Subsampling OceanSODA-ETHZ-HR

Based on a request by Helen Findlay, who is on the ESA Ocean Health - Ocean Acidification project with Niki and me.

• Date : 19-01-2024
• Creator : Luke Gregor
• Contact : gregorl@ethz.ch

Contents

Figures showing maps and occurrences for three species

In this directory, there are the following files:

• figures: some plots to show the subsampling and random sampling worked
• variable_metadata.yaml: a file that tells everything about the variables in the CSVs (units, description, source project, etc.)
• random_ocean_locations_sampled10k.csv: sampled from 10k random points in the ice-free ocean, see the map.

Species subfolders contain three files:

• [spp_name]-matched.csv: the matched version of the original input - may include NaNs and duplicate locations from the OceanSODA-ETHZ-HR product if the locations are close in time and space.
• [spp_name]-matched-no_duplicates.csv: a copy of the file above, without any missing values or any duplicate locations
• [spp_name]-sampled10k.csv: random samples from any time at the locations in no-duplicates.csv

Original email

Below is Helen’s original email.

Attached three csv files to start with for Magallana (Mg), Pinctada (Pm), and T. squamosa (Ts) occurrences. They have the lat and long given as decimals, but there are also rounded up/down lat longs to nearest 0.5 degree, which I used for matching with OceanSODA-ETHZ dataset. I’m not sure what you’ll need for the higher res datasets, if you can do nearest available position to the original lat longs that would be ideal. Then, as I said, I need another csv file for each species which gives background data, of 10,000 data points, randomly chosen from the global dataset at any time or location (actually this can be the same for all species). I also used a restricted background dataset of 10,000 points randomly chosen from within the same 1 x 1 degree box around the occurrence points but from any time point. A similar restricted background would also be useful here, maybe relating to the 0.25 x 0.25 degree box around each occurrence point but from any time point. Does that make sense? Can you generate that, or is it easier for me to just give you more CSV files to run through? This would need to be for each species. Ideally variables required: temp, salinity, DIC, alkalinity, pH, arag, pCO2