Temporal dynamics of primary productivity in the north-eastern Arabian Sea: an evaluation of ocean color models
Vinaya Kumar Vase,Nakhawa Ajay,Rajan Kumar,Sreenath Ramanathan,Jayasankar Jayaraman,Prathibha Rohit
DOI: https://doi.org/10.1007/s12517-021-07688-x
2021-07-01
Arabian Journal of Geosciences
Abstract:<p class="a-plus-plus">We examined the spatial and temporal variations in monthly primary productivity (PP; mg C m<sup class="a-plus-plus">−2</sup> day<sup class="a-plus-plus">−1</sup>) in the northern Arabian Sea from 2015 to 2017 using the PP model (Vertically Generalized Production Model, VGPM) to identify and tag productivity zones and seasons. The major objective was to validate the existing satellite algorithms and compare them with the regional parameterization and further define the most accurate model for the region. The models use both in-situ chlorophyll-<em class="a-plus-plus">a</em> (Chl-<em class="a-plus-plus">a</em>), SST, and euphotic depth (Z<sub class="a-plus-plus"><em class="a-plus-plus">eu</em></sub>) (derived from Chl-<em class="a-plus-plus">a</em>), and satellite-retrieved (photosynthetically available radiance, PAR and day length, DL) variables as input parameters. The measured PP values showed significant intra-annual variations, and the maximum was during December (3689.2 ± 505.3 mg C m<sup class="a-plus-plus">−2</sup> day<sup class="a-plus-plus">−1</sup>) and the minimum during August (2207.7 ± 202.2 mg C m<sup class="a-plus-plus">−2</sup> day<sup class="a-plus-plus">−1</sup>). The linear regression depicted that the input variables (Chl-<em class="a-plus-plus">a</em>, PAR, Z<sub class="a-plus-plus"><em class="a-plus-plus">eu</em></sub>, DL) together explained 48.68% variation in the PP. The chlorophyll-<em class="a-plus-plus">a</em> showed significant variability in PP, followed by DL, PAR, and Z<sub class="a-plus-plus"><em class="a-plus-plus">eu</em></sub>. We tested three modified models with minor modifications in input variables of basic VGPM. Among the models validated in the region, the VGPM-KI could explain 38.3% of the variance with the in-situ PP data, followed by other models (range of variance explained from 18.7 to 38.3%). The average model precision, as determined by RMSD and the bias, was lowest for the modified model (VGPM-KI, VGPM-BF, and VGPM-E), but highest in case of models with satellite as a sole source of input variables (VGPM and EVGPM). In the northern Arabian Sea, VGPM-KI and VGPM-E performed better than the other models. However, VGPM-KI overestimated the PP values when compared to in-situ estimates. The biogeochemical cycles and ocean processes like coastal upwelling and winter convection by winds, which are the key determiners for perennial productivity in the region, will also affect PP in the region.</p>
geosciences, multidisciplinary