DeCola 1970). Eggs are deposited in redds dug by adult to migrate to the sea. In the United States, the hatchery females at the downstream end of riffles where water stock consists of seven different strains or developing percolates through the gravel or at upwellings of ground strains reared at six federal and four state hatcheries (Kane water. One or more males fertilizes the eggs as they are 1989). The HSI model was developed for the stock of deposited, and the female then completes the redd by Atlantic salmon inhabiting streams in New England and covering the eggs with 10-25 cm of gravel displaced from the southern Canadian maritimes. Suitability indices (SI's) upstream. The eggs are slightly adhesive and stick to the have been produced for Atlantic salmon in Newfoundland substrate until they are covered. (Scruton and Gibson 1993), and a workshop to develop The eggs incubate over winter buried in gravel. The models was conducted in 1992. incubation period varies with temperature. Eggs hatch after 175-195 days under normal winter conditions of Age, Growth, and Food Maine (Jordon and Beland 198 1). The incubation time of 110 days cited by Leim and Scott (1966) was for a tem- Juvenile Atlantic salmon grow relatively slowly in perature of 3.9' C, typical of a hatchery drawing hypolim- fresh water, whereas adults grow rapidly at sea. Juveniles nionic water from a lake. may spend 2-3 years in fresh water to reach 125-150 min After hatching, the eleutheroembryos (alevin or yolk- length in New England and 4-8 years to reach 180 mm in sac larvae) remain buried in the gravel for about 6 weeks, Ungava Bay, Canada (Schaffer and Elson 1975). The until their yolk sac is depleted of nourishment. The resul- young salmon grow fastest at temperatures of 15-190 C tant fry begin foraging while still in the substrate, then (DeCola 1970). The lower temperature limit of growth, emerge at night from mid to late May in Maine (Gustaf- which varies with nursery stream conditions, ranges from son-Marjanen 1982) and from late May to early June in 5 to 10' C (Jensen and Johnsen 1996). Survival is POsi- Canada (Scott and Crossman 1973; Lear 1993). In one tively correlated with water discharge from streams river in Finland, emergence was as late as July (Mills (Frenette et al. 1984; Gibson et al. 1993). Growth in fresh 1989). Survival from fertilization through hatching was water also is limited by availability of food, interspecific 74%, and only 2% from fertilization through emergence and intraspecific competition, and a range of other factors. (MacKenzie and Moring 1988). Chadwick (1982) found Growth is fastest in habitat with water velocities of about that the survival rate was depressed during a year when RABrrAT SurrABiLrry INDEX MODELS: NONMIGRATORY FRESHWATER LiFE STAGES OF ATLANTic SALMON 3 winter air temperatures and water discharge were both for first-year and yearling parr demonstrate that Atlantic low. After emergence, fry disperse, mostly downstream, salmon have the highest temperature limits for feeding and establish territories, Dispersal is highest at night (Crisp (22.5' C) and survival (27.8' C) among brook trout 1991). (Salvelinusfontinalis), brown trout (Salmo trutta), and five species of Pacific salmon (Elliott 1991). Specific Habitat Requirements The information on temperature relationships from the literature was adequate for creating SI's for spawning During its anadromous life cycle, the Atlantic salmon temperatures, egg incubation, upper tolerances, and sum- completely changes its habitat from freshwater streams to mer growth. The information was inadequate for develop- the sea. This report emphasizes the freshwater segment of ing models for cold lethal temperatures. the life cycle. Atlantic salmon require cold, clear streams that flow freely to the ocean. Because many such streams Dissolved Oxygen in the northeastern United States and Canada arise in areas with granitic bedrock, they are subject to acidification. Oxygen concentrations near saturation are needed for Atlantic salmon streams are generally well oxygenated, optimal development and growth. Embryo and larval de- except in a few rivers receiving industrial or domestic velopment requires a minimum of 6 mg/L of dissolved pollutants. Impoundments on many rivers not only block oxygen (Elson 1975). Mortalities occur if embryos are migration but also create long reaches of still water that exposed to oxygen concentrations of less than 6-7 mg/L result in increased water temperatures. The environmental (DeCola 1970). Juvenile salmon do not occur in streams requirements of Atlantic salmon were thoroughly re- in which dissolved oxygen regularly drops below 5 mg/L viewed by Gibson (1993). (Elson 1975). In the laboratory at 14.5' C, Atlantic salmon juveniles select the highest oxygen concentration avail- Water Teinperature able-7.5 mg/L or 72% saturation (Trial and Stanley Temperature is a key variable in determining habitat 1984). At 15-160 C, lethal concentrations are 1.5 mg/L for suitability for Atlantic salmon. All stages of the life cycle juveniles during the first summer of life and 1.9 mg/L require cool temperatures. Spawning occurs between 4.4 during the second year (DeCola 1970). and 10' C; the optimal temperature for fertilization and We converted all dissolved oxygen concentrations to incubation is about 6' C (Danie et al. 1984). Development units of saturation using standard tables to construct an SI proceeds, but at a slower rate, at temperatures as low as for minimum summer values. -0.5' C (Peterson 1978). Incubation temperatures above Acidity 12' C cause direct mortality, whereas temperatures be- tween 8 and 12' C may cause secondary mortality because The pH of salmon streams in granitic, sandy, or boggy of fungal infections (Garside 1973; DeCola 1975). areas may be depressed by melting snow or heavy rains Newly hatched larvae are exposed to and tolerate rising that contain acid. Episodes of low pH are often accompa- water temperatures in spring. Although they have little nied by high concentrations of metal ions that ]each from opportunity for selecting temperature, if given a choice the soil, especially aluminum (Lacroix and Townsend they will move to the coldest temperature available (Peter- 1987). Thus, toxicity may not be caused by acidity per se; son and Metcalfe 1979). At about 250 degree-days after nevertheless, pH may serve as a convenient indicator of hatching, when the fry establish territories in streams, they water quality. For many areas in New England and the prefer a temperature of 14' C. Juveniles (13-16 cm total Canadian maritimes, organic compounds chelate alumi- length) select a temperature of 14.5' C (Trial and Stanley num. Rivers in Nova Scotia with a mean annual pH of less 1984). They need a growing season of about 100 days with than 4.7 have lost their salmon runs, in rivers with pH stream temperatures above 6' C (Power 1969). between 4.7 and 5.0, runs declined, and in those with pH The optimal water temperature for growth and produc- above 5.0, runs were normal (Watt et al. 1983). In these tion-] 5 to 19' C (DeCola 1970)-seems to be slightly same streams, juveniles were most numerous where the higher than the preferred temperature. Growth seemed to mean annual pH was above 5.4, were much reduced be- be fastest at 16.6' C (Siginevich 1967). In the laboratory, tween pH 4.7 and 5.0, and were absent below pH 4.7. p4rr grew better at 13-19' C than at colder temperat Iures Eggs develop normally at pH 6.7 (Peterson et al. 1980). (Dwyer and Piper 1987). Young Atlantic salmon can tol- The embryo has a lower lethal level of about pH 3.5 during erate temperatures up to 27' C for short periods but seek early cleavages and pH 3.1 just before hatching (Peterson cooler water as these temperatures are approached (De- et al. 1980). However, a pH of 4.0-5.5 delays hatching. A Cola 1970). Juveniles withstand 32' C briefly (Huntsman pH less than 5.0 inhibits enzymes necessary for hatching, 1942), and the lethal temperature under laboratory condi- and reproduction fails (Haines 1981). Yolksac fry in situ tions is about 320 C (Garside 1973). Tolerance polygons had 100% mortality at pH 5.1 with high aluminum and 4 BIOLOGICAL SCIENCE REPORT 3 65% mortality at PH 5.7 with lower aluminum (Norrgren Canadian streams to be 32 cnits, andjuveniles were absent and Degerman 1993). in areas with a mean column velocity exceeding 120 cm/s. Juvenile Atlantic salmon are often exposed to low pH Fry were most abundant in stream sections where mean along with other stresses, such as swift water, toxic metals, column velocity was less than 30 cm/s, but some fry were and turbidity. The low PH causes edema of the gill lamella observed at mean column velocities up to 60 cm/s (Heg- and may disrupt respiration and excretion at times when genes et a]. 1990). metabolic demands are high. In the laboratory, juveniles The velocity in microhabitats selected by fish was less had a'lower lethal PH of 4.0 after 28 days of exposure than the mean column velocity. The mean velocity in the (Daye and Garside 1977, 1980). Concentrations of so- microhabitat where fry hold was 17 cm/s in one stream and dium, calcium, and chloride declined in the plasma after 5 cm/s in another (Trial and Stanley 1984). In Canadian exposure to PH 4.6 but not at PH 5.0 or 5.5 (Fanner et al. streams, fry preferred 0-5 cm/s in one stream and 17- 1989). 21 cm/s in another (DeGraff and Bain 1986). The mean An SI was developed with this literature information in velocity where fry held position was 12 cm/s; most selected which frequency of acidic episodes and their intensity are velocities of 5-19 cm/s (Morantz et al. 1987). Gibson (1993) key variables. Although aluminum and other metals are reported that the mean velocity where fry held position was obviously important in the manifestation of acidic effects, 13 cm/s in riffles and pools. In the laboratory, the velocity we could not incorporate such information into a simple where fry can no longer hold position was 150 cm/s at 6-8' model. In general, HSI models do not consider toxic C and 190 cm/s at 12-14' C (Heggenes and Traaen 1988). substances. The intent is for the models to predict the Trial and Stanley (1984) reported that 3-month-old Atlantic quality of habitat in the absence of specific contaminants salmon could maintain position in a flow tank in a velocity and myriad other confounding factors. of 50 cm/s at 16' C and PH 4.5-6.0 cm. Below pH 4.0, Velocity however, they were unable to maintain position in velocities faster than 42 cm/s. Even in this flow tank, the fry were able Atlantic salmon in fresh water require flowing water, to find pockets with currents that were 60-70% of those although they will occupy slow-moving or lentic habitats measured. (Einarsson et al. 1990; Cunjak 1992). Adults select spawn- Competition between size groups is reduced by habitat ing sites in riffles where the average velocity is about 50 segregation (Gibson et al. 1993). Yearling Atlantic salmon cm/s (Elson 1975; Beland et al. 1982). The lower limit of Parr have about the same velocity preferences as fry. The velocity was 15-20 cm/s, and the upper limit was related older and larger Parr take the best territories, usually in to female size (e.g., a 50-cm female would be limited to midstream. Habitat selection toward the middle of the velocity less than 100 cm/s; Crisp and Carling 1989). stream was evident from measuring where Parr occurred Egg incubation requires upwelling of ground water or relative to the shore-18 fish were 0-0.2 in from shore, percolation of stream water through the gravel substrate, 121 were 0.3-0.7 in from shore, and 266 were 0.8-1.5 in which is measured with a standard permeability test in from shore (Hesthagen 1988). In some streams fry prefer which the rate of water dispersal from apipe is determined. riffles and Parr prefer runs (Tremblay et al. 1993). As Permeabilities of 1.3-1.4 L/h are typical (Gustafson-Mar- juveniles grow, they are able to cope with the faster water janen and Moring 1984). and thus benefit from more drifting food (Morantz et al. Newly emerged fry occupy areas with current but 1987). They minimize energy expended on swimming by select microhabitat with slower water. In an artificial utilizing low-velocity areas, hiding among rocks in riffles stream, newly emerged fry dispersed fastest when in a low and darting into the swifter current only to feed. The most velocity of about 8 cm/s (Crisp 1991). More stayed in the favorable territories in some streams were in faster water, stream at velocities of 25-70 cm/s, implying that the faster in others slower, possibly because of other factors, such as velocity was more suitable for fry. There was considerable cover. Trial and Stanley (1984) reported that the velocity variation in how velocity was measured, reported, and in the microhabitat occupied by yearling Parr in one stream interpreted. The prevailing velocity may be as important was 9 cm/s, which was slower than in the areas occupied as that where thejuvenile actually rests. The mean velocity by fry, and 6.5 cm/s in another stream, which was faster in the water column in areas preferred by first-year fish is than in areas occupied by fry. 50-65 cm/s (Symons and Heland 1978). Knight et al. The gradients of streams where juvenile salmon occur (1981) reported that the preferred habitat had a mean range from 2 to 12 m/km (Elson 1975). In rivers in Nova column velocity of 14 cm/s. The preferred mean column Scotia, the highest densities of Parr were at a gradient of velocity was 10-31 cm/s in one Canadian stream and 1.2-1.4% (Amiro 1993). Such gradients generate mean 10-46 cm/s in another (DeGraff and Bain 1986). Morantz column and microhabitat velocities within the preferred et al. (1987) reported the mean column velocity in eight ranges for fry and Parr. Knight et al. (1981) found that HABiTAT SurrABiLiTy INDEX MODELS: NoNAaGRATORY FRESHwATER LIFE STAGES OF ATLANTIC SALMON 5 yearlings occupied stations where the mean velocity was 14 were abundant, mean depth ranged from 10 to 31 cm (Fran- cm/s. Based on distribution in one stream, most Parr pre- cis 1980). The preferred depths for fry in one stream in ferred a mean column velocity of 10-24 cm/s; in another Newfoundland were 13-25 cm and in another, 20-60 cm stream, Parr preferred 16-57 cm/s (Degraff and Bain 1986). (DeGraff and Bain 1986). In Newfoundland, fry occurred at The mean column velocity selected by small Parr (about 85 17 cm depth in riffles and 32 cm in pools (Gibson 1993). In mm long) was 40 cm/s and by large Parr (about 120 mm.) eight streams in Nova Scotia and New Brunswick, the 35 cm/s (Morantz et al. 1987). Gibson (1993) reported, how- average depth used by fry was 35 cm; most individuals were ever, that small (6-10 cm) and large Parr (>10 cm) were found between 20 and 40 cm (Morantz et al. 1987). In a located at a mean velocity of 20 cm/s in riffles. In pools, Norwegian stream, fry were mostly in waterless than 60 cm however, small Parr selected velocity of 6 cm/s and large deep (Heggenes et al. 1990). In England, in water less than Parr 13 cm/s. The most preferred holding velocity of Parr, 20 cm deep, fry outnumbered Parr, whereas in water deeper was 0-5 cm/s in one stream and 16-21 cm/s in another than 20 cm, Parr were more abundant (Kennedy and Strange (Degraff and Bain 1986). Large Parr were found in velocities 1986). In another English stream, the number of fry was of 22 cm/s but preferred velocities of 10-20 cm/s (Morantz positively correlated with depth up to 20 cm and inversely et al. 1987). Larger Parr occupied microhabitats with veloci- correlated with depth greater than 20 cm (Egglishaw and ties of 0-25 cm/s in streams with mean velocity of overlying Shackley 1985). water of 0-75 cm/s (Heggenes et al. 1991). Parr seem to prefer deeper water than fry (Gibson 1993; In fall, 2-year-old juvenile Atlantic salmon moved from Gibson et al. 1993), which is usually found midstream. the riffle area of streams into slower water, where they Selection within streams is affected by availability of deep remain during winter, whereas I -year-old juveniles did not water with suitable velocities related to stream morphol- move (Rimmer et al. 1984). Huntingford et al. (1988) found, ogy. As with fry, the average depth selected by Parr varies however, that all fish sought areas of low flow in fall. In among streams and therefore among studies. The mean winter, Parr hide under rocks in riffle areas with overlying depth of preferred areas for Parr in one New England velocities of 38-46 cm/s (Cunjak 1988). Parr destined to stream was 29 cm (Knight et al. 1981), 49 cm in another, become smolts the following year selected faster currents in and 33 cm in a third (Trial and Stanley 1984). In some an artificial stream than did Parr destined to remain in fresh Canadian streams, Parr preferred depths of only 10-15 cm .water for 2 years (Huntingford et al. 1988). (Symons and Heland 1978), whereas in two other streams, Suitability indices were developed for velocities meas- preferred depths were 22-42 cm and 14-48 cm (DeGraff ured at 0.6 of total depth of the water column or below. The and Bain 1986). The range of depths preferred by Parr lower portion of the water column is where the fish spend differed in eight Canadian streams; most occurred between most of their time, and velocity at 0.6 of the total depth 30 and 60 cm (mean 47 cm; Morantz et al. 1987), In approximates the average velocity for the water column Newfoundland, small Parr (6-10 cm) were found in 22 cm (Hamilton and Bergersen 1984). We had insufficient veloc- of water in riffles and 42 cm in pools (Gibson 1993). Large ity data to develop a fall or winter SI for velocity. Parr (>10 cm) used slightly deeper areas, 24 cm in riffles and 57 cm in pools. In Europe, yearlings were most abun- Depth dant at a depth of 35 to 40 cm in one stream (Kennedy and Spawning sites are selected at the tails of pools that are Strange 1986) and deeper than 25 cm in another (Eg- near the beginning of riffles. The depth depends on the size glishaw and Shackley 1985). Most yearling or older Parr of the stream and the size of the fish. A 50-cm female occupied habitats with depths less than 90 cm (Heggenes requires depths of 10-40 cm (Crisp and Carling 1989). In et al. 1990). Larger Parr used depths between 25-and Maine rivers, the average depth over spawning redds was 40 85 cm, but a few were in water deeper than 100 cm cm (Beland et al. 1982); in New Brunswick, it was 20 cm (Heggenes et al. 1991). (Peterson 1978). Suitability indices were developed for depth over spawn- After hatching, fry disperse and establish territories. Fry ing areas and at summer low flows for fry and Parr. We had establish residence in shallower water nearer shore. The insufficient data to develop a winter S1 for depth over redds. depths where fry reside in each stream are related to stream However, because most fish descend into the substrate in morphology, which determines the depth of near-shore areas winter, we believe that the summer depth SI's for fry and at low flows. Thus, average depth selected by fry will vary Parr are also applicable to winter low flows. In winter, Parr among streams. Knight et al. (198 1) reported that fry habitat occupy riffles in water 41-49 cm deep (Cunjak 1988). in New England streams averaged 25 cm deep (range 9-39 Substrate, Sedhnent, and Turbidity cm). In Maine, fry preferred water 34 cm deep in one stream and 27 cm in another (Trial and Stanley 1984). In Canada, Adults select spawning sites at the tails of pools that have for 62 sites on New Brunswick streams and rivers where fry substrate composition reflecting sorting by the swift currents 6 BioLoGicAL ScniNcE REPoRT 3 that move over this habitat. In Peterson's (1978) study, the interference with sight feeding and growth is possible. In particle size composition was 0-3% fine sand (0.06- the laboratory, coho salmon (Oncorhynchus kisutch) and 0.5 mm), 10-15% coarse sand (>0.5-2.2 mm), 40-50% steelhead (0. mykiss) grew fastest in clear water; growth pebble (>2.2-22 mm), and 40-60% cobble (>22-256 mm). was inhibited at 45-50 NTU (Sigler et al. 1984). In some Spawners preferred gravel of 20-30 nun diameter (Crisp tests, 38-49 NTU did not inhibit growth, and in other tests and Carling 1989). The substrate composition of the redds turbidity as low as 25 NTU inhibited growth. When coho of landlocked Atlantic salmon included higher percentages salmon were exposed to turbidities of 30 and 60 NTU, of intermediate-size particles (Warner 1963). The land- territoriality deteriorated and prey capture rates declined locked fish, which are smaller than the sea-run fish, may not (Berg and Northcote 1985). Coho salmon avoided turbid- be capable of moving the larger particles. ity of 70 NTU (Bisson and Bilby 1992). Episodes of high Because juvenile salmon occur in the riffle area of turbidity seem to do no harm, and turbidity alone corre- streams, they are likely to be found above substrate contain- lated poorly with effects of suspended sediments on fish ing sand, gravel, and cobble rather than silt. In one stream, (Newcombe and MacDonald 1991). Relatively low tur- Atlantic salmon fry selected a substrate classified as 4.8, bidities over long periods caused reduced feeding in sev- based on an index in which 3 represents fines and detritus; eral species of salmonids (Newcombe and MacDonald 4, sand; 5, gravel; and 6, cobble (Trial and Stanley 1984), @n 1991). Turbidities exceeding 1,150 standard units, meas- two Canadian streams, the most preferred substrate had an ured with a photometer during fall freshets, did not injure index of 4.5-5.5 for fry and parr (DeGraff and Bain 1986). or kill Atlantic salmon fry or parr (McCrimmon 1954). In eight other Canadian streams, this index was 5.6 for fry, The SI for turbidity was based on effects on other 5.9 for small parr, and 6.4 for large parr, indicating selection species of salmonids, primarily as reported in the review of a coarser substrate as juveniles grow (Morantz et al. by Newcombe and MacDonald (1991). 1987). During their first year, juveniles preferred gravel substrate (16-64 mm), whereas yearling parr preferred a Habitat Suitability Index (HSID boulder and rubble substrate where diameters were greater Models than 260 mm (Symons and Heland 1978). Gibson (1993) concluded that fry are most common where there is a pebbly bottom, and parr over coarser substrate. In a Norwegian APPlicability of the Models stream, firy were observed over a gravel to boulder substrate, Potential users of the HSI model for Atlantic salmon and parr occupied a wider range of substrate types (Heg- can have confidence in applying the model in habitats genes et al. 1990). where the model was developed and tested. Considerable Depth, velocity, and substrate are interdependent. Sub- effort has gone into validating the model, especially for the strate is related to velocity, and velocity is affected by more important variables. The model is equally applicable depth. It is difficult to determine whetherjuveniles, as they to anadromous and landlocked populations of juvenile grow, select larger substrates, faster velocities, or deeper Atlantic salmon. We recommend the use of this HSI model areas with similar substrate and velocity. However, all to help formulate expert opinion on habitat quality. How- variables seem to be differentially selected by fry and parr. ever, we caution that the model is a hypothesis describing Our SI's were constructed so that pebble-size substrates a simplified version of complex interrelationships within were best for fry and cobble substrates best for parr. a seasonally dynamic environment. In addition, the model Sedimentation into the spaces between pebbles and concerns a species with shifting habitat requirements, cobble interferes with the use of this space as shelter for complex behaviors, and a long life cycle. Furthermore, this young Atlantic salmon and decreases their survival rate in HSI model does not consider toxic chemicals, which if summer (McCrimmon 1954). In winter, siltation and sus- present may limit the application to predicting what habitat pended debris within the substrate are also important be- quality would be if the contaminant were removed. cause fish hide in spaces under rocks (Cunjak 1988). Such sedimentation obviously also affects benthic production Geographical Area and reproductive success. . The HSI model was designed for Atlantic salmon in Atlantic salmon typically occur in clear streams and streams of New England and the Canadian maritimes of depend on transparent water for site-feeding. Turbidities temperate North America. The model applies to embryos, of 40 nephelometric turbidity units (NTU) or less are fry, and parr in streams and to adults only in regard to the considered to represent clear water that is highly suitable selection of spawning sites. European populations share for feeding. Survival of fry and parr was highest in stream many of the same characteristics as the North American segments with the lowest base turbidities (McCrimmon populations, and the model could probably be applied to 1954). As turbidities increase to 100 NTU, progressive European populations with little modification. HAMAT SurrABELny INDEX MODELS: NONMIGRATORY FRESHWATER LWE STAGES OF ATLANTIC SALmoN 7 Season 1990). Parr then disperse gradually over the summer to The water quality, fry, and parr components of the HSI occupy all suitable stream habitats. In winter, older parr model are designed to evaluate the summer habitat ofjuve- move from the riffles in streams into slower waters (Rim- nile Atlantic salmon during base flow, when the extent of mer et al. 1984). Dispersal was faster for parr planted in deep, slow water than for parr planted in their preferred the available habitat is limited. The reproductive component obviously applies during the fall period. Winter habitat may habitat of fast-moving water (Heggenes and Borgstrom be particularly important to the survival of Atlantic salmon; 1991). For most of these dispersal phases, the extent of for example, low winter discharge significantly affects ju- movements is unknown. venile survival (Gibson and Myers 1988). Except for tem- Ve6fication Level perature and ice, the habitat occupied by juveniles differs only slightly from summer habitat (Rimmer et al. 1984; Originally, the SI's and HSI model presented here were Cunjak 1988). Although survival is correlated with winter derived from literature values and initially tested in Maine air temperatures and water levels (Chadwick 1982), no streams (Trial and Stanley 1984; Trial et al. 1984). Suitabil- measurements link specific winter conditions to embryo or ity indices for water depth, velocity, and substrate were juvenile survival. independently developed and tested in Canadian streams (Morantz et al. 1987). A third test for validation was done in Habitat Types Maine streams (Trial 1989). A fourth test was done in which The HSI model applies to embryos and juveniles in Trial (1989) analyzed data collected in New Brunswick by freshwater, riverine (lotic) habitat. The model describes the Francis (1980) and Trial (1989). Recently, SI's developed area where spawning and egg incubation occur, as well as for Newfoundland rivers were tested and found to consis- nearby nursery areas for juveniles. The model does not tently predict standing crop of fry (Scruton and Gibson consider the lake and marine feeding grounds of adults or 1993). any habitat characteristics critical to successful downstream Trial (1989) formulated four alternative HSI models or upstream migration in estuaries or freshwater streams. based on an evaluation of SI's related to velocity, substrate, The model has the most validity when applied to the streams and depth. Two of the HSI models used SI's from Morantz in which it was tested or to similar streams. Tests were done et al. (1987) from the fry and parr components, and two used in streams ranging from small brooks to the mainstem of SI's from Trial and Stanley (1984). Trial (1989) determined major rivers, such as the St. John River in New Brunswick. the goodness of fit between the measured variables for habitat selected by juveniles and the S1 values. In other Minimum Habitat Area words, the cumulative frequency distribution (CFD) of suit- In HSI models, the n-dnimum habitat area usually in- ability based on habitat selection by fish was compared with cludes egg incubation areas, nursery and juvenile feeding the hypothetical CFD. This process was done stepwise for grounds, and adult feeding grounds. For Atlantic salmon, the individual SI's and the life stage component indices pro- usual definition of minimum habitat area does not apply duced from either the product of the three SI's or their because the habitats for the different life stages usually are geometric mean. Trial's (1989) test of fry and parr compo- not contiguous. Of critical importance to Atlantic salmon nents indicated that the CFD from the SI's in Trial and populations is free passage between the different habitats, Stanley (1984) and Morantz et al. (1987) had a more gradual unobstructed by dams or interception by excessive fisheries. rise to 100% than the CFD from her data on fry in Maine The area of habitat used by Atlantic salmon varies streams. The apparent lack of fit of the components was considerably. Some stocks of landlocked Atlantic salmon expected because the Sl's were overestimates of the optimal exist within a single river system in which spawning, range of each habitat variable (Trial 1989). nursery, and feeding areas are within a few kilometers of In a second test, Trial (1989) found that thejoint prob- each other, for example, the West Branch of the Penobscot abilities and geometric means for the fry component index River in Maine (Warner and Havey 1985). At the opposite were correlated with the density of fry. The two ways of extreme, some populations have nursery grounds in small calculating component indices did not affect the ranking streams in Portugal, and the adults feed in Arctic waters of the sites-the ranks of the alternative component indi- off the coast of Baffin Island in North America (Netboy ces were correlated with the rank of fry density at 16 sites. 1974). The minimum habitat area for the juvenile life In contrast, none of the parr component models correlated stages is poorly defined, in part because little information with parr densities, probably because other variables, such is published on distances for the dispersion of fry and parr. as cover, were important. For total numbers of juveniles, Dispersal occurs rapidly in spring as fry emerge from the three of the four HSI models were correlated with popula- redd and move predominantly downstream (McKenzie tion density, and only the model based on joint probability and Moring 1988; Gustafson-Greenwood and Moring and the SI's by Morantz et al. (1987) was not correlated, 8 BIOLOGICAL SCIENCE REPORT 3 These tests of HSI models verified the HSI approach most conservative approach for modeling life stage suitabili- toward evaluating habitat and validated some of the SI's, ties. Because there was no difference in the statistical fit of especially for the water velocity, depth, and substrate of component indices calculated using the joint probability or fry. These tests measured density, abundance, or site se- the geometric mean (Trial 1999), we chose to use the model lection as an indicator of carrying capacity. In general, the that is easiest to use. Bain and Robinson (1988) expressed microhabitat used within any one stream was narrower concern that numerous variables in a geometric mean would than predicted by the models, whereas the range of habitats result in an unrealistic degree of compensation for the lowest used among streams was predicted accurately. Scruton and values. Therefore, we used a joint probability approach to Gibson (1993) noted that SI's are more useful if derived calculate component indices. from macrohabitat measurements (e.g., stream width) rather than microbabitat (e.g., variables measured at loca- Water QuAty tion of individual fish). The water quality component was modeled by a mini- In tests of HSI models by Trial (1989), the reproductive mum value. Fry (1971) and Bren (1979) recommended this component was based on water quality and stream order. model for limiting and lethal factors. The water quality Thus, the complete reproductive component, which consists component in Trial and Stanley's (1984) model consisted of of variables for depth, velocity, spawning temperature, in- water temperature, pH, turbidity, and n-dnimum oxygen. cubation temperature in winter, stream order, and dominant This component was not correlated with observed densities substrate (Trial and Stanley 1984), was not tested ade- for fry or parr (Trial 1989). Because the two temperature quately. variables, maximum and average temperature, were within The models tested by Trial (1989) did not include food the tolerance range for juvenile Atlantic salmon at all sites, availability because of the difficulty in sampling food the component index did not discriminate differences. How- abundance for an animal with opportunistic feeding habits. ever, temperature might profoundly affect biomass or A surrogate measure for food might be possible, based on growth rate, should these be used as end point measurements variables related to the productivity of food of salmonids, for testing comp such as alkalinity and conductivity (McFadden and Coo- onent indices. Growth ofjuvenile salmon is per 1962; Cooper and Scherer 1967). highly dependent on temperature (Egglishaw and Shackley 1985). Model Description Suitability Index Graphsfor Model The implicit assumption of HSI models is that habitat Variables with high HSI values has high carrying capacity and high productivity potential. These models were developed to The SI for each variable, as a function of the environ- predict the effects of environmental changes by relating mental range for that variable, is shown graphically in this environmental conditions to carrying capacity (U.S. Fish section. Habitat suitability indices can be computed with the and Wildlife Service 198 1). The aquatic and fish species HSI following SI's, which we modified from Trial and Stanley models provide a systematic method for evaluating projects (1994), based on new literature and Trial (1989). Trial that may alter the habitat of indicator species. (1989) discussed the assumptions associated with construct- ing the SI's, and assumptions are also discussed in sections Life Stage Component Indices below. The previously published model (Trial and Stanley 1984) Field use of these SI's requires measurements of key defined parr component suitability as the geometric mean of environmental variables. Methods for sampling habitat are the SI's for velocity, depth, and substrate. A geometric mean described in detail in Terrell et al. (1982), along with some increases the most when the individual variable with the shortcuts applicable to less rigorous studies. A multimillion lowest value is increased. In contrast, an arithmetic mean dollar project with great potential for widespread damage changes the same amount for a fixed amount of increase in might warrant a full-scale study with multiyear sampling. A a single variable, regardless of which variable is increased. local project with probable minimal impact might require We describe component models in which the quality of only a single visit to the site during summer base flow. Users holding or redd sites is based on multiplication of three must decide on the level of sampling required but should not variables (velocity, depth, and substrate) with values on a compromise on the methods recommended by Terrell et al. scale of 0 to 1.0. If the SI for a variable is considered to be (1982). Alternative methods forgathering data on individual aprobability ofhabitat utility, then the component suitability variables were discussed by Hamilton and Bergersen would be a product of the individual variable values. We (1984). As with the overall sampling plan, the method believe this "joint probability" approach for combining ve- selected to measure a variable may be dictated by the scope locity, depth, and substrate suitabilities is biologically the of the project. HABrrAT SurrAwLrry INDEX MODELS: NONNUGRAToRy FRESHwATER LiFE STArEs OF ATL_4NnC SALmON 9 Water Quality Component VI: Mean maximum daily water temperature for the warmest contiguous V2: Mean water temperature for the growing season or summer, 3-day period of summer during bass flow, preferably taken from a preferably taken with a hydrothermograph. continuous temperature record O.e., hydrothermograph). 1.0- -------------- 1.0 -------- X T SI (P X 13 is 1.0 20 1.0 31 0.5 - 25 0.25-0.75 8 0.25 D.5 30 0.0 14 1.0 IZ 40 0.0 U) 18 1.0 20 0.6 0.0 8 to 12 14 is IS 20 0.0 - Temperature CC) 15 20 25 30 35 40 Temperature CC) V4: Mean minimum daily oxygen saturation for the 3-day period with the lowest percent saturation during the summer, Ideally monitored continuously. 1.0 ------------ V& Mean turbidity based on monthly measurements over as much of the year as possible. X a) 1.0 ------------- % saturation at Al' 0.5 - 0 0.0 X 40 0 so '@O a) ft NTU I C0 100 1.0 0.5- 0 .0 40 .0 0.0 100 0.2 0 20 40 so so IDD CO) % Saturation 0.0 0 20 40 so 80 100 NTU V& Minimum pH -The frequency at which critical pH levels are reached, as measured during episodes of acid runoff over 3-day periods. 0.0 A B c D pH Category Description S.I. A pH below 4.0 at 0.05 least once annually B pH 4.0 to 5.5 at 0.40 least once annually I C pH occasionally 0.70 falls below 5.5 but never Wow 5. 0 D pH always 5.5 to 8.8 1.00 S I I 10 BIOLOGICAL SCIENCE REPORT 3 Fry Component If mean stream depth is greater than 50 cm, divide the stream into fourths. Because fry occur mostly in the shallower sections, average the variables for the two shallowest fourths of the section to arrive at a mean value for each SI of the fry component. In streams shallower than 50 cm, simply average the entire stream. V6: Mean column velocity for fry during base summer flow. VT Dominant substrate for fry. Measuring at a point 0.6 x total depth from the surface approximates mean column velocity. 1.0 ............ ............ ... ........ .... ....... . ........ . . ........ . .......... . ............ . ........ . 1.0- ------- ............ .......... X .... ... ...... .. ..... .. .... ..... X Velocity Sl ... ........ ........ .... ...... . . .. .... ...... ... ...... ... ........ .. 0 C: 0.5 - .......... . ........ .. .. ........ .. . .. 0.5 - 10 ........ .............. .. ..... .. 11-80 ......... ..... . ... .. . I...... .... ...... . Z 30 1.0 ......... I... ..... . ........ ... ........ 40 0.9 ........ ... ...... ... .. .. ........ .... ...... .. ......... .................. ..... ...... I., .. .....I..... ..... .. ..... ... .... ........ .... ... ........ ... ............. .. Cl) so 0.1 .. ........ ............. . .. ..... ... .... 1 0.0 ........... ...... .. .. ..... .... .. 0.0 .... :::::, 0.0 1* 1 1 1 0 10 30 40 8WO 1 0 1 2 3 4 5 Velocity (cm/s) Substrate Substrate Size Code Type (MM) sl 1 Fines < 0.5 0.1 1 2 Sand 0.5- 2.2 0.5 3 Pebble-gravel > 2.2 - 22.2 1.0 4 Cobble > 22.2 - 256 0.8 i 5 Boulder > 256 0.1 VS: Mean depth for fry during base summer flow. 1.0 ------ Depth Sl 0 0.0 10 1.0 X 13 40 1.0 70 0.2 0.5- 90 0.0 100 0.0 0.0 0 10 20 30 40 50 60 70 80 90 100 Depth (cm) HABrrAT SurFABu-rry INDEX MODELS: NONMIGRATORY FREsHwATER LEFE STAGES OF ATLANTIC SALMON I I Parr Component If mean stream depth is over 50 cm, divide the stream into fourths, and average the variables in the two deepest fourths to arrive at the mean value for each SI. In streams shallower than 50 cm, use the mean values for the entire stream. Vg: Mean column velocity for parr during bass summer flows. V10: Dominant substrate for parr. 1.0- ----------- ...... ..... ........... x x Va ........ ... Velocity Sl C ........ ..... 0.5 0 0.8 0.5- ............ . 0 L"00 I. ........... .0 ........... so ... ..... 0.5 ......... ....... ............ loo 0.1 ..... ... ... ........... I ..... - .... ... ... ..... ..... ... .. . ... . ..... ............ .......... ........ ... .. . .... ... .. O.D ...... ......... . ...... .... 0 10 20 30 40 @0 60 @0 @0 @0 1@0 0.0 ... ... Velocity (cm/s) 2 3 4 5 Substrate Substrate Size Code Type (MM) sl 1 Fines < 0.5 0.0 2 and 0.5-2.2 0.2 3 Pebbl&gravel > 2.2 - 22.2 0.7 4 Cobble > 22.2 - 256 1.0 5 Boulder > 256 0.3 V1 1: Mean depth for parr during base summer flows. 1.0 ---------- Depth Sl 0 0.0 x 20 1.0 so 1.0 70 0.2 0.5 - 90 100 00.'00 0.0 0 2 0 @0 60 810 100 Depth (cm) 12 BIOLOGICAL SCIENCE REPORT 3 Reproductive Component Evaluate at the head or tail of pools only if the substrate material is > 2.2 to 256 mm in diameter and water is at least 15 cm deep. The best time to conduct the field work would be in the fall, when Atlantic salmon are selecting spawning areas. Otherwise, attempt to estimate fall conditions by historical information on seasonal variation. V1 2: Mean depth for reproduction at spawning time. V1 3: Mean column velocity for reproduction during tail, or at flow conditions approArnating those occurring during fall. 1.0 ------- Depth 1.0- 0 D.0 % a 0.0 a) 10 0.5 20 1.0 x 40 1Z Q) 90 0.0 velocity @l 0.5 100 0.0 S 0 0.0 015 - 25 0.0 :3 .0 35 0.5 Cl) .9 00 1.0 :3 80 U) ':o 90 16 0.0 0 1 @0 @0 @0 so 100 100 0.0 Depth (cm) 0.0 0 20 @O 60 80 100 Velocity (cm/s) V14: Spawning temperature - If water temperature reaches then declines V15: Embryo Incubation temperature - Average meAmum below 112T In late October and early November, SI=1.0. Spawning daily temperature for the warmest 2-day perlod will follow the date that water temperature reaches and maintains a betmen November 1 Sand May 1, prefilrably taken temperature between 12* and rC. with a hydrothermograph left In the stream over winter. 1.0 -------- *C sl 1.0 ------------- 0.0 0.0 x x a) 2.6 0.0 0) T sl 13 7.0 1.0 0 0.0 12.0 1.0 0.8 0.5 - 11 0.5- >12.0 0.0 a 1.0 .93 7 1.0 :3 9 0.7 co U) I 10.1 tI 12 0.0 0.0 0.0 11 0 5 10 15 20 0 2 4 a8 10 12 Temperature CC) Temperature CC) V16: SUeam order, based on stream branches having V1 7: Dominant substrate for spawning and permanent water flow. embryo Incubation. ........... ........... 1.0- 1.0 .. .. . ..... ..... . ........... . . ......... ... . ... .... ... ..... x .......... .......... . ..... . ... .. .. ...... .. 0,5 ... ... 0.5 - .......... . .......... C0 ......... ... ......... ... .......... 0.0 ...... ......... . 1 2 3 4 Stream Order 1 2 3 Stream substrate substrale Size Order SI Code Type (MM) 81 1 0.5 1 Flnee @ 0.5 0.00 2 0.8 2 Sand 0.5. 2.2 0.00 3 Pebble-gravel > 22 - 22.2 0.95 3 1.0 4 cobble > 7.2 - 25a 1.00 4 1.0 s Boulder 258 0.00 HABrrAT SurrAffliriy INDEX MODELS: NoNmiGRAToRY FRESHWATER LWE STAGES OF ATLANTIC SALMON 13 HSI Determination fry and parr had suitability of 1.0 for a more restricted range Water Quality Component of velocities than SI's for mean column velocity. Similarly, CWQ = lowest of V I, V2, V3, V4, or V5 Rimmer et al. (1984) found that velocities selected by fry Fry Component and parr were lower than velocities in the overlying water. CFry = V6 x V7 x V8 We recommend measuring column velocity at 0.6 of the Parr Component depth (i.e., at a point 60% of the way from the surface of the CParr = V9 x VIO x V1 I water column to the bed of the stream). Reproductive Component Field Application of the Models CR = lowest of V14, V15, V16, or V17 Habitat Suitability Index For the mean maximum daily temperature for the HSI = (CWQ x CFry x CParr x CR) warmest contiguous 3-day period of summer (VI), the Some environmental situations might alter the appropri- SI was based on maximum summer temperatures of ateness of the HSI approach. For example, the presence of streams with Atlantic salmon populations and upper aluminum in combination with high acidity may affect incipient lethal temperatures. The significance of maxi- survival. Users may modify SI's based on local conditions, mum temperature is confounded becausejuvenile Atlan- carefully documenting the rationale for the changes. Like- tic salmon avoid high temperatures. The 3-day period of wise, the structure of the component and final HSI model exposure corresponds to lethal exposure periods of some may be modified. laboratory experiments. We recommend that hydrother- mographs be placed in streams being evaluated. If by- Rationale and Assumptionsfor drothermographs are not available, visit the stream dur- Suitability Indices (SIs) ing periods of hot weather to get daily temperatures. Embryo incubation temperature (V15) was based on The SI's for Atlantic salmon represent the relation be- the acute lethal temperatures for Atlantic salmon eggs tween habitat variables and the population density of fry and and embryos. The period of exposure reported in the parr in freshwater streams; population density is assumed to literature varied from 24 to 72 h. Thus, some considera- indicate stream productivity. Another assumption is that tion of the duration of lethal temperature was included animals select suitable habitats, when available, that en- in the variable label. The ideal way to measure this hance individual survival. Optimal habitat has a suitability variable would be to place hydrothermographs in or near value of 1.0. Where optimal habitat is unavailable or already the redds over winter. Without the equipment, the user fully occupied, the number of survivors in suboptimal habi- may have to measure stream temperatures, guessing at tat indicates habitat suitability (with a value less than 1.0). the times, based on episodes of warm weather (probably Suitability indices are proportional to population density, in spring). The mean temperature (V2) should be for the which is determined by distribution and survival of the growing season or summer-again we recommend a species. The mean value of a habitat variable indicates hydrothermograph. In addition, remember that V14 is habitat suitability for the area where the measurement was the temperature during spawning and that if water tem- taken. peratures reach, then decline below 12' C in late October We assumed that contaminants are absent. If contami- and early November, spawning will follow when tem- nants occur in the stream and their effects can be docu- peratures are between 12' and 7' C. mented, then the user might assign an HSI of zero. Calculate mean turbidity (V3) by month over as much Alternatively, the HSI might be calculated as if no con- of the year as possible. High turbidity during freshets is taminant were present, with a qualification added that not as important as chronic exposure, which has a sus- this is the value if pollution remediation were imple- tained inhibiting effect on feeding. The months during the mented. growing season are most important. If available turbidity While we recognize that competition between species data are expressed as concentrations of suspended sedi- and between life stages might be important, competition ments (C) in mg1L, convert to NTU units with the formula is not modeled. The HSI model is based on physical factors NTU = 10 + 0. 178 C (Sigler et al. 1984). only. The HSI value is stated as if competition were absent. Mean minimum oxygen saturation (V4) was based on The velocity that Trial (1989) reported was measured acute tolerances of Atlantic salmon to low dissolved oxygen where fish were actually located, which was usually less concentrations at several temperatures and the average con- than the average column velocity. Morantz et al. (1987) and ditions of streams that have populations. If low percent DeGraff and Bain (1986) measured mean velocity of the saturation of oxygen persisted for no more than I day, then water column and the velocity selected by each individual. the suitability of the habitat would be higher than if oxygen In both papers, the SI's developed for velocities selected by was low for an extended period. We recommend taking daily 14 BIOLOGICAL SCIENCE REPORT 3 measurements in the early morning on a cloudy day during The HSI should be treated as a linear index to the carrying the warmest summer periods. A recording oxygen probe in capacity of the particular habitat-the higher the HSI value, the stream would be ideal. the more fish the habitat should be able to support. However, Minimum pH (V5) was based on the acute lethal pH for only physical and chemical characteristics of a habitat are Atlantic salmon eggs and embryos. The period of exposure considered. Biological interactions and the effect of barriers reported in the literature varied from 24 to 72 h. Thus, some are ignored. Predation, human harvest, competition, and consideration of the duration of lethal pH was included in nutrition are not considered, even though they would obvi- the variable label. In addition, V5 includes consideration of ously affect survival, density, or carrying capacity. For frequency of low pH events. Other variables, such as alurni- Atlantic salmon, habitat is often uninhabited because dams num, affect survival to acid exposure-users are encouraged block access. Nevertheless, the model could assign a high to modify the SI if appropriate data are available for specific HSI, indicating the potential to support salmon spawning if sites. there were no barriers. The area and time for calculating the mean for several of All models represent simplifications of complex sys- the variables were stated in Terrell et a]. (1982), but we will tems. The purpose of HSI models is to help predict the repeat them to avoid confusion. We recommend that meas- responses of key species to development projects or urements be made at five transects, 10 in apart, across the management practices. These models may be useful for stream. Establish the position of the first transect at random, presenting simple alternatives to managers so that ra- such as at the position where a ball lands. Measure current, tional decisions can be made with data gathered at mod- depth, and substrate at I-m intervals across the stream, or at est costs. The HSI model approach has been widely 0.25, 0.5, and 0.75 of the width if streams are less than 2 in accepted because it provides a rational approach for wide. For spawning riffles, take a measurement at the head, evaluation of habitat that does not depend solely on the tail, and center of the riffle. Bottom substrate is calculated opinion of experts. We prefer that our model be used by by summing the linear amount of each type. experts as an aid to systematically applying their knowl- Lack of suitable water quality at a site or in a river edge to complex problems. system can limit the distribution of the species. The disappearance of Atlantic salmon from European and Sources of Additional Models Canadian rivers coincident with decreased pH is evi- dence of a limiting factor (Haines 1981; Watt et al. The National Biological Service, Midcontinent Eco- 1983). Thus, the water quality component was included logical Science Center, in Fort Collins, Colorado, main- in the HSI as potentially limiting. However, habitats for tains a library of SI's for use with the Instrearn Flow all life stages are not necessary at each site for reproduc- Incremental Methodology. ing populations because interspersion of habitats may The Canadian Department of Fisheries and Oceans, provide for all of the species' needs. The Atlantic St. John's, Newfoundland, has developed SI's to evalu- salmon's life cycle requires that some habitats exist ate habitat for the various freshwater life stages of At- within the drainage for each riverine life stage. The lantic salmon in Newfoundland (Scruton and Gibson model user must be aware of the mix of different quali- 1993), and a workshop was held in 1992 to develop a ties of life stage habitats within the study area and model. The SI's were based on data on the density of drainage. Calculating a species HSI that combines all life juvenile Atlantic salmon and on the habitat from 242 stage components into one index obscures information stations on 18 rivers on the island of Newfoundland. but may be needed in an assessment. Thus, the model Suitability indices consistently predicted densities of fry includes component indices for water quality, reproduc- based on important habitat variables. Suitability indices tion, fry, and parr and a formula for the species HSI. were most useful if based on macrohabitat measurements - of stream dimensions and characteristics rather than on Interpreting Model Outputs the microhabitat for the location of individual fish. The U.S. Army Corps of Engineers has developed an There are numerous possible applications of the HSI Aquatic Habitat Appraisal Guide, based on the Habitat model for Atlantic salmon. Potential users might differ Evaluation Procedures, using 16 habitat variables and HSI widely in their understanding of the premises and limi- scores. tations of the model. Only the most naive would take it Two other National Biological Service ecological 11 off the shelf," make a few measurements in a target science centers (located in Leetown, West Virginia, and habitat, and attempt to make far-reaching conclusions. Columbia, Missouri) are evaluating habitat requirements Experts on life history and biology should be able to use for adult Atlantic salmon and developing an improved the model in the more realistic context described below. approach for evaluating water quality requirements. HABrrAT SurrABiLrry INDEX MODELs: NONMIGRATORY FREsHwATER LiFE STAGES OF ATLANuc SALMON 15 Acknowledgments Curijak, R. A. 1988. Behavior and microhabitat of young Atlantic salmon (Salmo salar) during winter. Canadian Journal of We thank James W. Terrell, Jeanette Carpenter, and Fisheries and Aquatic Sciences 45:2156-2160. Henry E. Booke for reviewing earlier drafts of this manu- Cunjak, R. A. 1992. Comparative feeding, growth and move- script. Jeanette Carpenter produced the final SI graphs. ments of Atlantic salmon (Salmo salar) parr from fiverine and estuarine environments. Ecology of Freshwater Fish 1:26-34. Danie, D. S., J. G. Trial, and J. G. Stanley. 1984. Species profiles: Life histories and environmental requirements of coastal fish Cited References and invertebrates (North Atlantic)-Atlantic salmon. U.S. Fish and Wildlife Service FWS/OBS-82/11.22. U.S. Army Amiro, P. G. 1993. Habitat measurement and population estima- Corps of Engineers TR EL-82-4. 19 pp. tion of juvenile Atlantic salmon (Salmo salar). Pages 81-97 Daye, P. G., and E. T. Garside. 1977. Lower lethal levels of pH in R. J. Gibson and R. E. Cutting, editors. Production of for embryos and alevins of Atlantic salmon, Salmo salar L. juvenile Atlantic salmon, Salmo salar, in natural waters. Canadian Journal of Zoology 55:1504---1508. Canadian Special Publication in Fisheries and Aquatic Sci- Daye, P. G., and E. T. Garside. 1980. Development, survival, and ences Number 118. structural alterations of embryos and alevins of Atlantic Bain, M., and C. L. Robinson. 1988. Structure, performance, and salmon, Salmo salar L., continuously exposed to alkaline assumptions of riverine Habitat Suitability Index models. levels of pH from fertilization. Canadian Journal of Zoology Alabama Cooperative Fisheries and Wildlife Research Unit. 58:369-377. Aquatic Resources Research Series 88-3. Auburn, Ala. 20 pp. DeCola, J. N. 1970. Water quality requirements for Atlantic salmon. U.S. Department of the Interior Federal Water Qual- Beland, K. F., R. M. Jordan, and A. L. Meister. 1982. Water depth ity Administration, Northeast Region, Boston, Mass. 42 pp. and velocity preferences of spawning Atlantic salmon in DeCola, J. N. 1975. Atlantic salmon restoration and the question Maine rivers. North American Journal of Fisheries Manage- of water quality. International Atlantic Salmon Foundation ment 2:11-13. Special Publication Series 6:24-28. Berg, L., and T. G. Northcote. 1985. Changes in territorial, DeGraff, D. A., and L. H. Bain. 1986. Habitat use by and gill-flaring, and feeding behavior in juvenile coho salmon preferences ofjuvenile Atlantic salmon in two Newfoundland (Oncorhynchus kisutch) following short-term pulses of sus- rivers. Transactions of the American Fisheries Society pended sediment. Journal of the Fisheries Research Board of 115:671-681. Canada 42:1410-1417. Dwyer, W. P., and R. G. Piper. 1987. Atlantic salmon growth Bisson, P. A., and R. E. Bilby. 1982. Avoidance of suspended efficiency as affected by temperature. Progressive Fish-Cul- sediment by juvenile coho salmon. North American Journal turist 49:57-59. of Fisheries Management 2:371-374. Egglishaw, H. J., and P. E. Shackley. 1985. Factors governing Bley, P. W. 1981. Age, growth, and mortality ofjuvenile Atlantic the production of juvenile Atlantic salmon in Scottish salmon in streams: A review. U.S. Fish and Wildlife Service streams. Journal of Fish Biology 27A:27-33. Biological Report 87(4). 25 pp. Einarsson, S. M., D. H. Mills, and V. Johannsson. 1990. Utiliza- Brett, J. R. 1979. Environmental factors and growth. Pages tion of fluvial and lacustrine habitat by anadromous Atlantic 599-675 in W. S. Hoar, D. J. Randall and J.R. Brett, editors. salmon, Salmo salar L., in an Iceland watershed. Fisheries Fish physiology, Volume VIII. Bioenergetics and growth. Research 10:53-71. Academic Press, New York. Elliott, J. M. 1991. Tolerance and resistance to thermal stress in Chadwick, E. M. P. 1982. Stock-recruitment relationship for juvenile Atlantic salmon, Salmo salar. Freshwater Biology Atlantic salmon (Salmo salar) in Newfoundland rivers. Cana- 25:61-70. dian Journal of Fisheries and Aquatic Sciences 39:1496-1501. Elson, P. F. 1975. Atlantic salmon rivers, smolt production and Chadwick, E. M. P. 1985. Fundamental research problems in the optimal spawning: An overview of natural production. Inter- management of Atlantic salmon, Salmo salar L., in Atlantic national Atlantic Salmon Foundation Special Publication Se- Canada. Journal of Fish Biology 27A:9-25. ries 6:96-119. Cooper, E. L., and R. C. Scherer. 1967. Annual production of Farmer, G. J., R. L. Saunders, T. R. Goff, C. E. Johnston, and brook trout (Salvelinus fontinalis) is fertile and infertile E. B. Henderson. 1989. Some physiological responses of streams of Pennsylvania. Pennsylvania Academy of Science Atlantic salmon (Salmo salar) exposed to soft, acidic water Proceedings 41:65-70. during smolting. Aquaculture 82:229-244. Crisp, D. T. 1991. Stream channel experiments on downstream Francis, A. A. 1980. Densities of juvenile Atlantic salmon and movement of recently emerged trout, Salmo truna L., and other species, and related data from electroseining studies in salmon, S. salar L@111. Effects of developmental stage and the Saint John River system, 1968-78. Canadian Data Report day and night upon dispersal. Journal of Fish Biology on Fisheries and Aquatic Sciences Number 178. 95 pp. 39:371-381. Frenette, M., M. Caron, P. Julien, and R. J. Gibson. 1984. Inter- Crisp, D. T., and P. A. Carling. 1989. Observations on siting, action entre le d6bit et les populations de tacons (Sahno salar) dimensions and structure of salmonid redds. Journal of Fish de la rivi&e Matamec, Qu6bec. Canadian Journal of Fisheries Biology 34:119-134. and Aquatic Sciences 41:954-963. 16 BIOLOGICAL SCIENCE REPORT 3 Fry, E. F. J. 1971. The effect of environmental factors on the parr: Effects of predation risk, season and life history strategy. physiology of fish. Pages 1-98 in W. S. Hoar and D. J. Ran- Journal of Fish Biology 33:917-924. dall, editors. Fish physiology. Vol. 6. Environmental relations Huntsman, A. G. 1942. Death of salmon and trout with high and behavior. Academic Press, New York. temperature. Journal ofthe Fisheries Research Board ofCan- Garside, E. T. 1973. Ultimate upper lethal temperature of Atlantic ada 5:485-501. salmon, Salmo salar. Canadian Journal Zoology 51:898-900. Jensen, A. J., and B. 0. Johnsen. 1986. Different adaption Gibson, R. J. 1993. The Atlantic salmon in fresh water: Spawn- strategies of Atlantic salmon (Salmo salar) populations to ing, rearing and production. Reviews in Fish Biology and extreme climates with special reference to some cold Norwe- Fisheries 3:39-73. gian rivers. Canadian Journal of Fisheries and Aquatic Sci- Gibson, R. J., and R. A. Myers. 1988. Influence of seasonal river ences 43:980-984. discharge on survival of juvenile Atlantic salmon, Salmo Jordan, R. M., and K. F. Beland. 198 1. Atlantic salmon spawning salar. Canadian Journal of Fisheries and Aquatic Sciences survey and evaluation of spawning success. Atlantic Sea-Run 45:344-348. Salmon Commission, Augusta, Maine. AFS-20-R. 25 pp. Gibson, R. J., D. E. Stansbury, R. R. Whalen, and K. G. Hiffier. Kane, T. R. 1989. Atlantic salmon brood stock records. Pages 1993. Relative habitat use, and inter-specific and intra-spe- 30-31 in H. L. Kincaid and J. G. Stanley, editors. Atlantic cific competition of brook trout (Salvelinus fontinalis) and salmon brood stock management and breeding handbook. juvenile Atlantic salmon (Salmo salar) in some Newfound- U.S. Fish and Wildlife Service Biological Report 89(12). land rivers. Pages 53-69 in R. J. Gibson and R. E. Cutting, Kennedy, G. J. A., and C. D. Strange. 1986. The effects of intra- editors. Production ofjuvenile Atlantic salmon, Salmo salar, and inter-specific competition on the distribution of stocked in natural waters. Canadian Special Publication in Fisheries juvenile Atlantic salmon, Salmo salar L., in relation to depth and Aquatic Sciences Number 118. and gradient in an upland trout, Salmo trutta L., stream. Gustafson-Marjanen, K. 1. 1982. Atlantic salmon (Salmo salar Journal of Fish Biology 29:199-214. L.) fry emergence: Success, timing, distribution. M.S. thesis, Knight, A. E., G. Marancik, and J. C. Greenwood. 198 1. Atlantic University of Maine, Orono. 72 pp. salmon production potential of the Mad River, New Hamp- Gustafson-Greenwood, K. I., and J. R. Moring. 1990. Territory shire 1975-1980. U.S. Fish and Wildlife Service, Laconia, size and distribution of newly-emerged Atlantic salmon N.H. 14 pp. (Salmo salar). Hydrobiologia 206:125-13 1. Lacroix, G. L., and D. R. Townsend. 1987. Responses ofjuvenile Gustafson-lvla@anen, K. A., and J. R. Moring. 1984. Construc- Atlantic salmon (Salmo salar) to episodic increases in acidity tion of artificial redds for evaluating survival of Atlantic of Nova Scotia rivers. Canadian Journal of Fisheries and salmon eggs and alevins. North American Journal ofFisherties Aquatic Sciences 44:1475-1484. Management 4:455-456. Leim, A. H., and W. B. Scott. 1966. Fishes of the Atlantic coast Haines, T. A. 198 1. Acidic precipitation and its consequences for of Canada. Bulletin of the Fisheries Research Board of Can- aquatic ecosystems: A review. Transactions of the American ada Number 155. 485 pp. Fisheries Society 110:669-707. Lear, W. H. 1993. The management of Canadian Atlantic salmon Hamilton, K., and E. P. Bergersen. 1985. Methods to estimate Fisheries. Pages 151-176 in L. S. Parsons and W. H. Lear, aquatic habitat variables. Colorado State University, Ft. Col- editors. Perspectives on Canadian marine Fisheries manage- lins. 260 pp. ment. National Research Council of Canada, Ottawa, Ont. Heggenes, J., and R. Borgstrom. 199 1. Effect of habitat types on MacKenzie, C., and J. R. Moring. 1988. Estimating survival of survival, spatial distribution and production of an allopatric Atlantic salmon during the intragravel period. North Ameri- cohort of Atlantic salmon, Salmo salar L., under conditions can Journal of Fisheries Management 8:45-49. of low competition. Journal of Fish Biology 38:267-280. McCrimmon, H. R. 1954. Stream studies on planted Atlantic Heggenes, J., A. Brabrand, and S. J. Saltveit. 1990. Comparison salmon. Journal of the Fisheries Research Board of Canada of three methods for studies of stream habitat use by young 11:362-403. brown trout and Atlantic salmon. Transactions of the Ameri- can Fisheries Society 119: 101-111. McFadden, J. T., and E. L. Cooper. 1962. An ecological com- Heggenes, J., A. Brabrand, and S. J. Saltveit. 1991. Microhabitat parison of six populations of brown trout (Salmo trutta). use by brown trout, Salmo trutta L. and Atlantic salmon, Transactions of the American Fisheries Society 91:53-62. S. salar L., in a stream: A comparative study of underwa- Mills, D. 1989. Ecology and management of Atlantic salmon. ter and river bank observations. Journal of Fish Biology Chapman and Hull, London and New York. 351 pp. 38:259-266. Morantz, D. L., R. K. Sweeney, C. S. Shirvell, and D. A. Lon- Heggenes, J., and T. Traaen. 1988. Downstream migration and gard. 1987. Selection of microhabitat in summer by juvenile critical water velocities in stream channels for fry of four Atlantic salmon (Salmo salar). Canadian Journal ofFisheries salmonid species. Journal of Fish Biology 32:717-727. and Aquatic Sciences 44:120-129. Hesthagen, T. 1988. Movements of brown trout, Salmo trutta, Netboy, A. 1974. The salmon: Their fight for survival. andjuvenfle Atlantic salmon, Salmo salar, in a coastal stream Houghton-Mifflin Company. Boston, Mass. 594 pp. in northern Norway. Journal of Fish Biology 32:639-653. Newcombe, C. P., and D. D. MacDonald. 1991. Effects of Huntingford, F. A., N. B. Metcalfe, and J. E. Thorpe. 1988. suspended sediments on aquatic ecosystems. North American Choice of feeding station in Atlantic salmon, Salmo salar, Journal of Fisheries Management 11: 72-82. HABrrAT SUrrABiLrry INDEX MODELS: NoNbfflGRAToRy FRESHWATER LiFE STAGES OF ATLANTIC SALMON 17 Noffgren, L., and E. Degerman. 1993. Effects of different water parr in streams. Journal of the Fisheries Research Board of qualities on the early development of Atlantic salmon and Canada 36:1408-1412. brown trout exposed in situ. Ambio 22:213-218. Symons, P. E. K., and M. Heland. 1978. Stream habitats and Peterson, R. H. 1978. Physical characteristics of Atlantic salmon behavioral interactions of underyearling and yearling Atlantic spawning gravel in some New Brunswick, Canada, streams. salmon (Salmo salar). Journal of the Fisheries Research Canadian Fisheries and Marine Service Technical Report Board of Canada 35:175-183. Number 785. iv + 28 pp. Terrell, J. W., T. E. McMahon, P. D. Inskip, R. F. Raleigh, and Peterson, R. H., P. G. Daye, and J. L. Metcalfe. 1980. Inhibition K. L. Williamson. 1982. Habitat Suitability Index models: of Atlantic salmon (Salmo salar) hatching at low pH. Cana- Appendix A. Guidelines for riverine and lacustrine applica- than Journal of Fisheries and Aquatic Sciences 37:770-774. tions of fish HSI models with the Habitat Evaluation Proce- Peterson, R. H., and J. L. Metcalfe. 1979. Responses of Atlantic dures. U.S. Fish and Wildlife Service FWS/OBS-82/10.A. salmon alevins to temperature gradients. Canadian Journal of 54 pp. Zoology 57:1424-1430. Power, G. 1969. The salmon of Ungava Bay. Arctic Institute of Thompson, D. 1993. Status of the Atlantic salmon, Salmo North America Technical Report Number 22. Calgary, Al- salar L., its distribution and the threats to natural populations. berta. 72 pp. Pages 303-306 in J.G. Cloud and G.H. Thorgand, editors. Rideout, S. 1989. History of the Atlantic salmon restoration Genetic conservation of salmonid fishes. Plenum Press, New program. Pages 1-4 in H. L. Kincaid and J. G. Stanley, edi- York and London. tors. Atlantic salmon brood stock management and breeding Thorpe, J. E. 1988. Salmon enhancement: Stock discreetness and handbook. U.S. Fish and Wildlife Service Biological Report choice of material for stocking. Pages 373-388 in D. Mills 89(12). and D. Piggins, editors. Atlantic salmon, planning for the Riley, S. C., G. Power, and P. E. Ilissen. 1989. Meristic and future. Timber Press, Portland, Oreg. morphometric variation in parr of ouananiche and anadro- Tremblay, G., F. Caron, R. Verdon, and M. Lessard, 1993. Influ- mous Atlantic salmon from rivers along the north shore of the ence des parametres hydromorphologiques sur l'utilisation de Gulf of St. Lawrence. Transactions of the American Fisheries ]'habitat par les juveniles du Saumon atlantique (Salmo Society 118:515-522. salar). Pages 127-137 in R. J. Gibson and R. E. Cutting, Rimmer, D. M., U. Paim, and R. L. Saunders. 1984. Changes in editors. Production ofjuvenile Atlantic salmon, Salmo salar, the selection of microhabitat by juvenile Atlantic salmon in natural waters. Canadian Special Publication in Fisheries (Salmo salar) at summer-autumn transition in a small and Aquatic Sciences Number 118. river. Canadian Journal of Fisheries and Aquatic Sciences Trial, J. G. 1989. Testing habitat models for blacknose dace and 41:469-475. Atlantic salmon. Ph.D. thesis, University of Maine, Orono. Sayers, R. E., Jr. 1990. Habitat use patterns of native brook trout 129 pp. and stocked Atlantic salmon: Inter-specific competition and salmon restoration. Ph.D. thesis, University of Maine, Orono. Trial, J. G., and J. G. Stanley. 1984. Calibrating effects of acidity 125 pp. on Atlantic salmon for use in habitat suitability models. Schaffer, W. M., and P. F. Elson. 1975. The adaptive significance Completion Report A-054-ME. Land and Water Resources of variations in life history among local populations of Adan- Center, University of Maine, Orono. 37 pp. tic salmon in North America. Ecology 56:577-590. Trial, J. G., C. S. Wade, and J. G. Stanley. 1984. HSI models for Scott, W. B., and E. J. Crossman. 1973. Freshwater fishes of northeastern fishes. Pages 17-56 in J, W. Terrell, editor. Pro- Canada. Bulletin of the Fisheries Research Board of Canada ceedings of a workshop on fish Habitat Suitability Index Number 184. iv + 996 pp. models. U.S. Fish and Wildlife Service Biological Report Scruton, D. A., and R. J. Gibson. 1993. The development of 85(6). habitat suitability curves for juvenile Atlantic salmon (Salmo U.S. Fish and Wildlife Service. 198 1. Standards for the develop- salar) in riverine habitat in insular Newfoundland, Canada. mentofHabitat Suitability Index models. 103 ESM. U.S. Fish Pages 149-161 in R. J. Gibson and R. E. Cutting, editors. and Wildlife Service, Division of Ecological Services, Wash- Production ofjuvenile Atlantic salmon, Salmo salar, in natu- ington, D.C. 168 pp. ral waters. Canadian Special Publication in Fisheries and Warner, K. 1963. Natural spawning success of landlocked Aquatic Sciences Number 118. salmon, Salmo salar. Transactions of the American Fisheries Siginevich, G. P. 1967. Nature of the relationship between in- Society 92:161-164. crease in size of Baltic salmon fry and the water temperature. Warner, K., and K. A. Havey. 1985. Life history, ecology and Gidrobiologicheskii Zhurnal 3:43-48. Fisheries Research management of Maine landlocked salmon (Salmo salar). Board of Canada Translation Series Number 952. 14 pp. Maine Department of Inland Fisheries and Wildlife. Augusta. Sigler, J. W., T. C. Bjornn, and F. H. Everest. 1984. Effects of 127 pp. chronic turbidity on density and growth of steelbead and coho salmon. Transactions of the American Fisheries Societ Watt, W. D., C. D. Scott, and W. J. White. 1983. Evidence of y 113:142-150. acidification of some Nova Scotian rivers and its impact on Sosiak, A. J., R. G. Randall, and J. A. McKenzie. 1979. Feeding Atlantic salmon, Salmo salar. Journal of the Fisheries Re- by hatchery-reared and wild Atlantic salmon (Salmo salar) search Board of Canada 40:462-473. 18 BIOLOGICAL SaENcE REPORT 3 Appendix. Model Evaluation Form The habitat suitability index (HSI) model for juvenile Atlantic salmon is intended for use in the habitat evaluation procedures (HEP) developed by the U.S. Fish and Wildlife Service. This model for nonmigratory freshwater stages of Atlantic salmon is the third generation of a model that was developed originally from a review and synthesis of existing information on Atlantic salmon. The model was modified based on field testing in Maine in 1984 and further evaluated by comparison of alternative model outputs with a long-term data base from Canada and habitat selection data gathered in Maine. Despite the testing of this HSI model, further improvement and revision could result in an even better and more useful model. Please complete this form following application or review of the model. Feel free to include additional information that may be of use to either a model developer or model user. We also would appreciate information on model testing, modification, and application, as well as copies of modified models or test results. Please return this form to Landscape and Habitat Analysis Section National Biological Service Midcontinent Ecological Science Center 4512 McMurry Avenue Fort Collins, CO 80525-3400 Thank you for your assistance. HABITAT SurrABILnY INDEX MODELS: NONMIGRATORY FRESHWATER LiFE STAGES OF ATLANTIC SALMON 19 Species Atlantic salmon Location Habitat or cover type(s) Baseline- Other Variables measured or evaluated I was the species information useful and accurate? Yes No If not, what coffections or improvements are needed? Were the variables and curves clearly defined and useful? Yes No If not, how could they be improved? Were the techniques suggested for collection of field data: Appropriate? Yes - No Clearly defined? Yes - No Easily applied? Yes - No If not, what other data collection techniques are needed? 1 Were the model equations logical? Yes - No Appropriate? Yes No How could they be improved? Other suggestions for modification or improvement (attach curves, equations, graphs, or other appropriate information) Additional references or information that should be included in the model: Model evaluator or reviewer Date Agency Address Telephone Number: Comm: -Fax: A list of current Biological Science Report@ follows: 1. Reproduction'and Distribution of Bald Eagles in.Voyageurs National Park, Minnesota, 1973-1993, by Leland H. Grim and Larry W. Kallemeyn. 1995. 29 pp. 2. Evaluations of Duck Habitat and Estimation-of Duck Population Sizes with a Remote-Sensing-Based Systein, by Lewis A Cowardin, Terry L. Shaffer, and Phillip M. Arnold.. 1995. 26 pp. NOTE: The mention of trade names does not conititute endorsement or recommindation for use by the Federal GoXemment. U.S. Department of the Interior National Biological Service As the Nation's principal congervation agency, the Department of the Inierior has re- sponsibility for most of our nationally owned, public lands and natural resource5. This respon- sibility includes fostering the sound use of our@ lands and water resources; protecting our fish, wildlife, and biological diversity; preserving the @:-environmental and cultural values of our na- iional parks and historical places; and providing for the enjoyment of life through outdoor rec- reation. The Department assesses our energy and mineral resources and works to ensure that their development is in the best interests of all our people by encouraging stewardship and citi- zen participation in their care. The Department also has- a'major respons@bifity for American Indian reservation communities. 3 6668 00003 9687