Analisis Daerah Penangkapan Ikan Pelagis Kecil Menggunakan Model GAM dan Model MaxEnt di Laut Jawa

Abstract

Small pelagic fish are a group of potential fish resource types in the Java Sea. The high fishing effort over the last few years has led to an increase in utilization rate until it reaches an optimum level and tends to be over fishing. Purse seine and bauke-ami vessels are the dominant fishing vessels in these waters. Both types of vessels use light aids in the fishing process. The main obstacle to the success of fishing operations, apart from the dynamic movement of fish, is also caused by oceanographic factors in the waters. Fish respond differently to oceanographic factors, making it difficult for fishermen to carry out fishing operations. To help make it easier for fishermen to catch fish in the Java Sea, it is necessary to develop a better fishing area information system, so that fishing efforts can be more effective and efficient. The aim of this study was to develop a small pelagic fishing area prediction model based on oceanographic parameters, namely sea surface temperature and salinity. This study uses Generalized Additive Models (GAM) and Maximum Entropy (MaxEnt) for prediction modeling of fishing grounds in the Java Sea. The data used are Visible Infrared Imaging Radiometer Suite Boat Detection (VBD) images to determine the position and distribution of fishing vessels, sea surface temperature (SST) images, and salinity data as parameters for DPI modeling. VBD data assumed to be the location of fishing vessels to catch fish is stated as a response variable and oceanographic data is stated as a predictor variable in the development of the GAM model and the MaxEnt model. Data for 2018 is used as training data and data for 2019 is used as data validation. The analysis of the GAM model and the MaxEnt model was carried out using the same software, namely RStudio. The predictive performance of the model is assessed using the area under cover (AUC) matrix. The DPI prediction map generated from the GAM and MaxEnt modelers is expressed by the habitat suitability index (HSI) value. The GAM model for DPI prediction based on SST and salinity parameters shows that the presence of small pelagic fish is in the SST range 28 – 28.7 oC and salinity 32 – 33 psu. The MaxEnt model shows that the highest probability of fish distribution is in the range of SST 27.6 – 31.0 oC, and salinity 32 – 34 psu. Salinity is the most influential parameter in the distribution of small pelagic fish. Potential fishing areas, either based on the GAM model or the MaxEnt model, are found more in the central and eastern parts of the Java Sea closer to the southern waters of Kalimantan Island. The GAM model and MaxEnt model show accurate results in predicting small pelagic DPI in the Java Sea.