MUR SST

A global, gap-free, gridded sea surface temperature analysis at roughly 1 km resolution — and the containerized pipeline that produces it.

The Multi-scale Ultra-high Resolution (MUR) SST analysis blends infrared and microwave satellite retrievals with in-situ buoy observations into a single daily global field with no gaps. This site describes how the pipeline is put together, how to run it, and how the underlying analysis works.

The pipeline at a glance

Three preprocessing stages convert external data into compact binary observation files. The MRVA analysis stage then fits a multi-scale field to all of them at once and writes the NetCDF product.

flowchart TD
    subgraph sources["Data Sources"]
        OSI["OSI-SAF<br/>Ice Concentration"]
        PODAAC["PO.DAAC<br/>L2P Satellite Files"]
        IQUAM["NOAA iQUAM<br/>Buoy Data"]
    end

    subgraph containers["Preprocessing Containers"]
        LANDICE["Land/Ice Container<br/>makeicefiles.m"]
        L2P["L2P Container<br/>l2p2bic.m"]
        BUOY["iQUAM Container<br/>makedailyiquam.m"]
    end

    subgraph outputs["Preprocessed Data"]
        GDS[".gds files<br/>Land/ice masks"]
        BIP[".bip files<br/>Ice SST points"]
        BIC[".bic files<br/>Satellite observations"]
        BII[".bii files<br/>Buoy observations"]
    end

    subgraph analysis["Analysis"]
        MRVA["MRVA Algorithm<br/>Multi-scale variational analysis"]
        NETCDF["MUR SST Product<br/>NetCDF L4 output"]
    end

    OSI --> LANDICE
    PODAAC --> L2P
    IQUAM --> BUOY

    LANDICE --> GDS
    LANDICE --> BIP
    L2P --> BIC
    BUOY --> BII

    GDS --> MRVA
    BIP --> MRVA
    BIC --> MRVA
    BII --> MRVA

    MRVA --> NETCDF

    style MRVA fill:#e1f5ff,stroke:#0288d1,stroke-width:2px
    style NETCDF fill:#e1f5ff,stroke:#0288d1,stroke-width:2px

Where to go next

Reference material

Everything below is part of this site — the full detail, not a summary of it:

• Configuration — every key in config.json, stage by stage, plus troubleshooting
• Explicit input contract — the named-flag and manifest schema all four containers implement
• Land/ice encoding — bit-level layout of the mask files
• MRVA internals — the MATLAB and Fortran implementation in depth
• Future enhancements — what has been considered and deliberately not built
• Per-container reference: land/ice, L2P, iQUAM, MRVA

The published method

The analysis behind the product is described in:

• Chin, T. M., Vazquez-Cuervo, J., and Armstrong, E. M. (2017). “A multi-scale high-resolution analysis of global sea surface temperature.” Remote Sensing of Environment, 200, 154–169.

The product itself is distributed by PO.DAAC, and the pipeline source lives at podaac/mur.