SOM Composites

Enter any set of years and this shows how often each circulation pattern occurred in those summers, compared with the rest of the record (1959–2025). It is a lightweight version of the composite tools at NOAA/PSL, run against the self-organizing maps from my own work: pick the years of an event you care about — fire seasons, ENSO phases, whatever you like — and see which patterns were more or less common.

Fill in the second field to compare two sets of years directly — an enhanced composite, where each value is how many more days per season a pattern occurred in set A than in set B. Leave it blank and set A is compared with the full-record climatology instead.

The maps are 4×5 SOMs of daily 500-hPa Z and SLP over Alaska and the North Pacific, defined in Hostler et al. (2025) — cite that paper rather than this page. The training code is in synoptic-SOM (v1.0). Values are days per season, so +2.5 means that pattern occurred two and a half days more often than an average summer.

The baseline period sets what “average” means. It defaults to the whole record; narrow it if you need the climatology to match the period of some other dataset you are working with. It also governs which years the significance test draws from, so the comparison stays internally consistent.

The checkboxes above the download links choose what goes in the figure — often the useful pair is just the node patterns beside the anomaly.

Computing…

Node patterns

Figures and data from this tool are CC BY 4.0 — reuse them freely, including in papers and talks, and attribute with the citation marked ★. Patterns derive from ERA5, generated by the Copernicus Climate Change Service. CC BY 4.0

@article{Hostler2025,
  author  = {Hostler, Joshua and Bhatt, Uma S. and Bieniek, Peter and
             Ballinger, Thomas and Borries-Strigle, Cecilia and Burgard, Mitchell and
             Chriest, Jonathan and Fischer, Elizabeth and Lader, Rick and
             Stevens, Eric and Strader, Heidi and Thoman, Richard and
             Waigl, Christine and Parish, Zachary},
  title   = {Daily {SLP} and 500-hPa {Z} {SOM} Patterns Explain Lightning
             Variability in {A}laska},
  journal = {Artificial Intelligence for the Earth Systems},
  year    = {2025},
  volume  = {4},
  number  = {4},
  doi     = {10.1175/AIES-D-24-0122.1},
  url     = {https://doi.org/10.1175/AIES-D-24-0122.1}
}