A comparative evaluation of satellite and bio-argo-based net primary productivity algorithms for the north Indian Ocean

We estimate the net primary productivity (NPP) based on bio-argo measurements using three algorithms: Vertically Generalised Production Model (VGPM), Eppley-VGPM and the original Carbon-based Production Model (CbPM). To do this, we made a dataset of Chlorophyll-a (Chl-a), temperature, and salinity from argo measurements and then computed the NPP for the north Indian Ocean (NIO, 5° S–25° N, 40°–100° E). These new data are compared with satellite-based NPP, and the Root Mean Square Error (RMSE) suggests a better performance of argo-based standard VGPM with an RMSE of 0.20 g C m−2 day−1. We have calculated the bias for two seasons: winter (December, January and February: DJF) and summer (June, July, August and September, JJAS), which are characterised with high productivity in NIO. The bias is smaller for standard VGPM and Eppley-based VGPM, in which the former has a bias of − 0.1 g C m−2 day−1 in DJF and − 0.08 g C m−2 day−1 in JJAS. The correlation between satellite and argo-based VGPM is 0.33, suggesting that this is the best among the estimates and can be used for NPP studies in NIO. Furthermore, the individual argo-based NPP of the Arabian Sea (AS, 5°–25° N, 50°–75° E), Bay of Bengal (BoB, 6°–18° N, 76°–95° E) and Equatorial Indian Ocean (Eq IO, 5° S–5° N, 60–90° E) is tightly related to wind stress (WS), mixed layer depth (MLD) in AS and BoB.