Evaporation From The Desert

EVAPORATION

FROM THE DESERT: SOME PRELIMINARY RESULTS OF HEIFE (Research Note) JIEMING

Lanzhou

Institute of Plateau Atmospheric

WANG Physics, CAS, Lanzhou,

P.R.O.

China

and YASUSHI Disaster Prevention

MITSUTA

Research Institute, Kyoto University,

Uji, Kyoto, Japan

(Received in final form 4 November, 1991) Abstract. As part of a feasibility study for HEIFE, a Sino-Japanese Cooperative Research Program on Atmosphere-Land Surface Processes in Heihe River Basin, fluxes of water vapour are estimated above and below the sand surface. It is found that during the daytime under clear skies, the water vapour flux is directed towards the surface from both the atmosphere above and the sand below.

1. Introduction Evaluation of evaporation from the desert is one of the most important but unsolved tasks in meteorology of arid areas. Evaporation from an evaporation pan exceeds a few thousand millimeters a year in the desert area. However, this is apparently far from the true evaporation from a desert surface. To make some comprehensive studies of ‘air-‘land surface interactions over an arid area, the SinoJapanese Cooperative Research Program on Atmosphere-Land Surface Processes (HEIFE) is being planned, and some feasibility studies have been undertaken in the Heihe River Basin area in northwest China. The Heihe River is fed by a glacier in the Quilian Mountains, flows through the arid area of the Hexi Corridor and disappears in the Mongolian desert. There is a mixture of Gobi and sand desert, oasesand irrigated farm land in the experimental area in the Hexi Corridor. Air-land surface interactions are being studied at five observing stations of various surface conditions together with river and soil water budgets in the area (Mitsuta, 1988). 2. Water Vapor Flux Measurement A pilot experiment of turbulent flux measurement was made over the Gobi desert surface in September, 1988. This location (100”06’E, 39”09’N, 1480 m MSL) is one of five proposed stations of the project. The surface of the Gobi desert consists of sand grains and small pebbles with very sparse vegetation. About 2 km to the north is the Linze oasis. Turbulent flux measurements were made with a threeBoundary-Layer Meteorology 59: 413-418, 1992. @ 1992 Kluwer Academic Publishers. Printed in the Netherlands.

414

JIEMING

WANG

AND

YASUSHI

MITSUTA

dimensional sonic anemometer-thermometer (Mitsuta, 1974), a 12 pm tungsten wire thermometer and a fine wire thermocouple psychrometer (Tsukamoto, 1986) installed at a height of 2.5 m. The output signals were digitized at the rate of 10 Hz and recorded on tape for 30 min of every hour. The sonic anemometer has sufficient dynamic response to measure atmospheric turbulence. But the dry- and the wet-bulb of the fine wire thermocouple psychrometer have different time constants (0.19 and 0.7 set, respectively). A correction has been made to improve the dynamic response of the wet bulb (Tsukamoto, 1986), but it is not fully compensated on the high frequency side. However, as the high frequency side is not effective in turbulent transport processes (Panofsky and Dutton, 1983), it does not produce large errors in water vapor flux measurements. The diurnal variation of net radiation and sensible and latent heat fluxes (Wang and Mitsuta, 1990) observed on clear summer days in 1989 are shown in Figure 1. The variations of net radiation and sensible heat flux are quite similar to many other observations over bare ground on clear days (e.g., Panofsky and Dutton, 1983). However, there is an unexpected phenomenon, viz., that the latent heat flux was nearly always negative (condensation) in the daytime (7 to 18 h) becoming positive (evaporation) in the late evening. The averaged downward water vapor transport in the daytime was about 10 g/m2/h. The positive upward transport of water vapor in the night-time was about the same order of magnitude. Therefore, these data show that there is almost zero net daily flux of water vapor on a fine day over the Gobi desert. To confirm the existence of this phenomenon, turbulent flux measurements at the same position were repeated together with profile measurements in August, 1990. In this latter period, it rained lightly or moderately several times. The phenomenon of downward water vapor transfer was not so obvious but still occurred in more than half of the daytime runs. Especially after dry weather lasting several days, the diurnal variation of the energy balance components was quite similar to that of 1989 shown in Figure 1. An example is shown in Figure 2 (August 20, 1990). The soil heat flux is measured by a flux plate at 5 cm depth. Around local noon, incoming net radiation is almost balanced by outgoing sensible heat and heat conduction into the ground. Therefore a phase change at the ground surface cannot be seen. A downward water vapor flux was also revealed by the profile.measurements of specific humidity presented in Figure 3, which shows humidity and temperature at 11 and 12 h on the same day. The data below 8 m (shown by open circles) were measured by a Humicap and a resistance thermometer on a tower; the other measurements were obtained from a psychrometer on the TS-3A tethered balloon sensor. A low-level humidity inversion is clearly seen in both cases.The downward water vapor flux is the normal Fickian type diffusion down a mean gradient. A temperature inversion in the layer from 15 to 50 m at this time may be a mesoscale phenomenon but the cause is unknown. Large incoming solar radiation and the low thermal conductivity of the sand

EVAPORATfON

FROM

l-HE

DESERT

41.5

Fig. 1. Diurnal variation of net radiation, Rn, sensible heat flux, H and latent heat flux, LX on the Gobi desert surface on September 13-16, 1988, after Wang and Mitsuta (1990).

W/El2 500 r

Local

time

Fig. 2. Hourly variation of the surface-layer energy budget components; net radiation, Rn, sensible heat flux, H, latent heat flux, LE and heat flux into the ground, G at the same place as Figure 1 on August 20, 1990. It was a clear day and in the period 8-18 h, the wind direction was mostly NNW with wind speed of 4-7 m/s, but the wind reversed to S at 20 h with lower speed.

416

JIEMING

WANG

q (g/kg)

AND

YASUSHI

MITSUTA

T (‘C)

Fig. 3. Profiles of air temperature (right) and specific humidity (left) at the same place as Figure 2 at 11 and 12 h, August 20, 1990. Open circles were measured on a tower and solid circles by a tethered balloon system.

cause the surface temperature to increase and to dry the surface sand in the daytime. Therefore relative and specific humidity inversions may occur in the air layer near the surface. This phenomenon could be understood easily if the water vapor evaporated at the sand surface were transported downward and stored underground. 3. Water Vapor Transport in Sand The water content under the surface was observed together with profiles in the air in a preliminary study at the sand desert near the Gobi station on a fine day in March 1990. Air temperature and humidity were measured up to 1.5 m above the sand surface, and soil temperature, sand water content and inter-particle humidity down to 0.5 m below the surface were also observed. Sand water content was measured by the drying method and underground inter-particle humidity was measured with a Humicap in a porous metal shielded cap buried in the sand. The observed results (Sahashi et al., 1990) around noon of March 17, 1990 are shown in Figure 4. As is clear from this figure, above the surface, the air temperature profile shows highly unstable conditions, but specific and relative humidity show a slight inversion, as in the casesdiscussed earlier.

EVAPORATION

FROM

THE

417

DESERT

m

0 0

10

0

50

5

20-c IOmb 100%

T R,f S

Fig. 4. Profiles of air temperature, T, specific humidity, q, relative humidity, RH and soil moisture, S over and in the sand desert near the Gobi desert shown in Figures 1-3, around noon on March 19, 1990, after Sahashi et al. (1990). The underground humidity was measured by a Humicap in a porous metal shielded cap, and soil moisture was measured by the drying method.

The underground temperature decreases rapidly with depth, and sand water content and the inter-particle humidity increase rapidly with depth. There is no evidence of a water phase change at the surface. Also the inter-particle humidity is almost saturated below 0.2 m. This means that evaporation is occurring not at the ground surface but at underground evaporation surfaces at 0.2 m depth, below which the sand is wet. This underground profile of humidity near the surface shows that there must exist upward water vapor transport through sand above about 0.2 m. This means that in the daytime, water vapor is converging to the air-ground interface in the desert area, because water vapor in the air is transported downward and water vapor in the sand is transported upward.

4. Conclusion The question remains: what happens to the water vapor which is flowing towards the air-ground interface in this desert in the daytime? More detailed studies of water and heat transport in the air and in the sand will be undertaken in the Intensive Observation Periods of HEIFE, in which the humidity sensor will be changed to an infrared hygrometer with higher resolution. An answer to the question must await analysis of these further observations. However, we can say

418

JIEMING

WANG

AND

YASUSHI

MITSUTA

that the total water vapor flux from the desert surface on a fine day is extremely small and that evaporation occurs at a level below the sand surface. References Mitsuta, Y.: 1974, ‘Sonic Anemometer-Thermometer for Atmospheric Turbulence Measurements’, in R. B. Dowel1 (ed.), Flow, Vol. II, Instrument Society of America, Pittsburgh, pp. 341-348. Mitsuta, Y .: 1988, ‘Sino-Japanese Cooperational Program on the Atmosphere-Land Surface Processes’, Tenki 35, 501-505. Panofsky, H. A. and Dutton, J. A.: 1983, Atmospheric Turbulence, John Wiley, New York. Sahashi, K., Tsukamoto, O., and Wang, J.: 1990, ‘Vertical Distribution of Humidity in the Sand’, HEIFE Report 5, 123-129. Tsukamoto, 0.: 1986, ‘Dynamic Response of the Fine Wire Psychrometer for Direct Measurement of Water Vapor Flux’, J. Atmos. Ocean Technology 3, 453-461. Wang, J. and Mitsuta, Y.: 1990, ‘Peculiar Downward Water Vapor Flux over Gobi Desert in the Daytime’, J. Meteoroid. Sot. Japan 68, 399-401.