ZDC rate dependence of luminosity monitoring?

To monitor the luminosity run by run the random accept portion of the L2W trigger was used.  The same BHT3+coin software conditions as are used in the vernier scan analysis are used on these random accept events to count the events satisfying the BHT3+coin trigger for each run from the st_W dataset.  Below I look at different thresholds for the awayside requirement of the BHT3+coin software trigger and compare the low (0-100kHz) and high (100-160kHz) ZDC rate samples.

Figure 1: Awayside sum distribution (opening angle of 5 TP) for low ZDC rate (left) and high ZDC rate (right)

The nominal value for the awayside sum threshold is 98 DSM ADC.

Figure 2:  Ratio of events above awayside sum threshold vs threshold value (low/high ZDC rate)

The ratio is decreasing with higher threshold which indicates that the shape of the distribution is different for the two ZDC rate samples.  This could possibly be pointing to an intensity dependent effect in the luminosity monitoring that isn't currently accounted for.  More investigation is needed to determine what effect this might have on the final integrated luminosity, or a possible systematic uncertainty.