120 



Methods 



j<4 '^'■■'.vThe data eire ann ual observations of at least four kinds, ^i.ll 

 " meafeares of Chinook spavming escapements; redd counts on . natural 

 spawning areas, both foot and aerial, dam and weir counts, and 

 hatchery rack counts at all available localities in that portion of 

 the Columbia Basin within the states of Idaho, Oregon, and 

 Washington. The records were obtained from an excerpt of the 

 National Environmental Database, NED, provided by Mr. Duane 

 Anderson (NPPC, Northwest Power Planning Council, Portland, OR) . 

 The NED is coordinated by Mr. Tom Pansky (Bonneville Power 

 Administration, Portland, OR) . . 



Each locaJ-ity was examined in order to eliminate those annual 

 time series of less thaui twenty five yeaxs of continuous records. 

 For each series retained, a fifteen-year moving average was 

 calculated for each year as the eirithmetic mean of the year and the 

 fourteen preceding years, starting with the fifteenth year of data, 

 and progressing to the latest year available, usually 1990. 

 Working with the fifteen-year moving averages, the parameters of a 

 linear regression were calculated for. the annual time series of 

 each locality, ajid those localities having negative slopes that 

 tested significantly different (alpha = 0.05) from zero were 

 identified as, badly damaged and decreasing. Finally, the most- 

 recent annual abundance observation, usually that of 1990, was 

 taken as a percentage of the maximum annual observation in its time 

 series in order to provide a rough measure of the most recent level 

 of seeding of the spawning grounds. A spawning level that is 10% 

 of the maximum observed is said to be at 10% of seeding. The ten 

 percent level is arbitrarily defined as the threshold for being 

 critically damaged. 



