Data and code for prioritizing hydropower catchments for drought monitoring in Brazil

Description

This dataset accompanies the study “Prioritizing hydropower catchments for drought monitoring in Brazil using conditional risk and external evidence”. It links 204 hydropower entities to 109 upstream catchments and combines monthly generation records for 2000–2025 with corrected ERA5-Land runoff for 1991–2025. The package contains frozen analytical inputs, the complete catchment atlas, the reconciled entity inventory, reference tables and Python code. The analyses estimate runoff–generation associations, reconstruct conditional-risk magnitudes and rankings, assess sensitivity to hydrological periods and response specifications, compare runoff with ONS natural energy inflow and ANA gauge observations, and evaluate the contribution of prior stored energy to retrospective generation-deficit concordance. Quality-control records document source corrections, catchment assignments and the treatment of initial zero-valued ENA records. A Google Colab notebook provides an entry point for running the analytical reproduction. The scripts re-estimate the retained models, reconstruct numerical results and selected figures, and compare regenerated outputs with the supplied reference tables. Documentation includes execution instructions, dependency versions, variable definitions, file descriptions and source checksums. Reproduction starts from the supplied analytical inputs. Original source acquisition, initial gridded-data processing and the complete raw provider archives are outside the package. The documentation specifies which outputs are regenerated and which upstream controls are retained as archived records.

Steps to reproduce

1. Download Online_Resource_2_Data_and_Tables.zip and Reproduce_in_Colab.ipynb from this dataset. 2. Open Google Colab and upload Reproduce_in_Colab.ipynb. Use a Python CPU runtime; GPU access is unnecessary. The release was verified locally with Python 3.12. 3. Run the notebook cells in order. When prompted, upload Online_Resource_2_Data_and_Tables.zip without extracting or modifying it. 4. The notebook installs the specified dependencies, checks the distributed files and runs the analytical modules. These re-estimate the retained models, reconstruct the catchment atlas and hydrological comparisons, and regenerate selected figures. 5. Inspect the numerical verification results and execution logs. Confirm that all reported checks pass, then download the results archive using the final notebook cell. Reproduction starts from the supplied frozen analytical inputs. Original source acquisition and initial gridded-data processing are documented upstream steps outside this package. README.md, REPRODUCIBILITY_SCOPE.md and MANUSCRIPT_OUTPUT_MAP.csv describe the execution requirements and output coverage.

Institutions

• Universidade de São Paulo

  São Paulo, São Paulo

Categories

Funders

• Coordenação de Aperfeicoamento de Pessoal de Nível Superior

  Ministry of Education

  Federal District, Brasília

Licence