deletions | additions       diff --git a/Conclusions.tex b/Conclusions.tex  index 90db298..af8d8b5 100644  --- a/Conclusions.tex  +++ b/Conclusions.tex 
...
\section{Conclusion}  Ultrasonic measurements have been done on Cairnside and Covey Hill samples of the sedimentary basin of the St.\ Lawrence Platform in southern Quebec. The results show that $P$-wave are sensitive to pore fluid substitution at laboratory scale. Moreover, the measurements demonstrate that it is possible to detect the phase transition from gaseous CO$_2$ to liquid or supercritical CO$_2$, for both velocity and amplitude attributes. \\  The laboratory measurements were used for calibrating numerical models that were used to simulate the time-lapse seismic response to CO$_2$ injection. A number of scenarios were considered to evaluate the performance of a time-lapse seismic monitoring program for the particular context of the St.\ Lawrence Platform. We were particularly interested in comparing the response of a blocky model (commonly used in comparable stuies) with that of a model reflecting the natural variability found in nature. We were also interested in assessing the effect of the low permeability and porosity observed at B\'ecancour on the time-lapse seismic response, and comparing the performance the seismic monitoring with respect to more favorable storage conditions. conditions.\\  For all scenarios, a series of VSP synthetic seismograms imaging the CO$_2$ evolution during 15 years of injection and 35 year of CO$_2$ migration have been modeled. The results show quite well the effect of CO$_2$ which is indicated by a time-delay of arrivals, corresponding to a decrease in velocity when supercritical CO$_2$ replaces brine in pore spaces. The geostatistical approach used to generate our synthetic model allowed us to obtain more realistic seismograms compared to a traditional blocky model. The time-lapse amplitude response of the blocky and stochastic models also differ significantly in the lower part of the reservoir. Using over simplistic models thus yields misleading results. results.\\  For the Bécancour-like model, there is almost no evolution of the CO$_2$ plume after 5 years of injection and even 35 years after injection stopped, and it is therefore unlikely to detect any long-term variation in the seismic response. However, in the optimistic case the CO$_2$ has been partially dissolved during the migration period, and the VSP response after 50 years results in an early arrival of the reservoir bottom signal compared to the 15 years data. Therefore, carefully considering the behaviour of the reservoir in modeling the seismic response appears critical for accurate appraisal of the time-lapse response, which in return will help interpreting correctly real monitoring data. \\  These results highlight the fact that details in the models can have significant impacts on the expected reponse to a monitoring campaign. One element that was not considered in this work is pressure changes induced by injection of CO$_2$ in the reservoir. As reported in the literature, pressure effects can affect the seismic reponse and change the results modeled in this study. In particular, as pressure redistributes over time, the long-term seismic response modeled in the B\'ecancour-like scenario should also change over time, contrary to what is predicted without considering pressure change in the simulations. Work is under way to take such effects into account.