Subtropical Pacific Cooling Driven by Land Precipitation Response to CO2 Forcing [Data and Code]

# South American precipitation and South Pacific SST This directory contains the postprocessed data and Jupyter notebooks used to reproduce the main and extended-data figures for manuscript "Subtropical Pacific Cooling Driven by Land Precipitation Response to CO{2} Forcing" The plotting notebooks read the packaged NetCDF files directly. ## Directory structure ```text South_American_Precip_SP_SST/ ├── README.md ├── environment.yml ├── data/ │ ├── masks/ │ │ └── cmip_LandMask.nc │ └── source_data/ │ ├── Fig1/ ... Fig5/ │ └── ExtendedData/Fig1/ ... Fig7/ └── scripts/ ├── main_figure1_plot.ipynb ... main_figure5_plot.ipynb ├── extended_figure1_plot.ipynb ... extended_figure7_plot.ipynb └── diabatic_heating_forcing/ ``` ## Quick start The notebooks require Python packages including NumPy, Xarray, netCDF4, Matplotlib, SciPy, Basemap, `cmaps`, and Jupyter/nbconvert. `environment.yml` is currently a placeholder rather than a complete reproducible environment specification. ## Main-figure data and scripts ### Figure 1 `data/source_data/Fig1/` contains: - ERSST and HadISST observational SST trends. - C1-LE4CO2 relative-SST responses for months 0–4 and year 0–1. - Subtropical South Pacific ensemble-mean SST evolution for 24 months and 10 years. - First-year individual-ensemble SST evolution. - Wind-driven latent-heat SST contribution and lowest-model-level winds. Plot with `scripts/main_figure1_plot.ipynb`. ### Figure 2 `data/source_data/Fig2/` contains: - C1-LE4CO2 and C1-pSST4CO2 precipitation responses over land. - Prescribed and Q-slab diabatic-heating fields. - Q-slab SST response, Monte Carlo spread, latent-heat contribution, and lowest-model-level winds. - 250 hPa velocity potential and divergent winds for C1-pSST4CO2 and Q-slab. Plot with `scripts/main_figure2_plot.ipynb`. ### Figure 3 `data/source_data/Fig3/` contains 850 hPa streamfunction and wind responses for C1-LE4CO2, C1-pSST4CO2, Q-slab, and the stationary-wave-model experiment Q-swm. Plot with `scripts/main_figure3_plot.ipynb`. ### Figure 4 `data/source_data/Fig4/` contains: - cmip4CO2 SA-wet and SA-dry composite relative-SST maps. - amip4CO2 SA-wet and SA-dry composite precipitation maps. - Model-level SST and precipitation indices. - C1LE4CO2 and C2LE4CO2 South Pacific relative-SST ensemble spreads. Plot with `scripts/main_figure4_plot.ipynb`. ### Figure 5 `data/source_data/Fig5/` contains: - ERA5, GPCP, and CMAP precipitation trends for 1979–2014, expressed per 40 years. - Historical SST and precipitation composite differences with bootstrap significance. - cmip4CO2 SST and amip4CO2 precipitation composite differences. - Historical and AMIP precipitation-trend indices for models with and without large ensembles, including spread where available. Plot with `scripts/main_figure5_plot.ipynb`. ## Extended-data figures ### Extended Data Figure 1 `ExtendedData/Fig1/` contains C1-LE4CO2 and C2-LE4CO2 SST maps for years 0–2, 0–10, and 80–100, with Monte Carlo uncertainty and equatorial Pacific SST-gradient indices. Plot with `scripts/extended_figure1_plot.ipynb`. ### Extended Data Figure 2 `ExtendedData/Fig2/` contains fast and slow SST and land-precipitation responses for both CESM1 and CESM2 large ensembles. Plot with `scripts/extended_figure2_plot.ipynb`. ### Extended Data Figure 3 `ExtendedData/Fig3/` contains the subtropical South Pacific SST-budget decomposition and uncertainty for C1-LE4CO2, C2-LE4CO2, and Q-slab. Plot with `scripts/extended_figure3_plot.ipynb`. ### Extended Data Figure 4 `ExtendedData/Fig4/` contains longwave, shortwave, and shortwave cloud-radiative SST contributions for C1-LE4CO2 and Q-slab. Plot with `scripts/extended_figure4_plot.ipynb`. ### Extended Data Figure 5 `ExtendedData/Fig5/` contains pressure-thickness-weighted 300–500 hPa diabatic-heating maps and regional vertical profiles for C1-pSST4CO2 and C2-pSST4CO2. Plot with `scripts/extended_figure5_plot.ipynb`. ### Extended Data Figure 6 `ExtendedData/Fig6/` contains: - Land precipitation and Monte Carlo uncertainty for C1/C2-pSST4CO2. - C1/C2-LE4CO2 land precipitation for months 0–4, years 0–10, and years 80–100. - amip4CO2 and cmip4CO2 South American precipitation indices using the updated 13-wet/11-dry classification. Plot with `scripts/extended_figure6_plot.ipynb`. ### Extended Data Figure 7 `ExtendedData/Fig7/` contains: - C2-LE4CO2 relative-SST maps. - Shortwave and wind-driven latent-heat SST contributions. - Subtropical South Pacific monthly, yearly, and individual-member SST evolution. Plot with `scripts/extended_figure7_plot.ipynb`. ## Masks and conventions `data/masks/cmip_LandMask.nc` supplies the common 1° land–sea mask: - `landmask = 1`: land - `landmask = 0`: ocean Common units are: - SST and SST contributions: K - Precipitation: mm day⁻¹ - Winds: m s⁻¹ - Streamfunction: m² s⁻¹ - Velocity potential: m² s⁻¹ - Diabatic heating: K s⁻¹ or the explicitly documented scaled units in each file Each NetCDF file includes variable attributes and global metadata describing its averaging period, region, uncertainty calculation, masks, scaling, and source files. Significance and robustness criteria differ by experiment; use the metadata stored with each product rather than assuming one common threshold. ## Diabatic-heating forcing example `scripts/diabatic_heating_forcing/` contains: - `nudging.F90` and `physpkg.F90`: CESM1 CAM source modifications. - `generate_diabatic_heating_forcing.py`: example forcing-file generator. - A dedicated README with installation and usage instructions. For a CESM1 case, place the two Fortran files under `<CESM1_CASE_ROOT>/SourceMods/src.cam/` and build the case.