This counts, in seconds, the periods where the OTUC3 signal's error density exceeded the severely-errored-second threshold, indicating heavy, bursty corruption rather than occasional single-bit errors. Admins watch this separately from plain errored seconds because it better predicts service-impacting events on a 300G-class channel. A spike here coinciding with a known switching event, such as a protection-switch trigger elsewhere in the network, would help correlate cause and effect.