interface field generation fix on base

.gitignore

@@ -5,6 +5,8 @@ __pycache__
 .env
 .cache/
 cache/
+data/
+fastmip_env/
 
 #Tempfiles
 *~

CHANGELOG.rst

@@ -17,6 +17,17 @@ The changes listed in this file are categorised as follows:
 [Unreleased]
 ---------------------
 
+### Fixed
+- Fixes to generate annual and monthly gridded ensembles with unified noise and preserving more of the variance.
+
+### Changed
+
+- Now possibly to send variable length temperature scaling timeseries, fixed noise generator for wrong ordering of base data dimensions
+
+### Fixed
+
+- Variable length timeseries now works also when don't have "year" as time dimension.
+
 
 [Version 1.6.0]
 -----------------------------

FastMIP_README.md

@@ -0,0 +1,44 @@
+
+Steps to produce the FastMIP Phase 2 Tier 1 METEOR contributions.
+April 2026 by Maura Dewey (maura.dewey@cicero.oslo.no)
+
+	1. Train METEOR for all the models and scenarios. This is done with the METEOR venv (for example, within a tmux session on nac)
+		a. cd to METEOR folder and activate venv with: "source venv/bin/activate"
+		b. cd to scripts folder and run: "python FastMIP_phase2_trainMETEOR.py > output.txt 2>&1" - The list of ESMs that are emulated, the scenarios that are run, and the choices for subsetting the FAIR ensemble are made in the top of this python script. (NOTE: If you need to change the training settings, ie. which scenario or ESM, it needs to be done in this script. Possible update/to-do would be to put that in a yml file.)
+		c. Disconnect from the tmux session and let it run. This will train METEOR and create all the raw METEOR output for the given list of ESMs and scenarios (takes ~10 hrs for Tier 1, if also training new METEOR instances)
+
+	2. Create FastMIP specific output files. This is done with a conda env fastmip_env
+		a. setup modules, deactivate METEOR venv, and activate conda environment with: ". FastMIP_env_setup.sh"
+		b. cd to scripts folder and create FastMIP specific output files with: "python FastMIP_phase2_makeoutput.py"
+		c. disconnect from the tmux session and let it run (takes a couple hours). 
+		d. Once the output is done, update netCDF attributes with: "python FastMIP_phase2_atts.py"
+
+	3. Create plots 
+		The notebook FastMIP_phase2_plots.ipynb will create all the plots from the Fastmip phase2 example notebook (https://github.com/sarasita/fastMIP/blob/main/requested_output/tour_through_requested_output.ipynb) with METEOR results. 
+
+
+Files on fastmip branch:
+
+./FastMIP_README.md 
+./conda_env_fastmip.yml (conda environment file for processing output, needed to use for python-cdo for regridding)
+./FastMIP_env_setup.sh (to clear modules and activate conda environment on nac server)
+
+./scripts/FastMIP_phase2_trainMETEOR.py (trains METEOR for all Tier 1 models and creates 200 member ensembles - 20 FAIR ensemble members x 10 METEOR noise model realizations. The subselection of FAIR members used is saved in a pickle in ./data/FASTMIP_phase2/FAIR_data)
+./scripts/FastMIP_phase2_makeoutput.py (makes requested FastMIP output files - subset 10 member ensemble and bulk statistics across full 200.)
+./scripts/FastMIP_phase2_attrs.py (updates netCDF attributes to FastMIP protocol)
+
+./notebooks/FastMIP_phase2_plots.py (makes all plots as in https://github.com/sarasita/fastMIP/blob/main/requested_output/tour_through_requested_output.ipynb)
+
+
+Data files are not tracked with git, but the following structure is required and/or created with above scripts:
+
+./data/FASTMIP_phase2/FAIR_data (contains full FAIR ensemble files (climate_assessment_forced.csv), grid file for regridding (g025.txt), and the pickles of the FAIR subset.)
+./data/FASTMIP_phase2/scenario_data (contains the emission and concentration data files needed for the SCM)
+./data/FASTMIP_phase2/METEOR_emulations/raw (where METEOR output is saved initially)
+./data/FastMIP_phase2/METEOR_emulations/aggregated (where regridded and combined tmp files are saved)
+./data/FASTMIP_phase2/METEOR_emulations/processed (where FastMIP output is saved)
+
+Initial emulations where done for 4 scenarios (L, M, H, VL)
+The full CMIP7 scenario list and short_name markers are:
+scenarios_list = ['SSP1 - Very Low Emissions', 'SSP2 - Low Emissions', 'SSP2 - Medium-Low Emissions', 'SSP2 - Medium Emissions', 'SSP3 - High Emissions', 'SSP5 - Medium-Low Emissions_a', 'SSP2 - Low Overshoot_a']
+scenarios_short = ['VL','L','ML','M','H','HL','LN']

FastMIP_env_setup.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+
+# load modules and conda environment to run FastMIP_phase2_makeoutput.py, making the FastMIP phase 2 output files for METEOR. 
+# The same conda environment can be used as a kernel for FastMIP_phase2_plots.ipynb.
+# Note: this will deactivate the METEOR venv!
+# module versions as of 2026-04
+# the environment can be created from conda_env_fastmip.yml 
+
+module purge
+module load Anaconda3/2023.09-0
+source deactivate
+conda deactivate
+conda activate /div/no-backup-nac/users/maurad/METEOR/fastmip_env/

conda_env_fastmip.yml

@@ -0,0 +1,16 @@
+name: /storage/no-backup-nac/users/maurad/METEOR/fastmip_env
+channels:
+  - defaults
+dependencies:
+  - xesmf
+  - ipykernel
+  - matplotlib
+  - pandas
+  - dask
+  - cftime
+  - xarray
+  - cartopy
+  - netcdf4
+  - regionmask
+  - python-cdo
+prefix: /storage/no-backup-nac/users/maurad/METEOR/fastmip_env