Some rambling

Assign event variables to the vms pings

What is all the magic behind the trip_assign function?

code

rtip

Here we make a comparison of assigning things like gear from eflalo (logbooks) data to tacsat (vms) data using the approach provided in the ICES VMS datacall vs. an alternative approach strictly witin dplyr. Done both to verify the understanding of what the trip_assign function does as well as making steps towards the future aim of running things via the duckdb-framework.

Einar Hjörleifsson

Calculating effect of areal closure

code

rtip

One may be of interested in a particular area closure. The first question that springs to mind is what effect does it have on fishing? Here we explore some algorithm that may be of use when trying to answer that question.

Einar Hjörleifsson

Correcting/replacing logbook positions

code

rtip

Captains fill in the logbooks, including the geographic position of each activity. There are some errors in them like fishing points being recorded on land. Also the reported precision may be poor like recording to the nearest minute, leading to striations when rasterizing at a reasonably high resolution. Here an attempt is made to obtain ‘better’ values from the ais-data.

Einar Hjörleifsson

On duplicates in the new logbooks

code

rtip

The total number of records in the trip table is 257579 while the number of distinct vessel, departure and arrivals are 254111, a diference of some 3468 records.

Einar Hjörleifsson

On trip overlap

code

rtip

While using/testing ‘1.3.7 Remove trip with overlap with another trip’ in the ‘1_eflalo_tacsat_preprocessing.R’ I was dropping more trips than expected. Ran a little test on synthetic data to try to understand why. Turns out that the code in the workflow can be quite generous when dropping trips, even though only one trip start/end time may be an issue. added an alternative approach for testing time overlaps limited to sequential trips - apologies for not using data.table. What to do about overlaps is then of course another matter.

Einar Hjörleifsson

Metiers and things

code

rtip

In my ignorance I assumed that the metiers, targets and gears in the ICES vocabulary were hierarchical. In the sense that if one has a valid metier level 6, one could derive valid metier level 5, target and gear from that. Here a little script is provided that shows that this is not quite the case.

Einar Hjörleifsson

on_door_spread

code

rtip

library(tidyverse)
library(duckdbfs)
gear <- 
  nanoparquet::read_parquet(here::here("data/gear/gear_mapping.parquet")) |> 
  select(gid_agf = agf_gid, veiðarfæri) |> 
  m…

Einar Hjörleifsson

Dagur á sjó

code

rtip

Krókaveiðar standa nú yfir á fullu. Einn af þeim dögum þar sem hvað flestir bátar voru á sjó var 10. júní. Fyrir myndbirtingu á dagsverkinu er fyrst fundnir þeir bátar sem að landað hafa krókafiski á yfirstandandi sumri og sá listi síðan notaður til að finna stk-ferlana á umræddum degi.

Einar Hjörleifsson

STK data

code

rtip

A short overview of the STK data in MFRI database

Einar Hjörleifsson

Bækats datacall

code

rtip

Einar Hjörleifsson

Jigger-Fishtrawl overlap

code

rtip

Was asked about the overlap between jiggers and fishtrawlers. Here is a quick hash based on data from 2008-2024 with the effort (time) expressed on a log scale.

Einar Hjörleifsson

Points in polygon via duckdb.

code

rtip

Point-in-polygon analysis solves the question if a given point in a plane is within a specified polygon. If we have multiple polygons we can also solve the question in what polygon a point belongs to. Solving this within R is tried, tested and straight forward if the data is not too big. But what if we have a dataset that is so big that the processing challenges the available memory? Here we will put duckdb to its test.

Einar Hjörleifsson

On gridding

code

rtip

This post originated in that moi wanted to check if the code ‘x = lon %/% dx * dx + dx/2’ would give the same results as CSquare gridding as implemented in {vmstools}. Short answer is no. The former is more consistent, always resulting in binning such that if a value is exactly at ‘grid’-boundary it is included if on the lower boundary, excluded if on the upper boundary. In the {vmstools}-csquare case this depends on the global quarters. Now, somebody may argue that the practical consequences is miniscule and I would fully concur :-) Say so because a very small portion of real numeric data will be exactly on the grid boundary as focused on here. Leaving the devil in the details aside, the main message is that using ‘x = lon %/% dx * dx + dx/2’ is OK to use within R, arrow or duckdb code workflow. Why is this important? Because in the latter two workflows a csquare algorithm is not available/implemented.

Einar Hjörleifsson

On anonymity

code

rtip

library(tidyverse)
library(duckdbfs)
library(mapdeck)
library(tictoc)
library(h3)
tic()
tic()
q <-
  open_dataset(here::here("data/ais/trail")) |> 
  filter(.cid > 0,      …

Einar Hjörleifsson

Distribution of seaweed and kelp harvesting

code

rtip

I was asked what AIS data can inform us about distribution of seaweed and kelp harvesting. Normally one would filter out information from the logbooks and landings database with respect to vessels and dates of trips and then plot all records that fall within such trips. But since that is lacking here we, for the time being we filter the AIS for some suspeted vessels as a first go. If only to show what is feasable. For now the statistics shows the number of distinct years a hexagon grid was visited by a vessel.

Einar Hjörleifsson

Working with AIS data using {duckdb} and {h3}

code

rtip

Must admit that I have not fully grasped the duckdb-concept but what I have explored is magical. The main R-package is {duckdb} but {duckdbfs} provides a nice wrapper for making connection to files easier (similar to the {arrow}) package. On top of it {duckdbfs} has some intriguing spatial function. Here a little exploration of the AIS data is made to put the feature to a test.

Einar Hjörleifsson

Seaweed, fishing and beautiful creatures

code

rtip

I was asked to provide a map of net fishing in shallow waters, sea-weed bed distribution, seal rutting areas and coastal bird distributions. Here is the code behind that delivery.

Einar Hjörleifsson

Fishtrawl footprint - by proportion of total catch

code

rtip

One of the characteristics of fish bottom trawling is that the catch is taken in a very limited area of the total trawl footprint. E.g. about 60% of the catch is taken in 10% of the total area. Here we ‘look’ at this from the other angle and show show the areas were less than 5%, 10%, .. percentage of the total catch is taken. These kind of acrobatics may be a foundation for some intriguing discussions.

Einar Hjörleifsson

About depth

areas

I was asked today how many square kilometers of the Icelandic EEZ was below 1000 meters depth. Here a short code-note of the answer is provided.

Einar Hjörleifsson

leaflet addGeotiff

code

rtip

Just a quick note on using geotiff file in leaflet

Einar Hjörleifsson

Norwegian AIS data

code

When it comes to commerical fisheries data, the Norwegians are quite liberal. Actually extremely liberal. They basically have everything accessible. Here a quick code hash is provided as a case in point, showing the longline fleet individual hauling operation in 2022 to 2024.

Einar Hjörleifsson

Lumpsucker fisheries - effort distribution

code

rtip

A little overview of the distribution of lumpsucker fisheries and area closures. Turns out that there is not much of an overlap!

Einar Hjörleifsson