Description: Annual-resolved sulfur and non-sea-salt sulfur concentrations and inferred volcanic sulfate depoistion rates from the six ice cores NEEM-2011-S1 (Sigl et al., 2013), NGRIP1 (Plummer et al., 2012), NGRIP2 (McConnell et al., 2018), TUNU2013 (Sigl et al., 2015) and B19 between 699 and 1001 CE and annual-resolved non-sea-salt sulfur concentrations from a four ice-core stack (NEEM-2011-S1, NGRIP1, TUNU2013, B19) between 1731 and 1996 CE including volcanic samples and with volcanic samples replaced by a 11-year running median. Volcanic event detection is based a 91-year running median (RM) was used on the annually averaged nssS records on periods unaffected by strong changes in volcanic background emissions to estimate the natural background sulfate levels; a Median of Absolute Deviation (MAD), calculated from the RM, was used for volcanic peak detection over the background period. Between 700 and 1000 CE sulfur peaks were considered volcanic if they passed an upper threshold (K=3, estimated as RM plus 3* MAD). The duration of the event was determined when it passed the lower threshold (K=1, estimated as the RM plus 1 * MAD). These upper and lower thresholds were selected by validation on well-known historic eruptions; volcanic peaks were then removed to calculate the non-volcanic background (S RRMi). To further calculate the amount of sulfate deposited, S RRMi was subtracted from the average annual nss-S and then multiplied by the accumulation rate of the drill site; finally, volcanic flux was calculated for each event by summing the sulfate deposited across the total duration of the event. All ice cores are presented on the NS1-2011 chronology (Sigl et al., 2015) except for NGRIP2 which was on the NGRIP2-DRI chronology (McConnell et al., 2018).