Laserfiche WebLink
i <br /> i <br /> I <br /> 304 HYDROCARBON CONTAMINATED SOILS AND GROUNDWATER t GROUNDWATER REMEDIATION LIMITS 305 <br /> evidence for leveling off, attributable to the longer data record in the asyrnpmu, <br /> region After day 600 it yielded a slope of —0 068 ppb/day, not significantly <br /> 1200 • raw data used .n rrgrets o" dillcicnt from sero and an&,yniplotK.conccaltration of 73 ppb(see Figure 17 6) <br /> o - rem data not used in regression I he upper bound on the difterefte between the 90% confidence interval and the <br /> 1000 - fit - - asymptbte best fit was 15 ppb Based on this analysi%, the criteria spec.ilicd by Florida <br /> CL <br /> - prediction inler„al ' guidance are met These results are similar to the example analysis provided in <br /> a the Florida guidance <br /> a 800 0 <br /> S - <br /> 6 l7 ' 1 <br /> O �{ <br /> W 1500 12000 • = raw data used in regression <br /> U i <br /> d 0 O O = raw data net used in regrrssfon <br /> V <br /> 400 ANALYSIS 1000 rpt asymptote <br /> � <br /> PERFORMED = predLClion interval <br /> O O r n <br /> i <br /> 200 0 _ O 0 n <br /> O 0Sao O O <br /> 0 O O z <br /> O <br /> 120 240 360 480 600 720 Cr 600 <br /> Z <br /> 11 <br /> TIME (DAYS) � U 0 <br /> Z <br /> Figure 17 4 Florida example linear asymptotic regression after day 600 400 <br /> t O O <br /> O <br /> r ZOO � O <br /> D <br /> 1200 • = row data used in regression O b O 0 _ <br /> -------------------o----------------- • <br /> 0 +. row data not used in regrrss�on <br /> F.{ - asymptote F 120 240 360 480 600 770 <br /> 1000 <br /> ap s i TIME (DAYS) <br /> CL <br /> CL Figure 17 6 Florida example linear asymptotic regression after day 660 <br /> Z 800 • • ; <br /> o � ' <br /> Ix <br /> 4 • <br /> W 600 Actual Site Data Set <br /> V <br /> O <br /> The second example represents actual site data provided by API After 570 <br /> '\"o <br /> 400 days of pump-and-treat remediation,concentrations appeared to have leveled off <br /> 0 The nonlinear regression yielded C = 2300 e-0030 i f 300 (r"' = 0 91, see <br /> 200 • • 0 Figure 17 7) The linear asymptotic procedure identified the period from 420-570 <br /> • • days as tate asymptotic region (see Figure 17 8) The result from 1day 540 (900 <br /> • ppb) was refected as an outlier insofar as 900 ppb exceeded the asymptotic region <br /> • mean by more than two standard deviations The best estimate slope was —0 29 <br /> 120 240 iso Oso soo 720 ppb/day, which was not significantly different from zero The asymptotic con- <br /> t <br /> TIME (DAYS) f centratton was estimated to be 270 ppb At that time the linear regression predicted <br /> Figure 17 5 Florida example nonlinear regression performed after 660 days of rpmediatron concentrations to remain below 300 ppb for the next quarter <br />