# Running TSMP-PDAF #

The execution of TSMP-PDAF and the necessary input is closely related
to the one of TSMP. For executing TSMP, a ParFlow database file
`*.pfidb` (see ParFlow documentation for more details), an input file
for CLM (`lnd.stdin`) and several input files for OASIS-MCT and COSMO
need to be present in the run directory. TSMP-PDAF follows a similar
approach. The only difference is that for each realisation a different
set of ParFlow/ CLM/ COSMO input files needs to be present which
follow a certain naming convention (see Sections [Input files for
CLM](./input_clm.md#input-files-for-clm), [Input files for
ParFlow](./input_pfl.md#input-files-for-parflow) and [Input files for
COSMO](./input_cos.md#input-files-for-cosmo)). Additionally, a control
file for the data assimilation ([Control file
`enkfpf.par`](./input_enkfpf.md#control-file-enkfpfpar)) and
observation files ([Observation
files](./input_obs.md#observation-files)) need to be present in the
run directory.  Furthermore, some command line options ([Command line
options](./input_cmd.md#command-line-options)) need to be specified
when TSMP-PDAF is executed.

See the Virtual Machine download on webpage
<https://datapub.fz-juelich.de/slts/tsmp-vm/index.html>.

The data on this website is quite large - `12.8 GB`.

Only download this, if you really want to run the Virtual Machine!.

```{toctree} 
---
maxdepth: 3
caption: Inputs for TSMP-PDAF
---
input_clm.md
input_pfl.md
input_cos.md
input_oas.md
input_enkfpf.md
input_obs.md
input_cmd.md
```