Defining hydrological categories — United Kingdom (UKCEH)


Working with the UKCEH Hydrological Outlooks team, this case study explored how to categorise hydrological data to provide an overview of the status and outlook. It also explored how to display status and forecast information in graphs.



UK Centre for Ecology & Hydrology


Hydrological Outlook UK


UKCEH's Hydrological Outlooks products are already categorised into hydrological categories (e.g. above normal and below normal). The table below shows two categorisations used by products that have been used for the HydroSOS case studies, and demonstrates how different providers use different categorisations. For status information (such as the USGS WaterWatch), this issue is less of a challenge as there is only one value to categorise. However with ensemble data determining a “dominant” category from the ensemble forecasts to display on the map is challenging when the categories are not equal in size. Engagement with end users indicated that five categories is preferable to seven, and only three should be displayed on the map. Therefore the HydroSOS team developed a five category system that could be condensed into three categories (by combining the bottom and top two categories), providing three equal sized categories for the map whilst retaining information on more extreme conditions for the graphs.

UK Hydrological Outlook USGS WaterWatch HydroSOS
Exceptionally high : >95% High
Notably high: 87–95% Much above normal: >90% High: 90–100%
Above normal: 72–87% Above normal: 76–90% Above normal: 67–90%
Normal: 28–72% Normal: 25–75% Normal: 33–67%
Below normal: 13–28% Below normal: 10–24% Below normal: 10–33%
Notably low: 5–13% Much below normal: <10% Low: 0–10%
Exceptionally low: <5% Low

The UKCEH Hydrological Outlook team therefore provided raw daily status and outlooks data, which was reprocessed into monthly data via a Python script. This script also output the latest climatology data according to the HydroSOS categories, and quantified the forecasts against these categories. It is worth noting that the UK Hydrological Outlooks forecasts are accumulated (the 3 month forecast predicts the average flow over the next three months, not just the flow for the month in 3 months’ time), and so the outlooks climatology is therefore different to the status climatology.

HydroSOS integration steps

Access UKCEH Outlook raw data
Convert daily flow to monthly accumulated data
Process simulated historic data to generate flow categories
Quantify outlook data according to flow categories
Display data as map points and graphs

Modelled data

Status data are hosted by the National Hydrological Monitoring Programme at UKCEH. UK outlook data are taken from the Historical Climate method of the Hydrological Outlook UK. Modelled using the “Ensemble Streamflow Prediction (ESP)” method using the GR4J model. 64 of the 315 modelled catchments are displayed here. Simulated observations are used for six months’ status assessment prior to the initiation of the forecast. The forecast runs for twelve months, of which six months are shown here. Five categories are used to define the flow values in the forecast, reduced to three on the map.