Response of a Brook Trout Population and Instream Habitat to a Catastrophic Flood and Debris Flow

—In June 1995, a massive flood and debris flow impacted fish and habitat along the lower 1.9 km of the Staunton River, a headwater stream located in Shenandoah National Park, Virginia. In the area affected by debris flow, the stream bed was scoured and new substrate materials were deposited, trees were removed from a 30-m-wide band in the riparian area, and all fish were eliminated. In the area that was unaffected by debris flow, habitat was moderately altered by the flood and fish populations persisted at decreased densities. Basinwide fish population and habitat surveys provided data to compare (1) the preand postevent population densities of brook trout Salvelinus fontinalis and instream habitat conditions and (2) postevent population density, brook trout growth, and instream habitat in the debris-flow-affected and unaffected areas. By June 1998, brook trout had recolonized the entire debris-flow-affected area, and population density exceeded preevent levels. Brook trout growth was significantly greater in the debris-flow-affected area than in the unaffected area through fall 1998, but it was not significantly greater in 1999. Population density appeared to have a negative influence on fish growth. A 1995 postevent habitat survey revealed increases in the number of pools and riffles and substrate size and decreases in pool and riffle surface area and depth. By 1999, the total number, surface area, and depth of pools and riffles had returned to near preevent levels and substrate size had decreased. Between 1995 and 1999, the amount of large woody debris increased in the debris-flow-unaffected area, where riparian trees had remained intact, and decreased in the affected area, where riparian trees had been eliminated. A number of factors, including a relatively intact watershed and nearby source populations, allowed the Staunton River to quickly respond to this dramatic natural event. Given the proper conditions for recovery, such events are less catastrophic than activities that lead to chronic stream degradation. Streams occupied by salmonid populations are subject to natural disturbances that lead to changes in instream habitat and fluctuations in population abundance. These disturbances range from moderate and predictable events over seasonal or annual intervals, to more extreme, so-called catastrophic events, with return intervals measured in decades or centuries. Both moderate and extreme * Corresponding author: [email protected] 1 Present address: U.S. Forest Service, Southern Research Station Center for Aquatic Technology Transfer, 1650 Ramble Road, Blacksburg, Virginia 24060, USA. 2 Present address: U.S. Fish and Wildlife Service, 17629 El Camino Real, Suite 211, Houston, Texas 77058, USA. Received June 1, 2001; accepted January 14, 2002 events play important roles in shaping lotic ecosystems (Sousa 1984; Reice et al. 1990). In the Appalachian Mountains, extreme flooding can alter both instream habitat and fish populations and produce more localized reaches of mass wasting where the stream channel and riparian areas may be drastically altered (Hack and Goodlet 1960). While changes brought about by mass wasting events can appear devastating, lack of preevent data often hampers the ability of investigators to quantify changes in fish populations or habitat (Lamberti et al. 1991; Dolloff et al. 1994). A handful of studies with preand postevent data have indicated that salmonid populations can be quite resilient to ‘‘catastrophic’’ disturbances. Typically, disturbed populations returned to preevent densities within 3 years (Lamberti et al. 1991; 719 BROOK TROUT RESPONSE TO A CATASTROPHIC FLOOD Propst and Stefferud 1997; Swanson et al. 1998) unless the event created unsuitable habitat conditions (Elwood and Waters 1969). When the habitat after a catastrophic event is suitable, abundance and density can actually rebound beyond preevent levels (Lamberti et al. 1991; Thorpe 1994). Although limited in their ability to detect changes, studies without preevent data can still provide valuable insight into how fish populations and instream habitat respond to conditions created by episodic events. For example, several researchers have observed greater growth rates by fish that recolonized areas affected by mass wasting than in areas of the same stream that were not affected by mass wasting (Lennon 1961; Elwood and Waters 1969; Lamberti et al. 1991; Letcher and Terrick 1998; Swanson et al. 1998). Debris flows, one of the more spectacular types of mass wasting events, occur in high-gradient watersheds that also commonly support salmonid populations. Debris flows occur when landslides combine with floodwaters to produce a mobile slurry of rocks, soil, water, and trees moving downstream at up to 10 m/s, tearing trees from the stream banks, scouring stream channels, depositing new materials, and eliminating entire instream faunas (Hack and Goodlet 1960; Gecy and Wilson 1990; Swanson et al. 1998). These events are rare on any given stream, with recurrence times of more than 50 years in the Pacific Northwest (Swanson et al. 1987) and more than 100 years in the Central Appalachian region (Hack and Goodlet 1960). The unpredictable and infrequent nature of such events, coupled with a lack of knowledge of preevent conditions, has led to a poor understanding of how episodic events influence habitat and populations, as well as to the perception that such events are catastrophic. Studying the response of stream habitat and fish populations following extreme events provides the information needed to make and defend decisions that promote and preserve natural response mechanisms. Over a 5-d period in June 1995, soils in the Shenandoah National Park (SNP), Virginia, became saturated by nearly continuous rainfall, culminating with more than 20 cm of rain over the final 24 h of the period (Karish et al. 1997). The result on SNP’s Staunton River was a streamwide flood, coupled with landslides and a massive debris flow along the lower 1.9 km of the stream to its confluence with the Rapidan River (Eaton 1999; Figure 1). Changes caused by the debris flow were dramatic and included the scouring of the stream channel and deposition of new substrates, the elimination of trees from a 30-m-wide band in the riparian area, the piling of large woody debris (LWD) in massive stacks on the stream banks, and the complete extirpation of fish from the debrisflow-affected area (U.S. Forest Service, Center for Aquatic Technology Transfer [USFS–CATT], unpublished data). Fish populations remained intact, although at decreased densities, in the debris-flowunaffected area of the stream. The existence of preevent data (Newman 1996) provided a rare opportunity to examine how the brook trout Salvelinus fontinalis population and instream habitat were changed by and responded to the flood and debris flow.