FLEXPART–ERA-Interim simulations with 3 million parcels globally (1979–2019)

Description
This data set contains FLEXPART simulations driven with the reanalysis ERA-Interim from 1 February 1979 to 30 August 2019. The simulations cover the entire globe (90°S to 90°N and 180°E to 179°W) with 3 million homogeneously distributed air parcels representing the entire mass in the atmosphere. 

Software
The following software versions were used: 

flex_extract v.7.1.2-dev (this version); see Tipka et al. (2020) for details. 

FLEXPART v10.4 (this version); see Pisso et al. (2019) for details. 

Forcing data
FLEXPART was driven with global reanalysis from ERA-Interim (Dee et al., 2011). The ERA-Interim reanalysis was downloaded and pre-processed using flex_extract v.7.1.2 (Tipka et al., 2020) using 6-hourly reanalyses at 00, 06, 12, and 18 UTC supplemented with 3- and 9-hourly forecasts (03 UTC and 09 UTC are taken from the forecasts initialised at 00UTC; and 15 UTC and 21 UTC are taken from the forecasts at 12 UTC). The settings are detailed in CONTROL_EI_1deg.global. The output of flex_extract were EIgYYYYMMDDHH files for each time step from 1 February 1979 to 30 August 2019 on a global 1° grid and contain the following variables: u, v, etadot, t, q, qc, sp, sshf, ewss, nsss, ssr, lsp, cp, sd, msl, tcc, 10u, 10v, 2t, 2d, z, lsm, sdor, cvl, cvh, sr; where the first 6 variables are available on 60 model layers. 

Simulations
Simulations with FLEXPART v10.4 were performed in forward mode and subdivided into annual chunks; i.e. simulations for each year were started on December 1st 00 UTC of the previous year and run through to January 5th 18 UTC of the following year in order to run all years in parallel. For example, simulations for the year 2000 were started on December 1st 1999 at 00 UTC and run through to January 5th 2001 18 UTC. The overlap in December guarantees a coherent evaluation of each year with trajectories expanding up to 30 days into the past. 

Table 1 shows the main settings of the FLEXPART simulations as set in options/COMMAND and options/RELEASES.

Table 1: Main simulation settings (set in options/COMMAND; flags not listed here are set to 'off' [0] and in options/RELEASES; note that dates have been omitted for comprehension).

options/COMMAND

LDIRECT
1

Simulation direction in time; 1 (forward)

LOUTSTEP

10800

 Interval of model output (s)

LOUTSAMPLE

900

Interval of output sampling  (s), higher stat. accuracy with shorter intervals

LSYNCTIME

900

 All processes are synchronized to this time interval (s)

CTL
2

CTL>1, ABL time step = (Lagrangian timescale (TL))/CTL, uses LSYNCTIME if CTL<0

IFINE

4

Reduction for time step in vertical transport, used only if CTL>1

IOUT
1

Output type: [1]mass 2]pptv 3]1&2 4]plume 5]1&4, +8 for NetCDF output

IPOUT

1

Particle position output: 0] no 1] every output 2] only at end 3] time averaged

LSUBGRID

1

Increase of ABL heights due to sub-grid scale orographic variations; 1] on

LCONVECTION

1

Switch for convection parameterization; 1] on

MDOMAINFILL

1

Switch for domain-filling, if limited-area particles generated at boundary

options/RELEASES

NSPEC

1

Total number of species

SPECNUM_REL

1

Species numbers in directory SPECIES

LON1

-179.00

Left longitude of release box -180 < LON1 <180

LON2

180.00

Right longitude of release box, same as LON1

LAT1
-90.00

Lower latitude of release box, -90 < LAT1 < 90

LAT2
90.00

Upper latitude of release box same format as LAT1

Z1
0.00

Lower height of release box meters/hPa above reference level

Z2
10000.00

Upper height of release box meters/hPa above reference level

ZKIND
1

Reference level 1=above ground, 2=above sea level, 3 for pressure in hPa

PARTS

3000000

Total number of particles to be released

COMMENT

"globalRELEASE"

Comment, written in the outputfile  

Stored outputs
Parcel positions and characteristics were stored at each 3-hourly time step in binary files names as partposit_YYYYMMDDHH0000. These files contain the following variables for each of the 3 million parcels:

parcel id (a unique parcel identity number)

longitude

latitude

height

time step

orography

potential vorticity

specific humidity

density

height of the surrounding boundary layer

height of the surrounding troposphere

temperature

mass.

One partposit_YYYYMMDDHH0000 file is 172 MB. All files for a month are compressed in monthly archives (as fp-eraint-global_v104_rYYYY_partposit_YYYYMM.tar.gz) and sum up to 393 GB for each annual simulation. The output of the entire simulation spanning 1979–2019 sums up to 17 TB. 

Metadata 
The metadata of each annual simulation is stored in a compressed fp-eraint-global_v104_rYYYY_meta.tar.gz archive. Metadata archives contain all namelists (all files from the options folder), invariant simulation outputs (such as header), and job outputs. For easy reference, the atmospheric mass at the time of initialisation is stored in a separate text file atm_mass.txt). The average atmospheric mass is 5.09205935e18 kg; each parcel thus represents approximately 1.697353e12 kg of air. 

Post-processing
The simulations were setup and their outputs were stored in a way so that post-processing with the Heat- And MoiSture Tracking framEwoRk (HAMSTER, Keune et al., 2022) is possible without further changes. HAMSTER constructs multi-day backward trajectories for a region of interest and facilitates the identification of, e.g., precipitation origins. 

Contact and data access
Only metadata is uploaded to this repository due to space constraints; access to the full data set can be obtained by contacting the corresponding author Jessica Keune via jessica.keune[at]ugent.be. 

References
Dee, D. P., Uppala, S. M., Simmons, A. J., Berrisford, P., Poli, P., Kobayashi, et al. (2011). The ERA‐Interim reanalysis: Configuration and performance of the data assimilation system. Quarterly Journal of the Royal Meteorological Society, 137(656), 553-597.
Keune, J., Schumacher, D. L., & Miralles, D. G. (2022). A unified framework to estimate the origins of atmospheric moisture and heat using Lagrangian models. Geoscientific Model Development, 15(5), 1875-1898.
Pisso, I., Sollum, E., Grythe, H., Kristiansen, N. I., Cassiani, M., Eckhardt, S., et al. (2019). The Lagrangian particle dispersion model FLEXPART version 10.4. Geoscientific Model Development, 12(12), 4955-4997.
Tipka, A., Haimberger, L., & Seibert, P. (2020). Flex_extract v7. 1.2–a software package to retrieve and prepare ECMWF data for use in FLEXPART. Geoscientific Model Development, 13(11), 5277-5310.