Laserfiche WebLink
0 <br /> 306 HYDROCARBON CONTAMINATED SOILS AND GROUNDWATER Y GROUNDWATER REMEDIATION LIMITS 307` <br /> 1 <br /> 3000 <br /> The following quarter confirmed the predictions All data fell within the predict-s <br /> • . row data used ,n regreas,or, 6 tion interval of the linear regression Nonlinear regression of data to day 660 <br /> p raw data not used .n regression <br /> yielded C = 2300 e-"'K""' + 280 The linear asymptotic procedure confirmed <br /> . <br /> 2500 that the asymptobi.period began, statistically,at day 420,and estimated the slope <br /> t,t asymplolr at —0 17 ppb/day, and the asymptotic concentration at 270 ppb (see Figure 17 9) <br /> = The maximum differem c between the 90% confidence interval and tfic befit Fit <br /> 4 <br /> z000 was 19 ppb At that time(660 days), the asymptotic analysis met the criteria es- <br /> • tablished by the State of Florida and could be used to petition the state to ter- <br /> 4 nmiate the pump-and-treat system under Rulel7-770, F A C <br /> 1500 • , f <br /> Z 3000 <br /> W <br /> u ANALYSIS • raw data used .n regression <br /> • PERFORMED <br /> a 1000 O 0. row data not used in regress-nn <br /> 2500 . fit asymptote <br /> • • • O <br /> prediction interval <br /> Soo <br /> ------------------- ------ • p 2000 <br /> ! _ <br /> 120 240 ,360 480 600 720 0 O <br /> TIME (DAYS) ry 1500 O <br /> } 7 O <br /> Figure 17 7 Actual site nonlinear regression performed after 570 days of remediation e <br /> Z <br /> 0 <br /> ' 1000 O <br /> 3000 0 O <br /> • • raw data used in regression <br /> O . row data not used ,n regresaton S <br /> 250Q500 0 0 O O <br /> Fit • asymptola 0 O 0 <br /> ` w prediction interval - <br /> rr <br /> 2000 120 240 360 480 600 720 <br /> 0TIME (DAYS) <br /> a 1500 0 ° Figure 17 9 Actual site linear asymptotic regression performed after 660 days of reme <br /> 0 ANALYSIS + <br /> rWj PERFORMED p p dratlon <br /> U 1000 � <br /> 0 0 SUMMARY OF FINDINGS <br /> t <br /> Soo 00 Statistical regression procedures to test for an asymptotic condition in ground- <br /> 0 water quality data records have been developed The linear asymptotic technique <br /> ---------------- <br /> o------ <br /> --- -•--+r-- -•-- =t7 - a--- - <br /> defines an asymptotic condition as one in which some concluding time period <br /> _LLin a time series data set exhibits a negligible trend or slope that is not sig11 nificantly <br /> 120 240 360 4Ro 600 72p different from zero Additionally, both the upper and lower bounds(95% confi- <br /> TIME (DAYS) dente) on the slope should be small, so that the data is virtually trendless As <br /> Figure 17 8 Actual site linear asymptotic regression performed after 570 days of reme- previously reported,' this method was compared with an alternative exponential <br /> dlatlon regression technique over seven characteristic data Sets and was found to be <br />