Sea ice concentration (SIC) plays an important role in the Antarctic polynyas, especially biological changes through chlorophyll-a (CHL) concentration phenology. To understand how much SIC changes affect CHL, the regime shift index (RSI) was applied to SIC anomaly and we obtained three regimes: 1997–2004 for Regime 1 (R1), 2005–2016 for Regime 2 (R2), and 2016–2019 for Regime 3 (R3). Based on reconstructed daily remotely sensed CHL observations using a random forest algorithm with 11 physical predictors, our analysis revealed that: (1) the earlier the SIC melting time, the earlier the CHL bloom peaks for each regime resulting from sea ice melting duration and physical forces; (2) SIC changes account for 10%-12.5% of CHL variation, which is the first estimation in the Ross Sea polynya; and (3) although they are significantly influenced by SIC changes, the 11 physical predictors for CHL were more dominant for each regime. Thus, these results could explain why CHL decreased even in with unprecedent decrease in Antarctic SIC from 2016. Therefore, this quantitative analysis can estimate the contribution of SIC formation and physical environmental parameters to CHL changes based on the random forest machine learning algorithm.