Appendix 5. Hydrology

Jump to navigation Jump to search

Navigate to: Hydromorphology // Macroinvertebrate // Vegetation // Fisheries // All Appendices

1. Definition of common hydrological terms

Gauging stations

Many catchments have a permanent gauging station usually located towards the bottom of the catchment which records flows. The station generally records a water level which is then converted to a flow using a specific equation unique to that particular gauging station. Some gauging stations, such as ultrasonics, records flow directly (see Figure A5.1). The flows/levels are recorded every 15 minutes with the start of day being 09:00 hours. From the 15 minutes time series a mean daily flow time series is calculated by taking an average of all the 15 minute flows between 09:00 on one day and 08:45 on the next day. Plate I illustrates some examples of different types of gauging station. The flow and level data is stored in the Agency's WISKI archive both as 15 minute levels and flows and mean daily flows.

Figure A5.1: Examples of Different Gauging Station Types, from left to right : Crump Weir, Flat V Weir, Essex Weir, Ultrasonic

Spot gauging

Spot gaugings (also known as current meter gaugings) are individual flow measurements taken at sites along a river (See Figure A5.2). Generally this type of monitoring is done at sites in a catchment where there are no gauging stations set up. Often a series of spot gaugings are done along the length of a catchment on the same day and this information is plotted up as an accretion diagram (see Figures A5.3 and A5.4) which shows how a river gains and loses water along its length. This example is from a Chilterns chalk stream (the River Misbourne)

Figure A5.2: Examples of Spot Gaugings Being Carried Out

Figure A5.3 shows that the river starts to gain water from its source at Mobwell down to Shardeloes Lake, then it starts to lose water, being dry from Chalfont St Giles to Chalfont St Peter. It the gains again from downstream of Chalfont St Peter to the gauging station at Denham Lodge. Figure A5.4 shows that in 2003 the river is flowing along its full length, though it still loses water between Lower Bottom Farm and Waterhall.

Mean daily flows

As explained above the 15-minute flow or level data recorded at gauging stations is converted to mean daily flows and stored on WISKI. The mean daily flows are the average flows between 09:00 on one day and 08:45 on the next, i.e. the average over a 24 hours period which starts at 9’clock in the morning. Figures A5.5 to A5.7 show the mean daily values plotted up as hydrographs. Figure A5.5 shows a hydrograph for a groundwater dominated catchment with flows increasing through the winter months generally reaching a maximum in March/April.

This hydrograph can be compared to the illustrated in Figure A5.6 where the graph is very spiky and there is not a great deal of difference between summer and winter baseflows. The flashy nature of the hydrograph shows the rapid response to rainfall events whereby the water flows over the ground or within the soil layer and is delivered to the watercourses shortly after the rainfall has occurred. Thus, the flows increase rapidly soon after the rainfall event and fall away relatively quickly once the rainfall has ceased. This shows as a spike on the hydrograph. With groundwater dominated catchments, rainfall percolates through to the aquifer once the catchment is saturated in winter. As the aquifer fills and groundwater (the water table) levels rise, spring flow outputs increase and the flows gradually rise over the winter reaching a peak in early spring. Once the warmer weather arrives the saturated soils begin to dry out, thus summer rainfall is held within the soil layer rather than percolating through to the aquifer. With the cessation of this movement of water downwards, the aquifer stores begin to deplete and groundwater levels start to fall over the summer months reaching a minima in September/October and it is not until the soils are saturated again by autumn and winter rains that recharge can start and the groundwater stores start to replenish. Figure A5.7 shows hydrograph with both groundwater and clay component.

Percentiles and flow duration curves

Flow time series are often represented at flow duration curves (FDCs) which is demonstrated in Figure A5.8 The FDC is a graphical representation of a flow time series which has been converted to percentiles. The flows are converted to percentiles by ranking the flow data in descending order and assigning a rank number to each flow. The equation;

Perc = rank/ (total+1)*100

Where rank is the rank number assigned to the flow time series which has been sorted in descending order, and total is the total number of flows in the time series. Table A5.1 shows the first 18 percentile calculations for a flow time series. The percentile calculation is done automatically in the Excel spreadsheet provided with this report.

Then the percentiles is calculated using Figure A5.8 shows the Q5, Q50 and Q95 percentiles position as blue vertical lines. The Q95 for the observed (gauged) flows is 1.22 cumecs, this means that for this given example, within the given time series of flows which ranged from 1992 to 2007, 95% of the time the flow was 0.22 cumecs or greater. In other words Q95 is a representation of low flows. Similarly the Q50 percentile means that 50 percent of the time flows are 0.119 cumecs or greater. Q50 is similar to, but not identical to the mean (average) flow. The Q5 percentile represents the high flow end of the time series range.

Table A5.1: Example of Percentile Calculation for a Flow Series

The Figure A5.8 also shows a naturalized time series (the green line) and the 90%, 80% and 50% of natural are also illustrated. The current actual glows are plotted in dark blue and a direct comparison can be made between the flows which actually occur in a river and those which would naturally occur if there was no abstraction or discharge going on the catchment. In this example it can be seen that at low flow the actual is less than half of what should be occurring naturally and even at high flows (Q5), the flow is well below what would naturally be occurring. Thus it can be demonstrated that the flow duration curves area good way of comparing what actually happens in a river with what the flows would be naturally and in the example it is clear that flows are significantly impacted by abstraction.

Figure A5.8: Example of a Flow Duration Curve

Volumetric descriptions of flow

There are a range of different descriptions of flow. Commonly hydrologists use cubic meters per second (cumecs). Hydrogeologists generally use megaliters per day (Ml/d), but can also describe flows in terms of thousands of cubic meters (TCMD). Flows can also be described in liters per second. Table A5.2 sets out all the common volumetric flow descriptors and gives the conversion rates for each. It should be noted that there are 1000 liters in a cubic meter, there are 1 million liters in a megaliter and 1000 cubic meters in a megaliter.

Table A5.2: Common Volumetric Flow Descriptors

Naturalised flows

Most rivers in England and Wales are not natural. They are influenced by a number of factors including abstractions, treated effluent discharges (in terms of both quality and flow), structures (weirs and online reservoirs) and management (dredging, weir operation, weed cutting etc).

Naturalised flows in terms of the amount or water flows in rivers can be calculated by using a rainfall-runoff model which has been calibrated for the current measured flow conditions and the abstraction and discharge files, which are part of the input time series, can be switched off to produce a naturalised series. Alternatively gauged flows can be naturalised by subtracting known discharge quantities and adding back known abstraction quantities. This is a slightly simpler method of naturalisation compared to the rainfall-runoff model, but can often be quite effective and gives a reasonable approximation.

Figure A5.9 shows the results of a calibrated rainfall-runoff model. The gauged flows recorded at a gauging station are in grey and the dashed line is the simulated flows from the model. The modelled flows are a good representation of the gauged flows. By running the model again with the abstraction and discharge files switched off the naturalised flows are then simulated. Figure A5.10 shows an example of the same model illustrated in Figure A5.9, but with naturalised flows. From the naturalised results (dashed line) it is clear that the catchment is significantly influenced by abstraction (comparing the dashed line with the grey infilled curve).

2. Available hydrological data and its uses

It is recommended that the ecologists ask the EA/SEPA/NIEA area hydrology team to produce these time series for them. A good deal of data is also available for download from the National River Flow Archive ( The flows should be presented in an excel spreadsheet with date in column A and the flow time series in cumecs (cubic metres per second) in column B. The hydrology team should be asked to infill any missing data and to calculate any derived flows at ungauged sites. If derived flow are required (i.e. a flow time series at a site where there is no gauging station), discuss with the hydrologist where you would like the time series to be and how good, i.e. how accurate the time series should be. The hydrologist should also be able to calculate a percentage error of the derived time series and give a description of the accuracy of the gauging station from which the derived series will be calculated.

It should be noted that a hydrological assessment tool is available within the Environment Agency in the form of an Excel spreadsheet called Flow Statistics.xls in which flow data can be input and monthly Q95 and monthly mean statistics can be calculated.

Deriving flows at ungauged sites using spot gaugings

In both hydrology and ecology it is often very useful to have a flow time series at sites other than just at sites where there is a permanent gauging station. If there are sufficient spot gaugings at the site of interest then a regression assessment can be carried out (see Figure A5.11'). In this example gauged flows are regressed against spot gauging flows from a site upstream of the gauging station and a regression equation derived. The equation (y = 0.6106x – 0.0368) can then be used to determine a time series at the target site (in this case the spot gauging site). So for each mean daily flow from the gauging station the flow in cumecs is substituted into the equation (x) to derive a flow at the spot gauging site. The R2 value in Figure A5.11 gives an indication as to how well the equation can reproduce the spot gauging values when the gauging station flows are substituted into it. A perfect fit would give an R2 of 1. Preferably your R2 value should be 0.8 or greater to give reasonable result.

Figure A5.12 shows the flow time series derived from substituting the gauging station flows into the regression equation. The spot gaugings are also plotted up and it illustrates that they fit on top of the derived flows (pink line) very well with only two significant outliers on 19/2/2001 and 20/1/2004.