snapclim and snapesm

The two snapshot backends build the climate from a reference state plus climate snapshots (e.g. present day, pre-industrial, LGM), blended in time by scalar indices (e.g. a glacial index). Both supply the atmosphere (monthly tas, pr, lapse-rate corrected to the current surface) and the ocean as depth profiles (to_ann, so_ann), which the marine shelf interpolates to the shelf base. Both run on grid_clim and read the [snap] group.

Program: yelmox (make yelmox), backends in libs/snapclim.f90 and libs/snapesm.f90, behind libs/yelmox_climate.f90.
Select: [comps] climate = "snapclim" or "snapesm".
Output: with [output] write_clim, snap.nc: the annual air temperature and precipitation; with snapesm also its driving indices (idx_<name>), the surface elevation, the monthly atmosphere (tas, tsl, pr), ta_sum, tsl_ann and the ocean profiles (to_ann, so_ann on depth).

snapclim

The default backend of every regional config (yelmox/yelmox_<domain>.nml). [snap] sets the forcing method for the atmosphere (atm_type) and the ocean (ocn_type), the index files (fname_at, fname_ao, …), the lapse rates and the scaling parameters; up to four snapshots are read from [snap_clim0] to [snap_clim3], plus [snap_hybrid] and [snap_recon] for those methods.

Method	Climate
`const`	the reference snapshot (`clim0`)
`anom`	reference + a homogeneous anomaly (`dTa`/`dTo`/`dSo`): the transient forcing’s, else from the index
`snap_1ind`, `snap_1ind_new`, `snap_1ind_abs`, `snap_1ind_miocene`	two snapshots blended by one index
`snap_2ind`, `snap_2ind_abs`	snapshots blended by two indices
`hybrid`	reference + monthly temperature anomalies from a prescribed series (`[snap_hybrid]`), no snapshots
`recon`	reconstructed snapshots, interpolated in time (`[snap_recon]`)
`fraction` (ocean only)	an ocean anomaly of `f_to` × the mean atmospheric anomaly

The transient forcing ([tsforcing]/[tsgen], see yelmox) reaches snapclim only in the anom methods.

snapesm

A rewrite of snapclim on the fesm-utils varslice reader and tsgen series: the snapclim methods become configurations of one blend model, over any number of snapshots, fields and indices. Example: yelmox/yelmox_Greenland_snapesm.nml, which reproduces snapclim snap_1ind_new / fraction (yelmox_Greenland_snapclim.nml).

&snap
    var_defs  = "input/greenland_clim.nml"   ! database of the input files
    combine   = "anomaly"                    ! snapshots as anomalies from the reference
    manifold  = 1                            ! number of indices spanning the blend (0/1/2)
    ref_name  = "pd"                         ! the reference snapshot
    snapshots = "pd" "picontrol" "lgm"
    fields    = "tas" "pr" "to" "so" "zs"
    indices   = "at" "ap"
    ! lapse, f_p, f_p_ne, f_stdev, f_to, f_hol, dTa_const, dTo_const, dSo_const
    ! as in snapclim
/

Three kinds of group follow:

Fields, [snap_field_<name>]: kind (atm_temp, atm_precip, ocn, ocn_salt, elev), blend (linear, ratio, fraction, const) and the driving index ("" = none).
Snapshots, [snap_<name>]: monthly, time, idx_coord (the snapshot’s position on the index manifold) and, per field, the var_defs group(s) holding it (one = monthly; two = annual + summer, from which the seasonal cycle is built). See Database namelists.
Indices, [snap_idx_<name>]: a tsgen series (method, series_file, sigma).

Restart bundles hold the state of each index, snapesm_idx_<name>_restart.nc, with the configuration (snapshots, fields, reference) as attributes.

The transient forcing is added on top in every configuration.