This project contains Python implementations of various climate index algorithms which provide a geographical and temporal picture of the severity and duration of precipitation and temperature anomalies useful for climate monitoring and research.
The following indices are provided:
- SPI, Standardized Precipitation Index, utilizing both gamma and Pearson Type III distributions
- SPEI, Standardized Precipitation Evapotranspiration Index, utilizing both gamma and Pearson Type III distributions
- PET, Potential Evapotranspiration, utilizing either Thornthwaite or Hargreaves equations
- PNP, Percentage of Normal Precipitation
- PCI, Precipitation Concentration Index
This Python implementation of the above climate index algorithms is being developed with the following goals in mind:
- to provide an open source software package to compute a suite of climate indices commonly used for climate monitoring, with well documented code that is faithful to the relevant literature and which produces scientifically verifiable results
- to provide a central, open location for participation and collaboration for researchers, developers, and users of climate indices
- to facilitate standardization and consensus on best-of-breed climate index algorithms and corresponding compliant implementations in Python
- to provide transparency into the operational code used for climate monitoring activities at NCEI/NOAA, and consequent reproducibility of published datasets computed from this package
- to incorporate modern software engineering principles and scientific programming best practices
This is a developmental/forked version of code that was originally developed by NIDIS/NCEI/NOAA. See drought.gov.
| Python Version | Status | Notes |
|---|---|---|
| 3.10 | Supported | Minimum supported version |
| 3.11 | Supported | |
| 3.12 | Supported | |
| 3.13 | Supported | Latest supported version |
All versions are tested on Linux (ubuntu-latest). Python 3.10 and 3.13 are additionally tested on macOS. Both latest and minimum declared dependency versions are tested in CI.
This project provides 12 months notice before dropping support for a Python version. When a version approaches end-of-life, removal will be announced via the CHANGELOG and a GitHub issue, and implemented no sooner than 12 months after announcement with a version bump.
Python 3.9 support was dropped in v2.2.0 (August 2025) due to scipy>=1.15.3 requiring 3.10+.
| API Surface | Status | Guarantee |
|---|---|---|
NumPy array functions (indices.spi, indices.spei, indices.pet) |
Stable | No breaking changes in minor versions |
xarray DataArray functions (spi(), spei(), pet_thornthwaite(), pet_hargreaves()) |
Beta | No breaking changes in patch versions |
Stable API: The NumPy-based computation functions follow strict semantic versioning.
Beta API: The xarray adapter layer provides automatic parameter inference, coordinate
preservation, CF metadata, and Dask support. While beta, computation results are identical
to the stable NumPy API — only the interface surface (parameter names, metadata attributes,
coordinate handling) may evolve. Beta features are tagged with BetaFeatureWarning and
marked in docstrings.
Breaking Change: Exception-Based Error Handling
Version 2.2.0 introduces a significant architectural improvement in error handling. The library now uses exception-based error handling instead of returning None tuples for error conditions.
Before (v2.1.x and earlier):
# Old behavior - functions returned None tuples on failure
result = some_internal_function(data)
if result == (None, None, None, None):
# Handle error case
passAfter (v2.2.0+):
# New behavior - functions raise specific exceptions
try:
result = some_internal_function(data)
except climate_indices.compute.InsufficientDataError as e:
# Handle insufficient data case
print(f"Not enough data: {e.non_zero_count} values found, {e.required_count} required")
except climate_indices.compute.PearsonFittingError as e:
# Handle fitting failure case
print(f"Fitting failed: {e}")DistributionFittingError(base class)InsufficientDataError- raised when there are too few non-zero values for statistical fittingPearsonFittingError- raised when L-moments calculation fails for Pearson Type III distribution
- Direct API users: No changes needed - the public SPI/SPEI functions handle exceptions internally
- Library integrators: If you were checking for
Nonereturn values from internal functions, update to use try/catch blocks - Benefits: More informative error messages, better debugging, and automatic fallback from Pearson to Gamma distribution when appropriate
Version 2.2.0 also addresses floating point comparison issues (python:S1244) throughout the codebase:
Floating Point Comparisons:
# ❌ OLD: Direct equality checks (unreliable)
if values == 0.0:
handle_zero_case()
# ✅ NEW: Safe comparison using numpy.isclose()
if np.isclose(values, 0.0, atol=1e-8):
handle_zero_case()Benefits:
- Eliminates floating point precision issues in statistical parameter validation
- Improves test reliability and numerical robustness
- Follows scientific computing best practices for floating point arithmetic
- See
docs/floating_point_best_practices.mdfor comprehensive guidelines
You can cite climate_indices in your projects and research papers via the BibTeX
entry below.
@misc {climate_indices,
author = "James Adams",
title = "climate_indices, an open source Python library providing reference implementations of commonly used climate indices",
url = "https://github.com/monocongo/climate_indices",
month = "may",
year = "2017--"
}
