D-Risk rationale


Although typically considered a ‘wet’ region for food cropping, UK production of high-value agricultural and horticultural crops, such as potatoes and field-vegetables, is not immune to drought. Supplemental irrigation is an essential component of production for many crops, used to buffer short-term periods of rainfall scarcity, and is a highly productive use of water for crops that are very sensitive to drought stress. In England, small depths of irrigation applied to high-value crops generate substantial financial benefits not only to individual farming enterprises, but also in terms of ‘added value’ to the supply chain and those businesses engaged in processing and distribution of fresh fruit and vegetables.

The UK has been affected by several severe drought episodes including 1975-76, 1988-1992, 1995-1997, 2003, 2004-2006, 2010-2012 and most recently in 2018 and 2019. Whilst the 1975-76 drought is widely recognised as being the most severe, all other recent drought events have also caused major impacts to agribusinesses and the rural economy. For example, previous research estimated the financial impacts of restrictions on irrigated production in East Anglia to be £144 million, excluding downstream impacts on the distributional supply chain.

A previous study as part of the NERC Drought and Water Scarcity programme showed that almost three-quarters of respondents considered droughts to be a very important risk to their irrigated business, and nearly half believed it was highly likely that drought impacts will become more serious in the future, partly due to proposed abstraction reform that will reduce their licensed headroom.

In my view, you cannot plan exactly the same area and hope that is going to rain and you get the full crop. You have got to reduce the economic risk to your business

Although government abstraction reforms are ongoing, the likely impacts will be profound and the consequences raise critically important questions regarding:

  • how changes in future licensed allocations might impact on growers’ ability to cope with future dry years?
  • how businesses can maximise productivity and economic output?
  • how best to plan strategically for changes in water allocation and reliability?

Working with the farming community, we have developed D-Risk, a web-based decision support tool to help agribusinesses and water and catchment managers evaluate the impacts of abstraction licence changes on their short-term irrigated cropping programmes, and functionality to inform longer-term strategic options for reconciling drought risk with water availability including collaborative water sharing or co-management of water allocations.

Using D-Risk


This section provides a brief explanation of the D-Risk webtool, including the farm data needed and how to interpret the key outputs. The explanation is divided into three sections:

  1. Description of input data for D-Risk;
  2. Description of D-Risk analysis, and;
  3. Interpretation of D-Risk outputs

Data input


Farm data

Data input to the D-Risk webtool starts with entering your location and period of analysis. This will be used to retrieve weather data and hydrological data for the area. D-Risk conforms to Cranfield University’s legal requirements for data protection and confidentiality - no farm or licence data are stored on our server.

Initially, you select the river catchment in which you are interested, which are ordered by country and county in the drop down menu. A link above the drop-down list (“see location”) opens a new browser window and takes you to an interactive map on the National River Flow Archive (NRFA) website where you can easily identify the gauging station ID or catchment name. You can select between an analysis for multiple farms in the catchment or for a single farm (in which case you have to enter the farm postcode to retrieve weather data).

The modelled “Baseline” weather data within D-Risk is consistent with the observed climate of 1975-2004 and has no additional influence of climate change. The “Near Future” data for 2020-2049 assumes that the world continues along a path of high greenhouse gas emissions which changes the UK climate. Unless you specifically wish to evaluate the effect on the changing climate on your risk, you should select the “Baseline” period.

Next, it is necessary to provide D-Risk with information regarding the irrigated crops and the soil types on the farm(s). This requires you to follow a simple process for each irrigated crop:

  1. Select an irrigated ‘Crop type’ from the dropdown menu
  2. Select the ‘Soil type’ from the dropdown menu
  3. Select the ‘Typical planting month’ from the dropdown menu
  4. Enter the ‘Irrigated hectares’ for this crop-soil combination
  5. Click the ‘Add’ button

The same crop can be added multiple times to allow for differences in soil type, crop variety or irrigation plan. A crop can be removed using the ‘delete’ button at the end of their row (cross shaped button).

D-Risk automatically estimates your ‘design’ dry year and 'average year' irrigation needs (depths applied, mm) for each crop, assuming the irrigation is scheduled to maximise both yield and quality. These are ‘net’ requirements so exclude any system application efficiency losses.

If required, you can modify the estimated ‘design’ irrigation needs by a percentage factor to match actual application depths. The percentage factor can be modified in each crop’s row either by typing it in or by clicking on the up and down arrows that appear in the box when placing the cursor over it.

Finally, the volumetric irrigation demand (m3) for the farm(s) is calculated based on the defined soil type, irrigated area, irrigation need for each crop type and D-Risk ‘design’ dry year data for your location.

Note: A ‘design’ dry year is defined as a year with an irrigation need with an 80% probability of non-exceedance or roughly equivalent to the 16th driest year in 20 years.

Abstraction licence and storage data

Individual abstraction licence data needed for the farm(s) are: water source (surface water, groundwater); purpose (direct or storage); annual and daily licence limits; the specified start and end month for each licence; and any Hands Off Flow (HOF) conditions. These can be found in your EA abstraction licence.

Depending on how a HOF is specified, the Hands Off Flow condition can be set by either (1) selecting from the drop down list of flow percentiles (in which case the flow threshold is displayed in m3/s) or (2) you can manually enter ‘Your actual HOF (m3/s)’and the webtool automatically calculates the equivalent HOF percentile. HOF are applicable only for surface water licences and hence ‘No HOF’ is specified for all groundwater licences.

Entering an abstraction licence data into D-Risk requires you to complete the following using drop-down menus or data input boxes:

  1. Select the appropriate ‘Water source’
  2. Select the ‘Licence purpose’
  3. Enter the ‘Annual licence volume’ in m3
  4. Enter the ‘Daily licence limit’ in m3
  5. Select the ‘Start month’
  6. Select ‘End month’
  7. Select ‘Hands Off Flow percentile’ or enter ‘Your actual HOF (m3/s)’
  8. Click on the ‘Add’ button

Licences can be removed using the delete button at the end of their row (cross shaped button).

You can optionally select to ‘Include emergency drought restrictions (e.g.S57)’ in the analysis, based on idealised Section 57 restriction rules in England.

The total combined usable storage capacity for all irrigation reservoirs available on the farm(s) can be entered under ‘Total usable storage capacity (m3)’. If no data are entered in this box, D-Risk assumes there is no storage capacity available.

Back to top

Summary description of D-Risk analysis


D-Risk uses a simple water balance to calculate the potential soil moisture deficit (PSMD) using monthly rainfall (P) and reference evapotranspiration (ETo) data, and the maximum PSMD (PSMDmax) is identified for each year. Using the crop and soils data provided by the user and the annual PSMDmax, D-Risk then calculates the theoretical annual irrigation needs for each crop-soil combination using a set of equations derived from previous research. By combining data on irrigation needs (depths water applied, mm) with the irrigated areas (ha) data, D-Risk calculates total water demand for each year.

A monthly time-step water balance model is then used to assess how irrigation demands compare against the annual licensed abstraction volume for the farm(s), taking into account abstraction licence constraints and on-farm water storage availability. In D-Risk, it is assumed that licenses dependent on surface water are used before licensed groundwater sources. Similarly, direct abstraction is preferred before storage, leaving farm reservoirs as the last source of water. Licence use is only possible between the specified start and end months.

From the water balance modelling, annual irrigation deficits and licensed abstraction ‘headroom’ are calculated. An irrigation deficit is assumed to be any proportion of annual demand that was not met either (1) due to annual or monthly licence limits, (2) abstraction constraints imposed by local flow conditions, and/or (3) not being able to supply water from reservoirs. Licensed ‘headroom’ is defined as the proportion of licensed volume that is not abstracted in any given year. It is calculated from the sum of all available licences (both direct and storage).

Because droughts are both rare and always different, the small number of past ‘real’ drought events are insufficient to provide a realistic assessment of future drought risk. Consequently, the MaRIUS project used a very complex regional climate model to produce a dataset of 100 series of ‘synthetic’ weather that could have occurred in the past. The data have been checked for robustness against actual historical records. Each weather series, of 30 years length, is equally probable weather. The D-Risk tool uses this data in the method described above in two ways:

  • Firstly, we have assumed that the future probability or likelihood of a given level of irrigation need is based on all of these time series and years - this is equivalent to basing the probability of rolling a given number on a dice on 3000 (100 x 30) throws of the dice. We have termed this analysis ‘long term average’ conditions;
  • Secondly, we have assumed that the future probability or likelihood of irrigation need is based on the years in a single time series - this is equivalent to basing the probability of rolling a given number on a dice on just 30 throws of the dice. In this analysis, some of these time series will be drier (or have a greater number of extremely dry years) than others. These ‘driest’ and ‘wettest’ time series then provide an uncertainty boundary (envelope) around the long term ‘average’.
Back to top

D-Risk outputs and interpretation

D-Risk results are presented in the form of probability distribution functions both in graphs and tables. Two sets of key outputs of relevance to growers, irrigation deficit and headroom, are produced.In the first set of two graphs, the upper graph shows the annual probability or risk of exceeding a given level of irrigation deficit without considering any abstraction constraints imposed by local river flow conditions. The lower graph takes full account of constraints imposed by volumetric licence limits, storage volumes and the local flow conditions, providing a fuller insight into drought risk. Providing the two graphs allows you to understand the relative importance of constraints due to volumetric (licence and storage) limits and local river flow conditions. The second set provides similar graphs for headroom.

1. Irrigation deficit (m3)

Irrigation deficit is defined as the total volumetric irrigation demand not supplied in a year. This is due to either monthly and/or annual licence limits being reached and/or insufficient water stored in reservoirs (upper graph) and with abstraction constraints imposed by local flow conditions (lower graph).

The figure to the right shows an example graphical output for irrigation deficit, showing the probability or likelihood of exceeding a certain level of irrigation deficit in any year. The solid black line represents the long term probability (from the 3000 years of synthetic weather). So, for example, there is a 30% probability that the irrigation deficit will exceed 30,000 m3. Conversely, a deficit of 30,000 m3 will be exceeded in 30% of years over the long term.

Picture 1

The figure also contains a solid grey area that represents the uncertainty region associated with the long term results, with the dashed lines representing the minimum (from the ‘wettest’ 30 year weather series) and maximum (from the ‘driest’ weather series) boundaries of the region. This means that for a single probability value, the deficit can vary around the average and be anywhere between the lower and upper boundaries of the uncertainty region. In the figure below, for a 30% probability of exceedance, the irrigation deficit could be between 20,000 m3 and 55,000 m3 (dashed yellow arrows) depending on whether the coming years are significantly wetter or drier than the long term average. Conversely, for example, following the dashed blue lines shows that the probability of exceeding a deficit of 25,000 m3 ranges between 25% (best case) and 48% (worst case), with a long term exceedance probability of 37%.

2. Headroom (%)

Headroom represents the volume of licensed water not used in a given year. Since annual licence limits are assumed to be constant, headroom is represented as a proportion (%) of total licensed volume. A headroom of 0% means that the business has reached its licensed volume limit and cannot abstract any further water for irrigation or for refilling reservoirs.

The graph to the right can be read in the same way as described above. For example, the red line shows that there will be 85% annual probability of exceeding a headroom of 10% in a given year or conversely that there is a 15% annual probability of having less than 10% headroom.

Back to top
Picture 2

D-Risk reservoir


The reservoir sizing module in D-Risk (here-after called ‘D-Risk reservoir’) is designed to help agri-businesses assess how the construction of an on-farm reservoir could reduce their drought risk. D-Risk reservoir determines the irrigation reservoir sizing based on the D-Risk’s calculated irrigation deficit that takes account of abstraction constraints imposed by local flow conditions, and provides an indicative estimate of the construction cost.

Based upon the irrigation deficit profile from D-Risk, D-Risk reservoir calculates and tests a range of reservoir volumes. The maximum and minimum volumes are prepopulated, but can be modified by the user. For each, a new winter (October to March) surface water storage licence equivalent to 100% of the useable reservoir volume is assumed– but this value can be modified by the user. The daily abstraction limit for the storage licence is set to 1/90th of the licenced annual volume, so that the default assumption is that an empty reservoir can be filled during 3 of the 6 months of the winter storage licence. Using the original D-Risk data and the additional winter storage licence and reservoir storage capacity, D-Risk reservoir calculates the new farm water balance and risk profile. Both the original annual irrigation deficit probability and the new probability profiles for each tested reservoir option are plotted together to facilitate an easy assessment of the risk reduction provided by each tested reservoir volume.

Indicative cost estimates for each reservoir volume for both clay-lined and (plastic- or HDPE-) lined designs are provided. The reservoir costs are calculated based on volume-cost functions derived from Weatherhead et al. (2014)1, adjusted for inflation to 2019 prices, and recent data provided by Hawes Associates and HYDREAU Consulting Engineers. The indicative cost estimate provided includes only civil engineering works, earthworks and lining costs, and does not include additional capital costs such as environmental assessments, planning and permitting applications.

D-Risk reservoir can be used to provide rapid indicative guidance to businesses to help determine whether the likely monetary investment in reservoir construction for a given level of reduction in annual drought risk merits further investigation. D-Risk reservoir is not intended to replace professional advice from reservoir design engineers.

Process for using D-Risk reservoir

After running D-Risk to produce the irrigation deficit and headroom probability profiles, the option to run D-Risk reservoir appears at the bottom of the screen. By clicking on this:

  1. Based on the irrigation deficit that takes account of abstraction constraints imposed by local flow conditions (lower irrigation deficit graph) of your farm(s), D-Risk reservoir prepopulates the maximum and minimum reservoir volumes to be tested, under ‘Additional reservoir storage’. You can change them by clicking on the values.
  2. The entry of ‘Additional storage licence volume as % of additional reservoir storage’ calculates the annual storage licence from the reservoir storage. The prepopulated value of 100% gives a storage licence equal to the reservoir volume. You can change this entry by clicking on the value. By changing this value you can assess how your irrigation deficit profile changes with licence size for a given reservoir volume.

1 Weatherhead, E.K., Knox, J.W., Daccache, A., Morris, J., Kay, M., Groves, S., et al., 2014. Water for agriculture: collaborative approaches and on-farm storage. FG1112 Final Report to Defra, Cranfield University