/ACCESS-AMOS22

ACCESS workshop for AMOS22

Primary LanguageJupyter Notebook

ACCESS Training (2020)

please spend sometime before the workshop to read and watch the video from the 
previous ACCESS training workshop done by Chloe and Holger.

** note **

https://confluence.csiro.au/display/ACCESS/ACCESS+Training

**1. ACCESS-CM2 Training

**2. ACCESS-ESM1.5 Training https://confluence.csiro.au/display/ACCESS/ACCESS-ESM1.5+Training

**3. Setting up for ACCESS-ESM1.5 https://confluence.csiro.au/display/ACCESS/Setting+up+for+ACCESS-ESM1.5

**4. Archiver https://git.nci.org.au/cm2704/ACCESS-Archiver.git

AMOS ACCESS ESM Workshop 2022

** note ** Please Please Please make sure you can git clone and run the ESM before the workshop!!!

**1. Read throught the instruction but git clone the following link:

https://github.com/ars599/PI-DEMO

This is the one particullary used for the training!!!

ACCESS-ESM AMOS 2022 Usefull scripts

**1. conver the atmosphere output file name from Jan to 001 (not necessary) ...

  please use change_unext_esm_ars599.sh $job_id

**2. xconv Download

/g/data/access/projects/access/apps/xconv/1.94/xconv

**3. ACCESS Archiver

** note **

Check out from Tammas Loughran:
https://git.nci.org.au/tfl561/ACCESS-Archiver/-/tree/main
Usage is detailed unger nci git. (NCI account needed)

- to convert UM format to netCDF
- to move wrap data from archive to gdata (easy for OOD)
alternately can use from:
https://github.com/ars599/ACCESS-Archiver_tfl561

It was based on tfl561, and modify to fit AMOS 2022 demo activity

** to do list and a quick Q&A to help you setup and run Archiver**

Issue #4 #4

**3-0. Modify under wrap_ACCESS_Archiver.sh

**3-1. make sure the base outputs files (forexample user: ars599)

	base_dir=/scratch/nf33/ars599/access-esm/archive

**3-2. output directory

	arch_dir=/g/data/nf33/ars599/archive/CMIP6/ACCESS-ESM1-5/

**3-3. setup run type

	#access_version = [cm2, cm2amip, cm2chem, esmscript, esmpayu, om2]
	old ESM k shell script
	access_version=esmscript
	now payu ESM
	access_version=esmpayu
	
	basic knowledge of the namelist file for each module

**3-4. atm outputs need to convert to netCDF (xconv can only deal with single file)

	#ncexists = [true, false]
	#true: Copy netcdf version of file if it exists; false: Always use UM pp-file if it exists, whether or not netcdf version exists
	ncexists=true

**3-5. convert subdaily 3h 6h data ?

	#subdaily = [true, false]; convert subdaily atm files?
	subdaily=false

**3-6. multiple job list

	#loc_exps = list of local experiment names (stored in 'base_dir') to archive
	loc_exps=(
	HI-00
	HI-02
	)

python scripts (analysis data)

HI-00 origin need to read all files and then save to a single variable file
https://github.com/ars599/ACCESS-AMOS22/blob/main/scripts/plot_ts_concat_demo3.ipynb


Now just use existing files:
fn1 = 'sst_HI-00_output.nc'
fn2 = 'sst_HI-02_output.nc'

The files have been copied to 
idir = '/g/data/access/access_amos22/archive/'
ext1 = idir+"sst_HI-00_output.nc"
ext2 = idir+"sst_HI-02_output.nc"
ext3 = idir+"pr_HI-00_output.nc"
ext4 = idir+"pr_HI-02_output.nc"

https://github.com/ars599/ACCESS-AMOS22/blob/main/scripts/plot_ts_concat_demo.ipynb

Cheat sheet for PI-DEMO interactive run for ACCESS workshop

** Before submitting your job:

Login via ssh

	ssh -X UserID@gadi.nci.org.au 

Need to switch project to nf33 (check with echo $PROJECT)

	switchproj nf33 
	echo $PROJECT 

Create and cd into /scratch/nf33/userid/access-esm (mkdir)

	mkdir /scratch/nf33/UserID 
	mkdir /scratch/nf33/UserID/access-esm 
	cd /scratch/nf33/UserID/access-esm 

** Load modules

	module use /g/data/hh5/public/modules 
	module load conda/analysis3-unstable 

*** Checkout code git clone

	git clone https://github.com/ars599/PI-DEMO 
	cd PI-DEMO 

** Explain settings in config file (Tilo/Arnold to explain settings)

	cat config.yml 

Also readme file, but please follow our instructions today

cat README.md

submit model simulation to queue

	payu run 

** After submitting your job:

Check status of run

	qstat -wau UserID 

** New folder (and link) will be created, contains output for 3 main components and coupler including information for each processor in atmosphere + copy of all configuration files (Tilo/Arnold to explain files and output in work folder)

/scratch/nf33/UserID/access-esm/work/PI-DEMO

cat /scratch/nf33/UserID/access-esm/work/PI-DEMO/atmosphere/atm.fort6.pe0

cat /scratch/nf33/UserID/access-esm/work/PI-DEMO/atmosphere/STASHC

** output files for atmoshpere created for monthly and daily output

aiihca.paa1jan and aiihca.pea1jan

New folder “archive” created when run complete, contains model output

/scratch/nf33/txz599/access-esm/archive/PI-DEMO/output000

** Look at monthly and daily output with xconv and/or ferret, convert one field to netcdf with xconv and then load with ferret (Tilo/Arnold to explain analysis of output)

	module use ~access/modules 
	module load xconv 
	module load ferret 
	xconv aiihca.paa1jan 

Please finish the Survey and if you have any qestion please leave your concern at the following link

** AMOS ACCESS ESM workhop discussion

https://github.com/ars599/ACCESS-AMOS22/issues

** AMOS ACCESS Survey