Releases | Open Sustainable Technology

Thermofeel - 2.3.0

Added a scientific validation campaign under validation/ for every method
new in this release: per-index methodology READMEs, runnable validation
scripts with pre-registered acceptance criteria, and versioned result tables
and plots. Evidence includes the ISO 7730:2005 Annex D table and 58k ASHRAE
Thermal Comfort Database II field records (PMV/PPD), NOAA SURFRAD ground
observations and the live IFS open-data pipeline (fdir estimators), the NREL
SPA ephemeris (Earth-Sun distance factor), and independent peer-reviewed
implementations/transcriptions (pythermalcomfort, pvlib, BoM/Asghari forms)
for the simple heat indices. Install pip install thermofeel[validation] to
run; see validation/README.md for the verdict table (all PASS).
Added calculate_discomfort_index: Thom's Discomfort Index (Temperature-
Humidity Index) from 2 m temperature and relative humidity, returning Kelvin.
Implements the Celsius/relative-humidity formulation of Thom's index given by
Giles et al. (1990, https://doi.org/10.1007/BF01093455),
DI = T - 0.55 (1 - 0.01 RH)(T - 14.5); the index originates with Thom (1959,
https://doi.org/10.1080/00431672.1959.9926960). Validated against analytic
reference values (including the RH=100% -> DI=Ta and T=14.5 °C -> DI=14.5 °C
identities) and pinned by a regression CSV. Adds a guide page. See also the
review Epstein & Moran (2006, https://doi.org/10.2486/indhealth.44.388), which
uses a different wet-bulb formulation.
Added a batch of new thermal indices (all vectorised, SI in/out, each cited
and validated against published/independent references; see the per-index
guide pages under docs/guide/):
- calculate_apparent_temperature_radiation: the radiation-inclusive
  Steadman/BoM Apparent Temperature from 2 m temperature, 10 m wind, relative
  humidity and a caller-supplied body-absorbed net radiation q
  (Steadman 1994, https://doi.org/10.1071/es94001; Australian BoM).
- calculate_relative_strain_index: the Relative Strain Index from 2 m
  temperature and relative humidity (Lee & Henschel 1966,
  https://doi.org/10.1111/j.1749-6632.1966.tb43059.x; hPa closed form and
  bands from Asghari et al. 2020, https://doi.org/10.2174/1874213002013010011).
- calculate_summer_simmer_index: Pepi's Summer Simmer Index (common 1987
  closed form, an affine transform of Thom's THI) from 2 m temperature and
  relative humidity (Pepi 1987, https://doi.org/10.1080/00431672.1987.9933356;
  equation reproduced from secondary sources — see the guide).
- calculate_pmv and calculate_ppd: Fanger's Predicted Mean Vote and
  Predicted Percentage of Dissatisfied (ISO 7730:2005; Fanger 1970), from air
  and mean-radiant temperature, body-level air velocity, humidity, metabolic
  rate and clothing insulation; validated against the ISO 7730 Annex D table.
Modernised the examples/ scripts and notebook to the ECMWF earthkit 1.0
stack (earthkit-data / earthkit-meteo / earthkit-plots): compute-thermal-indices.py
now fetches gridded ECMWF forecasts online (open data, Polytope or MARS, or a
local file) and writes GRIB/NetCDF; compute-obs.py computes indices from
near-real-time station METARs; the notebook draws maps with earthkit-plots.
Removed orphaned example data files. Install with pip install thermofeel[examples].
(Examples are not part of the shipped library.)
Added the thermofeel.approximations namespace (namespace-only, not
re-exported at the top level) with estimators of the direct solar radiation
fdir from global radiation ssrd + solar geometry, for datasets that lack
it: approximate_fdir_erbs (Erbs et al. 1982, with an optional Earth-Sun
distance correction) and approximate_fdir_disc (DISC; Maxwell 1987). Both
reproduce pvlib to machine precision (validation oracle only). The
compute-thermal-indices.py example exposes them via
--approximate-fdir[=erbs|disc].

Consumption - Buildings and Heating - Python
Published by tlmquintino about 1 month ago

Thermofeel - 2.2.0

Added inline type hints across the public API and a py.typed marker
(PEP 561), so downstream type-checkers pick up thermofeel's signatures
Exported fahrenheit_to_celsius at the top level, completing the C/K/F
converter set
Fixed silent integer-dtype truncation in
calculate_saturation_vapour_pressure_multiphase and approximate_dsrp: an
integer-typed input array no longer truncates the float result
Added the missing citation for calculate_relative_humidity_percent
(Magnus-Tetens; Murray 1967) and documented validity ranges in the UTCI, wet
bulb, simple WBGT, NET, apparent-temperature and humidex docstrings; documented
the array calling convention and that relative humidity is not clamped
calculate_bgt (and therefore the Stull calculate_wbgt) now returns the
mean radiant temperature at exactly zero wind — the analytic calm-air limit
(no convection ⇒ globe at radiative equilibrium) — instead of NaN. Values for
any positive wind are unchanged
Ran a first numerical-robustness pass (ROBUSTNESS.md): audited the
log/sqrt/power/division domains and added NaN-propagation and
error-contract tests
The published wheel now ships only the thermofeel package (examples/,
scripts/ and tests/ are no longer packaged)
Raised the minimum supported Python to 3.10 (requires-python = ">=3.10"),
refreshed the trove classifiers and added Python 3.14
Repaired the examples/compute-thermal-indices.py example to the 2.x API:
solar geometry now obtained from earthkit-meteo
(solar.cos_solar_zenith_angle_integrated), calculate_utci/calculate_wbgt/
calculate_bgt/calculate_normal_effective_temperature return Kelvin directly,
renamed/removed functions and the obsolete timing decorators dropped
Removed dead, overwritten code in scale_windspeed (kept the computed log-law
coefficient) and corrected the approximate_dsrp threshold comment
Added the thermofeel.excess_heat submodule: Excess Heat Factor (EXHF),
Excess Cold Factor (EXCF), heatwave severity, and the supporting daily mean
temperature, significance and acclimatisation indices, after Nairn & Fawcett
(2014). The factors are also exposed at the top level as
calculate_excess_heat_factor / calculate_excess_cold_factor. The functions
implement the per-day formulas; the temporal aggregations are left to upstream
tooling (e.g. earthkit-transforms)
Added calculate_wbgt_liljegren: physically based Wet Bulb Globe Temperature
using the Liljegren et al. (2008) method, validated against Liljegren's
reference implementation (github.com/mdljts/wbgt)
Added calculate_wind_speed_2m_liljegren: the KNMI/Liljegren
stability-dependent 10 m-to-2 m wind-speed profile, used by default in
calculate_wbgt_liljegren (selectable via its wind_scaling argument)
Added calculate_heat_force: the KNMI 0–10 heat-force (hittekracht) scale
derived from WBGT (KNMI Technical Report TR-26-04)
Removed the broken, unexported experimental_wbgt module, now superseded by
calculate_wbgt_liljegren
Fixed the calculate_bgt argument order in the scalar test suite

Consumption - Buildings and Heating - Python
Published by tlmquintino 2 months ago

Thermofeel - 2.1.7

ChangeLog:

Change project to single branch main

Full Changelog: https://github.com/ecmwf/thermofeel/compare/2.1.5...2.1.7

Consumption - Buildings and Heating - Python
Published by tlmquintino 7 months ago

Thermofeel - 2.1.5

Fix descriptions
Fix variable names
Updated humidex calculation and reference

Consumption - Buildings and Heating - Python
Published by cladinapoli 7 months ago

Thermofeel - 2.1.4

Fix setup of project
Removed unused dependency on earthkit-meteo

Consumption - Buildings and Heating - Python
Published by cladinapoli 9 months ago

Thermofeel - 2.1.3

Documentation revised and updated

Consumption - Buildings and Heating - Python
Published by cladinapoli 9 months ago

Thermofeel - 2.1.2

ChangeLog

Small fix in the heat index adjusted computation

Consumption - Buildings and Heating - Python
Published by cladinapoli 11 months ago

ChangeLog

Small bug fix in the UTCI computaiton

What's Changed

Rename utci variable by @tlmquintino in https://github.com/ecmwf/thermofeel/pull/39
release 2.1.1 by @tlmquintino in https://github.com/ecmwf/thermofeel/pull/40

Full Changelog: https://github.com/ecmwf/thermofeel/compare/2.1.0...2.1.1

Consumption - Buildings and Heating - Python
Published by tlmquintino over 1 year ago

Version 2.1.0

What's Changed

Fix computaiton of normal effective temperature
Fix documentation links to new repository URL
Better explanation of T0

Full Changelog: https://github.com/ecmwf/thermofeel/compare/2.0.0...2.1.0

Consumption - Buildings and Heating - Python
Published by tlmquintino almost 2 years ago

Version 2.0.0

Standardisation

I/O variables converted into International System of Units (SI) (e.g., K)
Docstrings standardised in their descriptions (e.g., inputs as float arrays) and references
Variable names standardised across function declarations (i.e., t2m vs t2k vs t_k → now t2_k)
Use of temperature converter functions
Input variables are limited to those explicitly needed in the function formula (see e.g., calculate_normal_effective_temperature function)

New Functions

scale_windspeed function to scale 10m wind speed to height h with h < 10m
calculate_nonsaturation_vapour_pressure to calculate vapour pressure at a given relative humidity and air temperature
calculate_wbgt_simple function renamed
calculate_normal_effective_temperature function renamed

Improvements

thermofeel library docstring lists computed variables in alphabetical order
the cosine of the solar zenith angle now computed via the earthkit-meteo <https://github.com/ecmwf/earthkit-meteo>_ library
calculate_bgt function is calculated via a 4x faster formula
calculate_saturation_vapour_pressure_multiphase formulas replaced with those used in the IFS
changeable threshold in approximate_dsrp function (set to 0.1 by default)
invalidity outside input variables range specified in calculate_wind_chill function docstring

Bug Fixes

fixed approximate_dsrp to avoid fdir being overwritten with dsrp when calculating MRT
fixed calculate_wbgt_simple constant value and vapour pressure calculated from non-saturated formula
fixed calculate_bgt wind speed at 1.1m in input
fixed calculate_mrt_from_bgt wind speed at 1.1m in input
fixed calculate_normal_effective_temperature wind speed at 1.2m in input
fixed calculate_apparent_temperature wind speed at 10m in input; vapour pressure calculated from non-saturated formula
fixed calculate_wind_chill wind speed in km/h and operation symbols in main formula
fixed calculate_heat_index_simplified a wrong sign and missing constant value in hiarray; hi set to 2m air temperature when the latter is below 20°C
fixed fahrenheit_to_kelvin converter function

Consumption - Buildings and Heating - Python
Published by tlmquintino almost 3 years ago

Thermofeel - 1.3.0

Multiple fixes and documentation

Consumption - Buildings and Heating - Python
Published by tlmquintino almost 4 years ago

Introduced optional Numba JIT optimisation of functions for significant speedup.
Improved some functions to make them compilable and vectorisable by Numba.
Remove __wrap function, as such many functions now require numpy arrays as input.
Fixed calculate_heat_index_adjusted.
Added calculate_wbt_dj
More and improved examples.
Improved documentation.

Consumption - Buildings and Heating - Python
Published by tlmquintino over 4 years ago

Thermofeel - 1.1.3

Update to the Documentation
Small changes to the an example file in the examples folder and updating the documentation for read the docs.

Full Changelog: https://github.com/ecmwf-projects/thermofeel/compare/1.1.2...1.1.3

Consumption - Buildings and Heating - Python
Published by chlobrim over 4 years ago

Thermofeel - 1.1.2

Improvements in the examples

Consumption - Buildings and Heating - Python
Published by tlmquintino almost 5 years ago

Thermofeel - 1.1.1

What's Changed

Temporarily pin the Python version until new NumPy wheels are available by @dvuckovic in https://github.com/ecmwf-projects/thermofeel/pull/10

New Contributors

@dvuckovic made their first contribution in https://github.com/ecmwf-projects/thermofeel/pull/10

Full Changelog: https://github.com/ecmwf-projects/thermofeel/compare/1.1.0...1.1.1

Consumption - Buildings and Heating - Python
Published by tlmquintino almost 5 years ago

Thermofeel - 1.1.0

Consumption - Buildings and Heating - Python
Published by tlmquintino almost 5 years ago