Laserfiche WebLink
FISH IN AQUEDUCTS <br />stream in 1988 to include the Harquahala (92 km) <br />and Hassayampa pumping plants (193 km). The <br />expansion made a 24-h sampling regime logisti- <br />cally impractical; thus, all collections were taken <br />within 1 h before and 1 h after dawn, noon, and <br />dusk. <br />Fish eggs, larvae, and fry (<2 cm total length) <br />were separated from other planktonic material and <br />identified with the use of a dissecting microscope <br />(3-30 x magnification) and appropriate larval <br />keys (Hogue et al. 1976; Snyder 1981; Wang <br />1981). Young fish, hereafter referred to as larvae, <br />were identified to family, and eggs were classified <br />as either adhesive (<1.8 mm) or nonadhesive <br />(>1.8 mm, percichthyidae). <br />Fish surveys.-Relative fish abundance was de- <br />termined with electrofishing, experimental gill <br />nets, and angling. Surveys were conducted in the <br />pumping plant forebays and the canal reaches im- <br />mediately upstream of all five pumping plants. <br />Sampling started at the Bouse Hills pumping plant, <br />and as water service expanded, additional down- <br />stream sites were added to the sampling schedule. <br />The Hassayampa Pumping Plant was added in <br />1987, and the Salt-Gila and Brady pumping plants <br />were added in 1988. Sites were sampled during <br />the spring and fall of 1986 and 1987. Sampling <br />frequency was increased from semiannual in 1987 <br />to quarterly in 1988 and 1989. <br />A johnboat equipped with a Coffelt RF-10 unit <br />was used for electrofishing. The forebay and a 0.5- <br />l-km reach of canal immediately upstream of each <br />pumping plant were sampled after dark. Initially, <br />the majority of fish were observed avoiding cap- <br />ture by swimming in front of and below the elec- <br />trical field. To increase capture efficiency, six to <br />eight gill nets were set diagonally within the elec- <br />trofishing area just before electrofishing began to <br />block and reduce fish escape. Nets and entangled <br />fish were removed immediately after electrofish- <br />ing. Fish captured by electrofishing and the block- <br />ing nets were combined and expressed as numbers <br />and biomass captured per minute of electrofishing <br />(fish/min, g/min). <br />Standardized sampling with gill nets provided <br />another index of relative fish abundance and bio- <br />mass. Standard, five-panel (bar measures: 1.2, 1.9, <br />2.5, 3.8, 5.1 cm), 2.4-m-tall by 38.1-m-long mono- <br />filament experimental gill nets were set overnight. <br />Nets were set well upstream (1 km) of the elec- <br />trofishing effort to minimize possible bias. Nets <br />were set in pairs (surface and bottom sets) and <br />effort was expressed in fish numbers and biomass <br />captured per 305 m2 of net for each 12-h set. Stan- <br />797 <br />dardized gillnetting was used at all sampling sites <br />with the exception of the Salt-Gila Pumping Plant, <br />where a large upstream siphon did not allow the <br />canal space needed (at least 1 km) to accommodate <br />both electrofishing and standard gillnetting. <br />Fish were also collected by angling with bait <br />and artificial lures. Numbers of fish caught per <br />angler-hour (fish/h) and the creel composition were <br />recorded. All fish collected by the various sam- <br />pling methods were identified and counted and to- <br />tal lengths (mm) and weights (g) were recorded. <br />In Reach 1, standing crop was estimated with <br />the Petersen mark-recapture method (Ricker 1975) <br />in October 1986 and December 1989, when the <br />canal was partially drained for maintenance pur- <br />poses. Approximately 90% of the canal's normal <br />volume was drained, which reduced depths from <br />5.5 m to 1-1.5 m. A 3-km reach was blocked with <br />2.5-cm-bar mesh nets and then electrofished. Fish <br />were weighed, fin-clipped, and then redistributed <br />throughout the sampled area. Threadfin shad Dor- <br />osoma petenense were not marked because their <br />handling mortality was high. In 1986, the reach <br />was resampled 24 It after marked fish were re- <br />leased. Time constraints in 1989 only allowed a <br />3-h period between sampling. The marked popu- <br />lation was resampled with a combination of elec- <br />trofishing and gill nets. Estimates could not be <br />made in other portions of the aqueduct because <br />water was maintained at normal levels. <br />Estimates of standing crop were calculated for <br />the actual water volume sampled (fish/m3 and <br />g/m3) and then expanded volumetrically to normal <br />conditions. This compensated for a disproportion- <br />ate reduction of surface area relative to water vol- <br />ume (surface area dropped by 50% and water vol- <br />ume declined by 90%). Volumetric values were <br />then converted to more conventional surface-area <br />density and biomass units (fish/ha and g/ha). Scuba <br />diving was used periodically to determine if cen- <br />trarchids were constructing redds and spawning <br />and to make general underwater observations. <br />Results <br />Water Quality <br />Water temperatures during the ichthyoplankton <br />surveys ranged from 15°C to 26°C in 1987 and <br />from 18°C to 26°C in 1988. During the entire 4 <br />years, water temperature ranged from 9.1°C to <br />29.9°C; dissolved oxygen, from 6.0 to 12.0 mg/L; <br />pH, from 7.3 to 10.0; and conductivity, from 820 <br />to 946 µS/cm.