Met Office Hadley Centre observations datasets
Home >

HadISDH - gridded global land surface humidity dataset - version

HadISDH is a global gridded monthly mean land surface humidity dataset based on the quality controlled sub-daily HadISD dataset which is in turn based on the ISD dataset from NOAA's NCEI. Quality controlled and homogenised monthly mean anomalies (relative to a 1981-2010 base period) are provided alongside uncertainty estimates (station [climatological, measurement and homogeneity] and gridbox sampling). Actual values, climatological mean and standard deviation and no. stations per gridbox are also provided. The dataset begins in January 1973 and is updated annually.

Gridded products are available for six humidity variables in addition to temperature -*.

LATEST VERSION: The current version of HadISDH is and continues until December 2018. Only one update will be made this year so this is a final (f) version. For previous versions please contact the dataset maintainers. Please read the Update Document.

NEW MARINE VERSION: HadISDH-marine. is now available as a TEST product. PLEASE NOTE THAT THIS HAS NOT YET BEEN THROUGH PEER REVIEW!!! Feedback on any issues/errors found is much appreciated - please contact Kate Willett. You can find out more about this dataset and download it from the HadISDH-marine. webpage.

Brief description of the data

HadISDH utilises simultaneous subdaily temperature and dew point temperature data from >8000 quality controlled HadISD stations that have sufficiently long records. Further information on the quality control tests and HadISD can be found here. All humdiity variables are calculated at hourly resolution and monthly means are created.

Monthly means are homogenised to detect and adjust for features within the data that do not appear to be of climate origin. While unlikely to be perfect, this process does help remove large errors from the data an improve robustness of long-term climate monitoring. We have used NCEI's Pairwise Homogenisation Algorithm directly on DPD and T. We have designed an indirect PHA method (ID PHA) whereby changepoints detected in DPD and T are used to make adjustments to q, e, Tw and RH. Changepoints from DPD are also applied to T. Td is derived from homogenised T and DPD. Further information can be found here. Stations with very large (>5 °C in T and Td, > 3 g kg-1 in q and > 15 %rh in RH) adjustments applied are removed.

Measurement, climatological and homogeneity adjustment uncertainty is estimated for each month. Further in formation on uncertainty estimates can be found here.

Climatological averages over the 1981 to 2010 period are calculated and monthly mean climate anomalies obtained. These anomalies (in addition to climatological mean and standard deviation, actual values and uncertainty components) are then averaged over 5° by 5° gridboxes centred on -177.5°W and -87.5°S to 177.5°E and 87.5°N.

Given the uneven distribution of stations over time and space, sampling uncertainty is estimated for each gridbox month. Further in formation on uncertainty estimates can be found here.

Keep in touch

Follow us on twitter: @metofficeHadOBS for updates, news and announcements.

For more detailed information, follow our HadISDH blog. Here we describe bug fixes, routine updates and other exploratory analysis.

2018 ANNUAL ANOMALIES (1981-2010 climatology)








Go to download page Data are available from the download page without charge, please read the licence conditions. Data are also available from the CEDA Catalogue.

Commercial and media enquiries

You can access the Met Office Customer Centre, any time of the day or night by phone, fax or e-mail. Trained staff will help you find the information or products that are right for you.
Contact the Met Office Customer Centre


When using the dataset please use the following two citations and state the version used:

Willett, K. M., Dunn, R. J. H., Thorne, P. W., Bell, S., de Podesta, M., Parker, D. E., Jones, P. D., and Williams Jr., C. N.: HadISDH land surface multi-variable humidity and temperature record for climate monitoring, Clim. Past, 10, 1983-2006, doi:10.5194/cp-10-1983-2014, 2014.
Main Text PDF file 2.6Mb
Supplementary Material PDF file 1.8Mb

Smith, A., N. Lott, and R. Vose, 2011: The Integrated Surface Database: Recent Developments and Partnerships. Bulletin of the American Meteorological Society, 92, 704\u2013708, doi:10.1175/2011BAMS3015.1.
Available from BAMS

We strongly recommend that you read the Willett et al. (2014) paper before making use of the data. Additionaly information can be found in the reference for version 1.0.0:

Willett, K. M., Williams Jr., C. N., Dunn, R. J. H., Thorne, P. W., Bell, S., de Podesta, M., Jones, P. D., and Parker D. E., 2013: HadISDH: An updated land surface specific humidity product for climate monitoring. Climate of the Past, 9, 657-677, doi:10.5194/cp-9-657-2013.
PDF file (5.4MB)

Useful Diagnostics

Figures from the Willett et al. 2014 paper.

Figures from the Willett et al. 2013 paper.

Figures from State of the Climate and IPCC monitoring plots.

Decadal Trend maps for each variable.

Annual and monthly average timeseries for the globe, hemispheres and tropics, including uncertainty estimates.

Annual anomaly maps from 1973 onwards for specific humidity.

Annual anomaly maps from 1973 onwards for relative humidity.

Annual anomaly maps from 1973 onwards for vapour pressure.

Annual anomaly maps from 1973 onwards for dew point temperature.

Annual anomaly maps from 1973 onwards for wet bulb temperature.

Annual anomaly maps from 1973 onwards for temperature.

Annual anomaly maps from 1973 onwards for dew point depression.

We have adapted our versioning system from CRUTEM4, and so the dataset numbering is of the form HadISDH.type.X.Y.Z.1234i. 'type' refers to the variable (e.g., landq=specific humidity). 'X' is for a major change and would be accompanied by a peer-reviewed paper or Met Office Technical Note. 'Y' is a more minor change, e.g., in one of the QC tests or homogenisation algorithms and would be described in a tech-note. 'Z' is a small change, for example addition or changes to data in the past. The last complete year of the dataset is given by '1234', and the final character shows if the dataset is f-final or p-preliminary. Therefore HadISDH.landq. is the preliminary version of the dataset containing data up to the end of 2012.

Dataset and Diagnostic Creation Code

The IDL and python code used (excluding quality control, homogenisation and regional average uncertainty estimates) was written by Kate Willett and can be found here. We do not intend to provide in-depth support for this code. We do appreciate constructive feedback on this code.

Dataset produced in collaboration with: