Narrative:

I was working controller in charge (controller in charge) at the time of the event. It was the [late morning] departure push; and I had opened the second ground control (ground control) in anticipation of over 100 operations in the coming hour. Aircraft X had exited 28L at tango; local control (local control) then crossed aircraft Y and aircraft Z on twy T and K; respectively. I watched all three aircraft come to a stop in the K/T/B intersection; with aircraft Y and aircraft Z unable to clear 28L. GC1 (ground control) was very busy with several aircraft without gates; numerous outbounds and the added complexity constraint of having part of twy zulu closed; necessitating complex staging to having repositioning aircraft cross the 28s. I walked over to local control and told him to send aircraft a around; which was on short final rwy 28L. Local control had a southbound VFR transition; aircraft B in the missed approach corridor; so myself; local control and lca were watching that to ensure the missed approach and the transition were separated; I was also continuing to watch when 28L became clear from the previous arrivals that had stopped. The controller being relieved at GC2 noticed the non-acquisition in the upwind for the 1s. At this point I had stepped back to start logging the missed approach. I looked at the radar and initially thought there was only one target; since they were so close together; in fact; neither of the departures (aircraft C and aircraft D) had acquired. I was monitoring local control and not rc; so I did not hear the coordination between lca and ci-1. Lca and local control had difficulty discerning which aircraft was which. Local control first reiterated the departure procedures that each aircraft should have been on. He then instructed aircraft C to turn right 30 degrees; which that aircraft did not acknowledge until the second time local control issued the turn. At no point did the stars conflict alert sound with the two upwind departures.I could have kept clearance delivery (clearance delivery) open to have an extra set of eyes in the tower. However; that person had already worked for an hour; would have needed a break during this push anyway (otherwise would have exceeded 2 hours time on position); and there was no one else coming up from their break due to our short staffing. I could have suspended transitions at the beginning of the departure push; which would have precluded the go-around traffic conflict in the air with the aircraft B from occurring; and hopefully would have meant someone would have seen the upwind departure conflict sooner. Ultimately it's up to the pilots to fly their planes. They should not be so dependent on their cockpit automation that they don't notice themselves that the aircraft turned left on departure instead of right; as per the SID. The old adage I learned in flight training applies here: aviate; navigate; communicate. It appears the pilots of aircraft C didn't do any of those three things very well. Had the stars ca sounded with the upwind pair; perhaps someone would have noticed what was happening sooner. Overtime should have been scheduled or assigned to provide adequate facility staffing.

Google
 

Original NASA ASRS Text

Title: SFO Tower controllers reported a loss of separation. One of the reporters told the Local to turn one of the aircraft to get them separated.

Narrative: I was working CIC (Controller In Charge) at the time of the event. It was the [late morning] departure push; and I had opened the second ground control (GC) in anticipation of over 100 operations in the coming hour. Aircraft X had exited 28L at Tango; LC (Local Control) then crossed Aircraft Y and Aircraft Z on Twy T and K; respectively. I watched all three aircraft come to a stop in the K/T/B intersection; with Aircraft Y and Aircraft Z unable to clear 28L. GC1 (Ground Control) was very busy with several aircraft without gates; numerous outbounds and the added complexity constraint of having part of Twy Zulu closed; necessitating complex staging to having repositioning aircraft cross the 28s. I walked over to LC and told him to send Aircraft A around; which was on short final Rwy 28L. LC had a southbound VFR transition; Aircraft B in the missed approach corridor; so myself; LC and LCA were watching that to ensure the missed approach and the transition were separated; I was also continuing to watch when 28L became clear from the previous arrivals that had stopped. The controller being relieved at GC2 noticed the non-acquisition in the upwind for the 1s. At this point I had stepped back to start logging the missed approach. I looked at the radar and initially thought there was only one target; since they were so close together; in fact; neither of the departures (Aircraft C and Aircraft D) had acquired. I was monitoring LC and not RC; so I did not hear the coordination between LCA and CI-1. LCA and LC had difficulty discerning which aircraft was which. LC first reiterated the departure procedures that each aircraft should have been on. He then instructed Aircraft C to turn right 30 degrees; which that aircraft did not acknowledge until the second time LC issued the turn. At no point did the STARS Conflict Alert sound with the two upwind departures.I could have kept CD (Clearance Delivery) open to have an extra set of eyes in the tower. However; that person had already worked for an hour; would have needed a break during this push anyway (otherwise would have exceeded 2 hours time on position); and there was no one else coming up from their break due to our short staffing. I could have suspended transitions at the beginning of the departure push; which would have precluded the go-around traffic conflict in the air with the Aircraft B from occurring; and hopefully would have meant someone would have seen the upwind departure conflict sooner. Ultimately it's up to the pilots to fly their planes. They should not be so dependent on their cockpit automation that they don't notice themselves that the aircraft turned left on departure instead of right; as per the SID. The old adage I learned in flight training applies here: Aviate; Navigate; Communicate. It appears the pilots of Aircraft C didn't do any of those three things very well. Had the STARS CA sounded with the upwind pair; perhaps someone would have noticed what was happening sooner. Overtime should have been scheduled or assigned to provide adequate facility staffing.

Data retrieved from NASA's ASRS site and automatically converted to unabbreviated mixed upper/lowercase text. This report is for informational purposes with no guarantee of accuracy. See NASA's ASRS site for official report.