Assessing the skill of a biogeochemical model tailored to a global coupled forecasting system: a case study for the Bay of Bengal carbon cycle

To achieve real-time prediction of surface ocean carbonaceous variables in the Indian Ocean, we integrate an extended-range coupled forecast system with a global ocean biogeochemistry model incorporating an explicit carbon cycle. The physical ocean state is provided by the Indian Institute of Tropical Meteorology Extended Range Prediction for Applications to Society (IITM-ERPAS) coupled model, which drives the Ocean Tracer Transport Model (OTTM) with phosphate-dependent biology and OCMIP-II-type carbonate chemistry (hereafter referred to as OTTM-ERPAS-BGC). Daily mean outputs from IITM-ERPAS are used to drive the biogeochemical simulations. Model skill is evaluated against in situ observations of surface partial pressure of CO2 (pCO2) and pH from the Bay of Bengal Ocean Acidification (BOBOA) mooring (90∘E, 15∘N) over the period 2013-2018. The model successfully reproduces observed variability, with strong and statistically significant correlations (r = 0.65 for pCO2 and r = 0.80 for pH; p < 0.001). Based on this performance, a coherent region of influence (87.5∘E–92.5∘E, 12.5∘N–17.5∘N) is identified to further assess model skill against BOBOA time-series observations, climatological products and other derived datasets of pCO2 and pH in terms of seasonal cycle, subseasonal variability and long-term trends. Seasonal cycle analysis reveals that the updated Takahashi climatological pCO2 product is systematically underestimated by ∼18±3μatm compared to BOBOA observations, OTTM-ERPAS-BGC simulations and other derived products. Subseasonal variability exhibits pronounced intraseasonal signals, with significant spectral peaks at 13–41 days for pCO2 and 17–41 days for pH, reflecting dominant intraseasonal modulation of the Bay of Bengal carbon cycle. Long-term trends indicate a robust increase in pCO2 (2.02 μatm yr−1) and a concomitant decline in pH (−0.00253 yr−1), consistent with progressive acidification of the basin. The decomposition analysis demonstrates that dissolved inorganic carbon (DIC) and total alkalinity (ALK) exert dominant and complementary controls on pCO2 and pH variability, respectively, while the sea surface temperature (SST) plays a secondary role. Overall, this study demonstrates that coupling extended-range physical forecasts with ocean biogeochemical models provides a viable pathway for the operational prediction of carbon variables. Such capability is critical for the early detection of ocean acidification and hypercapnia hotspots, with significant implications for marine ecosystem management and climate services in the Indian Ocean.