MODELING DISTRIBUTION OF SAURY CATCHES IN RELATION WITH ENVIRONMENTAL FACTORS
Известия ТИНРО(2019)
Abstract
Pacific saury Cololabis saira is widely distributed in the North Pacific, with commercial harvesting in the area between 140–172о E. Relationship of its commercial catches distribution with environmental factors is investigated using the daily SST data, the daily data set of multivariate ocean variational estimation system (MOVE) produced by Meteorological Research Institute (Japan) for the area between 140–159о E (about 95 % of all catches and 100 % of the Russian catches of saury were landed in this area in 1994–2017), and the daily set of saury catches position with 1 km resolution collected by the Russian vessel monitoring system. Spatial resolution for all data sets is upscaled to the resolution of MOVE system (0.1 x 0.1 degree). Contribution and permutation importance for the catch distribution are estimated for 184 possible combinations of SST and MOVE products with the lags of 0–7 days and moving average window from 0 to 7 days using the method of maximum entropy (MaxEnt). For synchronic relationships, the best results are found for SST, water temperature at 50 m depth and its spatial gradient, moreover, SST provides high contribution with the lag up to 2 days and the temperature at 50 m and its gradient — with the lag 3–7 days. The same sets of environmental parameters are used for the catches distribution modeling with GAMs and Random Forest techniques; the latter method shows better accuracy and other indices of the confusion matrix. Year-to-year changes of the total area with predicted conditions favorable for the saury fishery within the EEZ of Russia and Japan correlate strongly (r = 0.96, p < 0.05) with the total annual catch of saury, in particular for the extreme years (high catches in 2008, 2014, and 2018, low catch in 2017).
MoreTranslated text
Key words
pacific saury,catch,north pacific,sdm software (species distribution modeling),sea surface temperature (sst),method of maximum entropy (maxent),generalized additive model (gam),random forest technique
AI Read Science
Must-Reading Tree
Example
![](https://originalfileserver.aminer.cn/sys/aminer/pubs/mrt_preview.jpeg)
Generate MRT to find the research sequence of this paper
Chat Paper
Summary is being generated by the instructions you defined