SCOPE 35 - Scales and Global Change

ABSTRACT

Quantification of atmospheric systems has established processes that occur on a wide range of time and space scales. Numerical models of these processes are used for various applications, such as simulation of mesoscale processes, global weather predictions, and projections of global climate change. The climate system has experienced large changes over geological times, forced by changes in atmospheric composition, location of continents, and earth's orbital parameters. Climate changes on time scales of seasons to centuries are a current focus. The contemporary research agenda includes the questions of seasonal climate prediction from an observed initial state and the projection of decadal climate changes. Treating the wide range of spatial scales of the climate system is difficult, both in the atmosphere and especially at the terrestrial surface. Current societal issues include the question of the gradual warming occurring from increasing concentrations of trace gases and CO₂ and questions of climate change from land-use changes, such as desertification and deforestation.

INTRODUCTION

We all have an intuitive sense of the temporal variability of the atmosphere through our personal experiences as residents in one or more locations. Thus, we see individual clouds float by, sometimes in 5 to 10 minutes or so, and cloud systems build up and dissipate for times of less than an hour to several days. We experience the absence of solar radiation at night and night-time declines in temperature whose magnitude increases with increasing distance from oceans, increasing dryness, and absence of clouds.

We are familiar with the various precipitation systems that produce rain or snow, thunderstorms lasting for tens of minutes to a few hours at most, and the extratropical cyclonic storms that can last up to several days. Some of us have experienced nature's severest weather-tornados lasting for a few minutes and hurricanes or severe downslope mountain-lee windstorms. On longer time scales, we notice the variations of temperature and precipitation with seasons. We also sometimes note excesses or deficits in rainfall and abnormally warm or cool temperatures over a month or a season. Such information on the time scales of atmospheric processes has been available to humankind since the dawn of history.

With the availability of instruments, we have been able to extend and quantify our description of atmospheric processes at individual points in the globe. In particular, at the surface we can now measure turbulent fluctuations of wind, temperature, moisture, and trace gases on time scales of seconds to minutes, and through correlation techniques infer the fluxes between surface and atmosphere of momentum, sensible and latent heat, and trace gases. We can measure how much solar energy is incident and absorbed, as well as the fluxes of thermal infrared energy, and so deduce the balance of energy for various terrestrial surfaces.

                                         
Figure 5.1 Regions of the atmosphere and vertical distribution of temperature (redrafted from Banks and Kockarts, 1973)

By probing the atmospheric column above us using radars, balloons, rockets, and satellites, we see the vertical layering of the atmosphere (Figure 5.1). Most of the humidity and clouds is found in the first several kilometres of the atmosphere. The cumulonimbus clouds of thunderstorms, frequently extending above 10 km, are exceptions. Cirrus ice-crystal clouds also can extend above 10km. Over the first 10 to 15 km or so in the troposphere, temperatures decrease with altitude, typically by about 5 to 7 °C km^-1 (Figure 5.2). None of the systems that produces our 'weather' reach very far above this region. Above and extending to about 50 km is the stratosphere, where temperatures increase with altitude. The stratosphere has significant radiative, dynamic, and photochemical connections to the lower atmosphere. Perhaps of greatest practical significance is its storage of most of the atmosphere's ozone, hence controlling the levels of ultraviolet radiation incident at the earth's surface. Higher layers of the atmosphere contain many interesting processes, but significant connections to the lower atmosphere have not been identified and are not discussed further here.

Figure 5.2 Longitudinal average atmospheric temperatures (Newell  et al., 1972) (Copyright 1972 MIT Press)

Indeed, rapid vertical turbulent fluxes of various properties are largely confined to the planetary boundary layer, typically about 1 km thick, but extending up several kilometres or more during summer continental conditions and shrinking to less than l00 m over land during night-time and polar conditions.

SPATIAL SCALES, OBSERVING SYSTEMS, AND MODELS

Most atmospheric systems have spatial scales that are larger than the extent of human vision. For many centuries, sailors and geographers have noted the average patterns of surface winds and temperatures over the globe, in particular, the trade easterlies and mid-latitude westerlies, and the much stronger seasonal variation of temperature over continents than over oceans (Figure 5.3). Establishing the spatial patterns and scales of the day-to-day weather superimposed on the mean patterns has required spatially distributed systems of instruments.

The initial weather observation networks developed in the 1850s linked by telegraph the surface measurements of temperature, pressure, and incidence of rain over continental areas. These networks and the maps drawn from their data described and projected the generally eastward movement of mid-latitude cyclonic storm systems and anticyclones with spatial scales of, typically, 1000 km. About 50 years ago, meteorologists began to link data from upward balloon ascents to describe the vertical extent and structure of these weather systems.

Procedures for analysing maps nearly immediately after the acquisition of these data were developed in the 1950s, along with the initial digital computers and numerical weather forecast models for projecting the measured fields up to several days forward in time. These systems of synoptic weather-data acquisition and analysis have since then been greatly improved, now utilizing other sources of data that allow global coverage, such as satellite-temperature soundings.

The numerical models have likewise been refined to be integrated on a global mesh and to provide useful weather forecasts up to more than a week in advance. The initial thrust of these models was the solution by numerical methods of the three-dimensional, time-dependent equations for atmospheric hydrodynamics. The determination of atmospheric temperature is also crucial in these solutions, since winds are driven by pressure gradients whose vertical variation depends on temperature.

Atmospheric temperatures over the time and space scales of cyclonic disturbances are largely adiabatic, that is, they change primarily because of air movement and pressure variation. On longer time scales, atmospheric 'diabatic' heating terms become crucial for establishing temperature. These terms include release of latent heat during condensation of rainfall, atmospheric radiational processes, and convective heating from the earth's surface. General Circulation Models (GCMs) of atmospheric processes were first developed in the 1960s. These models began with the equations of the weather prediction models, but included equations for atmospheric water vapour, its evaporation, precipitation, and movement within the atmosphere, equations for atmospheric radiation, and frictional and thermal coupling to the earth's surface. The intent of these models was to simulate the average statistics of atmospheric circulation systems.

In the 1970s, the atmospheric science community, concerned with the question of climate change resulting from increasing concentrations of atmospheric carbon dioxide (CO₂) and other trace gases, first developed simple energy-balance and radiative-convective equilibrium models to study this problem. Previous GCMs had used fixed sea-surface temperatures taken from observations, but by assuming that oceans acted as wet surfaces and were locally in energy balance, Manabe and Wetherald (1975) applied a version of their GCM to study climate change from increasing CO₂. A decade of further such studies (e.g. as reviewed most recently by Bolin et al., 1986) has clarified the difficulties confounding detailed precise projections of future climate change from increasing trace gases. These difficulties largely involve problems of matching global climate change to other scales and to interfaces with other disciplines.

Many of the severest weather systems occur on the mesoscale, ranging from a few kilometres to a few hundred kilometres in horizontal extent (Figure 5.4). At the smaller end of this range are the individual thunderstorm cells; at the larger end are frontal phenomena, including rainbands, and mesoscale convective complexes. The conventional radiosonde observing systems are too coarse to capture these features. Weather radars have been deployed for two decades to provide early warning of dangerous mesoscale systems. At present, considerable effort is devoted to developing and deploying a variety of new radars, automatic weather stations, and other new wind measurement techniques, as well as fine-resolution, limited-area numerical models for simulating mesoscale motions over a time scale of a day or less. Such research and operational developments are needed to better describe the mesoscale atmospheric processes (e.g. Anthes, 1983).

Figure 5.3 Geographical distribution of winter and summer average ((a) December-February and (b) June-August) global surface temperatures (Washington and Meehl, 1984) 

Figure 5.4 Mesoscale weather system illustrated by U.S. National Weather Service radar summary at 1435 GMT, 22 June 1981 (Anthes, 1983). The contoured pattern shows large reflectivities from a severe storm over the states of Nebraska and Missouri

Figure 5.5 Structure of a GCM climate model

Global forecast models and the observing systems that provide them with data are also being improved. With improvements in computer power, the spatial resolution of global forecast models is shrinking to about 100 km or less. Weather forecasts can now provide useful information up to at least 10 days into the future, and some centres, e.g. the European Centre for Medium Range Weather Forecasts, are actually disseminating forecasts that far ahead. With the extension of global forecast models to finer space scales and to larger time scales, inclusion of many of the physical processes previously treated in climate GCMs (Figure 5.5) is also becoming important for forecast models. Likewise, considerable emphasis is placed on upgrading the treatment of physical processes in mesoscale models. In sum, to a certain extent, mesoscale, forecast, and climate models are converging to similar descriptions of atmospheric physical processes, and their primary distinction becomes their spatial resolution and time span of integration.

GLOBAL-SCALE CLIMATE CHANGES

Consideration of climate change and variations in the time frame of human experience benefits from the perspective of longer time-scale variations revealed only in the geologic record.

Climate change on geological time scales

Geological processes range over time scales of thousands to billions of years. Several intriguing questions over this range of time scales have been emphasized in recent investigations. On the billion-year scale, the atmosphere, oceans, and continents themselves evolved. The primary observation for climate theorists on this time scale is that climate must not have been so drastically different from that of today as to preclude the development and evolution of life. On the other hand, at the time of our planet's formation, four and a half billion years ago, the output of energy from the sun was about 30% less than it is today (Newkirk, 1983). With the composition of today's atmosphere but with 30% less solar radiation, land and oceans would have become covered with ice and snow, and, the resulting high albedos would have maintained this condition even with current fluxes of solar radiation. Thus, CO₂ and possibly other gaseous absorbers must have been more abundant. At least 100 times the current concentrations of CO₂ would have been required, if it were the only factor preventing such an 'ice-covered earth' catastrophe. Such concentrations could have been supplied by interactions between geological and climatic processes (e.g. Walker et al., 1981).

On time scales of 100 million years, sea floors are created from spreading centres and destroyed by subduction, rafting continents together and apart, and modifying their climate by changing their latitude. The Cretaceous period, centred about l00 million year ago, was much warmer than today, especially in high latitudes (Figure 5.6). GCM climate modelling studies (e.g. Barron and Washington, 1984) indicate that this warmth could not have been maintained only by differences in the size, shape, and location of continents. Rather, there must have been some large change in the concentration of the atmosphere's radiatively active gases; most likely, CO₂ concentrations were 4 to 10 times greater than at present-as also suggested by evidence from geochemical cycling (Budyko and Efimova, 1981; Berner et al., 1983).

 Figure 5.6 Estimated excess of Cretaceous temperature over present (Barron and Washington, 1984)

Subsequent to the Cretaceous, the global climate cooled until it entered the sequence of ice ages experienced over the last one and a half million years, with large fluctuations in temperature and ice cover over 100-, 40-, and 20- thousand-year periods. These periodic fluctuations, apparently, are largely driven by variations of the earth's orbital parameters (e.g. imbrie et al., 1984) and accompanied by fluctuations in atmospheric CO₂ levels. For example, during the peak of the last ice age 20000 years ago, CO₂ levels were less than 60% of their current levels. These changes in CO₂ levels may have been controlled by variations in ocean-uptake rates driven by biological processes (Knox and McElroy, 1984). On continental areas in the tropics, far from the ice sheets, the cover of tropical forests was greatly diminished (e.g. Ab'Saber , 1982; Haffer, 1982). Whether this decline in forest cover resulted mostly from more arid conditions as indicated or represented, in part, by a response to the different CO₂ levels is not known. Evidence for tropical ocean temperatures during the last ice age is conflicting (e.g. Rind and Peteet, 1985).

At the end of the last ice age, 10 thousand years ago, the earth passed closest to the sun in August. At present, it does so in January. Consequently, the Northern Hemisphere absorbed about 7% more solar radiation in July than it does today and about that much less in January. Summer temperatures in the Northern Hemisphere were warmer then than now and monsoon circulations were more pronounced (Kutzbach and Guetter, 1984), as has been simulated with GCM integrations.

The contemporary climate research agenda 

Of greatest interest for human affairs are climate changes on time scales of seasons to centuries. The current World Climate Research Programme divides research over this range of time scales into three 'streams' as follows (WCRP, 1984):

The first stream aims at. .. weather anomalies on time scales of one to two months, ...observing the initial value of the ocean surface temperature and sea ice fields, and making progress in the ability to predict the... amount of stored soil water and the rate of evaporation, ...precipitation and extended clouds, and in the formulation of radiative transfer in the presence of such clouds.

The second stream aims at ...variations of the global climate over periods up to several years ...particularly evident in the tropical regions . The largest contribution to the variations of the global atmosphere which may be predictable on interannual time scales is. ..the influence of the oceans and especially, the tropical oceans in which large-scale circulation and temperature anomalies can be forced by remote atmospheric events and propagate along the Equator. ...Stream 2 is . ..modelling the coupled atmosphere-ocean system using a truncated version of oceanic dynamics restricted to the tropical part of the ...oceans.

The third stream aims at variations of climate over periods of several decades and assessing the response of climate to either natural or man-made influences.

 In other words, the first stream is concerned with extending the time span of current numerical weather prediction approaches out to several weeks or longer by improving GCMs used for forecasting to better incorporate fluctuations in the hydrological cycle, radiation, and surface fluxes. On a time scale of seasons to years, much of the natural climate variability, especially in the tropics, is coupled with variations in ocean-surface temperatures. The most dramatic example of such coupling has been the El Niño-Southern Oscillation climate fluctuations that occur every 3 to 5 years or so and last for about a year (Wright, 1985). These fluctuations encompass most of the tropical Pacific and extend into the Indian and Atlantic Oceans with teleconnections to extratropical latitudes. Thus, the second stream emphasizes coupling of atmosphere, sea ice, and ocean dynamic models.

On a time scale of decades or longer, the largest projected climate changes are those resulting from the addition of CO₂ and other trace gases to the atmosphere. CO₂ has increased from concentrations less than 280 ppm a few hundred years ago (Neftel et al., 1985; Oeschger, 1987) to present values of nearly 350 ppm. Of most concern besides CO₂ are the chlorofluorocarbons CFC-12 (CCl₂F₂) and CFC-11 (CCl₃F) (Figure 5.7) and methane (CH₄) (Figure 5.8), tropospheric ozone (O₃) and nitrous oxide (N₂O), listed in order of their probable relative importance for future climate change.

The climate effect of future increases of the CFCs alone may be about 70% of that of CO₂, assuming no serious attempt is made to further regulate their production (e.g. Oickinson and Cicerone, 1986). The warming from CO₂ and other trace gases will probably exceed that implied by a doubling of CO₂  within the next 50 years.

The question of steady-state climate change from large increases of atmospheric CO₂ has been considered with the most advanced GCMs suitable for such study, including realistic continents and a seasonal cycle of solar radiation (Manabe and Stouffer, 1980; Hansen et al., 1984; Washington and Meehl, 1984; Wetherald and Manabe, 1986). These model studies indicate a steady-state global warming of from 2.0 to 4.8°C global-average temperature for a doubling of CO₂. The range of possible outcomes depends heavily on the extent of positive feedbacks from cloud radiative effects, which are now modelled with extreme uncertainty. It also depends significantly on the modelling of changes in the seasonally varying sea-ice distribution, in the vertical distribution of atmospheric water vapour, and in vertical lapse rates.

          
Figure 5.7 Projected future of the CFCs, CFC-l l, and CFC-12. Upper, lower, and middle scenarios are shown for each gas as inferred from an economic analysis. (Dickinson and Cicerone, 1986). (Reprinted by permission from Nature, 319, No.6049, 109-115, copyright @ 1986 Macmillan Magazines Limited.)

Figure 5.8 Past history of CH₄. The circles refer to data from Greenland ice cores, the triangles to South Pole ice core data, the diamonds to current concentrations. (Rasmussen and Khalil, 1984.)

Global-average temperature change is constrained by global-energy balance and is thus simplest to estimate approximately. However, there is no reason to expect a uniform change, and not only large regional variations in temperature change but also geographical and seasonal shifts in rainfall patterns are likely. Figure 5.9 illustrates the latitudinal variation of temperature calculated from one GCM simulation of climate change from CO₂ doubling. All GCM simulations have indicated a considerably larger surface-temperature increase in high than in low latitudes, especially in winter, apparently a result of the highly stable temperature stratification and the ice-snow albedo feedbacks. Figure 5.10 shows relative changes in soil moisture from another such model that suggests a strong summertime mid-latitude, mid-continental increasing dryness. However, not all modelling groups obtain this result. All models for climate change from increasing CO₂ have used a very simplistic treatment of soil moisture, e.g. the Budyko bucket method as illustrated in Figure 5.11.

Figure 5.9 Zonal average annual average temperature change from CO₂ according to Washington and Meehl (1984). Top frame shows values for December-February; bottom shows those for June-August. ( P/P* = pressurel surface pressure)

Figure 5.10 Geophysical distribution of soil-moisture change (in cm) during June-August for two model simulations. The upper frame is for relatively coarse resolution and the lower frame for a finer resolution. (Manabe et al., 1981)

One of the difficulties in projecting future climate change from increasing concentrations of trace gases is that the oceans require decades to centuries to equilibrate with changes in global-energy balance. There has already been a considerable increase of atmospheric trace gases since 1900, but the thermal inertia of the oceans is thought now to be reducing the realized warming to about half of that implied by a steady-state assumption. Still, global temperatures should have increased from between 0.3 and 1.0 °C over those of 1900. Indeed, global temperatures are now the warmest ever recorded and are about 0.5°C warmer than in 1900. However, temperatures were nearly as warm in 1940 as at present (Figure 5.12) .Evidently, natural fluctuations in global temperature occur with up to 0.5°C range and obscure the signal from increasing trace gases.

Figure 5.11 Budyko bucket model for GCM soil moisture

Figure 5.12 Record of smoothed global-average surface temperature (heavy line) relative to 1950-1979 mean value (light line) (Jones et al., 1986). (Reprinted by permission from Nature, 322, 430-434, Copyright @ 1986 Macmillan Magazines Limited.)

OUTSTANDING SCALE-DEPENDENT PROBLEMS

The atmospheric sciences now have an excellent conceptual and practical frame-work for considering the global atmosphere over time scales of days to years. We are learning to include in GCMs details of the diurnal cycle and to carry out simulations for up to several decades of time. Mesoscale models simulate the smaller-scale atmospheric processes out to a day or two. Some of the most serious current difficulties relate to time- and space-scale mismatches to this framework, involving either processes internal to the atmosphere or at the interface with other disciplines.

Scale problems internal to the atmosphere

The severest difficulties for modelling the atmosphere come from the need to parameterize many subgrid-scale atmospheric processes, that is, the net effect of processes occurring on scales of e.g., l0 m to l00 km, must be included in global models that resolve typical minimum horizontal scales of at least several hundred kilometres. Especially significant are the descriptions of the radiative properties of clouds and of the vertical transfer of heat, moisture, and momentum by moist and dry convection.

Cloud-property changes may be important for climate change on time scales of years or longer, but even the sign of the net effect on radiation is still difficult to establish. Subgrid-scale vertical fluxes of water vapour, sensible heat, and momentum are generally described by simple one-dimensional, conceptual models whose relationship to real three-dimensional processes is poorly known. Thus, the sign of these fluxes is generally expected to be correct, but their magnitudes for given atmospheric profiles may be far from correct. Considerable compensation and negative feedbacks may make the net effect of errors in subgrid-scale vertical flux descriptions less serious than they might otherwise be. Errors in the treatment of horizontal momentum fluxes from grid scale to subgrid scale may also degrade atmospheric simulations.

The above discussion has taken the perspective of a global-scale modeller frustrated by the difficulty in treating linkages to smaller scales. However , severer problems may be encountered by modellers on smaller scales attempting to link their models to larger-scale, up to global, processes. An example is Idso's difficulties in extrapolating a good understanding of local micro- meteorological processes to global climate problems (Idso, 1980; CO₂/Climate Review Panel, 1982). Most small-scale investigators of atmospheric problems are more timid and prefer to assume that larger scales are fixed. Since the coupling of the small-scale processes to larger scales may be their most significant aspect, the practical output of such research tends to be primarily support for improving parameterizations of larger-scale models.

 Scale problems at the terrestrial interface

 From the viewpoint of the atmosphere, the terrestrial surface has a simple role-its main functions are to determine how much solar radiation is absorbed versus reflected and how absorbed energy is partitioned into latent and sensible heat versus storage. Longwave radiation is also absorbed and emitted but varies less and is less sensitive to details of surface processes. However, when the surface processes actually involved in energy transfers are examined, they are found to be very complex with considerable small-scale variability in space and time.

 For example, a tree is not a simple flat surface but is a collection of a large number of surfaces, often idealized as plates or cylinders. These surfaces absorb solar radiation, depending on their orientations, shading from over- lying elements, and transfer of radiation through canopy elements (e.g. Dickinson, 1983). Such elements individually reach equilibrium temperatures that balance their solar heating with sensible and latent heat losses. These losses, in turn, depend not only on element temperature but also on the rate of air movements past the element. If the element is a photosynthetic surface, it loses water by transpiration through its stomates. Stomatal resistance depends on the plant species, but varies strongly with environmental variables, e.g. incident solar radiation, water-vapour pressure deficit, temperature, and leaf-water potential, which depends on the availability of water to the roots. Incident solar radiation, CO₂ mixing ratio, wind flow, and stomatal resistance, in turn, determine gross carbon uptake by the canopy.

Thus, one major modelling issue is the development of simple but effective and reliable parameterizations for the detailed energy-transfer processes within vegetation canopies that yield satisfactory descriptions of the exchange of heat and moisture with the atmosphere. A second question is how to include root resistance, that is, the limits to the rate at which water can be transferred from soil to leaves depending on soil properties and moisture (Dickinson, 1984; Bolin, 1988).

Soil properties can be included in one-dimensional soil models with reasonable accuracy. However, these properties vary widely on almost all horizontal scales and within the vertical column as well. To model the properties of soil, we must specify statistical properties of soils on the model grid and obtain the required global data. Holes by termites and worms can be important in many areas, and removal of vegetation may lead to compaction and pore closure, both by the effects of raindrops and of travel by animals and equipment.

Another issue of concern is the modelling of surface runoff, which depends not only on the vertical soil-column physics but also on steepness and extent of slopes, on vegetation including debris cover, and on the intensity of the rainfall. Rainfall intensities, especially in convective situations, can be much more intense locally than they are averaged over a model grid square. Thus, how to include the subgrid-square distributions of soil properties, slopes, and rainfall intensities is a key issue for representing land processes in global models.

 Figure 5.13 Scales of the atmospheric processes linked to surface processes (Dickinson, 1987) 

Trace-gas emissions, e.g. those of methane and nitrous oxide, presumably also have considerable small-scale variability as a consequence of the mosaic structure of land surfaces. Figure 5.13 illustrates the range of time and space scales involved in connecting the surface to the atmosphere.

The smallest-scale processes above the molecular level that determine the fluxes of moisture and gases to the atmosphere occur on the spatial scales of cells that are of the order of tens of microns; stomatal closure controls the fluxes through leaf surfaces, and microdecomposers, in metabolizing carbon, release various gases. Trees develop over spatial scales of tens of metres but have lifetimes of up to hundreds of years. In studying climate change and variations, we are concerned with spatial scales of at least several hundred kilometres and with time scales of seasons out to a century. Land surfaces are linked to this climate change through micro meteorological processes, which have individual eddies with 10 to l00 m spatial scales. One possible approach to linking large-scale model calculations to local ecosystems is that suggested by Gates (1985).

SOME KEY SOCIETAL ISSUES FOR ATMOSPHERIC MODELLING REQUIRING ATTENTION TO SCALES AND INTERDISCIPLINARY INTERFACES

We have already discussed the development of atmospheric models from mesoscale to global and the difficulties in accounting for the wide range of scales of phenomena internal to the atmosphere and at the terrestrial surface. Some current societal issues are also of considerable scientific interest in forcing us to come to grips with the wide range of scales.

Climate warming from trace gas increases

Past and future increases in atmospheric trace gases may drive the climate system within a century to a warmer state than it has realized in the last ten million years (Dickinson and Cicerone, 1986). It is difficult to imagine the implementation of practical controls on these increases, since the global distribution of the trace gases will translate sources in one location to climate impacts in another location. In particular, the industrialized countries control most of the emissions, whereas the less-developed countries-who have many more immediate concerns-are likely to suffer the greatest costs. If societal adjustments are the primary recourse to this event (Kellogg and Schware, 1981), such adjustment should be less painful if we anticipate decades in advance of the likely climate changes. Approaches to ameliorating the trace-gas climate change should be developed in the context of all the other environmental stresses from modern industrialization (Clark, 1986).

As already discussed, projections of future climate change for a given trace-gas scenario are limited by questions involving coupling between scales. How will cloud radiative properties change in step with such climate change? What shifts in regional climate patterns including rainfall will accompany the overall global change? What will be the contribution of decreases in sea ice and snow cover to amplifying the warming? How will shifts in surface-energy balance couple with the dynamics of soil-moisture and plant-water usage to modify the growth patterns of natural vegetation and cultivated crops? Attempts have been made to treat all these questions, but the answers can yet be viewed with little confidence.

Desertification and droughts

Periods of rainfall deficit are one of the known consequences of natural climate variability. Furthermore, for given rainfall, vegetation and soils can be degraded such that runoff is amplified and moisture storage reduced. What is less obvious is whether changes of land cover can shift mean rainfall or change its seasonal and stochastic variability. We know that changing an isolated small-scale patch of land can have little effect. The water for rainfall over this patch comes from elsewhere, and such a change cannot significantly modify the patterns of atmospheric vertical motion. On the other hand, changing all continental land surfaces from wet to dry condition leads to large reductions in rainfall over land areas and widespread shifts in climatic patterns (Shukla and Mintz, 1982; Mintz, 1984). Up to half of the rainfall over continental areas is derived from evapotranspiration of land surfaces, and widespread changes in surface-energy processes necessarily affect the dynamics of the overlying atmosphere. On a smaller scale, irrigation over 100km scale areas apparently enhances warm-season precipitation in the presence of atmospheric uplift (Barnston and Schickedanz, 1984).

However, we cannot describe with any confidence the role of such actual changes of the land surface on rainfall patterns and amounts. The Sahel region of Africa is a dramatic example of a large area where both vegetation degradation and a major drought have evolved over the last several decades (e.g. Nicholson, 1983; Lamb, 1983; Dennett et al., 1985). The connections between the vegetation degradation and the drought span a wide range of space and time scales. Equally convincing arguments can be made that the drought is entirely natural or that it is largely man-made. Large and widespread albedo increases can promote drought (Charney et al., 1977), but Gornitz and NASA (1985) find that albedo over West Africa has increased by, at most, 0.5%. We must greatly improve our descriptions of the land-surface changes and their links to the atmosphere in parallel with improving our ability to model atmospheric response to such changes. Only by doing so can we hope to quantify the relative contributions to such droughts of natural fluctuation and environmental degradation.

Deforestation

Tropical forests are being degraded and destroyed with alarming rapidity and serious ecological damage. However, for reasons described above, at present we can just begin to evaluate what effects deforestation might have on local, regional, and global climate (e.g. Henderson-Sellers, 1987).

Sioli (1984) has speculated that, when too much forest is destroyed, water cycles will be interrupted by lack of sufficient water vapour, total annual rainfall will decrease, and the length of dry season will increase, in turn affecting the structure of the forest. He suggests that 'once a critical point has been reached an irreversible chain reaction will start. ' What are the most sensitive aspects of tropical forest ecosystems to climate change? Mori and Prance (1987) suggest that pollination and seed-dispersal processes may be disrupted. Fires may also cause considerable damage. How will the survival of seedlings be modified?

CONCLUSIONS

Atmospheric systems span a wide range of time and space scales. Changes in climate are studied using models that couple hydrodynamic and physical processes in the atmosphere and link these processes to the land and ocean systems. Improving descriptions of the links to land will require determining what factors best describe the fluxes of water and energy between land surfaces and the atmosphere. Establishing such factors will likely involve coping with the wide range of spatial scales of land processes and especially those that are small compared to the grid resolution of the atmospheric models. Treating the two-way interactions between climate change and ecosystems will also demand better descriptions of the influence of the atmospheric systems on the ecosystems.

 REFERENCES

Ab'Saber, A. N. (1982). The paleoclimate and paleoecology of Brazilian Amazonia. In Prance, G. T. (Ed.) Biological Diversification in the Tropics. Columbia University Press, New York 741.

Anthes, R. A. (Ed.) (1983). The National STORM Program-Scientific and Technological Bases and Major Objectives. UCAR Report, University Corporation for Atmospheric Research, Boulder, CO 520.

Banks, P. M., and Kockarts, G. (1973). Aeronomy Part A. Academic Press, New York 430.

Barnston, A. G., and Schickedanz, P. T. (1984). The effect of irrigation on warm season precipitation in the southern Great Plains. J. Climate Appl. Meteor., 23, 865-888.

Barron, E. J., and Washington, W. M. (1984). The role of geographic variables in explaining paleoclimates: Results from Cretaceous climate model sensitivity studies. J. Geophys. Res., 89, 1267-1279.

Berner, R. A., Lasaga, A. C., and Garrels, R. M. (1983). The carbonate-silicate geochemical cycle and its effect on atmospheric carbon dioxide over the last million years. Amer. J. Sci., 283, 641-683.

Bolin, B. (1988). Linking terrestrial ecosystem process models to climate models. (Chapter 7, this volume).

Bolin, B., Döös, B. R., Jäger, J., and Warrick, R. A. (Eds.) (1986). The Greenhouse Effect, Climatic Changes, and Ecosystems. A Synthesis of Present Knowledge. John Wiley and Sons, Chichester, London 541 pages.

Budyko, M. I., and Efimova, M. A. (1981). The influence of carbon dioxide on climate. Meteorol. Hydrol. 2, 15-17.

Charney, J., Quirk, W. J., Chow, S.-H., and Kornfield, J. (1977). A comparative study of the effects of albedo change on drought in semiarid regions. J. Atmos. Sci., 34, 1366-1385.

Clark, W. C. (1986). Sustainable development of the biosphere: Themes for a research program. In Clark, W. C., and Munn, R. E. (Eds.) Sustainable Development of the Biosphere, pp. 5-48. International Institute for Applied Systems Analysis, Laxen burg, Austria.

CO₂/Climate Review Panel (1982). Carbon Dioxide and Climate: A Second Assessment. National Research Council, National Academy Press, Washington, DC: 72 pages.

Dennett, M. D., Elston, J., and Rodgers, J. A. (1985). A reappraisal of rainfall trends in the Sahel. J. Climate., 5, 353-361.

Dickinson, R. E. (1983). Land-surface processes and climate-surface albedos and energy balance. Advances in Geophysics., 25., 305-353.

Dickinson, R. E. (1984). Modelling evapotranspiration for three-dimensional global climate models. In Climate Processes and Climate Sensitivity, Maurice Ewing Series 5, pp. 58- 72. American Geophysical Union, Washington, DC.

Dickinson, R. E. (Ed.) (1987). The Geophysiology of Amazonia. Vegetation and Climate Interactions. John Wiley & Sons, Inc., New York: 526 pages.

Dickinson, R. E. and Cicerone, R. J. (1986). Future global warming from atmospheric trace gases. Nature, 319, No. 6049, 109-115.

Gates, W. L. (1985). The use of general circulation models in the analysis of the ecosystem impacts of climatic change. Cimatic Change, 7, 267-284.

Gornitz, V ., and NASA (1985). A survey of anthropogenic vegetation changes in West Africa during the last century-Climatic implications. Climatic Change, 7, 285-325.

Haffer, J. (1982). General aspects of the refuge theory. In Prance, G. T. (Ed.) Biological Diversification in the Tropics. Columbia University Press, New York: 714 pages.

Hansen, J., Lacis, A., Rind, D., Russell, G., Stone, P., Fung, I., Ruedy, R., and Lerner, J. (1984). Climate sensitivity: Analysis of feedback mechanisms. In Hansen, J. E., and Takahashi, T. (Eds.) Climate Processes and Climate Sensitivity, Maurice Ewing Series 5, American Geophysical Union, Washington, DC: 368 pages.

Henderson-Sellers, A. (1987). Effects of change in land use on climate in the humid tropics. In Dickinson, R. E. (Ed.). The Geophysiology of Amazonia. Vegetation and Climate Interactions. pp. 463-496. John Wiley & Sons, Inc., New York.

Idso, S. B. (1980). The climatological significance of a doubling of the earth's atmospheric carbon dioxide concentration. Science, 207, 1462-1463.

Imbrie, J., Hays, J. D., Martinson, D. G., McIntyre, A., Mix, A. C., Morley, J. J., Pisias, N. G., Prell, W. L., and Shackleton, N. J. (1984). The orbital theory of pleistocene climate: Support from a revised chronology of the marine d ¹⁸O record. In Berger, A. L., Imbrie, J., Hays, J. Kukla, G., and Saltzman, B. (Eds.) Milankovitch and Climate, Part I. pp. 269-305.

Jones, P. D., Wigley, T. M. L., and Wright, P. B. (1986). Global temperature variations between 1861 and 1984. Nature, 322, 430-434.

Kellogg, W. W ., and Schware, R. (1981). Climate Change and Society. Consequences of Increasing Atmospheric Carbon Dioxide. Westview Press, Boulder, CO: 178 pages.

Knox, F., and McElroy, M. B. (1984). Changes in atmospheric CO₂: Influence of the marine biota at high latitudes. J. Geophys. Res., 89, 4629-4637.

Kutzbach, J. E., and Guetter, P. J. (1984). The sensitivity of monsoon climates to orbital parameter changes for 9,000 years BP: Experiments with the NCAR general circulation model. In Berger, A. L., Imbrie, J., Hays, J., Kukla, G., and Saltzman, B. (Eds.) Milankovitch and Climate, Part 2, 801-820.

 Lamb, P. J. (1983). Sub-Saharan rainfall update for 1982: Continued drought. J. Climat., 3, 419-422.

 Manabe, S., and Stouffer, R. J. (1980). Sensitivity of a global climate model to an increase of CO₂ concentration in the atmosphere. J. Geophys. Res., 85, 5529-5554.

 Manabe, S., and Wetherald, R. T. (1975). The effects of doubling the CO₂ concentration on the climate of a general circulation model. J. Atmos. Sci., 32, 3-15.

 Manabe, S., Wetherald, R. T., and Stouffer, R. J. (1981). Summer dryness due to an increase of atmospheric CO₂ concentration. Climatic Change, 3, 347-386.

 Mintz, Y. (1984). The sensitivity of numerically simulated climates to land-surface boundary conditions. In Houghton, J. T. (Ed.). The Global Climate, pp. 79-105. Cambridge University Press, Cambridge.

 Mori, S. A., and Prance, G. T. (1987). Species diversity, phenology, plant-animal interactions, and their correlation with climate as illustrated by the Brazil nut family (lecythidaceae). In Dickinson, R. E. (Ed.) The Geophysiology of Amazonia. Vegetation and Climate Interactions, pp. 69-89. John Wiley & Sons, Inc., New York.

 Neftel, A., Moor, E., Oeschger, H., and Stauffer, B. (1985). Evidence from polar ice cores for the increase of atmospheric CO₂ in the last two centuries Nature, 315,  45-57.

 Newell, R. E., Kidson, J. W., Vincent, D. G., and Boer, G. J. (1972). The General Circulation of the Tropical Atmosphere, Vol. I. MIT Press, Cambridge, MA: 258 pages.

 Newkirk, G., Jr. (1983). Variations in solar luminosity. Ann. Rev. Astron. Astrophys., 21, 429-467.

Nicholson, S. E. (1983). Sub-Saharan rainfall in the years 1976-80: Evidence of continued drought. Mon. Wea. Rev., 111, 1646-1654.

Oeschger, H. (1988). The ocean system-ocean/climate and ocean/CO₂ interaction. (Chapter 15, this volume.)

Rasmussen, R. A., and Khalil, M. A. K. (1984). Atmospheric methane in the recent and ancient atmospheres: Concentrations, trends, and interhemispheric gradient.  J. Geophys. Res., 89, 11,599-11,605.

Rind, D. and Peteet, D. (1985). Terrestrial conditions at the last glacial maximum and CLIMAP sea-surface temperature estimates: Are they consistent? Quart. Res., 24, 1-22.

Shukla, J ., and Mintz, Y. (1982). Influence of land-surface evapo-transpiration on the earth's climate. Science, 215, 1498-1501.

Sioli, H. (1984). Former and recent utilizations of Amazonia and their impact on the environment. In Sioli, H. (Ed.) The Amazon: Limnology and Landscape Ecology of a Mighty Tropical River and its Basin. W. Junk, Dordrecht: 763 pages.

Walker, J. C. J., Hays, P. B., and Kasting, J. F. (1981). A negative feedback mechanism for the long term stabilization of the earth's surace temperature. J. Geophys. Res., 86, 9776-9782.

Washington, W. M., and Meehl, G. A. (1984). Seasonal cycle experiment on the climate sensitivity due to a doubling of CO₂ with an atmospheric general circulation model coupled to a simple mixed layer ocean model. J. Geophys. Res., 89, 9475-9503.

Wetherald, R. T., and Manabe, S. (1986). An investigation of cloud cover change in response to thermal forcing. Climatic Change, 8, 5-23.

Wright, P. B. (1985). The Southern Oscillation: An ocean-atmosphere feedback system? Bull. Amer. Meteor. Soc., 66, 398-412.

WCRP (1984). Scientific Plan for the World Climate Research Programme. WMOICSU Joint Scientific Committee, WCR Publication No.2: 95 pages.