• English 
  • Spanish 

The Atmospheric River Archive

Reminiscent of a river's course seen from bird's-eye perspective, atmospheric rivers (ARs) are narrow structures of intense horizontal water vapour transport thousands of kilometers long (Zhu & Newell 1998). Bound to the wind-field and normally located over the sea, these structures usually travel north-eastward in the extratropcis and are responsible for a large fraction of the poleward water vapour transport. They are also known to trigger hydrological extreme events (precipitation, river-flooding and landslides) and are important contributors of snow accumulation over the polar ice-sheets (Lavers & Villarini 2013, Neimann et al. 2008, Gorodetskaya et al. 2014, Ramos et al. 2015).

This web-page is an exhaustive archive of atmospheric-river arrivals along the European Atlantic sea-board and the West Coast of North America. It was built to explore these events during the course of the entire 20th century and -by comparing the results from two distinct long-term reanalyses- to estimate dataset-related uncertainties.

Focusing on the extended winter season (October to March), AR-arrivals were detected separately for 8 target regions in Europe and 5 target regions in western North America (see panel on the right, ERA-20C's orography is shown). To this aim, 6-hourly instantaneous time-series from 3 distinct reanalysis products were used.

Two sub-archives are provided:

  1. AR arrivals in the historical past (1900-2010) using the two 20th century reanalysis datasets available to-date: NOAA-CIRES 20th Century Reanalysis v2 and ECMWF ERA-20C (Compo et al. 2011, Poli et al. 2013).
  2. AR arrivals in the recent past (1979-2014, the satellite era) using ECMWF ERA-Interim (Dee et al. 2011). This part of the archive is going to be updated once a year or on request from the interested user.

For each AR detected in at least one of the aforementioned datasets, the following variables are mapped:

  1. The intensity of the vertically integrated horizontal water vapour transport (IVT) in kg m-1 s-1. This measure is indicated by the colour shading and the length of the blue vectors.
  2. The direction of IVT, indicated by the orientation of the blue vectors.
  3. The atmospheric river track (the light blue line) detected by the method proposed in Brands et al. (2016) (under revision).

AR structures are required to have a minimum length of 3000 km. In the target region, IVT must exceed the 90th percentile threshold. Along the AR-track, the 75th percentile must be exceeded except in Morocco and southern Iberia where the 85th percentile is used instead. All thresholds are calculated separately for each grid-box and month considering the reference period 1979-2009.

Note: During the early 20th century, the correlation coefficient (r) between the seasonal AR-occurrence counts from the two long-term reanalyses is insignificant for western North America. This is contrasted by r > +0.6 in Europe (see Brands & Gutierrez, in revision) indicating that the dataset-induced uncertanties are much smaller in this region.

Contact: swen.brands@gmail.com

< >

The corresponding netCDF files, containing six-hourly AR presence/absence time series from four reanalyses, are available here.

References (click here)

This archive is the results of a joint research effort undertaken by:

  1. Meteorology Group (UC-CSIC). Instituto de Física de Cantabria (IFCA). Swen Brands was funded by the JAE-PREDOC programme.
  2. MeteoGalicia, Consellería de Medio Ambiente, Territorio e Infraestruturas - Xunta de Galicia
  3. Universidade de Santiago de Compostela, Grupo de Física Non Lineal

MeteoGalicia    Santander Meteorology Group (UC-CSIC)    USC    Predictia