Potential evapotranspiration derived from HadUK-Grid 1km gridded climate observations 1969-2021 (Hydro-PE HadUK-Grid)
THIS DATASET HAS BEEN SUPERSEDED The latest version is Potential evapotranspiration derived from HadUK-Grid 1km gridded climate observations 1969-2022 (Hydro-PE HadUK-Grid)
A number of missing data points were discovered in this dataset. The majority of the missing data points were occurring when the wind speed was set to 0 during the interpolation from monthly to daily, which occurred if the interpolation resulted in the wind speed dropping below 0. The zero wind speed resulted in a divide-by-zero in the PET calculation (when calculating aerodynamic resistance), therefore resulting in NaNs (missing data) in both the PET and PETI datasets. This was fixed by raising all 0 wind values to the minimum non-zero value found in the daily-interpolated wind dataset.
Further missing data was found over St Kilda. This was due to missing data in the original, monthly sfcWind variable in the HadUK dataset used in the PET calculations for Jan and Apr 2009. This was fixed by simply interpolating the monthly data linearly in time to obtain values for these months.
Two further minor bugs were identified that only affected the PETI dataset, resulting in a few further NaNs and a few incorrect values:
The first bug was due to the fact that the code calculates the amount of time it takes the canopy to dry out as the inverse of the potential interception. Even though this is done before it sets negative values to be zero, the potential interception rate (PEI) is sometimes equal to zero, which makes t_to_dry NaN, which then propagates through to the PETI as missing values. It’s rare that the PEI is numerically exactly zero, but it does happen! The simple fix was to set t_to_dry=D if PEI<0, so that those points worked out as expected (they end up as PETI=0). The other bug was that the code didn’t properly handle cases where PEI is negative (i.e., condensation, so the canopy never dries out). This meant there were some points which had positive PETI where it should actually have been zero. The impact of these two bugs was very minor.
If you need access to the archived version, please contact the EIDC
Download/Access
THIS DATASET HAS BEEN SUPERSEDED The latest version is Potential evapotranspiration derived from HadUK-Grid 1km gridded climate observations 1969-2022 (Hydro-PE HadUK-Grid)
A number of missing data points were discovered in this dataset. The majority of the missing data points were occurring when the wind speed was set to 0 during the interpolation from monthly to daily, which occurred if the interpolation resulted in the wind speed dropping below 0. The zero wind speed resulted in a divide-by-zero in the PET calculation (when calculating aerodynamic resistance), therefore resulting in NaNs (missing data) in both the PET and PETI datasets. This was fixed by raising all 0 wind values to the minimum non-zero value found in the daily-interpolated wind dataset.
Further missing data was found over St Kilda. This was due to missing data in the original, monthly sfcWind variable in the HadUK dataset used in the PET calculations for Jan and Apr 2009. This was fixed by simply interpolating the monthly data linearly in time to obtain values for these months.
Two further minor bugs were identified that only affected the PETI dataset, resulting in a few further NaNs and a few incorrect values:
The first bug was due to the fact that the code calculates the amount of time it takes the canopy to dry out as the inverse of the potential interception. Even though this is done before it sets negative values to be zero, the potential interception rate (PEI) is sometimes equal to zero, which makes t_to_dry NaN, which then propagates through to the PETI as missing values. It’s rare that the PEI is numerically exactly zero, but it does happen! The simple fix was to set t_to_dry=D if PEI<0, so that those points worked out as expected (they end up as PETI=0). The other bug was that the code didn’t properly handle cases where PEI is negative (i.e., condensation, so the canopy never dries out). This meant there were some points which had positive PETI where it should actually have been zero. The impact of these two bugs was very minor.
If you need access to the archived version, please contact the EIDC
The data are provided in gridded netCDF files. There is one file for each variable, for each calendar month.
These data were generated as part of NERC grant NE/S017380/1 (Hydro-JULES: Next generation land surface and hydrological prediction.)
Format
NetCDF
Spatial information
Temporal information
Provenance & quality
The HadUK-Grid meteorological variables used were: daily maximum and minimum air temperature at 1.5 m (°C), daily precipitation (mm d-1), monthly mean water vapour pressure (hPa), monthly mean sea level pressure (hPa), monthly total sunshine hours (hours) and monthly mean wind speed at 10 m (m s-1). Other variables used were the surface elevation (m) and Ångström coefficients a, b and c. A description of the methodology is available in the supporting documentation.
Licensing and constraints
THIS DATASET HAS BEEN SUPERSEDED The latest version is Potential evapotranspiration derived from HadUK-Grid 1km gridded climate observations 1969-2022 (Hydro-PE HadUK-Grid)
A number of missing data points were discovered in this dataset. The majority of the missing data points were occurring when the wind speed was set to 0 during the interpolation from monthly to daily, which occurred if the interpolation resulted in the wind speed dropping below 0. The zero wind speed resulted in a divide-by-zero in the PET calculation (when calculating aerodynamic resistance), therefore resulting in NaNs (missing data) in both the PET and PETI datasets. This was fixed by raising all 0 wind values to the minimum non-zero value found in the daily-interpolated wind dataset.
Further missing data was found over St Kilda. This was due to missing data in the original, monthly sfcWind variable in the HadUK dataset used in the PET calculations for Jan and Apr 2009. This was fixed by simply interpolating the monthly data linearly in time to obtain values for these months.
Two further minor bugs were identified that only affected the PETI dataset, resulting in a few further NaNs and a few incorrect values:
The first bug was due to the fact that the code calculates the amount of time it takes the canopy to dry out as the inverse of the potential interception. Even though this is done before it sets negative values to be zero, the potential interception rate (PEI) is sometimes equal to zero, which makes t_to_dry NaN, which then propagates through to the PETI as missing values. It’s rare that the PEI is numerically exactly zero, but it does happen! The simple fix was to set t_to_dry=D if PEI<0, so that those points worked out as expected (they end up as PETI=0). The other bug was that the code didn’t properly handle cases where PEI is negative (i.e., condensation, so the canopy never dries out). This meant there were some points which had positive PETI where it should actually have been zero. The impact of these two bugs was very minor.
If you need access to the archived version, please contact the EIDC
This dataset is available under the terms of the Open Government Licence 
Citations
Supplemental information
Correspondence/contact details
Wallingford
Oxfordshire
OX10 8BB
UNITED KINGDOM
https://orcid.org/0000-0003-1127-0279