Click for next page ( 13


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 12
based on the assessment of physiological signifi- selected using the balanced design, Treatment 1 was cance, which considered the significance of multi- added to the analysis to refine the results. A com- ple injuries and inspection of data for the occurrence plete statistical analysis is shown in Appendix E. of injury combinations. For example, the occurrence of two injuries categorized as Moderate was assessed CHAPTER 3 RESULTS to have physiological costs similar to that of one injury categorized as Mortal. Sound Exposures An RWI was calculated for each treatment fish A primary objective of this project was to test and each control fish. The formulas used were: the hypothesis that the magnitude of barotrauma m effects increased with increasing exposure to pile RWI ( Control ) = ( Wi Ci ) Equation 1.1 driving sounds as measured in terms of cumulative i sound exposure level (SELcum). Each experimental m animal was exposed to one of the pile driving treat- RWI ( Treatment ) = ( Wi Ti ) Equation 1.2 ments shown in Table 2. The maximum exposure i had an average SELss of 187 dB re 1 Pa2s and a SELcum of 219 dB re 1 Pa2s. This level of SELss had Where an average peak SPL of 213 dB re 1 Pa, which was RWI = response weighted index, the maximum amount of energy that could be safely i = injury type index, generated by the HICI-FT. Therefore, for Treatment m = 22, number of injury types (Table 1), 1's "pair," a SELcum of 219 dB re 1 Pa2s could not Ti = the proportion of the sample of fish exposed be generated with 960 strikes since this would have to a treatment that experienced injury type i, required a peak SPL above 215 dB re 1 Pa. The Wi = the Trauma Category weight (5, 3, or 1) for remaining treatments were conducted in pairs since injury type i, the same SELcum could be reached with both 1,920 Ci = the proportion of the sample of control fish for and 960 strikes by having a higher SELss (and thus a treatment that experienced injury type i. higher peak SPL) for 960 strikes than for 1,920. In each treatment pair, the goal was to present the same SELcum value but vary the number of pile strikes, Statistical Analysis which altered the SELss value. The response variable RWI was transformed, Holding the SELcum steady was done to imple- as shown in Equation 1.3, before analysis in order ment treatments that could be used to explore the to stabilize variance and linearize the response equal energy hypothesis while providing insight into curve. the relative importance of SELss and SELcum in deter- mining effects of sound on fish. yi = ln ( RWI i + 1) Equation 1.3 Each treatment pair was aimed at a specific SEL value. However, many variables affect the ability to present the precise signal level, and thus the range of In addition, cumulative energy was expressed as: SELcum treatments was continuous, rather than dis- SEL cum = SEL ss + 10 log10 ( number of strikes ) crete points. Each treatment blends with the next SELcum, such that each treatment is 1.5 dB of its Equation 1.4 specified value, i.e., 216 1.5. Because the data are continuous, an RWI was calculated for each individ- Analyses of covariance (ANCOVA) were per- ual fish. However, in Table 2, the average RWI is formed regressing yi against SELcum and assessing reported for each treatment. whether number of strikes (960 or 1,920) had an additional effect on fish response beyond that de- Barotrauma scribed by SELcum. Initial analyses were conducted on Treatments 2 through 11 to balance the design. Inspection of the log-transformed RWI values Treatments 2 through 11 were paired while Treat- show that fish with 960 strikes had a statistically ment 1 lacked a counterpart. Once a model was significant higher RWI value than fish exposed to 12

OCR for page 12
Table 2 Study exposure treatments and exposure details. Treatment Avg. Number of Avg. Avg. Peak Duration, Exposed Control Avg. No. SELcum Strikes SELss SPL min Fish, n Fish, n RWI 1 219 1920 187 213 48 44 33 15.318 2 216 1920 183 210 48 36 16 5.971 3 216 960 186 213 24 28 10 6.071 4 213 1920 180 207 48 26 5 2.346 5 213 960 183 210 24 31 7 4.323 6 207 960 177 203 24 24 8 1.042 7 207 1920 174 201 48 43 17 0.581 8 210 960 181 208 24 31 10 4.032 9 210 1920 177 204 48 30 11 3.433 10 203 960 174 201 24 32 11 0.656 11 204 1920 171 199 48 31 12 0.419 1,920 strikes at the same value of SELcum. This can cepts for fish exposed to either 960 or 1,920 strikes be seen when comparing the ln (RWI+1) regression (Table 3). lines for each 960- and 1,920-strike treatment. In Adding treatment 1 (i.e., SELss = 187, SELcum other words, for common values of SELcum, higher = 219, number of strikes = 1,920) to the analysis did values of SELss resulted in significantly higher val- not change the linearity of the data on the natural ues of RWI for 960 strikes (Figure 4). log scale or the regression relationships (Table 4, The ANCOVA demonstrated that ln (RWI+1) Figure 4). The final model was based on the use of all was linearly related to SELcum, and fish exposed to Treatments 1-11. a common value of SELcum using 960 or 1,920 The RWI values were calculated and plotted for strikes had statistically different RWI values. Using each fish as shown in Figure 5. a balanced design for Treatments 2 through 11 (i.e., Distributions of 1,920 and 960 strikes in Figure 5 these had pairs of common values; Treatment 1 did show an increase of RWI values correlated with an not), common slopes were found for the regression increase of exposure severity (SELcum). The increase of ln (RWI+1) versus SELcum, but different inter- in RWI was the result of both the number of injuries All Treatments: Model Ln(RWI+1) ~ SELcum + #strikes 4 960 Strikes 960 Strikes 3 1920 Strikes 1920 Strikes 2 Ln(RWI+1) 1 0 1 2 200 205 210 215 220 225 SELcum Figure 4 Scatterplots of SELcum vs. ln (RWI+1) for all treatments. Solid line shows predicted ln (RWI+1) values for 960 strikes and dashed line for 1,920 strikes. Red squares denote the 960 strikes and blue diamonds denote the 1,920 strikes. 13

OCR for page 12
Table 3 Sequential analysis of covariance (ANCOVA) results for Treatments 211. Response: ln (RWI+1). Best model contains the covariate SELcum followed by number of strikes. Treatments 211 Source df SS MS F-value P-value SELcum 1 103.22 103.22 281.240 <0.0001 Residuals 309 113.41 0.37 Source df SS MS F-value P-value SELcum 1 103.22 103.22 289.679 <0.0001 Strikes 1 3.95 3.95 11.077 0.0010 Interaction 1 0.07 0.07 0.196 0.6583 Residuals 307 109.39 0.36 Best Model Source df SS MS F-value P-value SELcum 1 103.22 103.22 290.437 <0.0001 Strikes 1 3.95 3.95 11.106 0.0010 Residuals 308 109.46 0.36 each exposed fish experienced as well as the biologi- defined by the SELss and number of strike variables cal significance of those injuries. The increase in the (Figure 8), Figures 8A and 8B, are presented sepa- number of injuries per test fish with increase in sever- rately to show the construction of 8C. ity of exposure is shown in Figure 6. Figure 8A is the background layer of the plot and Additional insight can be obtained by examining represents the sample space for the study. The x- and Figure 7. As SELss increases, RWI increases as well. y-axes are SELss and number of strikes, respectively, However, the indication is that as SELss increases, while the z-axis is SELcum. The blue dashed contours the severity of barotrauma response is differentially represent SELcum values, which were generated by amplified as the number of strikes increases. There calculating the relationship between the number of is a higher level of RWI values for the 960 strike strikes and the SELss (see Equation 1.4). For exam- treatments, which can be seen when comparing the ple, a SELcum of 208 dB is produced when there are ln (RWI+1) regressions lines for each 1,920 and 960 strikes at a SELss of 178 dB, or when there are 960-strike treatment (Figure 4). 1,920 strikes at a SELss of 175 dB. Figure 8C summarizes study findings. However, Figure 8B shows the RWI contour lines in black. because of the complexity of the response of fish to Axes are the same as for Figure 8A with treatment exposure over the SELcum treatments as they are RWI in the z-axis for this layer. Note that the top Table 4 Analysis of covariance (ANCOVA) from the best model from Table 6 applied to all Treatments 111. Response: ln (RWI+1). All Treatments Source df SS MS F-value P-value SELcum 1 201.30 201.30 542.039 <0.0001 Strikes 1 2.24 2.24 6.033 0.0145 Residuals 352 130.72 0.37 14

OCR for page 12
45 40 40 RWI 1920 35 RWI 1920 RWI 960 35 RWI 960 30 30 25 25 RWI RWI 20 20 15 15 10 10 5 5 0 0 203 206 209 212 215 218 221 170 173 176 179 182 185 188 SELcum SELss Figure 5 RWI values for each fish by SELcum for Figure 7 Individual RWI values by SELss. 1,920 and 960 pile strikes. 30 25 20 Number of Fish 15 10 5 0 0 2 Nu 4 mb 6 er T1 of T2 T3 inj 8 T8 T4 T5 uri 10 T9 es T7 T6 EL cum) 12 T11 T10 t (o rd e red by S en Treatm Figure 6 Frequency of barotrauma injury occurrence per fish. The number of test fish (z-axis) with number of unweighted-barotrauma injuries (y-axis) by each treatment (x-axis). For example, in the most severe exposure (Treatment 1 = T1), 1 fish had 13 injuries, and 10 fish had 8 injuries. Similarly, for the least severe exposure (T11), 6 fish had 1 injury, and 24 fish had 0 injuries. 15

OCR for page 12
Dotted Lines = SELcum = SELss + 10Log10(Num.Strikes) 2200 1800 Number of Strikes 1400 1000 20 20 20 20 20 21 21 21 21 0 2 4 6 8 0 2 4 6 600 170 172 174 176 178 180 182 184 186 188 A SELss Black Lines = RWI = exp(-30.050 + 0.149*SELcum - 0.000171*Num.Strikes)-1 2200 1 2 3 4 5 6 7 8 9 10 1800 Number of Strikes 1400 1000 600 170 172 174 176 178 180 182 184 186 188 B SELss Dotted Lines = SELcum = SELss + 10Log10(Num.Strikes) Black Lines = RWI = exp(-30.050 + 0.149*SELcum - 0.000171*Num.Strikes)-1 2200 1 2 3 4 5 6 7 8 9 10 1800 Number of Strikes 1400 1000 20 20 20 20 20 21 21 21 21 0 2 4 6 8 0 2 4 6 600 170 172 174 176 178 180 182 184 186 188 C SELss Figure 8 Panel A is the background layer plotting the SELcum contours (dashed lines) by SELcum, by SELss, and number of strikes within the treatment range. Panel B is a contour plot of ln (RWI+1) (the solid lines labeled 1-10), which illustrates values increase as SELss increase. The upper horizontal line indicates the 1,920 strike-line, and the bottom horizontal line indicates the 960 strike-line. Panel C is the composite of B on top of A, and shows where the RWI contours fall over the SELcum and SELss in relation to number of strikes. See text for further discussion.