Equilibrium response to a change in land surface albedo

As is well known, land makes up 30% of the Earth's surface, and only 50% of the land is covered with vegetation (meadows, forests, pastures and agricultural lands), the remainder comprising deserts (25-30%), continental ice sheets (11%), tundra (6-9%), and lakes, rivers and marshes (2-3%). Different types of land surface have different reflectance power (albedo) and react differently to incoming solar radiation. Albedo, in its turn, depends on the Sun's altitude, the type of vegetation, the nature and age of the snow-ice cover and the wavelength of the solar radiation.

Cloudiness has a pronounced effect on the albedo of the Earth's surface. Clouds not only absorb solar radiation but scatter it, causing an increase in the effective zenith angle of the Sun for small (<60°), and a decrease at large (> 60°), altitudes of the Sun. The influence of solar radiation wavelength has considerable impact on the value of the albedo of vegetative cover. This is connected with the fact that the absorption bands of chlorophyll are concentrated in the wavelength band from 0.4 to 0.7 pm. Within this band plants reflect three times less solar radiation than in the neighbouring IR band covering wavelengths from 0.7 to 4.0 pm. For snow and soil the ratio between albedo values in indicated bands amounts to on average 0.5 and 2.0 respectively.

The different nature of attenuation in two spectral intervals for almost equal incoming solar radiation at the upper atmospheric boundary determines the time-space variability of the net (integral over the spectrum) albedo of the underlying surface. Specifically, a decrease in the net albedo in winter compared with summer occurs as the result of weakening absorption and enhancement of scattering of solar radiation in the first interval, whereas a decrease in the net albedo in the tropics as compared with the midlatitudes in summer when the diurnal insolation is essentially independent of latitude occurs as the result of enhancement of absorption of solar radiation by water vapour in the IR band.

We also mention the trap effect (a decrease in albedo due to multiple reflection), which is created by undulations of the underlying surface and by elements of the vegetative cover. So, in changing vegetative height from 0.2 to 10 m, its albedo decreases from 0.25 to 0.1; for Scots fir woods this decrease is twice that for pine forests. It should be borne in mind that the albedo of the vegetative cover depends on the temperature of the environment controlling the intensity of evaporation from plants and, therefore, their colour. Perhaps it is these properties that explain the fact that the albedo of the vegetative cover at high latitudes is less than that at low latitudes.

Troubles in determination of ice and snow albedo are no less than in the case of vegetative cover. Indeed, according to Dickinson (1983) ice without air bubbles has the same albedo as sea water and, because of this, the ice albedo is controlled first of all by the size and distribution of air bubbles. Similarly, the snow albedo is determined by the size and distribution of ice crystals in air. The albedo of snow also depends on its age (the effect of'snow ageing'), non-homogeneity of the underlying surface, and on cloudiness when related to high latitudes.

The albedo of soil and sand surfaces varies within the limits from 0.1 for dark organic soils to 0.5 and more for white sands, and also depends on the size and colour of their constituent particles (when the latter increases, the albedo decreases). But the moisture of soils, their aggregate state, the degree of roughness, and, especially, the relative content of absorbing organics and mineral components completely overrule this effect. The dependence of albedo on the determining parameters is perhaps best known for the ocean surface, for which the net albedo for different values of the Sun's altitude and degree of roughness (waves) changes from 0.05 to 0.15. In addition, the influence of waves manifests itself more with decreasing altitude of the Sun.

At present, two methods for the determination of the surface albedo have gained wide acceptance: satellite and inventory methods. The first is based on the use of multichannel scanning radiometers recording the radiation field in different spectral ranges; the second is based on the thorough analysis of data from ground-based measurements of albedo with consideration given to the area of every type of underlying surface. The introduction of both methods has led to the construction of global maps of the surface albedo which have turned out to differ because of errors in the separation of the useful signal from atmospheric noise (cloudiness, water vapour and aerosols) in the first case, and subjectivism of estimates of the Sun's altitude, roughness of the underlying surface, soil humidity and density of vegetative cover in the second case. This results in a considerable scatter of estimates even for the zonal average values of the surface albedo obtained by different authors: maximum differences amounting to 0.3 occur in high latitudes and result from non-identical prescription of the sea ice area, melted water, leads and hummocks. But even in temperate and low latitudes the differences between available estimates of zonal average surface albedo can reach 0.06. It is clear that local peculiarities of spatial fields can differ even more.

We now direct our attention to a discussion of the results of numerical experiments on the equilibrium response of the climatic system to a change in the land surface albedo. In so doing we restrict ourselves to presentation of only the main facts. Discussion of this issue within the framework of three-dimensional models was initiated by Charney et al. (1977), who carried out three experiments. In the first experiment the albedo of the snowless land surface was assumed to be equal to 0.14; in the region of deserts in the Northern Hemisphere it was assumed to be 0.35. In the second experiment, values of the surface albedo inherent in deserts were extended to the region of the Sahel (the southern periphery of the Sahara desert), Rajputan (India) and the western part of Grand Plains (USA) which imitate the process of desertification. Finally, in the third experiment a similar change in albedo was realized in three other regions: Central Africa, the region of Bangladesh and in the Mississippi valley which imitated the process of destruction of vegetation. In all cases sea surface temperature and soil moisture were assumed to be constant.

The most remarkable result of this series of numerical experiments was the fact that, regardless of expectation, an increase in the surface albedo on average for the six regions examined led not to a decrease but to an increase in solar radiation absorbed by the underlying surface. This last circumstance is due to a decrease in cloudiness, which is caused by a decrease in the local evaporation and horizontal transport of water vapour from neighbouring regions. In turn, a decrease in the local evaporation is determined by attenuation of downward long-wave radiation and, accordingly, by a decrease in the value of radiation balance and temperature of the underlying surface; and a decrease in the horizontal transport of water vapour is caused by the formation of a circulation cell of the monsoon type, superimposed on the large-scale atmospheric circulation.

By the formation of a circulation cell of the monsoon type we mean the following. A decrease in the temperature of the underlying surface and the surface atmospheric layer results in a rise in atmospheric pressure and amplification of the downward vertical motions over the region of increased albedo if the background temperature in this region is less than that in the neighbouring region or, otherwise, in the attenuation of upward vertical motions over the region of increased albedo. Opposite changes in the intensity of vertical motions occur in the neighbouring region. This favours intensification of the process of cloud formation and condensation. Accordingly, the horizontal transport of water vapour into the area of increased albedo decreases.

Thus, both sources of water vapour for the atmosphere over the region of increased albedo - local evaporation and horizontal transport - decrease, that is, an increase in albedo favours the enhancement of aridity of the climate. This is the essence of the effect of self-amplification of deserts, discovered by Charney (1976). An increase in the recurrence of droughts in the Sahel region as a result of pasture spoiling serves as a manifestation of the above.

Further, since the horizontal temperature contrast determining the intensity and direction of circulation in a cell of the monsoon type depends on the location and extent of the area of increased albedo, then, naturally, the ratio between the change in local evaporation from the underlying surface and the change in the horizontal transport of water vapour in the atmosphere does not remain equal everywhere. So, according to Charney et al. (1977), in regions of the Sahel, Rajputan, and Central Africa a decrease in the horizontal transport of water vapour dominates over a decrease in local evaporation; in the western part of the Grand Plains the opposite is true; and in the Mississippi valley and in Bangladesh a decrease in local evaporation is even accompanied by an increase in the horizontal transport of water vapour.

Similar numerical experiments within the framework of a three-dimensional general circulation model were carried out by Carson (1982). In his work, as in the work of Charney et al. (1977), the sea surface temperature and soil moisture were fixed; in return, the albedo was changed from 0.1 to 0.3, not in restricted regions but, rather, over the whole surface of continents free from snow and ice. As a result it was established that an increase in albedo occurring everywhere leads to a decrease in evaporation, horizontal transport of water vapour and precipitation over continents, and to some increase in precipitation (due to the fixing of sea surface temperature and, to some extent, evaporation) over oceans.

It is clear that the assumption of constancy of soil moisture when the surface albedo is changed is not fulfilled in actuality. Thus the next step had to be its rejection. This step was taken by Potter et al. (1981) and Chervin (1979). In the first-mentioned work a zonal model was used as the basis; in the second, a three-dimensional general circulation model with a fixed sea surface temperature formed the basis. In both works the changes in the land surface albedo were prescribed as local: in the first work, the albedo increased from 0.16 to 0.35 for a region of area 9 x 106 km2 located in the vicinity of the 20° N parallel, and from 0.07 to 0.16 in two other regions (each of area 7 x 106 km2) in the vicinity of the equator and the 10 °S parallel; in the second work, the actual distribution of albedo in North Africa (from the Mediterranean Sea to the 7.5 °N parallel) and in the western part of the Grand Plains in the USA was replaced by a constant value equal to 0.45.

The data analysis presented by Potter et al. (1981) points to the fact that an increase in albedo in the vicinity of the 20 °N parallel determines a local decrease in absorbed solar radiation, temperature of the underlying surface, evaporation and precipitation. But the matter does not end there. It also causes an increase in the meridional temperature gradient, enhancement of evaporation in the northern Hadley cell and its displacement to the south. The latter leads to a fall in temperature and moisture content in the atmosphere, and, hence, to a decrease in the meridional sensible and latent heat transport to the pole, which in turn involves an increase in sea ice area and a further fall in temperature and moisture content in the atmosphere of the Northern Hemisphere. In the Southern Hemisphere changes in climatic characteristics become apparent by a shift of the southern Hadley cell and its associated decrease in cloudiness, increase in the absorbed solar radiation and a rise in temperature of the underlying surface and the atmosphere in the vicinity of the 30 °S parallel. As a result the temperature and moisture content of the Southern Hemisphere even increase a little. The global fall in surface air temperature amounts to 0.2 °C; the global decrease in moisture content of the atmosphere is equal to 0.04 g/kg.

Numerical experiments carried out by Chervin (1979) confirm Charney et £j/.'s (1977) conclusion that there is a change in the velocity of vertical motions and a decrease in temperature of the underlying surface and precipitation in the region of increased surface albedo. According to Chervin (1979), the velocity of upward vertical motions in North Africa decreases up to 2 mm/s; the temperature of the underlying surface decreases up to 2 °C; the precipitation decreases up to 5 mm/day, and the soil moisture content decreases up to 5 cm (for a soil moisture capacity equal to 15 cm). Simultaneously, in the southern part of the area examined (in the zone between the 12.5° and 7.5 °N parallels)

the temperature of the underlying surface does not fall but rises by about 0.5 °C. In other words, here, instead of the negative correlation between albedo and temperature of the underlying surface, a positive correlation takes place. The cause of such a change in sign is the suppression of the effect of evaporation due to a sharp decrease in soil moisture content.

But the changes in climatic characteristics discussed above are inherent only in local regions of increasing albedo. In neighbouring regions they have an opposite sign. As a result it turns out that, for the Earth as a whole, changes in the temperature of the underlying surface and of the surface atmospheric layer amount to only —0.25 °C. This estimate of temperature drop is close to the one obtained by Potter et al. (1981). But it should be remembered that the above-mentioned estimate corresponds to the local increase in land albedo, and the estimate of Potter et al. (1981) complies with a global increase in land albedo.

It is no mere chance that we focus our attention on the approximate character of the available estimates. By this we emphasize that these do not necessarily have to coincide with each other if for no other reason than the differences between the models and the steady states corresponding to them. But then would it be better, instead of comprehensive three-dimensional models, to use simple thermodynamic models which describe the meridional heat transport in the ocean explicitly? The fact that such a question is not unfounded is attested to by the results of a numerical experiment that were obtained within the framework of a 0.5-dimensional seasonal model of the ocean-atmosphere system (see Section 5.5), in increasing the albedo of the snowless land surface from 0.19 to 0.25.

Judging from the results of this numerical experiment an increase in albedo leads to a decrease in absorbed solar radiation at the underlying surface, as well as a decrease in temperature of the land surface, evaporation, precipitation and run-off. Accordingly the moisture content of the atmosphere and heat release due to phase transitions of water vapour are reduced. A decrease in the temperature of the land surface and heat sources in the atmosphere is accompanied by a drop in air temperature and, hence, by a weakening of downward long-wave radiation. It would have caused an increase in net long-wave radiative flux at the underlying surface but this flux decreases due to the increase in upward long-wave radiation flux.

A decrease in the net flux of long-wave radiation and in the absorbed solar radiation at the underlying surface do not counterbalance each other completely: the latter dominates. As a result the value of net radiation flux at the underlying surface decreases. Latent heat flux at the underlying surface changes in a similar way that determines a decrease in heat transfer from the ocean into the atmosphere in the area of cold deep water formation, and an increase in heat transfer from the atmosphere into the ocean in temperate regions and at low latitudes. The heat transport from these regions into the area of cold deep water formation also decreases. This is accompanied not by a rise but, rather, by a fall in temperature of the UML at temperate and low latitudes of the ocean. This last circumstance is the consequence of displacement towards the south of the boundary between the northern and the southern boxes, causing a reduction of ocean area in the southern box (see Section 5.5).

It should also be noted that a decrease in the meridional heat transport from temperate and low latitudes of the ocean results in a fall in temperature in the area of cold deep water formation, and this, in turn, results in weakening of the heat transport into the polar ocean, increase of sea ice area, and further enhancement of the effects of positive feedback between the albedo of the underlying surface and surface air temperature.

0 0

Post a comment