search

LBL-EESA / TECA

TECA, theToolkit for Extreme Climate Analysis, contains a collection of climate anlysis algorithms targetted at extreme event detection and analysis.
Other
56 stars 21 forks source link

implement ar ranking scale #554

Open burlen opened 3 years ago

burlen commented 3 years ago
Cat Strength Impact Max. IVT[a]
1 Weak Primarily beneficial ≥250–500
2 Moderate Mostly beneficial, also hazardous ≥500–750
3 Strong Balance of beneficial and hazardous ≥750–1000
4 Extreme Mostly hazardous, also beneficial ≥1000–1250
5 Exceptional Primarily hazardous ≥1250

Notes^ Maximum vertically integrated water vapor transport, 3-hour average, units of {\displaystyle {\frac {kg}{m\cdot s}}}

https://en.wikipedia.org/wiki/Atmospheric_river#:~:text=The%20Center%20for%20Western%20Weather,scale%20was%20developed%20by%20F.

taobrienlbl commented 3 years ago

Looks like that Wikipedia article needs to be corrected; that table isn't right. This article gives a correct table, which shows that the AR category is a function of both IVT and the duration of IVT over land: https://www.wunderground.com/cat6/Cat-1-Cat-5-Scale-Atmospheric-Rivers

The Ralph et al. (2019) method of applying the scale (this is the paper that defined the scale) would apply it independently at each point. Computationally, this scale is ideal for spatial parallelization, since it effectively amounts to an independent timeseries analysis for each gridpoint. Alan and Yang have both come up with variations that apply the scale in the context of ARs that have been detected spatially (e.g., detected with TECA BARD), and this would be more complicated to compute since it involves both AR detection and timeseries analysis.

That said, it's debatable whether this AR categorization scale makes sense for areas outside the western United States, since the scale is based on analysis of AR impacts on the West Coast. A very recent paper by Eiras-Barca et al. suggests that the scale should be different for Europe than for the western U.S. We should discuss further before going much further with an implementation.

Full references: