Data and Code for the manuscript: "Explaining monthly precipitation anomalies in northwestern South America by integrating vertical dynamics and energetics"

This repository contains the processed datasets and computational code required to reproduce the analysis, figures, and statistical validations. The study introduces the Buoyancy Work Rate (BWR) index and validates the physical energy balance mechanism controlling hydroclimatic extremes in the Northwestern South America (NWSA) region. File Inventory 1. monthly_data.nc (NetCDF format) This file contains the monthly time series (1981–2025) of the atmospheric and hydrological variables averaged over the study domain (NWSA adjacent to Niño 1+2 region: 81.25°W–78.5°W, 8°S–0.75°N). All variables are provided as standardized anomalies. w: Vertical velocity (ω) at 500 hPa [Pa s⁻¹]. Proxy for dynamic forcing. dT: Thermodynamic instability (ΔT) [K]. Averaged parcel buoyancy between 700–400 hPa. bwr: Buoyancy Work Rate index. Calculated as the vertical integration of −ωΔT. e: Local evaporation [mm day⁻¹] derived from ERA5. pp_pisco: Monthly precipitation [mm month⁻¹] derived from the PISCO v2.1 dataset. 2. monthly_sst.nc (NetCDF format) This file contains the monthly Sea Surface Temperature (SST) anomalies used as the oceanic boundary condition for the analysis. sst: Sea Surface Temperature anomalies [K] averaged over the Niño 1+2 region (90°W–80°W, 10°S–0°). 3. manuscript_calculations.ipynb (Jupyter Notebook) A Python notebook containing the complete workflow to reproduce the statistical results and figures of the manuscript. Causal Discovery: Implementation of the PCMCI+ algorithm (using Tigramite) to validate the causal control of vertical dynamics on precipitation. Persistence Analysis: Autocorrelation functions to evaluate the signal memory of BWR vs. ω. Tail Dependence: Calculation of the Upper Tail Dependence Coefficient (λU) using empirical copulas to assess robustness during extreme events. Usage Instructions To run the notebook, ensure both NetCDF files (monthly_data.nc and monthly_sst.nc) are located in the same directory as the notebook. The code requires standard Python libraries (xarray, numpy, matplotlib, scipy) and the tigramite package for causal inference. Data Sources The processed data in this repository were derived from: ERA5 Reanalysis PISCO v2.1 (Peruvian Interpolated Data of the SENAMHI's Climatological and hydrological Observations)