The Use of Generalized Additive Model (GAM) To Assess Fish Abundance and Spatial Occupancy in North-West Bay of Bengal

Abstract

The remote sensing data, such as sea surface temperature & chlorophyll concentration obtained from various satellites are utilized by Indian National Centre for Ocean Information Services (INCOIS) to provide Potential Fishing Zone (PFZ) advisories to the Indian fishing community which plays a vital role in national GDP. The data on Sea Surface Temperature (SST) is retrieved regularly from thermal-infrared channels of NOAA-AVHRR and chlorophyll concentration (CC) from optical bands of Oceansat-II and MODIS Aqua satellites for the identification of Potential Fishing Zones (PFZ) in Indian water. PFZ information has certain limitations, such as it can't predict the type of fish available in the notified fishing zone. In this dissertation, I have worked towards the development of short-term Hilsa shad predictive capabilities in a sustainable way. An effort has been taken to categorize all essential biological, environmental and climatic signals that have a direct or indirect impact on the Hilsa shad distribution. Remote sensing, ocean biogeochemical modelling, and statistical modelling approach have gained an increasing importance to study the marine ecosystems as-well-as for understanding the dynamics of the oceanic environment. Shad habitat has been studied from the geo-tagged fish catch data and oceanic/ecological indicators as predictor variables. For short-term prediction, the variables have been derived from a biophysical model, configured at INCOIS, using Regional Ocean Model System (ROMS) and remote sensing data. Using generalized additive model (GAM) Catch per Unit Effort (kg h?1) has been calculated as a response variable. Probability maps of predicted habitat with no fishing zone information have been generated using geographic information system.