define satellite anomaly products (what's weird)

[bbest]

From Satellite Team product development:

See:

• Satellite Team - Google Drive
• Biodiversity Team Spreadsheet - Google Drive

Satellite Level 3, 9 km (monthly):

• SST
• Chl-a
• nFLH/Chl-a ratio (nFLH: normalized Fluorescence Line Height [phyto proxy])
• PAR
• Adg: absorption of detritus and CDOM
• reflectance of remote sensing RRS (667 nm), or ratios for turbidity (how brown/red water is)
• Spectral slope backscatter (cellsize)

Others:

• HyCOM for mixed layer depth (see #9)
• OSU Ocean Productivity
  • VGPM, net primary production from OSU: remove nearshore data
  • CDPM may be easy

For combining anomalies into single super wierd environmental data product:

• slider to zoom in/out 9km at global, 1km at local
• Bopp et al (2013) Multiple stressors of ocean ecosystems in the 21st century: Projections with CMIP5 models

Physical parameters (Hycom or ECCO-2)

• Mixed Layer Depth
• Current velocity
• Salinity

Climatology: 2003 - 2007

Per-pixel anomalies based on: [Obs - Mean(2003-2007)]/SD(2003-2007)

Bopp et al. 2013 approach for combining variables for anomaly indices.

Critical: Data flow for on-the-fly analyses

Server Storage

Since Github only allows 1 GB of total file storage (and only 100 MB per file limit), we need to determine a space for storing files. Let's use the server and 2 TB of storage that Qiao Qiao (@asurinsaka) setup at mbon.marine.usf.edu. He can give you SSH/SFTP accounts to access the big folder there /mbon.

[eqmh]

MARCH 22: Satellite Level 3, 9 km (monthly) currently being processed:

• SST (night time)

• Chl-a (default)

• nFLH

• PAR

• Adg-443: absorption of detritus and CDOM

• Rrs (667 nm) - not good at global scale. Recommendation: use 1-km data for Sanctuaries

• Spectral slope backscatter

• Dan will generate separate raster files (NetCDF4) for anomalies (z factor and anomaty [Obs-mean]), averages, standard dev., and climatologies

• Data will go to this server: mbon.marine.usf.edu.

• Need to develop a concept for a data processing pipeline

• HYCOM products: Dan is figuring out how to download and deal with varying projections.

• NPP product: we can choose to remove data (1 degree window) from very near shore, and to use cloudless product using Kriging. Maria suggests including VGPM (where from?).

• Maria encourages us to focus on scholarly products stemming from these efforts.

• SeaWiFS?: we should have a SeaWiFS version and not a merged product. SeaWiFS data can be used to reconstruct seascapes for defined domains (with some level of error).

• Maria: consider having all the data hosted at NASA as opposed to locally. Dan suggests hosting the data locally as we will have products more quickly. ~2003 - 2007 climatology: this period is where MODIS and SeaWiFS overlap the most. ~Too early to talk about layouts.

[eqmh]

March 24, 2016 telecon:

Luke: up to 35 indicators. The group is using a tier system for prioritizing specific indicators across sanctuaries and those for international areas of interest.

Frank: Anomalies not really dependent on the quality of the variable. Primary productivity products near the coast would be OK to use near the coast.

Jenn: The group is working on a generalized set of infographics that can be used across ecosystems:

Gabrielle:

• something about forage species.
• How eDNA can help management in the sanctuaries?
• Make infographics accesible to a wide range of users
• What local example products will look like? Like for Galapagos.

Jenn Bosch:

• Make products with focus on local scale usable across various types of systems. Make them useful to the international community

[bbest]

Qiao upped memory on mbon server 1 to 8 GB, so now allows me to project and visualize the global rasters :)

Here's a quick test:

marinebon.github.io/sdg14/satellite/explore.html

[eqmh]

APRIL 06:

Satellite Level 3, 9 km (monthly) fully processed and anomalies calculated for: SST4 (night time) - MODIS Chl-a (default) - MODIS (SeaWiFS will follow - need to download and process [monthly available]) and should be quick. nFLH - MODIS all done PAR - MODIS all done SeaWiFS will follow - need to download and process [monthly available]) and should be quick. Adg-443: absorption of detritus and CDOM - MODIS (SeaWiFS will follow) Rrs (667 nm) - not good at global scale. Recommendation: use 1-km data for Sanctuaries Spectral slope backscatter - MODIS (SeaWiFS will follow): NASA General IOP algorithm (G-IOP; for updates may take up to a month for new data) Dan downloaded daily NOAA Coral Reef Watch products (5 km, global): SST, SST anomaly, Degree Heating Week (DHW). Dan will later bin to month but need to confirm with Frank (not happy about this product as it only goes back to 2013)

Dan is now downloading daily GHRSST (blended 1/4 deg grid starting 1981; via PODAAC) and should finish this week: this is for SST and sea ice fraction (0-1). Dan will bin to month and calculate anomalies and so on. We need to decide which SST product we will use for the SDG14 product

NCEP Met products (via OBPG): 1 degree grid, 6hr, wind u and v and 10 m above surface, since 1978 - Dan needs to download. For products: we could use this product to show wind intensity (not u and v components)

HYCOM: still need to figure out how to re-project with 3 different projections. Maria was going to talk to Behrenfeld's group (Ocean Productivity Group) about this. Can Ben deal with different projections?

Products will keep spatial resolution and projection of each product and we deal with how to display data later.

Need to define how we are going to blend anomaly products and what products will look like.

Seascapes: spatial assessment of seascapes with reference to climatological areas.

Frank suggest using thumbnails to display products generated on the fly.

We need to choose a single type of area, like LMEs or EEZs, from which anomalies and OBIS occurrence data (or list of taxa from Aquamaps) would be extracted. These areas can be obtained from WCMC. Aquamaps would provide a much larger list of taxa from ROIs that can be related back to anomalous environmental conditions to determine which species could be affected by them.

Ben needs to ask Kristin Kaschner (Aquamaps) for data and permission to use it.

Ben will get started with EEZs and SST and Chl anomaly products and linkages to biological datasets. The tool would allow to assimilate shapefiles for analyses.

OBIS role: taxonomic (provide a list of taxa) and functional group.

Next steps - Ben needs technical support with PostGIS applications. Qiao and Tiler can help with this.

Need to decide if we use WCMC shapefiles or those from other sources. WCMC marine regions are cumbersome to use.

SDG14 telecons will take place every two weeks from now on, starting again on Friday 21.

[eqmh]

APRIL 7: on EEZ, moved this comment to deal with country EEZ issue in global product · Issue #17 · marinebon/sdg14

[bbest]

more on EEZ, moved to this comment in #17

[eqmh]

Agenda - June 2, 2017 telecon:

• Show of hands: MBON SDG 14 team for August 8-9, 2017, in Monterey, CA (Venue: Moss Landing Marine Lab.) -Update from working groups (Jen, Luke, Enrique, Ben) -Possible test groups for product: Suggestion at May 26 MBON meeting: form a group to 'beta test' the MBON portal, with a goal to eventually incorporate into the MBON Axiom portal -International demonstration and engagement: timeline? (Australia/IMOS - Tim Moltman next few days, weeks) (St Lucia, soon?) (St. Vincent others - August?) (Costa Rica - AmeriGEOSS Aug 1-4)

From Dan: Satellite directories on the mbon server now have imagery with two different spatial resolutions. Along with the original set of products at 9km, there is now a set of products spatially binned to 0.25 degree. These are much smaller and should be better suited for on the fly calculations in polygons or other ROI that the user may define.

Grid sizes (global): 4320x2160 (9km) 1440x720 (0.25 deg) - FYI, this matches the GHRSST products

Within each product directory, there are now four sub-directories. Here is an example using chlor_a: mbon/data_big/satellite/chlor_a/anom_9km (anomalies at 9km) mbon/data_big/satellite/chlor_a/anom_27km (anomalies at 0.25 deg) mbon/data_big/satellite/chlor_a/clim_9km (climatologies at 9km) mbon/data_big/satellite/chlor_a/clim_27km (climatologies at 0.25 deg)

There are also some new climatology products. The original set uses 2003-2007. However, in looking at these, I noticed some gaps, so I created a set that uses 2003-2016, which are much smoother with many less gaps. The filenames tell you which years are being used. New sets of these can be easily generated, so if you have ideas on what years would be best, I can easily make them up. These climatology products are used to make the anomaly products, so reduced gaps in the climatologies leads to better-looking anom ("what's weird") products. We will use the 2003-2007 version when we incorporate SeaWiFS data, which is next on my list.

On MBON SDG14 and IMOS (Integ. Marine Observing System - Australia): · Tim Moltman may be in a position to provide constructive advice on our MBON SDG14 prototype product. · IMOS data, including biological data, can be discovered and accessed through the Australian Ocean Data Network (AODN) – see here. If you choose ‘biological’ parameters you’ll see what is there. Because of the complexities of biological data, we’re also developing a number of more specialised data tools which we can discuss if interested. · The Acoustic tracking data is handled in a separate database can be accessed here. Gabrielle, attached is an email from Xavier Hoenner reporting back on his trip to Barb Block’s lab. · Animal Borne Instrumentation is on the agenda at JCOMM Observations Coordination Group (OCG), and Lars Boehme gave a nice presentation on this at their most recent meeting – see here. · That said, in IMOS are not really on top of taxonomic data yet, and this is where we are hoping to work more closely with the Atlas of Living Australia (ALA). ALA have supported development of the FishMap app. ALA also supports OBIS Australia (through CSIRO).

NOTES FROM TELECON:

• Argie Kavvada: Deadline for submission of a dedicated SDG14 side event to the GEO Plenary is June 15.
• Ben will work with Argie in the development of SDG14 relevant storymaps.
• Kathleen's group has already built an eDNA storymap.
• Ben: Larval dispersion models being developed by Ben's students on Shiny with focus on sources and sinks - where particles come from and using mget tools for broad geographical applications. The model uses AVISO or HYCOM fields. Frank: from the Blue Planet meeting it seems that there is interest in Small Island nations on these type of models.
• Luke: Monique has been working on PP time series. Codes for analyzing RVC fish abundance/dominance are ready and can be used for other datasets. Luke met with Jerry Ault last week. (Something about a conversation with a group in Seattle working on citizen science).
• Abby: colelcted data from EUROBIS, about 3000 species (123K records), from the Norwegian Sea with biomass and abundance. Abby is working with R code from Luke and Megan. Integration (Ben): Abby can send data with species classified based on traits as Megan is doing with RVC datasets, EUROBIS would allow to build fish abundance time series.
• NOAA (NESDIS) has agreed to host seascape data via Coastwatch.
• Ben: working with Tylar for getting the MBON server going. Also working on SBC & DataOne-OBIS-ERDDAP translations of the integrated fish dataset using your Vocab mappings: Translate Dataset from DataOne to OBIS columns. This should lead to mapping the interplay between OBIS, DataOne and ERDDAP.