Presentation Information
[U11-P04]Significant Ecological Reorganization of Planktonic Diatom Communities Across the Mid-Pleistocene Transition in the Okhotsk Sea
*Li Lo1, Yusuke Okazaki2, Liang-Chi Wang3, Chih-Lin Wei4 (1.Department of Geosciences, National Taiwan University, Taipei, Taiwan, 2.Department of Earth and Planetary Sciences, Kyushu University, Fukuoka, Japan, 3.Department of Environmental Sciences, National Chung Cheng University, Chiayi, Taiwan, 4.Institute of Oceanography, National Taiwan University, Taipei, Taiwan)
Keywords:
Mid-Pleistocene Transition,Okhotsk Sea,diatom diversity,paleoecology
High-latitude marine ecosystems are highly sensitive to climate variability due to strong feedbacks among sea ice, ocean stratification, and nutrient supply. However, long-term ecological responses to major climate transitions remain poorly constrained, particularly across the Mid-Pleistocene Transition (MPT; ~1.2–0.8 Ma), when glacial cycles shifted from 41-kyr to dominant 100-kyr periodicity. Here we present a high-resolution reconstruction of planktonic diatom biodiversity and community structure from the central Okhotsk Sea over the past 1.5 million years, providing new insights into ecosystem dynamics across this critical interval.
We analyzed 383 diatom assemblage samples from sediment core MD01-2414 and quantified biodiversity using coverage-standardized Hill numbers, decomposing diversity into alpha (within-sample), beta (between-sample), and gamma (regional) components. Diversity patterns were examined across Marine Isotope Stages, including glacial, interglacial, and super-interglacial intervals. Beta diversity was further partitioned into species turnover and nestedness. Environmental controls were assessed using generalized additive mixed-effect models (GAMM) and machine-learning approaches, including Random Forest (RF), integrating sea surface temperature, nutrient proxies, productivity indicators, and MPT-related temporal structure.
Our results reveal a pronounced reorganization of diatom communities across the MPT. Prior to and during the MPT, interglacial periods were characterized by higher alpha, beta, and gamma diversity, indicating more heterogeneous assemblages under warmer conditions. In contrast, this pattern reversed after the MPT, with glacial periods exhibiting higher diversity, suggesting a fundamental shift in ecological structure and environmental constraints. Super-interglacial intervals (e.g., MIS 5, 9, 11, 31, 47, and 49) consistently displayed the lowest diversity values, implying strong ecological filtering during extreme warm conditions.
Decomposition of beta diversity shows that species turnover dominated community dissimilarity throughout most of the record, particularly during the MPT, reflecting rapid ecological restructuring. Nestedness increased after the MPT, indicating enhanced species loss and reduced ecological resilience. GAMM analyses demonstrate significant non-linear relationships between alpha diversity and sea surface temperature, nutrient availability, and productivity. Among the machine-learning models, RF consistently outperformed parametric approaches (mean R² ≈ 0.75) and identified MPT timing, temperature, and nutrient proxies as the primary drivers of diversity changes.
Overall, our results document a major ecological reorganization of surface-ocean ecosystems in the Okhotsk Sea across the MPT, likely driven by changes in sea-ice extent, water-column stratification, and nutrient supply. This study highlights the sensitivity of high-latitude marine biodiversity to long-term climate forcing and demonstrates the value of combining paleoecological records with advanced statistical and machine-learning techniques to better understand ecosystem resilience under future climate change.
We analyzed 383 diatom assemblage samples from sediment core MD01-2414 and quantified biodiversity using coverage-standardized Hill numbers, decomposing diversity into alpha (within-sample), beta (between-sample), and gamma (regional) components. Diversity patterns were examined across Marine Isotope Stages, including glacial, interglacial, and super-interglacial intervals. Beta diversity was further partitioned into species turnover and nestedness. Environmental controls were assessed using generalized additive mixed-effect models (GAMM) and machine-learning approaches, including Random Forest (RF), integrating sea surface temperature, nutrient proxies, productivity indicators, and MPT-related temporal structure.
Our results reveal a pronounced reorganization of diatom communities across the MPT. Prior to and during the MPT, interglacial periods were characterized by higher alpha, beta, and gamma diversity, indicating more heterogeneous assemblages under warmer conditions. In contrast, this pattern reversed after the MPT, with glacial periods exhibiting higher diversity, suggesting a fundamental shift in ecological structure and environmental constraints. Super-interglacial intervals (e.g., MIS 5, 9, 11, 31, 47, and 49) consistently displayed the lowest diversity values, implying strong ecological filtering during extreme warm conditions.
Decomposition of beta diversity shows that species turnover dominated community dissimilarity throughout most of the record, particularly during the MPT, reflecting rapid ecological restructuring. Nestedness increased after the MPT, indicating enhanced species loss and reduced ecological resilience. GAMM analyses demonstrate significant non-linear relationships between alpha diversity and sea surface temperature, nutrient availability, and productivity. Among the machine-learning models, RF consistently outperformed parametric approaches (mean R² ≈ 0.75) and identified MPT timing, temperature, and nutrient proxies as the primary drivers of diversity changes.
Overall, our results document a major ecological reorganization of surface-ocean ecosystems in the Okhotsk Sea across the MPT, likely driven by changes in sea-ice extent, water-column stratification, and nutrient supply. This study highlights the sensitivity of high-latitude marine biodiversity to long-term climate forcing and demonstrates the value of combining paleoecological records with advanced statistical and machine-learning techniques to better understand ecosystem resilience under future climate change.