## Measurement of open water evaporation

An indirect measurement of evaporation from open water can be made by taking the difference in storage of a body of water measured at two known times, which gives a measure of the evaporated water over the time interval. If rain has fallen during the time period, then the rainfall quantity must be taken into account. In practice, this water budget method is used on two widely differing spatial scales, by measurements at reservoirs and by measurements with specially designed instruments maintained at meteorological stations.

4.2.1 Water budget of reservoirs

The evaporation from a reservoir over a time period is given by: Eo = I - O ± A S

where I = riverflow into the reservoir plus precipitation on to the reservoir surface, O = outflow from the reservoir (i.e. drawoff to supply and overflow) plus subsurface seepage and AS = change in reservoir storage. Although water engineers are anxious to assess evaporation losses from their surface reservoir sources of supply, few impounding reservoirs are instrumented to give the measurements required for the water budget equation. In the UK, it is rare to find river flow gauging stations on the streams flowing into reservoirs. There are usually several feeder tributaries, which add to the complexity and cost of total inflow measurements.

The evaluation of outflow from a reservoir is, however, usually made regularly. Measurements of drawoff to supply compensation water releases to the river and overflow are made regularly and the water levels in the reservoir give changes in storage. Difficulties sometimes arise in the assessment of flood flows over spillways while the amount of leakage beneath the dam and through the sides and bottom of a reservoir can only be roughly estimated. The measurement of evaporation from an operating reservoir using the water budget method can only give a broad approximation to water loss unless a thorough knowledge of the different components is available.

A valuable, if old, study of reservoir evaporation was made by Lapworth for the Kempton Park Reservoir from 1956 to 1962 (Lapworth, 1965). During these years, there was no inflow and no outflow from this storage reservoir on the Thames flood plain. Hence, it was expected that Eo would be equal to AS, the change in storage with rainfall deducted. In addition to the necessary rainfall measurements, meteorological stations were set up to make the observations required for calculating evaporation (to be described later) so that comparisons of the different methods of evaporation evaluation could be made. The results showed that there were marked seasonal differences between the measured changes in storage ERes and the calculated Eo. The average annual evaporation total over the 7 years was 663 mm ERes and the monthly means are given in Table 4.1. In explanation, the values for ERes, the observed evaporation from the reservoir, have two components, Eo, calculated evaporation due to surface water conditions, plus ESt (i.e. ERes - Eo), calculated evaporation taking into account by the changing heat storage of the water in the reservoir. As seen in Table 4.1, during the autumn months, evaporation from the reservoir due to heat storage effects is enhanced. This results from heat diffusing and convecting to the surface from the lower water layers, which had absorbed the energy of the summer sun. In the spring, the temperature of the water body is low following the colder winter months and the evaporation is reduced as some of the available incoming energy is absorbed by the cold lower layers. The overall effects on reservoir loss by these seasonal fluctuations of stored energy are dependent on the dimensions of the reservoir. Wide shallow bodies of water are more readily affected by marked seasonal temperature changes, whereas in deep narrow reservoirs, the smaller seasonal fluctuations of stored energy will have less effect on water loss. For an operational impounding reservoir, heat storage is also affected by water temperatures of inflows and outflows. To obtain monthly estimates of ERes in practice, calculated values of Eo can be obtained by other methods, to which are added estimated values of ESt, either positive or negative according to season.

 J F M A M J J A S O N D ERes 15 18 28 48 76 94 107 94 74 58 33 18 £St +3 -3 -13 -20 -18 -23 -5 +5 + 15 +23 +20 + 10 Eo 12 21 41 68 94 117 112 89 59 35 13 8 Eo tank 5 13 30 56 89 109 103 84 58 33 15 8

4.2.2 Evaporation tanks and pans

Although there may be difficulties in relating the measurements of evaporation from small bodies of water to the real losses from a large reservoir, the advantages in using tanks and pans are numerous. These relatively small instruments, with either circular or square plan sections, are easily managed and can be transported to any required location for simple installation. Originally designed to be kept at meteorological stations where readings are made regularly at a fixed time each day, their operation has been improved by the attachment of self-filling devices and by the continuous measurement of the water level (Chow, 1994). However, the general opinion in the UK is that this method of evaporation measurement is unreliable and the data collected are incapable of being adequately quality controlled. The current tendency is to use calculated estimates of evaporation (Chapter 10).

Of the many evaporimeters used experimentally in the 1860s, the tank ascribed to Symons became the British standard instrument (Fig. 4.1a). It is a galvanized iron tank, 1.83 m square and 0.61 m deep and set in the ground with the rim 100 mm above ground level. The tank holds about 1.8 m3, the water level being kept at near ground level and never allowed to fall more than 100 mm below the rim. Measurements of the water level are made daily using a hook gauge attached to a vernier scale and any rainfall measured in the previous 24 h must be added. The depth of evaporation is evaluated as shown in the example in Table 4.2. Records compiled in this way from a British standard tank kept at Kempton Park are given in the last row of Table 4.1. It should be noted that rainfall observations made at 0900 h are normally allocated to the previous day.

The most widely used instrument nowadays is the American or US Class A pan (Fig. 4.1b). This is circular with a diameter of 1.21 m and is 255 mm deep. It is set with the base 150 mm above the ground surface on an open wooden frame so that the air circulates freely round and under the pan. The water level in the pan is kept to about 50 mm below the rim. The level is measured daily with a hook gauge and the difference between two readings gives a daily value of evaporation. Alternatively, evaporation can be obtained by bringing the water level in the pan back to a fixed level with a measured amount of water. Again any rainfall must be allowed for. Since the sides of the pan are exposed to the sun, the contained water tends to attain a higher temperature than in pans set in the ground and thus the measured evaporation is higher than otherwise. For example, in the Kempton Park study, a US Class A pan was installed in 1959, and for the 4 years of records, 1959-62, the average annual evaporation measured by the pan was 963 mm, compared with 673 mm from the reservoir and 625 mm from a Symons tank. On an annual basis, the reservoir evaporation was 0.7 times the pan measurement. This factor, 0.7, is known as the pan coefficient and its value varies slightly over different climatic regions. If seasonal evaporation values are required, the heat storage effects cause greater differences between US Class A pan and reservoir. Monthly pan coefficients must be obtained and used to give monthly estimates of reservoir evaporation from pan measurements.

Many experiments with modified installations of the US Class A pan have been made in attempts to inhibit the exaggerated evaporation due to the overheating of the water. In India, the outside has been painted white to increase radiation reflection and some studies have recommended setting the pan in the ground. In arid regions, the pan