Availability of yellowfin sole Limanda aspera to the National Marine Fisheries Service eastern Bering Sea trawl survey, rather than trawl sampling efficiency, is proposed as the primary reason for relatively high annual variability of biomass estimates in this region, including most recently, a 48% increase from 2015 to 2016. The main hypothesis presented here is that temperature-mediated differences in the timing of spring-summer spawning migrations to unavailable nearshore spawning grounds, affect survey biomass estimates. Colder bottom temperatures delay both migrations and spawning, causing higher proportions of mature individuals to reside in the unavailable nearshore grounds at the time of annual survey (June–July). Indicators of this scenario include decreases of mature fish proportions and decreases in mean overall fish lengths during colder years when biomass was less than expected. Further evidence includes differences in spatial distribution between warm and cold years, and spatial shifts away from nearshore areas between early June and July–August sampling during which catch per unit effort (CPUE) increased and proportion of females increased. That neither of these spatial shifts nor temperature-CPUE relationships occurred for northern rock sole Lepidopsetta polyxystra, a species of similar morphology and abundance, and overlapping spatial distribution, suggests that temperature-mediated trawl sampling efficiency was not a major contributing factor for yellowfin sole. We have also found a positive relationship between survey biomass estimates and survey start times, reinforcing that availability is a function of timing. The addition of survey start time to the catchability (q) parameter within the current stock assessment model significantly improved model fits to abundance data.