Identification and Forecast of Potential Fishing Grounds for Anchovy (Engraulis ringens) in Northern Chile Using Neural Networks Modeling

Round 1

Reviewer 1 Report

In this study, the authors present an application of a neural network model in predicting the fishing ground of Peruvian anchovy in northern Chile. The study is well designed, and the presentation of the results is good. The major issue I have with this manuscript is the lack of detailed specification of the model, specifically in section 2.2 of the text.

More details about the model need to be given. For example, which version of the Tensorflow was used, if it is used in this study. Keras is just a front-end. How many layers were used, what is the learning rate, what is the dimension of the input data, etc. The authors need to show the detailed specification of their model here so that others can replicate it. Preferably, the authors can add their python notebook in the appendix.

What is the temporal scale of all the environmental variables? Are they daily, weekly or monthly data? Figure 5 suggests they are not daily data, since daily satellite data usually do not have complete coverage over an area this large. If so, these are not instantaneous states of the ocean.

The discussion section can be further improved. Combine paragraphs between lines 187 and 199.

Also, please check journal guidelines on the use of the decimal point.

Author Response

Please see attahment

Author Response File: Author Response.pdf

Reviewer 2 Report

I consider the basic premise, method developed and questions asked by this study as valuable and interesting. The authors outlined a novel method to identify and predict potential fishing grounds for small-pelagics through neural networks.
The strength of the study is the novelty and broad usefulness of this approach. The methodological approach and the statistical methods are valid and correctly applied to the data. Results and data interpretation are coherent and well presented. Tables and figures are also well presented and titled. The discussion is detailed enough and the scope of the ms very well presented with high quality of reference and examples.

Author Response

Please see attachment

Author Response File: Author Response.docx

Reviewer 3 Report

General comments

This paper tries to predict fishing grounds of anchovy Engraulis ringens in northern Chile by implementing a neural network model using as information fishing days per industrial purse seine vessel. The fishery is managed by a type of TAC (Total Allowable Catch) called Maximal Annual Catch per Ship Owner (MCSO), and therefore this could help the fleet to decrease the fishing effort and reduce the operational costs by optimizing/reducing fishing trips. The objective of the manuscript, in general, is interesting from the fisheries management (e.g., for fisheries management authorities) as well as economic (e.g., for fishing sector) points of view especially for an important fishing country like Chile.

The paper is well-written with a good structure and good objectives, and shows some interesting results. However, the manuscript still lacks some background about the fishing area and the fishery itself especially for the readers that are not familiar with Chilean fisheries, and also to provide some reflections on the applicability of these results from management perspective. The manuscript should become acceptable for publication pending suitable minor revision in light of the comments appended below.

More specific comments:

Introduction:

Line 40 – 41: "This species provides 75%-80% of the annual catch of the purse seine fleet of northern Chile". Please add more information about the fishery. For example, the percentage of the purse seine fleet out of the total fleet in northern Chile. Also, whether the species is caught by other fishing gears and fleets and its contribution in each.

Line 49 – 52: "Since 2003, with the application of the Chilean administrative rule of Maximal Annual Catch per Ship Owner (MCSO)". Please add more information about how this fishery is managed. For example, what is the share of this MCSO of the estimated recruitment potential/annual yield.

Materials and Methods

Line 77 – 79: "The georeferenced catch data of E. ringens per vessel and the daily trajectories of each ship were obtained from the database of the fishing company Corpesca S.A. This database contains 1,5 million records for the 2003-2020 period, with a spatial resolution 79 of 3x3 nm2". First, how many vessels are composing the fleet of this company and the importance (either the share of catch or number of vessels) of this fleet in respect to the whole fishery of northern Chile? Secondly, is there difference in the data resolution/quality between the earlier years (e.g., first decade 2003-2010) and the recent years (e.g., first decade 2010-2020).

Line 124 – 126: "To validate the neural network model as a predictive tool we chose the 36 days (10% of the total days in a year) with greatest total captures in the study period (authors’ criterion)". Did the authors try to validate the model using another alternative set of data?

Results

Line 161 – 162: "Fig. 5 shows the results of applying the model to the four days with greatest catch of E. ringens in the entire time series". Again, why were 4 days chosen? This is not explained in the material and methods section.

Figures:

-        Figure 2: please add the source of the data.

Tables:

-        Table 1: please specify in the caption that these numbers are the number of cells.

Discussion

Line 200 – 208: In this paragraph the authors explained that the neural network model applied in this study has advantages and shows better performance than conventional statistical models such as GLM (Generalized Linear Models) and GAM (Generalized Additive Models). It would be also useful to specify the limitations and disadvantages that authors found in the model.

Suggestion: taking into account that the purpose of this study is to help the fleet to decrease the fishing effort and reduce the operational costs by optimizing/reducing fishing trips, the authors should explain, at the end of the discussion, the applicability in regards to the MCSO and how this can be applied from a management authority point of view.

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have done a good job revising the manuscript. No more comments.