deletions | additions       diff --git a/chapter_Methodology_section_Objective_1__.tex b/chapter_Methodology_section_Objective_1__.tex  index c29db36..4dead73 100644  --- a/chapter_Methodology_section_Objective_1__.tex  +++ b/chapter_Methodology_section_Objective_1__.tex 
...
\subsection{NLLoc location of detections}  After removing duplicate detections, P-picks for each of the newly detected events were made based upon cross correlation of the detecting template with the continuous data at the time of the detection. Template waveforms were correlated with network data over a 0.2 second window centered on the detection time. Picks were recorded at the time corresponding to the highest correlation value within that window. Time errors were then assigned to each pick based upon its correlation value with the template. Through visual review of a random sample of 100 events we assigned timing uncertainties of 0.01 seconds to picks with correlation values of greater than 0.90, 0.05 s for values of between 0.70 and 0.90, 0.10 seconds for values between 0.50 and 0.70, 0.30 s for values between 0.30 and 0.50 and 0.50 seconds for all other picks. We kept all events with more than 7 picks and discarded those with fewer. From an original total of nearly 30,000 detections, we were left with 15,471 detections after removing duplicates and discarding events with an insufficient number of picks.   For these 15,471 events, we then incorporated S-picks made my MSc student Stefan Mroczek \cite{mroczek2016}  using the method developed by Diehl and others \cite{Diehl_2009}. These events were then located with the nonlinear location program NonLinLoc \cite{Lomax_2014} using a 1-D model for Rotokawa used by former VUW MSc student Zara Rawlinson (\cite{rawlinson2011microseismicity} pp. 77). \subsection{HypoDD relocation of events}  As a final step, the entire catalog was relocated using the double-difference relocation program HypoDD \cite{Waldhauser_2000} with differential pick times generated using a cross correlation routine within the EQcorrscan Python package \cite{Chamberlain_2014}. Due to the size of the catalog and memory limitations in the software, events were split into separate Rotokawa and Ngatamariki catalogs before being relocated with HypoDD.