LlBRARIES llllllllllllllllllllllllllllllllll HESJ: 3 1293 10705 4883 1—— LIBRARY Michigan State University This is to certify that the thesis entitled THE IMPACTS OF IRRIGATION WATER WITHDRAWALS ON BROWN TROUT (Salmo trutta) AND TWO SPECIES OF BENTHIC MACROINVERTEBRATES IN A TYPICAL SOUTHERN MICHIGAN STREAM presented by Charles Gowan has been accepted towards fulfillment of the requirements for Master of Science degree in Fisheries and Wildlife mama Major professor Date 11-6-84 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution MSU RETURNIN§_MATERIALS: Place in book d?6p“io remove this checkout from your record. fiNfifi will be charged if book is returned after the date LIBRARIES “ 4. ‘ stamped below. \ . lw‘f‘l‘QiZ'f’ 5m”, 1 0 “dog ‘ l" I 1 a 071 0 e 9 . “VP-.3". @- ‘1‘! l - fl THE IMPACTS OF IRRIGATION WATER WITHDRAWALS ON BROWN TROUT (Salmo trutta) AND TWO SPECIES OF BENTHIC MACROINVERTEBRATES IN A TYPICAL SOUTHERN MICHIGAN STREAM By Charles Gowan A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Fisheries and Wildlife 1984 ABSTRACT THE IMPACTS OF IRRIGATION WATER WITHDRAWALS ON BROWN TROUT (Salmo trutta) AND TWO SPECIES OF BENTHIC MACROINVERTEBRATES IN A TYPICAL SOUTHERN MICHIGAN STREAM By Charles Gowan The Instream Flow Incremental Methodology (IFIM) was utilized for the first time in a Michigan stream. The objective was to test the method's applicability to the midwest, and to detail the impacts of irrigation withdrawals on a typical lower Michigan trout stream. The IFIM was found to accurately simulate the hydraulic characteristics of a midwestern stream, and to predict brown trout (Salmo trutta) habitat locations within the stream. Brown trout habitat losses were most critical in the month of July, with reductions up to 16 percent. It was demonstrated that these habitat reductions lead to a reduction in trout population levels. It was also shown that a greater percentage of the brown trout population remaining experienced negative growth rates as habitat availability was reduced. Benthic macroinvertebrate habitat for gydropsyche spp. and Ephemerella spp. was found to be less impacted by irrigation withdrawals. Habitat losses for_§ydropsyche spp. reached a maximum of 11.05 percent during irrigation periods in July of 1983. Habitat losses for Ephemerella spp. reached a maximum of 6.35 percent during the same period. Keywords: instream flow, habitat, brown trout, irrigation. ACKNOWLEDGEMENTS The research on which this report is based was financed in part by the United States Department of the Interior as authorized by the Water Research and Developement Act of 1978 (P.L. 95-467), and by the Michigan Agricultural Experiment Station. Much appreciation goes to my graduate committee, Dr. Niles Kevern, Dr. William Taylor, and Dr. Roger Wallace. Each provided invaluable assistance without which this project could not have been completed. Various other members of the Fisheries and Wildlife Department's faculty, especially Dr. Darrell King, provided helpful comments during the preparation of this manuscript. Members of the Instream Flow Service Group, especially Ken Bovee, provided much needed assistance throughout the course of this study. Finally, I'd like to express my sincere thanks to those who assisted me in the field: Dave Dowling, Mark Freeberg, Chris Cooper, Kirsten Nelson, Craig Milewski, Jim Wood, and Eric Alexander. ii TABLE OF CONTENTS Page LIST OF TABLES ..... ...... ....... ............. ..... ..................... v LIST OF FIGURES .......... ..... ..... .................................... vi INTRODUCTION............ ..... ............ ..... ......................... 1 METHODS AND MATERIALS.................................................. 5 Habitat Estimation.................... ..... ....................... 5 Preference Curve Construction..................................... 9 brown trout.................................................. 10 benthic macroinvertebrates................................... 17 A Test of the IFIM................................................ 22 velocity adjustment factors.................................. 22 fish locations............................................... 23 depth, velocity, substrate, and cover predictions............ 24 The Relationship Between Habitat and Population Density........... 25 The Relationship Between Habitat and Growth Rate.................. 26 RESULTSOO..........OOIOOOOOOOOOOOOOOO......OOOOOOOOOOOOOOOOOO0.00.00... 27 Fish Preference Curves............ ..... ........................... 27 Fish Habitat Availability......................................... 28 Insect Preference Curves.......................................... 41 Insect Habitat Availability....................................... 41 A Test of the IFIM................................................ 59 The Relationship Between Habitat and Population Density........... 82 The Relationship Between Habitat and Growth Rate.................. 82 DISCUSSION..;....... ................. ............ ................... ... 87 Fish Preference Curves.............. ....... ....................... 87 Insect Preference Curves.......................................... 92 A Test of the IFIM................................................ 93 Fish Habitat Estimations.......................................... 96 habitat losses in 1983....................................... 99 habitat losses in 1984.......................................104 Insect Habitat Estimation.........................................105 habitat losses in 1983 and 1984..............................106 The Relationship Between Habitat and Population Density...........106 The Relationship Between Habitat and Growth Rate..................111 The Future........................................................115 iii iv TABLE OF CONTENTS (continued) Page CONCLUSIONSOO ........ 0.0.0...0.000.00.000.00........OOOOO...0.0.0.0000119 LITERATURE CITEDOOOOO...0.0............OOOOOOOOOOOOO0.0.0.0....0......122 LIST OF TABLES Table Page 1. Cover and substrate codes used during utilization and availability data collection............................... 11 2. Example of utilization curve construction from raw field dataOOOOO......OOOOOOOOOOOOO......OOOIOOOOO0.0.0.0... 14 3. Brown trout utilization, availability, and preference cover by substrate matrices.................................... 18 4. Velocity Adjustment Factor ratings as given by Milhous et. a1. (1984)......................................... 77 5. A test for significance between the average WUA found at a transect and the number of fish captured in that transectOOOOOOOIO ...... ......OOOOOOOOOOIOOOOO ..... 00...... 81 6. Summary of the computer's ability to simulate the depths, velocities, substrates, and covers actually occurring in Fish Creek........................................ 83 7. Chi square interaction tests showing an example and the actual test reSUItSOOOOOOOOOOOOOOOOO......OOOOIOOOOOOOO 90 8. Summary of aquatic insect habitat losses for 1983 and 1984.....107 LIST OF FIGURES Figure Page 1. A map showing the location of Fish Creek and the StUdy area................................................. 4 2. Brown trout depth utilization, availability, and preference.... .7 3. Brown trout velocity utilization, availability, and prEferenceOOOOO........................‘........................ 29 4. Brown trout cover by substrate preference curve devaloped from Table 3C........................................ 31 5. Brown trout habitat area versus discharge curve................ 32 6. Hydrographfor1983and1984................................... 33 7. Brown trout habitat time series for 1983 and 1984.............. 42 8. Aquatic insect depth, velocity, and substrate availability curveSooooococo-0.000.000.00000.000000000000000... 50 9. Depth utilization and preference for Hydropsyche spp. . and Ephemerella 822............................................ 53 10. Velocity utilization and preference for Hydrqpsyche spp. and Ephemerella $22............................................ 55 11. Substrate utilization and preference for Hydropsyche spp. and Ephemerella 8220000000000000coo...000000000000000000000000. 57 12. Aquatic insect habitat versus discharge curve.................. 60 13. Ephemerella spp. habitat time series for 1983 and 1984......... 61 14. Hydropsyche spp. habitat time series for 1983 and 1984......... 69 15. Map detailing downstream, midstream, and upstream computer predicted habitat locations and actual fiSh locationSooooooooo0.000.000.0000000.000.000.0000000000.... 78 16. Regression of habitat availability versus population den81ty for 1983 and 1984......... ...... ....................... 84 vi vii LIST OF FIGURES (continued) Figure Page 17. Regression of habitat availability versus mean grOWth rate (1983 data only)...........................O....... 86 18. Habitat duration curve for June to September 1983 and 1984.....101 19. Chronological sequence of habitat availabilities and population levels of brown trout, 1983.........................109 20. Change in weight through time of four "typical" brown trout in1983..................................C...............113 21. Percentage of the brown trout population gaining weight versus habitat availability in the summer of 1983..............114 22. Mean monthly habitat availability in June, July and August versus probability of occurrence curve. The time span considered is twenty years............ ..... ..........116 INTRODUCTION The lotic environment is characterized by variation . The intermittent or seasonal stream is an extreme example. However, even the largest river exhibits changes in temperature, velocity, depth, width and other physical characteristics by the hour, day, and season. Natural flow variation can be considerable. In temperate zones the spring thaw can produce flow rates an order of magnitude greater than the flows occurring later in the summer. A heavy thunderstorm can change flow drastically in a short time. The attendant changes in velocity, depth, and width can be severe. Despite these rapid changes in the physical environment, the biologic community can thrive. However, a stream‘s biota do have a finite amount of tolerance. Human influence can exceed even the harsh natural regime. In the Western United States tremendous demands are being placed on the available water resources (Powledge, 1982). The threatened extinction of some species can be directly related to reduced flow (Davis, 1979), and once productive fisheries are being lost due to streamflow regulation (Anon., 1977; Graham, 1980). The midwestern U .S. has not had these water demand problems, and, in fact, is considered water rich (White, 1976). However, the situation is changing . Irrigation demands are rapidly increasing . Irrigation in Michigan counties (measured in acre-inches of water used) has increased an average of 268 percent from 1970 to 1977 . Some of the most 1 cultivated counties have had increases of over 11,000 percent. The prediction is for the trend to continue (Bedell, 1977) . Already the effects are being noticed . The Water Management Division of the Michigan Department of Natural Resources (DNR) in 1979 started a file to keep track of complaints made by riparian property owners concerning water use by their neighbors (Bedell, per. comm .) . Some streams are completely dewatered by irrigation demands (Doyle, per. comm .) . Currently, no law limits the amount of water an irrigator can remove from a stream in Michigan . The courts can order withdrawals stopped if one riparian can demonstrate that another is making "unreasonable use" of the resource, or if the DNR can prove that ”ecological damage" has resulted . ”Unreasonable use" and ”ecological damage" are not defined (Bedell, per. comm .) . Clearly, the stage is set for the midwest to start experienceing the same type of water use problems that the western states have faced . The midwest, however, is in a position to learn from its drier neighbors. Methods have been developed in the west to deal with the question of ”ecological damage.“ The first application in Michigan of one of these methods, the 0.8. Fish and Wildlife Service's Instream Flow Incremental Methodology (IFIM), is the topic of this report. The IFIM is designed to estimate the amount of habitat available to a particular species at any flow . Thus, the amounts of habitat lost through water removal can be estimated . There were two purposes in trying the IFIM in a Michigan stream . First, because the method was developed in the west, it was believed that this project would provide a test of the applicability of the method to midwestern streams. The second reason was to begin to establish an instream flow data base for the region in order to provide the information necessary for future water use policy . The objective of this study was to detail the possible impacts of irrigation water withdrawals in terms of fish and aquatic insect habitat loss, and fish population density and individual growth rate responses to habitat loss. The stream chosen for study was Fish Creek, a small first order stream that runs along the eastern border of Montcalm Co., Michigan (Figure 1). The study section is located in section 11 of Evergreen Township . This stream was chosen on two criteria. The first was that, in many ways, it is typical of Michigan's marginal, managed trout streams. It is regularly stocked with brown trout (§§l.1_“9 m) and has a native population of brook trout (Salveling’ W) . The second reason for selecting Fish Creek was that it is surrounded for most of its length by farmland producing corn, potatoes, and soybeans. As a result, it is the major irrigation water source for at least eleven farms (Cooper, 1984) . The section studied is located in section 11 of Evergreen Township . The section is just downstream from the heaviest irrigation withdrawals and is representative of the headwaters of Fish Creek . .Monm huspm on» cam xmmno swam no cowpmooa one mcwaonm mma < “P mmmem If oe_uenao azo coo.co ......u" , w a... .93.- ..k 44 \ METHODS AND MATERIALS Habitat Estimation The IFIM was used to determine the amount of brown trout habitat, measured in square feet of weighted usable area, present in the study section at flows ranging from six cubic feet per second (cfs) to thirty cfs. This constitutes the normal summer range of flows in Fish Creek . The amount of habitat available to two of the dominant benthic macroinvertabrates, Ephemerella _s_pp . and Hydropshyche spp., was determined for the same flow range in order to estimate the effects of flow reduction on these two trout food organisms. A brief summary of the theory behind the IFIM method is given here. A complete description is given by Bovee (1982) . The method involves four physical parameters: depth, velocity, substrate, and cover. In theory, if these variables could be measured at every location in a stream one would have a complete description of the physical characteristics of that stream . If this could be done at all flows of interest, one could say with certainty how that stream is affected by changes in discharge. The IFIM allows this with relatively little data . Depth, velocity, substrate, and cover are measured at specific points along carefully chosen transects. This is done at the exact same locations at three different discharges. A series of computer programs perform a task known as Physical Habitat Simulation (PHABSIM), relating changes in discharge to changes in the availability of combinations of depth, velocity, substrate, and cover. The utility of these programs is that they interpolate data and give predictions of the availability of these physical parameters at discharges other than those at which the data were collected . The estimate of fish habitat present is determined when the above predictions are combined with preference curves . Preference curves are probability density functions that describe the affinity a particular species has for various depths, velocities, substrates, and cover types. Preference ratings result from measurements detailing the availability of the physical characteristics and the utilization of these parameters by the species of interest (Figure 2a). Preference curves were constructed based on data collected on Fish Creek . A depth preference curve for brown trout is shown in Figure 2b. The construction of these curves will be addressed later in this section . As seen from Figure 2b, all intervals of a given parameter are rated from 0 to l . By multiplying the individual ratings for a given interval of each of the four physical parameters together, a composite rating for that particular combination of physical parameters is determined . This is the Joint Preference Function (JPF) . For example, if the ratings for a depth of 1 ft., a velocity of 0.5 ft/sec., a substrate of course gravel, and a cover of down timber were 0.2, 1 .0, l .0, and 0.5 respectively, the JPF would be 0.1 . This indicates that the combination of physical characteristics described would be rated as one-tenth as preferred as the optimum combination . .hpaaanmaam>m can soapMNaHHpc space pconp csonm "mm mmwam AMP“: ...-haun— N.~ N a; w; v; N; i we as To «5 o _ _ _ . _ _ _ _ _ _ o a... n ., «.9 u villi-o n .. z .0 v can u ... to o ... ..N ... v .u A .. 00° N ” .... ... v n m n. .... u o.o .; ... .1 »c.4~m¢4_¢>c ..... m 2255:5111 m l o.— _ .moconomonc gamma scone czoum "pm mmchm ...—h: Ihama o.“ 7— N." d o.o 0.9 v.0 «.0 _ . _ _ _ _ _ _ _ \ii. N6 To m6 30N383d38d To determine fish or aquatic insect habitat the PHABSIM program multiplies the number of square feet of a particular combination of depth, velocity, substrate, and cover present by that combination's JPF. This results in Weighted Usable Area (WUA). Thus, if a particular combination of physical characteristics has a JPF rating of 0.1, and there are 100 square feet of that combination available, then the WUA would be 10 square feet. This indicates that 100 square feet of this type of habitat is equivalent to 10 square feet of optimum habitat. The WUAs for all the habitat types present are summed to give the total WUA. This is usually reported as WUA per 1000 linear feet of stream . The term ”habitat area" will be used interchangably with WUA throughout the rest of this thesis . grifrerence Curve Construction The preference that an organism exhibits for a particular habitat component results from the interaction of two functions . The utilization function represents a frequency distribution of sites occupied by a population, given a certain variety of sites from which to select. The availability function describes the range of sites actually available to the population . The preference for any given interval of a physical parameter is defined by Bovee (1982) as: PREFERENCE = UTILIZATION/AVAILABILITY . (eq. 1) Thus, in order to accurately estimate preference, both utilization and availability must first be described . 10 The major concern in data collection is to avoid bias introduced by the sampling gear. The goal is to accurately assess the depth, velocity, substrate, and cover occupied by an individual of the species of interest. Visual observation of undisturbed individuals is the best method of data collection. However, this was impractical for brown trout due to the overgrown, inaccessable nature of Fish Creek . Observation of individual benthic macroinvertebrates would be equally as difficult. For these reasons, electroshocking was used to determine the locations of fish, and a pole mounted Eckman dredge was used to collect the invertebrates. A detailed description of each method is given below . Brown Trout A DC backpack electroshocker was used to make all of the brown trout observations. While moving upstream as quietly as possible, the probe was ”poked" into the water ahead of the observer. This poking motion helped to prevent fish from being disturbed from their resting places by the advancing electric field . Fish obviously disturbed were ignored. When a fish was shocked, the depth, velocity, substrate, and cover was noted at the point of first observation . No attempts were made to guess where the fish "should have been ." Depth and average velocity (measured at 0 .6 of the depth) were measured with a wading rod and pygmy water velocity meter, respectively. Substrate and cover were estimated visually . The codes used for this are given in Table l . Table 1: 11 Cover and substrate codes used during utilization and availability data collection. Cover Code * ... \DCDQO‘Ul#\JIN-b Description No cover Undercut bank less than 1 It. deep Undercut bank greater than 1 ft. deep Overhanging vegetation greater than 1 ft. above surface Overhanging vegetation within 1 ft. of the surface Emergent or submerged overhanging vegetation Down timber Half-log improvement structure Large rock or boulder Substrate Code Description 1 * thU'l-FUIN “’1 \D Rooted aquatic vegetation Fines (sand, silt) Pebbles or fine gravel (up to 1") Large gravel (1-3”) Cobble (3-12") Boulder (greater than 12”) Bedrock Detritus . Down timber imbedded in the substrate *eliminated due to impracticality or absence in the habitat +considered as down timber **considered as fines 12 These data constitute the information necessary to build the fish utilization curves. It should be noted that all of these data were collected on fish located out of the study section. This was necessary because a part of this study concerned testing the computer's ability to predict fish locations within the study section. Using datacollected on fish found within the study section for construction of the preference curves would have precluded the use of this test. Essentially, using these data would have forced the computer to predict the correct fish locations . In order to quantify the combinations of depth, velocity, substrate and cover available to the fish, random transects were run across the stream . Depth, velocity, substrate and cover were noted at l .5 ft. intervals. Three random transects were measured for aproximately every 25 fish collected . The transects were run during fish collection so the availability data reflect the conditions present at the time of utilization data collection . Data were analyzed using the procedure outlined by Bovee and Cochnauer (1977) . The parameters of depth and velocity were divided into intervals of 0 .1 ft. and ft. per second (fps), respectively .* The number of observations in each interval was tallied . Right-hand and left-hand clustering was performed to reduce the natural variation present. One of the two clusters was chosen on the basis of having the smallest variance and the presence of a single peak . Chi square analysis was performed to discern significant differences between the l3 totals of adjacent clusters. The average of the two clusters was used as the expected value. If two clusters were found to be not significantly different at the 0 .1 level, the next adjacent cluster was included and significant differences tested for among the three. Once significant differences between clusters were established, all clusters were scaled from 0 to l . The cluster with the most observations was rated as l and the other clusters were rated relative to it. This process was followed for both utilization and availability . Equation 1 was then used to determine preference. Preference curves were constructed by dividing the utilization rating (ranging from 0 to l) of a given cluster by its availability rating . The values resulting from this division were then normalized from 0 to 1. The preference curve construction procedure is demonstrated, for depth, in Tables 2a-c . The parameters of substrate and cover were handled differently for two reasons. First, unlike the continuous variables depth and velocity, they are discrete. Thus, the type of cluster analysis performed for depth and velocity would not be appropriate . Second, the computer simulation is set to handle only three physical variables (depth and velocity along with one of the user's choice). Since both substrate and cover can be important, they have been combined into one curve. The method used is adapted from Bovee (1982) . First, cover by substrate utilization and availability matrices were constructed . The cell containing the most observations was given a 14 Table 2a: Example of utilization curve construction from raw field data. Utilization De th Number observed eft-hand c uster Ri ht-hand cluster Ratin 0.0-0.09 O "0.1-0.19 0 0.2-0.29 0 0.3-0.39 O O 0 0.4-0.49 0 5 0.5-0.59 5 5 :15 0.6-0.69 O 6 0.7-0.79 6 . 1f .18 0.8-0.89 1O 23 0.9-0.99 13 34 1.0v 1.0-1.09 21 41 1.1-1.19 20 38 1.0 1.2-1.29 18 37 1.3-1.39 19 34 1.0 1.4-1.49 15 39 1.5-1.59 24 33 1.0 1.6-1.69 ‘ 9 27 1 07.1079 19 26 1.0 1.8-1.89 7 12 1.9-1.99 5 15 .45 2.0-2.09 1o 13 2.1-2.19 3 5 .18 2.2-2.29 3 5 2.3 + 3 .01 15 Table 2b: Example of availability curve construction from raw field data. Availabilipz pepth Number observed Left-hand cluster Eight-hand cluster Rating 0.0-0.09 not used ' 0.1-0.19 18 18 36 .19 0.2—0.29 18 T w 0.3-0.39 20 _____ 50 0.4-0.49 30 65 .77 0.5-0.59 35 77 0.6-0.69 42 82 .77 0.7-0.79 4O 99 0.8-0.89 59 95 1.0 0.9-0.99 36 _ 6S 1.0-1.09 29 47 .49 1.1-1.19 18 43 1.2-1.29 25 46 .49 1.3-1.39 21 __ 35 1.4-1.49 14 23 .24 1.5-1.59 9 ' 14 ' 1.6-1.69 s 8 .05 1.7-1.79 3 6 1.8-1.89 3 3 .05 1.9-1.99 0 4 2.0-2.09 4 5 .05 2.1-2.19 1 4 2.2-2.29 5 3 .05 2.5 + O 16 Table 20: Example of preference curve construction from raw field data. Preference Depth Decimal eguivalent of utilizationlavailabilitz Rating 0.0-0.09 0.0 0 0.1-0.19 0.0 0 0.2-0.29 0.0 0 0.3-0.39 0.0 " 0 0.4-0.49 0.0 0 0.5-0.59 .19 .01 0.6-0.69 .19 .01 0.7-0.79 .62 .05 0.8-0.89 .48 .02 0.9-0.99 1.0 .05 ' 1.0-1.09 2.04 .1 1.1-1.19 2.04 .1 1.2-1.29 2.04 .1 1.3-1.39 2.04 .1 1.4-1.49 4.17 .21 1.5-1.59 4.17 .21 1.6-1.69 20 _ 1.0 1.7-1.79- 20 1.0 1.8-1.89 20 1.0 1.9-1.99 9 .45 2 . 0-2 .09 9 . 45 2.1-2.19 3.6 .18 2.2-2.29 3.6 .18 203 + -.. ..- 17 rating of one, and all other cells were rated relative to it. From this procedure, a utilization and an availability matrix were constructed (as opposed to a curve). Equation (1) was used to construct the preference matrix in a manner similar to that used to construct the preference curves for depth and velocity . Because the computer will not accept input in the form of a matrix, a curve had to be constructed . Each cell in the preference matrix was given a number, 1 through 42. This number describes a unique combination of substrate and cover and thus can be used as the abscissa in a preference curve. The curve construction is demonstrated in Tables 3a-c . Benthic Macroin vertebrgges The construction of preference curves for the two species of benthic macroinvertebrates was done differently . As stated previously, utilization data were collected with a pole mounted Eckman dredge. Before going into the field a series of ten random numbers ranging from 1 to 100 was generated using a random number table. A second set of ten numbers ranging from 1 to 10 were similarly generated . These two sets of numbers were paired such that one number in the range 1 to 100 was matched with a number in the range 1 to 10. Once in the field a random point was selected as a starting spot. A number of steps equivalent to the lowest value in the 1 to 100 range were taken upstream from this point. The second number in the pair (in the range 1 to 10) determined Table 3a: Brown trout utilization cover by substrate 18 matrix. Utilization 92:2; 1 gw g: 5 6 .17 9 0 0 o 1 0 0 0 1 0 o 0 .01 o 0 0 0 4 4 0 7 65 O 80 O .05 .05 .08 .74 O 1 6 15 4 2 98 1 117 ”w .04 .07 .17 .05 .02 1.0 .01 ‘g 0 0 0 0 3 O 3 .§ 0 0 0 o 0 .03 0 a) 1 1 3 1 0 1 9 .01 .01 .03 -.01 0 .01 .02 0 0 0 0 0 o o 0 0 0 0 0 0 0 0 2 11 22 6 9 157 3 210 19 Table 3b: Brown trout cover by substrate availability matrix. Availabilit Cover 5 8 .3 x; 7. O 22 2 1 O 1 4 9 000000 00 00 :3 ,0 7. 2) no ,0 7 0549.0 00 06 4 1| 0 0 6: 06.2.0 00 08 4t Qj 31 O 6 .5 no as . 1, a no no no no .4 1. 4... O 0 so no 9. . 1. . no no no no a, 2:1 0 0 a; .u .4 . 1. . no no no no :9 63820 6 1.5 no .4 .8 ,o a n. no no 9. 4| 11011.0 10* 93 n4. 2; 4. .9 manhumnzm 20 Table 30: Brown trout cover by substrate preference matrix. .ZEEESEEEES 2222: t 1 A; __3 5 6 7 9 1 0 0 0 .01 0 0 0 o 0 0 .01 0 0 0 2 0 2.5 5 0 2 2.1 0 0 .15, .29 0 .1g_____;p;v o 3 .01 7 17 2.5 2 16.67 .01 {3 0 .41 1.0 .15 .1g__.__‘2§ ‘1 401 :g 4 0 0 | 0 0 0 .03 0 .fi 0 0 0 0 0 .405 0 5 .17 .01 .01 ‘ .01 0 o .02 .01 .01 .01 .01 0 0 ._£gg__ 9 0 o 0 0 0 0 0 0 0 0 0 0 __ 0 21 the number of steps to be taken from the right bank of the stream . A sample was taken at this point. This procedure was repeated using the remaining nine matched pairs of numbers. Thus, the stream was effectively divided into a grid with the matched pairs of random numbers determining the cell within the grid to be sampled . The samples were placed in plastic, sealable bags and taken back to the lab. Once there, the total numbers of Ephemerella spp. and Hydropshyche spp. in each sample were determined by hand-picking . At the point where the sample was taken, a measurement of depth, velocity, and substrate was taken in the same manner as described for the fish observations. Cover was not noted, as it was felt that depth, velocity, and substrate would be sufficient to describe macroinvertebrate habitat. Because the points at which the samples were taken were determined randomly in the lab, the data collected could be used to determine utilization, availability and preference. The procedure is similar to that described for brown trout, with the exception that the species in question was considered ”present" in a sample if more than 10 individuals were found in the sample. The number of samples found with the species of interest "present” was then used in the utilization curve construction . Total numbers of individuals found were not used because of the great variability this introduced into the calculations. The threshold limit of ten was used because it was judged that the few individuals found in some samples were artifacts of drift and not residents of the sampled area . 22 The data described above can now be used in the PHABSIM process to determine the amount of brown trout, Ephemerella spp., and hydropsyche pm. habitat present in the study area at all flows of interest. A 'Itest of the Egg The applicability of the IFIM to midwestern streams was tested in three ways . First, the canputer routines perform internal quality control checks as they are run. One of these checks is known as the Velocity Adjustment Factor (VAF) . If the VAFs are within acceptable ranges, the computer should be estimating the hydraulic characteristics of the stream accurately. The second way to test the accuracy of the model is to determine if fish within the study section are occupying the sites predicted by the corputer to be fish habitat. The final test was a check of the canputer's ability to predict depths, velocities, substrates, and covers as they occurred throughout the stream. The three tests are described below. velocity Adjustment Factors As stated earlier, the study section is mapped by taking depth, velocity, substrate, and cover data at specific points along carefully chosen transects. Fran these data a discharge estimate, (0). can be made at each transect. All transects should have the same discharge at any one time. However, due to errors made in the field, the discharge estimates for each transect invariably differ slightly. The computer 23 performs a mass balancing procedure in order to ”force“ all transects to have the same discharge. It does this by adjusting the velocities of each transect until all the discharges match (the user specifies what the actual discharge should be). The more adjusting necessary in order to reach the user specified 0, the less accurate the simulation. The VAFs give the user a measure of the amount of adjusting necessary. The VAF for any one transect is defined by Milhous et. a1. (1984) VAF=Q computed/Q trial . (eq. 2) where Q computed is the discharge specified by the user, and 0 trial is the discharge that results from the unadjusted data. A VAF of l .0 indicates that the field data need no adjustment. VAFs above or below 1 .0 indicate increasingly unreliable data. The VAFs resulting frcm the Fish Creek data were analyzed and a determination of the accuracy of the simulation was made based on the guidelines specified by Milhous et. al. (1984) . Fish Locations A second test of the accuracy of the simulation involved observation of brown trout in the study area. If the trout are occupying the areas described by the cauputer as habitat, the sinulation must be accurate . Brown trout were collected by electroshocking and their location triangulated using permanent reference stakes set up in the study area. These locations were then noted on a map of the study 24 section detailing the fish habitat locations as estimated by the computer. Comparison of actual fish locations to computer predicted habitat locations provides a qualitative test of the computer's ability to simulate brown trout habitat. A quantitative measure involves testing the relationship between habitat quality (as measured by WA) and the number of fish found in that habitat. This was done on a transect by transect basis . The average WUA found for each transect over the study period should correlate to the number of fish found at that transect over the same period. In other words, more fish should be found at transects the computer predicts to be high quality habitat compared to the number found at low quality transects . The average WA for each transect was generated by computer simulation and a test of significance between this value and the number of fish captured during the study at each transect was made using the nonparametric statistic, Kendall's Tau (Lehner, 1979) . A nonparametric statistic was necessary due to the non-homogeneous nature of the variances involved . D_epth, Velocity, Slbstrate, and Cover Predictions As noted previously, fish locations within the study section were determined bi-monthly. At each spot a fish was captured, a depth, velocity, substrate, and cover measurement was taken . The computer was then used to predict the depth, velocity, substrate, and cover present at that spot based on the streamflow present at the time of fish 25 collection . Thus, computer predicted depth, velocity, substrate, and cover could be directly compared to actual depth, velocity, substrate, and cover. These three tests should be sufficient toidetermine the applicability of the model to the midwest. The VAFs and the depth, velocity, substrate, and cover checks will give an indication of the model's ability to simulate the hydraulics of midwestern streams. The fish locations will test the model's ability to actually describe trout habitat. The Relationship of Habitat Availability to Population Density The IFIM is only designed to estimate the amount of habitat present at different flows. It does not estimate the number of fish or aquatic insects that can be supported by these varying amounts of habitat. In order to determine the relationship between available habitat and brown trout population levels, bimonthy population estimates were made. However, due to inclement weather, not all sampling dates were exactly 14 days apart. These estimates could then be regressed against the average amount of habitat present in the study section over the previous two week period and a test of significance made. POpulation estimates were made by making two shocking trials, without replacement, through the study section. Because of the small size of Fish Creek, these two trials were assumed to capture all of the trout present in the study area. 26 In order to estimate available habitat, a continuous water level recorder (Leupold and Stevens, Model F) was installed just above the study section. This recorder, once calibrated, supplied a continuous record of the discharges occurring in the study section. By taking the average daily flow in the study section over the two week interval between population estimates, and simulating this flow with the PHABSIM procedure, an estimate of average daily habitat over the interval was made. This could then be compared to the population estimate made at the end of that interval. The Relationship of Habitat Availability and Growth Rates Along with population levels, biologists and anglers are concerned with growth rates. In theory, the amount of habitat present in an area slould have a direct influence on the growth rate of the fish in that area. To determine if this relationship existed, bimonthly estimates of individual growth rates were made . . Fish were collected by electroshocking (at the same time the population estimates were made). Each individual collected was identified with a small numbered tag (dimensions: 10mm * 3mm * 1mm) attached to the dorsal surface of the caudal peduncle by use of a lightweight nylon thread. These tags were fairly permanent, with fish tagged in June, 1983 retaining their tags until the completion of the study in August, 1984 . (hoe tagged, the fish were weighed on a triple beam balance using a method of difference. A small bucket with 27 approximately six inches of water in it was placed on the scale and weighed. The fish was then added and the total weight determined. Fish weight was determined by substraction. In this manner fish weight could be determined to the tenth of a gram. Individual growth rates could be determined for fish caught on two successive sampling dates. Growth rate was defined as: Growth Rate = W2-Wl/Wl/days (eq. 3) where Wl= fish weight at sample time t-l, W2 fish weight at sample time t, and days= the number of days between sarpling times. The resulting value is in units of grams/gram/day. The average growth rate of recaptured fish was then calculated. This value was regressed against the average habitat availability over the previous two week period . RESULTS Fish Prgfgrgpce 92:29.5 A total of 210 individual fish utilization and 433 availability observations were made. The depth interval most utilized by brown trout was 1 .5 to 1 .59 ft., with 24 observations. The depth most available to fish was in the range 0.8 to 0.89 ft., with 59 observations. Combining depth utilization and availability results in the highest depth preference occurring between 1 .6 and l .79 ft. Because of tie extremely low flows during the time of data collection, information about the 28 preference for depths greater than 1.8 ft. is lacking. The assumption was made that depths greater than 1.8 ft. are as preferred as the interval 1.6 to 1.79 ft. Brown trout utilization, availability, and preference curves for depth are shown in Figures 2a and b. The velocity most utilized by fish was in the range 0.6 to 0.69 fps with 32 observations. The most available velocity was in the range 0.3 to 0.39 fps with 37 observations. The preferred velocity is in the range 0.1 to 0.69 fps. Brown trout utilization, availability, and preference curves for velocity are shown in Figures 3a and b. The most frequently utilized combination of cover and substrate was down timber and gravel with 88 observations. The most available combination was no cover and gravel with 162 observations. The most preferred combination was undercut banks greater than 1 ft. deep and gravel. The brown trout utilization, availability, and preference matrices are shown in Tables 3a,b, and c. The resulting preference curve is shown in Figure 4. Fish Habitat Availability The IFIM model used these preference curves along with the stream mapping data to arrive at estimates of brown trout habitat present at flows of 6 to 30 cfs. The weighted USable Area vs. discharge plot for brown trout is shown in Figure 5. The average daily flows by month as recorded by the water level recorder are shown in Figures 68-h. The estimate of average 29 .hpwaaomawmbm one soapmuaflaes hpfiooam> poonp ozonm ”mm mmDon an“: >....00..m> N; a To To 4.8 as o _ _ _ _ _ _ b _ o :SESEE ..... 1 . m. 28:33.51 N o I ...... z v u . ...................... . r To 0 .. ..N v .. A ... a To W n 1- “ uv .. m ... n m .. 0.0 .A .. A .. .......................... --i _ 3O o.~ m.— _ ¢.~ _ .mocepemono heaooao> pooh» ozonm "pm mmoon Amt“: >._..OO._u> N.~ _ u _ o.o b 0.0 _ ¢.o _ «.0 .P / N.o SONEHSdBUd 31 \‘ u\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\“\\\““ L\\\\‘ \\\\\V .\\\\\\\\\\\\\\\“\\\\\\‘\\\\\\\\\\\\\\\\\\\‘ \\\\\\\\\\\\\\\\\‘ “\V L\.\\\‘ \\\\\\\\\\\\‘ \\\\\\‘ l I T a: co V' 04 c: c: c: . c: c: HONEHEJSHCI I T l 20 25 30 35 40 15 SUBSTRATE AND COVER MATRIX CODES FIGURE 4: Brown trout cover by substrate preference curve developed from Table 3(0). 10 .9500 omngomao nachos. some 9039.0: 950.3 85.5 um .55on some. gotten a a on mm mm on «a ow g on 3 «a o” o 32 0cm emu can own oov owe [14038.13 '13 0001/ '13 '09) 8388 18.118314 33 .mmmw .eosb you nomamonohm new mmDon «www.mzas on 5N vN "N on m— Nu m _ P1 _ _ . _ Azouhcouzzu haozhux. waxczumuo cuhczubww.!:- .zonhcouzmu :e—x. wczczomuo 4c:~uc.lll r o— «g v“ m— @— ON Nu ow mm mm on ($30) BSHVHOSICI 34 .mmma .sase you eamnmonesm ”pm mmemHa 000 _. $.55 .m mm mm «w as o. no 8— a a .zof—gzm. 252:3. mamczoma oupczzwm..-i .29—.2333: 1:3. momczoms Joshua .II cm «a em on ow on ($210) SDHVHOSIG .mmma .pmsmsa you somnmonesm "om mmeuHa ammwfimacqz «a 2 2 2 S s 4 b L _ _ r b 35 Azouhccuzs 2.62:3. womczowuo 09—5:th .-..-- .zozcouzs 2:3. Neutron—o acahocl Na *— m— @— on «N vN mu ow on (3:13) SSHVHOSIO 36 mm mm _ _ .mmma .umpEmsamm you aamumonssm new mmecHa «www.cmmzmhamm mm mm on mu 0" b e . . _ _ _ gzoupccuzm_ paozb_z. womcruwmo ou~¢z-wm ..... Azo~h¢o~m1~ thus. mamczom_o oc:~o¢.lll r 0" Nu aw @— 0— ON NN em mm mm om mm vm (5:10) BSHVHOSIG .smma .mese poo eamnmonesm."mo mmeeHa emu w .mza. mm mu «a mo 8— as o_ s 37 _ _ o— N—. ¢~ m: m— on «N vN mm mm om. (S:l0)398VHOSIG 38 an em _ mm _ .4mm. .sase not aampmonaem new meeme com r 943—. NN mu m— m—. n: b _ _ _ _ . . Azozcouzs 2.53:1. womczuwa cubes—haw .-..-- 35:83.5 :va wczczuws ...cahucull A: N. z m— @— ON NN vN mu ow om (350)398VHOSIG 39 .emmp 303:4 no.“ nonsmonohm «we coop ....m:6:< mu cu m— o— . _ _ _ 556 HR. 22:33:: 253:: mom—$.02: awhczzmw 2...-.. 22.—5:5: 2:: Mammy-om; ...cahoc III o— .2 ON mu on (8:10) SSHVHOSIO 40 on .59 £38633 up... segments? 26 mamas vamp .mumzwhamm ON _ w— _ o— _ :— m— ON mN on (8:10) ESHVHOSIO 41 daily flow that would have existed without irrigation was made based on flows recorded when little or no irrigation was occurring). The combination of these curves yields the brown trout habitat time series (Figures 7a-h) . Insect Preference Cums A total of 209 insect samples were taken. The resulting depth, velocity, and substrate availability curves are given in Figures 8a-c. mpth utilization and preference curves for both species are given in Figures 9a and b. Velocity utilization and preference curves are given in Figures 108 and b. The depth preference curve for Hydropsyche _pp. shows three peaks . This was a result of insufficient data to accurately detail this species depth preference. It was judged that all depths in the range 0 .01 to 2.0 should be rated as l .0. The resulting curve (not shown) was used in all computer simulations. Sibstrate utilization and preference curves are given in Figures 118 and b. It should be noted that the insect utilization and preference curves may contain significant errors. The reason for this will be treated in the discuss ion section . Insect Habitat Availability The IFIM model used these preference curves along with the stream mapping data to arrive at estimates of Ephemerella spp. and Hydropsyche spp. habitat present at flows of 6 to 30 cfs. 42 .nmma .oese you magnum men» pandas: pooh» ezonm ”as mmeoHa flaw _. £22. on so cu am 9— m_ N— _ u» _ _ a .u “22:33:: 25:22. bah—mm: ouhctuhmm ..... 22.—coax»: 3:3. ...-:51: ..mzhomllu om— m- Dew mmN omw mum ovm mom can u: ( '.I.:IOOO |-°/'.l.:l'OS)V3th' .|.V.I.|8VH 43 .nmm. .sase you museum asap swoops: anon» nacho up» mmeeHa moor .>.=.=. 8 mm «a 2 m: 2 S s . a _ p _ _ _ _ b _ «28:3:me finer-:3. 2:35.. ouhcfhwm ..... ”22.—cozy: 1:3. kahuna... .Eahocll. om— mam Dew mew omN mam oem mom omm m: (13000 I» /'.I.:|'OS)VSHV lVlISVH 44 .mmmp .pmsms< no“ mmfinmm mad» papflpms anon» nzonm ”on mmpon. ”ma w Armani: fin cu mu «N m. a“ a“ o_ 5 ¢ _ _ _ _ _ _ _ _ . ... VQ‘ Azo_h¢o~m¢~ haozh_z. p¢_.m¢= ou~¢=_hmu ..... .ZDMPMMPMMM :wuz.IMMMPM¢: Ampwm¢.||. .08. EN 0¢N mmN 0mm an own man can m: (1:!000 l- /'.I.:I'DS)V38V .lVJJSVH ..- ~1- Amrtv 9.24!!- ~...—..., 45 on 0N . _ .mmmw .HmnEmpnmm you mmaymm mafia papapwn 9:099 czonm “up mmanm pomp.¢mmzmhamm mu «N m— mu mu o“ h _ _ _ _ _ _ _ .zouhco.mm~ haozhuz. kahuna: ouwcz—wmm ..... azouhcouzmn rhuzu hcpuocz 4c:~o¢.lll --A om— maw 0¢N mmw 0mm wwm D¢m mom can m~¢ ('l:I000 t /'.I.:I'OS)V38V lVlISVH 46 .4mmp .onsu you mmfinmm mag» pmpfipwn psop» nzonm “mp mmson mm b mu _ NN _ couplmzas mu _ o— _ m— _ o— F om— w—m oww me emu mum ovm mom can w: (11.5000 l- /'J.:I'OS)VSUV .LVLISVH ‘47 an .wmmp .haah Hon mmanmm mafia papfinm: p509» axonm "mp mmwam voaw .>.==. mm mm mm mg 0— ma O“ b v _ _ — _ . _ _ b _ .zo_~¢c~¢¢~ haozhuza bah—mm: ouhczuhmw ..... azo~h¢¢_¢xu thus. bah—mm: 4c:~o¢.lll .---.‘..~ om— w—N ovN www omN mum O¢m mum omm w~¢ ('l:|000 L /'..|.:I'OS)V38V .LVJJSVI-l 48 :39 $263 you $28 2.3 9333 39$ :3on “E. HESS . can 9 .hm=0=< on mu ON mu cu m _ _ p _ _ — 23:33:: 25:5: bazaar awbczuhmu ..... 22.—guzz— :.:: bazaar #535 Ill cow omN can own 90¢ ('15000 l /'.I.:I'OS) VBHV lVlIBVl-I 49 .39 .nopsopnom you $23 2.3 #332 393 :3on 2; $82 an wN _ v00 — .mmmsmbumm ON _ w— b c— L ooN omN can own 2.: (macaw/mans) vauv .Lvilavn 50 .m>n:o mpHHHpmHampm.npgmu woman“ oflpmsw< "mm mmchm ......mvzhama mé N m; _ m5 _ g _ _ .A N.o v.0 m.o 0.: AJJ'IISV'IIVAV 51 .m>pao hpaaapwaam>m apfiooam> pommna capm5w< "pm MMDme Amat>=004m> m; N m; g m... _ _ _ _ _ N.o ¢.o m.o 0.0 ALI'IISV'IIVAV 52 .obpao hpaaapmaflm>m mpmnvmnnm vmmmca capmsa< "om mmbch MDOO mh._..oo._m> m m.N N m; N o _ L _ . _ c I N.o I to r 9o 1 9o mozmzwummm ..... . zSENSE=II IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 1. fl BONBHSdSHd/NOILVZI'IILH 56 m m Hamnmama m How mesmNmmmnA cum soapMNNHHps apqoofim> Amauv>h.004m> m.— _ N. m.o .------C‘O----.-C-O-‘---'---'---.-‘---‘-‘----"----. uuzmzumwzm ..... zoubc-4~h:.lll .v.o o.o m.o "now mchHm SGNSHSdSHd/NOILVZI'IIJ.“ I_> l 3 SUBSTRATE CODE § ' [ W l 2 - unuznnou a PREFERENCE SONBHBdBHd/NOILVZI'IILH gydropsyche _s_p. for FIGURE 11a: Substrate utilization and preference .- unuznnou a PREFERENCE E N\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\ -'v- I 3 r—i l 2 EONSHSAEHd/NOIlVZI'Illn SUBSTRATE CODE FIGURE 11b: Substrate utilization and preference 92. E hemerella for '59 The weighted Usable Area vs. discharge plot for Ephemerella spp, and fiydropsyche spp. is given in Figure 12. These data were combined with the data given in Figure 6 to produce the Ephemerella.§pp. and Hydropsyche spp. habitat true series shown in Figures 13a-h, and l4a-h, respectively. Test of the IFIM velocity Adjustment Factors were generated for each transect for flows 6 through 30 cfs (in increments of 2 cfs). This yielded a total of 243 VAFs. Each VAF was then compared to the ratings given by Milhous et. al. (1984) (Table 4). Two hundred and thirteen or 87.7 percent of the VAFs fell into the "good" rating. Seventeen or 7.0 percent fell into the ”fair” catagory, 9 or 3.7 percent were "marginal”, and 4 or 1.6 percent were rated as ”poor." The IFIM predicts not only the total amount of habitat available to a particular species, but also where that habitat is found in the stream. If the model is working well, fish should be located in the places the computer predicts there is habitat. Figure lSa-c is a map of the study section detailing fish locations as they occurred throughout the summer and computer predicted habitat. This is a qualitative test of the computer's ability to predict brown trout habitat. A quantitative test between habitat quality at each transect and the number of fish found at that transect was made using the nonparametric statistic, Kendall's Tau. Table 5 details the test and 60 on .m>hoo mmuosomfio msmnm> wopfipo: «cocoa caposd< “NF mmDon .muo.moc<:om_o 3 8 .N NN 8 S 2 2 No 2 m m . _ _ _ _ _ _ _ _ _ a m ...r N ssssss T. o \‘I!\\‘\\ttt|.\ ..l m 1. S 1 : .am arbmmwzuzau ..... ..a £32292] . N_ m— 000l X ('.L:lOOO|-/'.|.:l'03)V33V lVllSVH 61 cm .nmm. .que you mmanmm wsN» pmpann . m maaonmaoemm ”an. mmstm mom _. .mzzfi bN eN “N o— m— N— . _ p _ P b .zo—»c¢~x¢_ paozhmzu pap—ma: ouhczuhmu ..... .zcuhccumz— :h—zu kahuna: Acshoc.lll 0— ~— N~ 000I- X ('15000l/'.L:|'OS)V38V .I.V.I.l8VH @— 62 .89 .33. n8 328 was. 3333 .mm mfimnmamgmm 3m. 55me amaw .>..==. S NN mN NN 8 S 2 S N. v . p .. _ _ P p _ _ .zozccams 242:5. .29—“ma: cupcxzwm ..... .zahguzs 3:3. hczmcz Jcahocl n: S N— m— 000 l- X ('ldOOO L /'.I.:I'OS)VSHV .LVlIEVI-I 63 ~m .mmmP .pmsmsa ON _ wN _ you mmwnmm mafia vopapm: . m mHHmHmsmnmm ”amp mmbch mam _. ....m:0:< NN m— @u m— D" h v ._ _ . — _ _ p b .zo_»¢¢~¢¢~ bacthuza hchumc: owhczuhmm ..... .zo_h¢¢~¢¢~.:huz. bah—ac: Acabo¢.llu Du u“ N— n— 000 l- X ('1d000 L /'.L:I'OS)VEHV .l.V.|.I8Vl-l 64 .mmm. .nmpampaom you moflnmm mafia ampNnmn . m wHHwNmamzom ”an. mmsuHe noaw .cmmsmbamm oN wN NN mu m— n— o— b v n _ _ — — — b p _ _ Azozguzs 25::3. .Ezocz amber—ram ..... I N— .zozgazg .33: ...—#30:: ..EP—oclll O) 000l X ('13000l/‘ld'OS)V38V .|.VJ.|8VH «— 65 .emm. .meza mom weapon mafia papapmn . m maaonmammmm «mm. mmpuHm VGGF .mzafi oN mN NN m— 9 mu 2 h v u _ b _ h _ r _ p _ m u .h 1 m r m h>ll||l4 I c— ...: 1 Nu .m— 000l- X ('15000l/‘ld'OS)V38V lVllSVH -‘. _5-':..e!a-_L;_‘_=_ _e _. 4 ‘~ n- 4—-..4. __ 41* 66 .emme .Naze you «magma oaN» pmpNnmn . m Redeemawemw ”on. mesaHm com .. .>.:=. S 8 3 NN 8 2 2 S N . _ _ _ p P _ P — _ 22.3.35: haozhuxv .Ezmc: OPE—Shaw ...--- “22.—5:5: :23. .29—acct acahuclll @— 3 N— ha 000l X ('.l.:|000|-/'.L:I'OS)VBHV lVlISVH 67 .39 $253. 8.... 82.8 2.3 “.333 .Mm «Sonoaosmm ”mm. $5ch coop ....m:o:< on B 3 m: 3 m _ . _ _ r r _ 2222252 .59.:: bah—ac: =93:me ..... 22:32:: 2:: .5205. #55:. .II... cu N— ¢— 05) VBHV lVLISVH com X ('ldOOOL/‘id .39 $3533 up... 828 25» 933a: .qm «3983 2mm 32 $5ch coop.¢mm:mhamm on mN oN mu cu m _ _ _ _ b 68 a: N— oam X ('ldOOOl/'_.|.:l'0$) vauv lVllGVI-l '69 .mmm. .oaze new mmanom mafia penance . m mnohwmoneam “we. amouHN 000.. .mza... on NN «N “N no we N, _ _ _ _ c _ «22.23qu .25:::. kahuna: omhcz:wu ..... ”20:35.5 ...:2. bazaar ...cahocll. Mu 000l X('.L:1000|-/'.L:I'OS)V38V .LVLIBVH 7O .mmmw gash. no.“ moanmm 93v amazon . m mnowmmononm 2:2. 3:83 Omar .>..=.=. um ON wN NN mu o— m— o— h v u — _ - P _ L F — — w I N. ... oo/Itn s. ol . T. D / uuuuuuu t It: I m I on I um Azo~bcaumma hzozhuz. bah—mm: Duhctahmw ..... 1 Nu 3207:5qu2 rpm... hchumcz ..Ezhocll mu 000 l- X ('l:I000 l- /'J.:I'OS)VSHV .LVJJGVH 71 .mmm. .pmsmze Noe «may». asap panacea .mm_mmmxmmmmmxm nee. NNNNHN 33.5303 _m oN mN NN m. o_ a" . o“ N v a _ _ _ _ b _ _ _ b o N o m . o“ - ~— .zo_N¢¢_¢¢_ N=o2N_:. N¢N.N¢= ouN¢=_Nmu ..... . No .zo_N¢¢_¢¢_ zN.z. N¢N_oc= JcaNuc.uuu mu 000l X ('15000l/‘ld'OS)V38V .LVLISVH 72 .89 ..Hmpampaom .8.“ 82mm 2.3 9.332 .mm «gouache? 33 ESE 0N _ GOO w .mmmimhnmm mN NN 2 .8 2 S .N N P _ _ — — b _ _ .zcuhguas .5292»: bah—mm: 99—52me .2... 35:33:: 3.2:. 233:: ...-5.8:! o— u— N— n— v“ @— 000t X ('.l.:|000|-/'.I.:.I'OS)V38V .LVLISVH 73 4m? .33. n8 333 2.3 9.8.33 .mw. onoummonexm 3.: $er eoaN.mz:a SN. .3 NN 2 8 2 8 N e _ _ _ p. _ _ _ _ — - 000L X ('15000L/‘15'09)V38V .LVJJSVH 74 3 4mm. :33. n8 338 2:3 ”.332 .mm meoaoonflm £2: .55on vamp .>.=.=. oN mN NN m— cm n— o— b _ p _ b _ _ r _ 35:53:: .5295... bah—om: ouhczzwu .-..-. 38:83:: :23. bazaar acahoclll 000l- X ('1:l000|-/'J.:I'OS)VBHV .I.V.|.|8VH _, .. 75 .emmN .vmomsa.nom oedema mad» povunon .mm mnonmoanm "meN mmbon vcaw.hm303< on wN oN m— a: _ _ — .b _ 28.23;”: 251:: 2:85.. awhczzmu ..... 22:33.5 3:: bah—ac: 4.250: I @— N— oom X ('laooou'u'OS) 'vaav lulavn 76 .¢m¢w .Hopaovmom How moanmm oaav vaapdn .mm onohmnohuwm un¢w HMDOHM coup .cmmfimhamm OO ON ON mu O— m _ . _ — p 5 <2 <1... O OOOI. X ('laooou'la'OS) vauv .LvuaVI-I . 77 Table 4: Velocity Adjustment Factor ratings as given by Milhous et. a1. (1984). Velocity Adjustment Factor Rating 0.9-1.1 good 0.85.009, 1.1.1015 fair 0.80-0.85, 1.15-1.20 marginal 0.70-0.80, 1.20-1.30 poor less than 0.70, greater than 1.30 very poor 78 closed circles: 1983 fish locations open circles: 1984 fish locations rectangles: habitat locations numbers indicate transects 1 FIGURE 15a: Map detailing downstream computer predicted habitat locations and actual fish locations. 79 closed circles: 1983 fish locations open circles: 1984 fish locations rectangles: habitat locations numbers indicate transects. FIGURE 15b: Map detailing midstream computer predicted habitat locations and actual fish locations. 80 18 closed circles: 1983 fish location- open circles: 1984 fish locations rectangles: habitat locations numbers indicate transects 15 FIGURE 15c: Map detailing upstream computer predicted habitat locations and actual fish locations. 81 N .¢.o umo.o nmnnam as use madness «nNm.o u mw_u Hmnmmmxuzma «oenmNumNNum mnoupobnompo confine Ho Hopes: mspuz ones: .A N IZ Zvl +mNum +mummaampo» mN O NP mw mamml‘mtfid'FPPO m 0 Pm.N¢ ou.mw 0N.NN om.m mm.m No.m mo.N Nme.o mom.o MNN.O NN.O mmo.o mo.o P PVI'V‘MLDKDNOC‘NMI‘Q 1-1- F gnaw oobnmmno no Hones: Hovoonm a spa: ombnmmpo porno msHmMonona poomnmnv mpoomsonv pooodmmpsm Ho Hones: swam no Hopes: ma doxgmn <53 .oma m.HHooooM pomaodmmooo oedemamnnoo on» ma coma poop one .voomoonv Pan» :« conspnoo noun no Hones: as» com poomoonv o no undo“ 40: omonobm one noozpop oooooamaomam pom pump 4 um Hands 82 results. As given in Table 5, there was a statistically significant relationship between average WUA and the nunber of fish (Tau= 0.513, N=13, alpha= 0 .05) . Table 6 details the canputer's ability to simulate the depths, velocities, substrates, and covers as they occur throughout the study section. As given in this table, the cauputer was not always able to accurately predict these parameters . The Relationship of Habitat Availabilty to Pbpulation Density A simple linear regression was used to relate the average daily amount of habitat present over the number of days prior to the sanpling time and the population estimate made at the sampling time. The resulting plots for the 1983 and 1984 data are shown in Figure 16a and b. As seen fran this figure, no statistically significant relationship could be demonstrated for the 1983 data (r2= 0 .09, n.s . at alpha= .05, d.f .=6) . A very highly significant relationship did occur during 1984 (r2= 0 .837, alpha= .01, d .f .=6) . The Relationship of Habitat AvailabiltLand Growth Rate . A linear regression was used to relate the average amount of habitat available over the number of days prior to the sampling time and the average growth rate of the fish captured at the sampling time. This could only be done for the 1983 data because insufficient recaptures were made during 1984 (Figure 17) . No statistically significant s3 -N.N¢ ¢.ap Nn m— an ¢m HOFOO ovauvnpso % -voonuoood usoquobuooao no won-so- n -aoouuoo onouaophoopo no Hannah- h¢w ¢.nN m .N.np ¢.ON NN ¢.Ov >.nm ¢.ON n. p.wm O.NN O n.ov n.¢n m.oN v. O.v¢ v.mN nN N.ov ¢.nN ¢.nN m ¢.mN v.mN NN N.Ov * .Iso % osouvo>uoopo % .650 x occuaopuouno no hopaao no Hopes: VEHOOAH> 29mm: .oou coo .vu cu consumes hauooaob no canoe Hooves use can hauoouoh no gvooo oovoucoua unannEoo as» soozuoa oocououuuo 0:» «coconnou :ESHoo vHoH you on» :m mooHo> .xoono snub :« mounnoooo haaosaoo nachos coo moaonvoaoo .oouauooam> .ozvooo on» oaoasaqo on hauadno o.no»:neoo one no hhoEESm um canoe 84 .mmmN you avamnmo soapmasmon mammob hpaawnwaam>m popanon Ho scammmnwmm «MON mmwam AOLOOO w \..Pu—.Om. avaaapoaaobm popwnon Ho :oammmnmmm “DON HMOme ...NuoooN\.N.._.om.m advanms yo :oammwnmom ...—Loco w \.Om.vee and Cochnauer, 1977) . Estimates of aquatic insect habitat availabilities are subject to error and will only be used to detect the possible impacts of irrigation. A Test 9f thg IEIM The Velocity Adjustment Factors provide strong evidence that the IFIM model is completely applicable to the midwest. Nearly ninety percent of the VAFs were in the ”good" range. The VAFs that were only I'fair" or "marginal" nearly all occurred at flow volumes below 10 cfs. This means that flows lower than this were not simulated as well as the higher flows . This is a reflection of the quality of the data supplied to the ‘conputer, not the computers ability to simulate low discharges. More evidence of the IFIM's applicability to the midwest is found in Figure 15 . Virtually all the brown trout captured throughout the summer were found in, or directly next to, spots predicted by the computer to be trout habitat. Descrepencies at transect 8 resulted from an error in stream mapping . A piece of down timber was missed and this misinformation fed to the computer. This resulted in many fish being captured in seemingly non-habitat areas. In actuality, if the stream 94 mapping had been accurate, the computer would have indicated habitat to be present where those fish were caught. The Kendall's Tau test provides quantitative evidence of the computer's ability to predict trout habitat. As given in Table 5, a statistically significant relationship did occur between the oorputer's measure of habitat quality (WUA) and fish numbers. This indicates that the conputer can predict with some accuracy not only the location of fish habitat, but that habitat's quality as well. The computer's ability to accurately predict depths, velocities, substrates, and covers is called into question from the data in Table 6. As given in this table, the computer's predictions are substantially inaccurate in some cases. However, in the majority of cases, the computer was able to estimate depth and velocity to within three-tenths of a foot and foot per second, respectively . The computer was also correct in predictions of substrate and cover in the majority of cases . The errors that do occur are from two sources . The first source of error is human. It is difficult to accurately measure depth and velocity in log jams and under undercut banks. Thus, the information used by the conputer to simulate flow conditions was, in some cases, inaccurate. Better data will result in more accurate sinulations . The second reason the predicted and actual values do not always match has to do with the way the data were collected . The depth, velocity, substrate, and cover data were collected where a fish was 95 located. Thus, this was not a random sampling . The fish are going to actively search for the best conditions possible. Thus, small variations in the stream channel morphology (too small to warrant detailed mapping) will influence where the fish is found. This point is demonstrated by the fact that computer predicted depths were, on the average, less than the actual depths the fish chose (1 .08 ft versus 1 .22 ft). Predicted velocities were, on the average, faster than the velocities the fish were able to find (0.703 fps versus 0.689 fps). The fish were also able to find small areas of undercut bank or down timber that were not mapped. A similar situation existed with substrate. Thus, the fish are able to key in on small variations in the physical conditions present which would require an excessive amount of time to map. This results in a systematic disagreement between predicted and actual depths, velocities, substrates, and covers. Given these errors, the computer is doing an adequate job of simulating the physical characteristics of Fish Creek . Aside from testing the computer's predictive reliability, the data in Table 6 provide insight into which physical parameters are the most important determiners of brown trout habitat . The parameters which the fish key on most strongly when selecting a resting area must be those the computer simulates the least accurately. The reasoning for this is as follows . It is impossible to accurately map every small variation in any of the physical parameters. However, the fish can very easily find these small areas of suitable habitat. The more strongly the fish are 96 searching for, as an example, a certain depth, the more likely it will be able to find a small area of suitable depth not mapped during data collection. Thus, the computer prediction will not match the actual depth where the fish was found. Using this logic, it is clear fromlTable 6 that fish are keying in on cover very strongly. Depth and velocity seem to be of somewhat lesser importance. Substrate does not appear to be an important parameter. In future applications of the method, it may be worthwhile to use only depth, velocity, and cover in the simulations. Naturally, if a different life stage is being considered (spawning habitat, as an example), substrate may become a very important parameter and stould be included in the simulation. There is no theoretical discrepancy in the IFIM model precluding its use in the midwest . The above evidence provides empirical support of the applicability of the method to the region. Indeed, Carlson (1979) reported that the computer's predictive reliabilty is a reflection of the reliability of the data it is supplied. As long as well trained people are responsible for the data collection, the model is applicable to virtually any region. Fish Habitat Estimations Given that the model is working well on Fish Creek, it is now possible to make estimates of fish habitat loss due to irrigation withdrawals . The following discussion of habitat losses is designed to 97 present the type of analysis possible given the data described in the Results section. The first step is the examination of the WUA.vs. discharge plot (Figure 5) . As given in.Figure 5, brown trout habitat in Fish Creek changes rapidly in the range 10 to 14 cfs. Relatively small changes in discharge will result in large changes in available habitat. Discharge records from a U.S.G.S. gaging station located in Carson City, downstream.from the study section, were regressed against discharges measured with the water level recorder installed in the study section. From this regression and the discharge records made at Carson City in previous years, it is possible to estimate the summer base flow in the study section over the years 1974 to 1980. This estimate is 12.1 cfs. This estimate falls in the range of discharges most critical in terms of habitat.(FTgure 5). Thus, in a normal year, any irrigation ‘withdrawals will have a significant impact on available habitat. The slope of the line between 10 and 14 cfs is roughly 34 square ft.flcfs. Irrigation rigs on Fisthreek have a capacity of 600 gallons per minute (gpm) or 1.3 cfs. Thus, each rig on Fish Creek removes not only 600 gpm but also 44 square ft..of habitat/1000 ft. of stream.during an average summer base flow period. The study section, while only approximately 950 ft. long, was chosen because it is representative of the headwaters of Fish Creek. The habitat estimates made in this section should represent a stream length of approximately 37,000 ft. of Fish Creek. Thus, it is possible 98 to estimate the total number of square ft. of brown trout habitat present in the headwater section of the stream.at the various flows of interest. Using this logic, the 44 square ft./1000 ft.of stream.removed by a single irrigation rig operating at full capacity, translates into a total habitat loss of 1,628 square feet. The above discussion is largely theoretical. The actual habitat losses will depend on how many rigs are running at any one time, at what percentage of full capacity the rigs are operating, and what the actual discharges would be without irrigation. While the discharge recorder will not give detailed information on the number of rigs operating or at what capacity they are running, it does provide a composite estimate of water withdrawals. Actual streamflow is measured by the recorder, and it is possible to estimate the discharge that would have occurred without irrigation. This estimate is based on discharges recorded during those times of the day when little or no irrigation is occurring . Both estimates can then be converted into a habitat estimate . From these estimates a habitat time series with, and without, irrigation can be generated. This was done by month for the summers of 1983 and 1984 (Figure 7a-h). Using these figures a fairly complete analysis of habitat loss can be made. To quantify habitat loss over time, the concept of a habitat-day will be used. A.habitat-day will be defined as 1 square foot of habitat available for a period of one day. Thus, if streamflow conditions are such that an average of 400 square ft ./1000 ft. of stream 99 of habitat are present on each of two consectutive days, it can be said that a total of 800 habitat-days were present. If irrigation withdrawals accounted for a loss of 200 square ft ./1000 ft. of stream of habitat on each day, a total of 400 habitat-days were lost. This type of analysis was performed with the data contained in Figure 7a-h. Fish Habitat Losses in 1983 Irrigation began on June 22 of 1983. Eton this time until the end of the month, a total of 159 habitat-days were lost. Averaged over the month, this is a l .7 percent reduction in habitat-days. This loss does not seem excessive. However, the average habitat loss that occurred at the time of irrigation withdrawals was 7 .1 percent. In other words, during those times that irrigation was occurring, habitat availability was reduced an average of over 7 percent. This value may be the more important one if fish are responding to short-term, low habitat availability events, rather than longer-term, average habitat availabilities . July was an extremely dry month. Irrigation occurred on all but six days during the period. The total habitat loss was 1,171 habitat-days. Averaged over the month, this is a 12.2 percent reduction in available habitat. Taking into account only those times when irrigation was occurring, a 16 .0 percent reduction in habitat occurred . August was also fairly dry, and irrigation occurred on 11 days of the month. The total habitat loss was 312 habitat-days. The monthly 100 average habitat loss was 3 .6 percent. The loss during actual irrigation times was 11 .1 percent. Irrigation ended for the season on September 4 . The habitat loss due to irrigation was 82 habitat-days. This is a 0 .9 percent reduction averaged over the month, and a 9 .0 percent reduction during actual irrigation events . A convenient way to summarize the effects of irrigation in 1983 is the habitat duration curve (Figure 18a). The habitat duration curve describes the amount of time that a particular amount of habitat availability was equalled or exceeded . For example, the actual amount of habitat that was equalled or exceeded 50 percent of the time during the summer of 1983 was 303 square ft ./1000 ft. of stream. The estimated 50 percent value without irrigation was 346 square ft ./1000 ft. of stream. As given in Figure 18a, irrigation substantially increased the occurrence of low habitat availability events. It is probable that low habitat availability events are an important determiner of carrying capacity. If this is true, irrigation must be lowering the brown trout carrying capacity of Fish Creek . This will be examined in detail later in this discussion. The type of data given in Figure 18a can also be valuable if a certain ”critical habitat availability" is known. The habitat duration curve will describe how often this habitat availability will not be present both with and without irrigation. A discussion of this follows later . 101 .89 £3838 8. 39 .m 28a cannon m5 n8 226 83956 332% 39 go: Omnmmoxm CO nm44<30m was... ...—0 m0¢=u 09—5:me ...--- m>¢=o 22.—$50 .53.“... III o cm— SN emu own now can an own cum com com a: ('LdOOO I- /'J.:I'OS)V38V .LVJJSVH 102 .emms .on nanampaom on emm. .. onus cannon can you osnso nonsense passes: Omnmmoxm CO ONAJ<=OW WEE. n—O .m0<.—.zw0¢ma oc— an ea or cm on ov on ow cu o is e _ _ w _ _ a u . x zcsh¢o~¢¢_ c=o2__z machzuuzua ..... ., zo.ucs_¢¢_ z».x macbzmozue.ulu name Habch cm— H" SN W omu M V own I. V arm 8 ....— omu .v cu 3m 0 m. can .I. / own .L m chm nu ... can mu adv 103 :vmmp end nmmw Ho mamas—5m on». How mega scavenge vmpdpmm “amp ape“; Omammuxm CO nm44<=cm m5: "—0 m6 no.9? >m.¢ .mm mnoflmmcnoae 00.0 Fm.w 0¢.w 00.0 0¢.0 mm.P u0.m vm.0 00.0 mo.m bw.u 00.0 em.e 0v.u Fm.e Pm.0 . m mHHonoEoAmm . m oaohmmonoam coevmmenue no mood vooonom memos on» Hobo mood unoonom emm. .nmpampaom some .pmsms< some .sess soar .onse meme .HonEouoom mom. .umsmsa mom. .ssss mom. .ocse : 0: .eemw one meow new moaned agape: accuse 033:5 no bums—saw um manna 108 Trapicyna, 1981; Eley et. a1., 1981) . At least two others have specifically attempted to relate the availability of certain physical characteristics to trout populations (Lewis, 1969; Wesche, 1974) . This study is one of the first to attempt to verify the IFIM model based on trout population responses to fluctuating habitat availabilities as predicted by the metrod. As stated in Results, a significant relationship was found to exist between habitat availability and population levels for 1984 . my then did this relationship not show up in 1983? I believe the reason was rapidly fluctuating habitat availabilities during the summer of 1983. This resulted in a lag period between the time a certain habitat availability was present, and the time the fish responded to this habitat availability. This is demonstrated in Figure 19, which traces both habitat availabilities and population levels through time. The lag period described becotes evident when the data are arranged in this manner. If this lag period is incorporated into the regression, the resulting r2 value falls just short of being significant at the 95 percent level (the calculated and test values are .555 and .570, respectively, with d.f .=5) . There is a significant relationship at the 90 percent level . This lends support to the contention that it was rapidly fluctuating habitat availabilities that obscurred the relationship between habitat and population levels in 1983 . Brook trout were also present in the study area at times. Assuming that this species has habitat requirements similar to brown trout, it .109 HSI:I :IO HSBWHN mao>oa noepmaoooo one moepeaeomaembm append: mo codename Hmoamoaosonno "mp mmeeHm .mmmF .psonp szonn no mh335 on mass 0 e _ _ eeu u 1 .. mum e 4 .. 8“ H so 00 v \s oo/ m m .4 a. .x - new m_ fill \s to. \II- ... .. ... s m t ,.. 1 8mm. OF .... / I MN” m m u. 1 . r can C SEES SS: 2:25 .3 59:5 3:3: finesse some“: I: e. 1 . - use 3 eee 110 may be that total trout population levels are more closely related to habitat availability than are brown trout population levels. In order to test this, total trout populations present at the sampling date were regressed against average habitat availability over the two week period prior to the sampling date. It was found that this significantly improved the habitat availability-population level relationship found for brown trout alone. In 1983 (with the lag period included) the r2 value increased from the previously reported .555 up to .834 . This value is highly significant (alpha= .01, d.f .=5) . The 1984 r2 value increases from .837 up to .904 . This value is very highly significant (alpha= .001, d.f .=5) . These data verify that the IFIM generated habitat estimations do have biological significance . A reduction in habitat as estimated by the method will result in a reduction in trout population levels . While the habitat reductions reported were temporary, I believe that a permantent reduction would lead to a similarly permanent reduction in trout population levels . This has probably already occurred to the extent that past irrigation withdrawals have increased the occurrence of low habitat availability events. While it seems that the fish are responding to habitat availability in part, temperature must also be exerting an influence . Low habitat availabilities result from reduced discharge, and this leads to a more rapid warming of tre stream. airing the lowest discharge periods in July of 1983, water temperatures reached 70 degrees. It is at this temperature that brown trout will begin to actively search for cooler 111- water . This may result in an upstream migration which could reduce population levels in the study section. However, during the summer of 1984 water temperatures never rose above 63 degrees . Most temperatures were below 60 degrees. Thus, the influence of temperature was probably not substantial . Without the influence of tetperature, the relationship of habitat availability to population levels was even stronger. Irrigation then, can impact fish in at least two ways: first, by reducing the amount of habitat available and second, by prototing a warming of the stream. Fish Habitat Availability and its Relationship to gm Bates Average growth rate over the period between sampling times was regressed against mean habitat availability over the same period. This could only be done for the 1983 data due to insufficient numbers of recaptures during 1984 . No significant correlation existed (Figure 17) . I feel the lack of a significant correlation resulted from the inability to obtain an accurate estimate of average growth rate . There were soretimes as little as three fish from which to make growth rate estimates. This small sample size, while virtually the entire population, allowed the growth rate estimates to be heavily influenced by single individuals. It appears that individual fish did not respond to the same average habitat conditions in the same manner. Sore individuals experienced positive growth throughout the season, sore simply maintained their weight, and others had negative growth rates 112 (Figure 20). This type of highly variable growth rate has long been recognized, and is probably a result of high water temperatures (Brown, 1946). Thus, the problems caused by small sample size were magnified by the highly variable nature of the fish themselves. This resulted in average growth rate estimates having little biological significance. This is unfortunate because, during the data collection, it seemed obvious that extended low flow conditions were causing reduced.growth rates. During these low flow periods, a majority of the fish‘were experiencing negative growth rates. During times of greater habitat availability most or all of the fish experienced positive growth. To demonstrate this point, the percentage of fish experiencing positive growth was regressed against average habitat availability (Figure 21). This regression was very highly significant (r2 =.874, alpha> 0.001, d.f.=7). Thus, it appears that reduced habitat availabilities result in a greater proportion of the population being unable to maintain positive growth rates. This is certainly of biologic importance. The "break even” point for the population is approximately 260 square ft./1000 ft. of stream (Figure 21). Below this habitat availability greater than 50 percent of the population will lose weight, while above it the majority of fish will grow positively. This is where the habitat duration curve (Figure 18a and b) becomes extremely useful. The 1983 data (Figure 18a) will be used to demonstrate this point. As given in Figure 18a, irrigation was responsible for greatly increasing the number of low habitat availability events in Fish Creek. 113 .mmmp an ozone axons =Hmoeaep= neon no mafia assess» games; as omamno “ow emsoHe mh42. em wze... L b _ _ on To. .2... \ . .-8 .\ .......... u on ‘.|‘I|I‘Il\ ‘ t\\ \ (swwfiuHelaM ..eo \\ .. 8 xii. om: -..- . - mew ...... e2 II: e: 114 .mmmP no assess ms» nu epeaanaammpa powwow: msmnob ”Ewen: moanemm moepmflseoo 950.3. 86.3. 23. .Ho ommpooonom "Pu mmeon ..hu0009\.bn_.0w.h_.=mhso mononnoooo mo heedenmoono moose» ease me hpeaepmaembm powwow: haspoos one: "euu mmeeHm mozmmceooo “.0 >._._.:eh_.:mw4 ze.h¢e.xm. hzuzzee :h.z «u.e.eueeeemm heh.e¢:.lll we. w.u mvu wbu mew mew mew mew euv ('.I.:!OOO l- /'.|.:I'OS)V38V .l.VJ.l8VH 119 without irrigation is nonexistent. In contrast, with irrigation, the probability of occurrence of this habitat availability is 0.381 . As more and more acres are put into irrigation, the probability of occurrence of very low habitat availabilities will continue to increase. As this happens, the low habitat availabilities previously present only in an extremely dry year will occur nearly every year. Fish populations will respond in the manner described earlier, and a once productive fishery will be lost. CONCLUSIONS The Instream Flow Incremental Methodology proved to be applicable to a typical Lower Michigan stream. Internal quality control checks referred to as Velocity Adjustment Factors indicated that the IFIM could accurately simulate tie hydraulic characteristics of a Michigan stream. Corparisons made between corputer predicted habitat locations and actual brown trout locations demonstrated the methodology's ability to accurately simulate brown trout habitat. The accuracy of the method in any instance is a reflection of the quality of the data supplied to it. Simulations performed by the IFIM indicate that, below 16 cfs, small reductions in discharge will result in substantial reductions in available brown trout habitat. A one cfs reduction in discharge was shown to result in a 34 square ft ./1000 ft. of stream reduction in habitat. Actual mean monthly habitat losses ranged from 0 .85 percent up 120 to 12 .2 percent. Habitat reductions during periods of irrigation ranged from 4 .0 percent to 16 .0 percent. In general, irrigation was found to substantially increase the frequency of low habitat availability events as demonstrated by the flow duration curve . July was found to be the most critical month both in terms of naturally low habitat availabilities, and irrigation demands. A statistically significant relationship was found to exist between habitat availabilities and brown trout population levels. Rapidly fluctuating habitat availabilities introduced a lag period, but did not alter the nature of the relationship. Tbtal trout populations (i .e. brook and brown trout) were found to be even more significantly related to habitat availabilities. Thus, habitat reductions resulting from water withdrawals for any purpose will reduce trout populations . The relationship between average brown trout growth rate and habitat availability was found to be not statistically significant. The reason may have been lack of sufficient data to accurately determine average trout growth rates. However, the percentage of the population experiencing a positive growth rate was found to be statistically significantly related to habitat availability. Thus, based on this preliminary data, it appears that reduced habitat availabilities will lead to reduced brown trout growth rates . Aquatic insect habitat losses were estimated utilizing a limited preference data base. The simulations show that irrigation withdrawals will have less of an impact on this portion of the stream community, 121 cotpared to fish populations . Mean monthly habitat losses for Hydropsyche spp. ranged from 0 .54 to 8 .51 percent. Losses during irrigation periods ranged from 4 .57 to 11 .05 percent. Mean monthly habitat losses for Epheterella spp. ranged from 0 .34 to 5 .02 percent. Losses during irrigation periods ranged from 1 .55 to 6 .35 percent. The most important conclusion to be drawn from this study is this: current levels of irrigation in Michigan are having a detrimental impact on inland trout fisheries in terms of habitat, population levels, and growth rate. As irrigation water withdrawals increase, the magnitude of the impact will also increase. If left unchecked, irrigation will result in the degredation or loss of a significant portion of our fishery resource. Only through documentation of these potential losses, followed by strong legislation designed to protect the fishery resource, can we hope to reduce the problem to acceptable levels.. The Instream Flow Incremental Methodology provides a useful tool by which these losses can be quantified.p§§9§g they actually occur. Wbuld it not be preferable to avert the losses before they occur, rather than mitigate them afterwards? LITERATURE CITED LITERATURE CITED Anonymous. 1977. Dams and river regulation deadly to fish. Earth- Sci. Rev. 13(4):393. Anonymous. 1979. Evaluation of the effects of different streamflow releases on trout habitat below hydroelectric diversion dams: two case studies. California-Nevada Wildlife, 1979:55-68. Armitage, P.D. 1978. Downstream changes in the composition, numbers and biomass of bottom fauna in the Tees below Cow Green Reservoir and in an unregulated tributary, Maize Beck, in the first five years after impoundment. Hydrobiologia, 58(2):145-156. Avery, E.L. 1980. Factors influencing reproduction of brown trout above and below a flood water detention dam on Trout Creek, Wisconsin. Wis. Dept. Nat. Res. Rep. 106. 26pp. Bedell, D. 1977. Irrigation in Michigan. Michigan Dept. Nat. Res. Rep. 79pp. Bedell, D. Pers. Comm. Mich. Dept. Nat. Res., Water Management Division, Lansing, Mi. Blahm, T.M. 1976. Effects of water diversions on fishery resources of the west coast, particularly the Pacific Northwest. Mar. Fish. Rev. 38(11):46-51. Bovee, K.D. 1982. A guide to stream habitat analysis using the Instream Flow Incremental Methodology. Instream Flow Information Paper 12. U.S.D.I. Fish and Wildlife Service, Office of Biological Services. FWS/OBS-82/26. 248pp. Bovee, K.D. and T. Cochnauer. 1977. Developement and evaluation of weighted critieria, probability of use curves for instream flow assessments: fisheries. Instream Flow Information Paper 5. U.S.D.I. Fish and Wildlife Service, Washington, D.C. FWS/OBS-77/63. 39pp. Briggs, J.C. 1948. The quantitative effects of a dam upon the bottom fauna of a small California stream. Trans. Am. Fish. Soc. 78:70-81. Brown, M.E. 1946. The growth of brown trout (Salmo trutta) II. The effect of temperature on the growth of two-year-old trout. J. Exp. Biol. 22:145-155. 122 123 Carlson, C.A. 1979. Evaluation of four instream flow methodologies used on the Yampa and White Rivers, Colorado. Prepared for The Bureau of Land Management, Denver, Colorado, and The Western Energy and Land Use Team, Fort Collins, Colorado. 53pp. Cooper, C. 1984. Instream Flow Needs: Fish Creek. Unpublished manuscript. Dept. of Civil Engineering, Michigan State University. 32pp. Corning, R.V. 1969. water fluctuation, a detrimental influence on trout streams. Proc. 23rd. Ann. Conf. SE Assoc. of Game and Covich, A.P., W. Shepard, E. Bergy, and C. Carpenter. 1978. Effects of fluctuating flow rates and water levels on chironomids, direct and indirect alterations of habitat stability. In: Energy and Environmental Stress in Aquatic Systems, J. Thorp and J. Gibbons, eds. Technical Info. Center, U.S. DOE, Oak Ridge, TN. 141-156. Cress, C. pers. comm. Dept. of Statistics and Probability, Michigan State University, East Lansing, MI. Davis, J.R. 1979. Die-offs of an endangered pupfish, Cyprinodon elegans. Southwest Nat. 24(3):534-536. Delisle, G.E. and T. Wooster. 1964. Changing the flow regime and its possible effects upon aquatic life and fishing-Middle Fork of the Feather River. Calif. Dept. Fish and Game, Water Projects Br. Admin. Report. 36pp. Doyle, T. pers. comm. Mich. Dept. Nat. Res., Fisheries Division, Lansing, MI. Eley, R., J. Randolph, and J. Carroll. 1981. A comparison of pre and post impoundment fish populations in the Mountain Fork River in southeastern Oklahoma. Proc. Okla. Acad. Sci. 61:7-14. Finnigan, R.J. 1978. A study of fish movement stimulated by a sudden reduction in rate of flow. Fish. Mar. Serv. (Can.). no. 98. 6pp. Fisher, S.C., and A. Lavoy. 1972. Differences in littoral fauna due to fluctuating water levels below a hydroelectric dam. J. Fish. Res. Bd. Can. 29(10):1472—1476. Fraser, J.C. 1971. Regulated discharge and the stream environment. Presented at the International Symposium on River Ecology and the Impact of Man, June 20-23, 1971, Univ. of Mass. at Amherst. 41pp. Giger, R. 1973. Streamflow requirements for salmonids. Oregon Wildlife Comm., Final Job Report, Project AFS 62-1. 117pp. 124 Gislason, J.C. 1980. Effects of flow fluctuation due to hydroelectric peaking on benthic insects and periphyton of the Skagit River, Washington. Ph.D. dissertation, Univ. Wash. 175pp. Gordon, R.N. 1965. Fisheries problems associated with hydroelectric developement. Can. Fish. Cult. 35:55-98. Graham, P.J. 1980. Impacts of Hungry Horse Dam on aquatic life in the Flathead River. Publ. by Montana Dept. of Fish, Wildlife, and Parks, Kalispell, MT. 97pp. Havey, K.A. 1974. Effects of regulated flows on standing crops of juvenile salmon and other fishes at Barrows stream, Maine. Trans. Am. Fish. Soc. 103(1):1-9. Holcik, J. and I. Bastl. 1976. Ecological effects of water level fluctuation upon the fish populations in the Danube River floodplain in Czechoslovakia. Prirodoved Pr. Ustauv. Cesk. Akad. Ved. Brne. 10(9):1-46. Holcik, J. and V. Hruska. 1981. The impact of hydrological regime upon the fish communities of a river. In: Topical Problems of Ichthyology, M. Penaz and M. Prokes, eds. Czech. Acad. Sci. Inst. Vert. Zool.,Brno. 39-40. Holden, P.B. 1980. Will outlet modification of Flaming Gorge Dam help the Colorado squawfish? Summ. Proc. Ann. Symp. Desert Fish Council. 10:37. Hooper, D.R. 1973. Evaluation of the effects of flow on trout stream ecology. Pacific Gas and Electric Co., Dept. Engineering and Research, Emeryville, Calif. 97pp. Hourston, W.R. 1958. Power developements and anadromous fish in British Columbia. In: The Investigation of Fish-Power Problems. H.R. Macmillan Lectures in Fisheries, Univ. of B.C. 15-23. Kalleberg, H. 1955. Observations in a stream tank of territoriality and competition in juvenile salmon and trout. Inst. Freshwater Res., Drottningholm, Sweden, no. 39:55-98. Kovalak, W.P. 1978. Relationship between size of stream insects and current velocity. Can. J. Zool. 56(2):l78-186. Kraft, M. 1972. Effects of controlled flow reduction on a trout stream. J. Fish. Res. Bd. Can. 29(10):1405-1411. Kroger, R.L. 1973. Biological effects of fluctuating water levels in the Snake River, Grand Teton N.P., Wyoming. Am. Mid. Nat. 89(2):478-481. 125 Lambert, T.R. and J.M. Handley. 1980.An instream flow study involving smallmouth bass (Micropterus dolomieui) in the San Joaquin River, California. Proc. Ann. Conf. West. Assoc. Fish and Wild. Agencies. 60:433-442. Lehner, P.N. 1979. Handbook of Ethological Methods. Garland STMP Press. New York. Lewis, S.L. 1969. Physical factors influencing fish populations in pools of a trout stream. Trans. Am. Fish. Soc. 98:14-19. Milhous, R.T., D.L. Wagner, and T. Waddle. 1984. User's guide to the Physical Habitat Simulation System. Instream Flow Information Paper 11. U.S. Fish Wild. Serv. FWS/OBS—8l/43 Revised. 47Spp. Minshall, G., W. Winger, and V. Parley. 1968. The effect of reduction in streamflow on invertebrate drift. Ecology 49(3): 580-582. Mullan, J.W. 1976. Factors affecting upper Colorado River reservoir tailwater trout fisheries. In: Proceedings of the Symposium and Specialty Conference on Instream Flow Needs. Publ. by the Am. Fish. Soc. p. 405, v.2. Orth, D.J. and O.E. Maugham. 1982. Evaluation of the Incremental Methodology for recommending instream flows for fishes. Trans. Am. Fish. Soc. 111(4):413-445. Ottaway, E.M. and A. Clarke. 1981. A preliminary investigation into the vulnerability of young trout (Salmo trutta) and Atlantic salmo (Salmo salar) to downstream displacement by high water velocities. J. Fish. Biol. 19(2):135-145. Parsons, J.W. 1955. The trout fishery of the tailwater below Dale Hollow Reservoir. Trans. Am. Fish. Soc. 85:75-92. Phillips, R.W. 1969. Effect of unusually low discharge from Pelton Regulating Reservoir, Deschutes River, on fish and other aquatic organisms. Oregon State Game Comm., Basin Investigations Sect. Spec. Rep. no. 1. 39pp. Phillipson, G.N. 1954. The effect of water flow and oxygen concentration on six species of caddisfly (Trichoptera) larvae. Proc. 2001. Soc. London. 124:547-564. Phillipson, G.N. 1969. Some factors affecting the net-spinning of the caddisfly Hydropsyche instabilis Curtis (Trichoptera). Hydrobiologica 34:369-377. Powledge, F. 1982. Water: The Nature, Uses, and Future of Our Most Precious and Abused Resource. Farrar Straus Giroux, New York. 423pp. 126 Scarnecchia, D.L. 1981. Effects of streamflow and upwelling on yield of wild coho slamon in Oregon. Can. J. Fish. Aquatic. Sci. 38(4):471-475. Schuck, H.A. 1945. Survival, population density, growth, and movement of the wild brown trout in Crystal Creek. Trans. Am. Fish. Soc. 73:209-230. Smoker, W.A. 1953. Streamflow and silver salmon production in western Washington. Wash. Dept. Fish., Fisheries Res. Papers 1(1):5-12. Solomon, D.J. and D. Paterson. 1980. Influence of natural and regulated streamflow on survival of brown trout (Salmo trutta) in a chalkstream. Environ. Biol. Fish. 5(4):379-382. Spence, J.A. and H.B.N. Hynes. 1971. Differences in fish pOpulations upstream and downstream of a mainstream impoundment. J. Fish. Spence, J.A. and H.B.N. Hynes. 1971. Differences in benthos upstream and downstream of an impoundment. J. Fish. Res. Bd. Can. 28(1): 35-43. Trjapicyna, L.N. and L.P. Kizina. 1981. Ecology of cyprinid fishes in the Volga Delta under regulated run-off conditions. In: Topical Problems in Ichthyology, M. Penaz and M. Prokes, eds. Czech. Acad. Sci. Inst. Vert. Zool. Brno. 143-147. Trotzky, H.M. 1974. The effects of water flow manipulation below a hydroelectric power dam on bottom fauna of the upper Kennebec River, Maine. Trans. Am. Fish. Soc. 103(2):318-324. Von Heinz, A. 1959. The significance of flow as an ecological factor. Schweizerishe Zeischrift fuer Hydrologie. Vol xxi Fasc. 2:133-270. Ward, J.V. 1976a. Comparative limnology of differently regulated sections of a Colorado mountain river. Arch. Hydrobiol. 78(3): 319-342. Ward, J.V. 1976b. Effects of flow patterns below large dams on stream benthos: a review. In: Instream Flow Needs Symposium, vol. II. J.F. Orsborn and C.H. Allman, eds. Amer. Fish. Soc. 235-253. Ward, J.V. 1979. Stream regulation in North America. In: The Ecology of Regulated Streams. Proc. lst. Inter. Symp. on Reg. Streams. Plenum Press, New York. Wesche, T.A. 1974. Relationship of discharge reductions to available trout habitat for recommending suitable streamslows. Water Resources Research Institute, Univ. of Wyoming. Water Res. Series 53. 73pp. 127 White, R.J. 1975. Trout population responses to streamflow fluctuation and habitat management in Big Roche-a-cri Creek, Wisconsin. Verh. Internat. Limnol. 19:2469-2477. White, R.J., E.A. Hansen, and G.R. Alexander. 1976. Relationship of trout abundance to streamflow in midwestern streams. In: Instream Flow Needs Symposium, Vol. 2. J.F. Orsborn and C.H. Allman, eds. Am. Fish Soc. 597-615. Young, R.D., and O.E. Maugham. 1980. Downstream changes in fish species composition after construction of a headwater reservoir. Zalumi, S.C. 1970. The fish fauna of the lower reaches of the Dnieper (USSR), its present composition and some features of its formation under conditions of regulated and reduced river discharge. J. Ichthyology, 10(5):587-596.