2. Materials and methods
Online databases (http://giovanni.gsfc.nasa.gov, http://apdrc.soest.hawaii.edu) provided data on sea surface temperature (SST), surface winds, and chlorophyll-a. Ferret (version 7) was used for generating the monthly composite maps of these parameters. The first in-situ data from 2010 – 11was obtained from 15 locations in the GoM for three seasons [March 2010 (Pre-Monsoon), September 2010 (Southwest Monsoon), and January 2011 (Northeast Monsoon)]. Field sampling for this was conducted on board a trawler along five cross-shore transects in the Indian sector of the GoM (GoM1 to GoM5) (Figure 1). The second in-situ sampling was in September (Southwest Monsoon) 2018 to verify whether the coastal upwelling process occurs in the GoM. This sampling was carried out onboard CRV Sagar Purvi with three sampling locations in each of the five cross-shore transects (SP1 to SP 5) from the western proximity of the GoM up to 12°N along the southwest coast of India (Figure 1). A portable Conductivity Temperature Depth (CTD) profiler (Seacat, Seabird Electronics, USA) recorded vertical distribution of temperature and salinity. Standard Winkler’s method13 was used to estimate the dissolved oxygen concentration from the surface, subsurface, and bottom waters in the first sampling during 2010-2011 and from many prefixed standard depths (1 m, 5 m, 10 m, 20 m, and 40 m) during the second sampling in 2018. We also used the Regional Ocean Model System (ROMS) outputs of the vertical distribution of relevant environmental variables for the Southwest Monsoon period, obtained from the Indian National Centre for Ocean Information Services (INCOIS), Hyderabad, India (www.incois.gov.in) 14,15. Vertical sections of four cross-shore transects (T1-T4) obtained from ROMS beginning from the southern tip of India (Cape Comorin) towards the GoM were used to comprehend how best the model projection follows the environmental signatures evident in the in-situ observations14,15. Univariate ANOVA identified the significance of the spatial difference in hydrographical parameters between transects sampled in during 2018 (SP1-SP5). Before the ANOVA test, the homogeneity and distribution of the environmental variables were analysed using XL Stat PRO software. Tukey’s HSD post hoc test was performed for pairwise data comparisons. The distribution of the ecological variables was analysed in XL stat Pro (License information – Academic version, Purchased by CSIR-National Institute of Oceanography, Regional centre, Kochi). ANOVA was analysed in PASTUiO software Version 4.02 (Free software downloaded from -https://folk.uio.no/ohammer/past/). Principle Component Analysis (PCA) was also performed to visualise the spatial differences in the distribution of relevant environmental variables. Correlation matrix-based scaling was used in PCA Biplots, having the first two principal components. The dissolved oxygen values were overlaid (dotted line) on the Biplots to indicate the spatial variation. PCA was performed using CANOCA 4.5 software purchased by the CSIR- National Institute of Oceanography, Regional centre, Kochi.