c

Agriculture and forestry-related issues

The work carried out under this task is discussed under the headings of:

Agriculture and forestry Natural terrestrial ecosystems Hydrology and water resources IV Agriculture and forestry

Senior author: C Rosenzweig (USA)

Authors D Maclver (Canada) P Hall (Canada) M Parry (UK) O Sirotenko (Russia) J Burgos (Argentina) IV Agriculture and forestry

I Agriculture and Pittock 1991). These changes may bring a nearly global increase in flood and high runoff events, leading to Progress since 1990 soil degradation and flood control problems, as well as reductions in the return period of droughts. Observations Since the publication of the first IPCC Impacts Assess• of hydrological variables, such as the recently published ment, progress in the study of agricultural impacts of soil moisture dataset by Vinnikov and Yeserkepova (1991), climate change has continued on many fronts. In this should be useful in testing understanding of current and update, we have adopted the structure of the original IPCC projected hydrological relationships. Still critically needed report on agriculture and forestry and for each section of is the development of improved regional scenarios for the original report,we identify first, what, if any, new monsoonal areas (Yoshino 1991). work has been done and second, whether this new work Another key point which is emerging is that the leads to any different conclusions from the 1990 report. emphasis of previous climate change studies on the New sections are added where necessary. arbitrary doubling of atmospheric CO2 levels (or their equivalent) may be limiting estimation of impacts (Cline 1992). Atmospheric CO^ concentration is projected to 1 Climatic scenarios increase beyond doubling to 1600 ppm by the year 2250 (Sundquist 1990) and GCM simulations have shown large While assumptions about climatic scenarios have not warming effects of COj at such levels (Manabe and Bryan changed for the purposes of the ongoing IPCC assessment, 1985). Greater warming beyond the equivalent doubled it is important to note that improvements in methods for CO2 level has implications for projections of agricultural generating climate change scenarios are under develop• impacts, because higher temperatures may lead to earlier ment. One such method is the nesting of meso-scale maturity and lower yields of determinate crops in meteorological models within general circulation models temperate regions, higher levels of potential evapo- (GCMs) to improve regional climate change prediction transpiration and water stress, and possible high tempera• (Giorgi and Meams 1991). Another method is the use of ture damage to yield quality (Wang and Hennessey 1991). stochastic weather generation models to generate synthetic While the climatic effects of CO2 may negatively affect daily time series with changes in variance, as well as agricultural production at concentrations higher than means (Wilks 1991; Meams et al. 1992). This is helpful doubled, the beneficial direct effects of COj on crop because most impact studies so far have been limited to the growth may level off at higher COj concentrations, as seen in experimental work on leaf photosynthetic rates (Akita mvestigation of the effects of mean changes in climate and Moss 1973), leading to potentially greater negative variables. impacts later. Some studies have begun to investigate the effect of potential changes in climate variance on the variability of crop yields, a key factor of agricultural stability. The question of agricultural sensitivity to climatic variability 2 Assumptions concerning technology and manage• has been researched by superimposing past variations on ment altered average conditions to assess impacts for crop yields, farm returns and risks, and regional production in Since the publication of the first IPCC Impacts Assess• Ontario, Canada (Brklacich and Smit 1992), and the ment, there has been a growing emphasis on the need for American midwest (Easterling et al. 1991). The effects of 'moving' pictures of how agriculture may respond to a range of temperature and precipitation variance changes, transient climate changes over time. Research has begun to as well as future changes in variance predicted by one grapple with how such transient projections of agricultural GCM, on simulated wheat yields in Kansas have been technology should be estimated (Rosenberg and Crosson analysed by Meams et al. (1992). Further investigation of 1991). For example, conventional plant breeding and climate and crop yield variabilities should be encouraged. biotechnology are each expected to improve yields on the Improved prediction of potential future hydrological order of 1-2%/year (Pimentel et al. 1992) and such esti• regimes (including estimates of changes in soil moisture mates should be factored into projections of future impacts and groundwater availability) continues to be a critical of climate change. Other probable future changes in need for agricultural impact studies. Recent work on agriculture include higher fossil fuel prices and greater use climate change scenarios suggests that the frequency of of crops for fuel and paper fibre. heavy rainfall events at almost all latitudes may increase The vulnerability of crops to climate change may be with global warming, while the numbers of raindays in either increased (by design) or diminished (inadvertently) mid latitudes may decrease (Pittock et al. 1991; Whetton by fiiture technological changes. If technological advances 46 CIV Agriculture and forestry narrow the optimal range of input conditions for high national debts may increase vulnerability to climate agricultural production (eg need for high levels of fertilis• change. er), and if climate change results in increased variability such as increases in frequency of droughts as well, produc• tion risks may also be expected to increase. 4 Direct effects of elevated COj and other greenhouse gases

3 Potential impacts on agriculture and land use 4.1 Combined effects of climatic and direct effects of CO, 3.1 Critical types of climatic change A fundamental question that remains despite continued In the area of direct climatic effects on crop growth, recent research since the publication of the initial IPCC assess• work shows that continued concern for future agricultural ment is the overall relative importance of direct and production appears warranted, a confirmation of the IPCC indirect effects of COj on future crop production. On an conclusions. Drought frequency may indeed increase in individual cultivar and crop basis these effects have been low and mid latitudes owing to increases in the demand of shown to vary considerably, depending on plant physiology the atmosphere for water (potential evapotranspiration) and the coincident levels of COj, temperature and water over the supply (precipitation) (Rind et al. 1990) and to availability. Some scientists weight the beneficial effects of higher frequency of occurrence of dry spells (Pittock et al. CO2 as the dominant factor, while others are sceptical that 1991), although direction of change in climate variability the large benefits seen in experimental settings will be seen of temperature and precipitation overall is still uncertain in farmers' fields under changing climate conditions. Free (Meams et al. 1992). air release of CO2 experiments on cotton are continuing in Other critical types of newly identified climate effects Phoenix, Arizona (Hendrey et al. 1992), and other free air are slight amounts of warming in the early stages of the release experiments are under way elsewhere (Miglietta projected climate change or increases in variability. These 1991). types of climate change may increase frost damage to While the separate effects of atmospheric CO2 concen• crops (demonstrated by simulating wheat growth in tration and temperature on specific plant processes have Kansas), as slight warming or augmented variability been studied extensively, the interactive effects of tempera• actually lessens winter hardiness while not eliminating hard ture and atmospheric COj concentration over the entire frosts altogether in some regions (Meams et al. 1992). growing season continue to be less well understood. Combined studies of direct and indirect effects of CO2 on crops have intensified to provide clarification. A recently 3.2 Regions of risk published review of these experiments showed that as temperature increased (levels not given), relative increases Conferences since the advent of the IPCC have echoed the in plant growth and photosynthesis were greater under CO2 concept of regions at risk, based on preliininary projections enrichment (Allen et al. 1990). The extent of the tempera- of adverse climate change and present-day resource ability ture-COj interaction differed considerably among species: to support existing populations. More detailed descriptions C4 species were less responsive (as they have been shown of vulnerable regions and methodologies for integrated to be to CO2 increases alone) than C3 crops, and sorghum regional assessment of the potential effects of climate showed no response. The relative decrease in stomatal con• change have been brought forward, with emphasis on ductance to water vapour due to CO2 was greater (as much sensitive regions where farmers lack financial resources, as 25%) as temperaUires rose. fertile land, irrigation, drought-tolerant seeds, alternative There is a further need for studies of temperature and employment options, or reasonable prices for inputs and COj interactions on crops and cultivars from diverse products (Houston Advanced Research Center 1991). Jodha agricultural systems. These studies should be designed to (1989) has shown that traditional farming systems encom• test a wide range of levels of both factors and to quantify pass a wide range of adjustments to loss and advocates the a full complement of crop physiological and phenological preparation of an inventory of current fanning practices responses. Preliminary work suggests that potential crop available for transfer to regions with changing climates. responses to combined higher CO2 and higher temperature However, Jodha also cautions that unequal social structures may counter, at least in part, negative effects of climate and international positions, as well as unequal land tenure, change through gains in water use efficiency. From a high numbers of landless rural dwellers, low incomes and regional perspective, however, water consumption per unit of land area may not change significantly due to plants CIV Agriculture and forestry 47 with larger biomass (Eamus 1991). Furthermore, given the crop yields caused by higher CO2 levels (Bazzaz and Fajer trajectories for projected increases in greenhouse gases 1992). (GHGs) other than CO2, greater warming may occur with lower amounts of CO2 and thus lower levels of COj fertilisation (Clme 1992). 5 Multiple stresses Because experimental settings are limited and costly, crop growth models are required tools for simulating the It is important to emphasise that the future is projected to combined responses of crops to elevated COj and climate bring simultaneous multiple stresses, such as elevated UV- change, especially at scales beyond one or several plants. B radiation and augmented tropospheric ozone, in combi• Some of the efforts under way are described in Hesketh nation with high COj, and that these may negate the and Aim (1992) and Reynolds et al. (1992). Simulation beneficial physiological effects in some crops. Higher models for wheat growth have demonstrated that higher levels of CO2 may partially offset damage from tropo• temperatures cause earlier maturation, which can cancel spheric ozone, sulfiir dioxide and other pollutants, through out any gain in yield from higher CO2, and that only stomatal closure (Allen 1990). Experiments reported by cultivars adapted to high temperatures may do well in a Teramura et al. (1990) show that no increase in wheat or warmer and COj-enriched environment (Wang and rice grain yield, occurred under conditions ,of combined Hennessy 1991). These authors also cite experimental work elevated UV-B and COj. Soybeans, on the other hand, showing that wheat quality can be significantly reduced by produced more yield under the same conditions. A review spells of high temperature above 35 °C. of experimental literature has shown that susceptibility to plant pathogens may increase with higher UV-B radiation (Pimentel et al. 1992). 4.2 Competition

A review of the implications of increasing CO2 and climate 6 Effects of climate change on soil properties change on competition in agricultural ecosystems has confirmed the original IPCC conclusion that differential New work has improved knowledge of the soil as a source growth responses to both CO, concentration and climate of GHGs and of the potential impacts of climate change on change will affect future competitive ability and fitness of soils (Buol et al. 1990; Bouwman 1990; Scharpenseel et al. plants (Patterson and Flint 1990). The relative importance 1990; US National Science Foundation 1992). Global of various weed species in agro-ecosystems may change, climate change will affect soils primarily through changes but selection of adapted crop varieties and management in soil moisture, soil temperature and soil organic matter. methods may minimise negative impacts. Weedy species Higher air temperatures will cause greater potential with broad ecological amplitudes are likely to prosper at evapotranspiration and higher soil temperatures, which the expense of endemic species or those already in margin• should increase solution chemical reaction rates and al habitats. In the tropics, important C4 crops such as diffiision-controlled reactions. Solubilities of soil and maize and sugarcane, which are adapted to hot, dry gaseous components may either increase or decrease, but conditions, may experience yield reductions because of the the consequences of these changes may take many years to improved performance of C3 weeds (Bazzaz and Fajer become significant. 1992). A study.of potential land degradation in New South Wales, Australia, predicted significant increases in soil erosion, sedimentation and salinity if rainfall intensities 4.3 Pest interactions increase (Aveyard 1988), while a study of soil erosion potential in the US using the Universal Soil Loss Equation While experiments have tested some isolated aspects of found that national average sheet and rill erosion changes pest infestations at elevated COj concentrations, it is still are projected to range from about -5% to 16% (Phillips et hard to project future pest-COj interactions. Some experi• al. 1991). ments have shown that some insect populations may Higher temperatures will accelerate the decay of soil decline owing to lower nutritional quality of leaves; this organic matter, resulting in release of COj to the atmo• could lead to reconfiguration of many ecological interac• sphere and decrease in carbon/nitrogen ratios, although tions. Other experimental work has shown that insects these two effects should be offset somewhat by the greater consume more leaves in order to compensate for lower root biomass and crop residues resulting from plant nutritional quality; this could work to negate benefits in responses to higher COj. The tendency of increased temperature to increase decomposition may also be offset 48 C.rV Agriculture and forestry by the negative impact of increased carbon/ nitrogen ratios 1991b; Parry et al. 1992a), Rhine basin (Wolf and van on decomposition, and the negative impact of drought on Diepen 1991), US (Adams et al. 1990), US midwest decomposition, if droughts become more frequent in a (Rosenberg and Crosson 1991; Easterling et al. 1991), warmer climate. Higher temperatures could also increase agricultural regions in Canada (Cohen 1991; Singh and mineralisation rates, improving availability of phosphorous Stewart 1991), East Asia (Yoshino 1991), Southeast Asia and potassium and speeding colloid formation. (Parry et al. 1992b), Caribbean region (Granger 1991), and Mexico (Liverman 1992; Liverman and O'Brien 1991). Fewer national assessments have been published, 7 Effects of climate change on the distribution of some exceptions being Hungary (Farago et al. 1990 and agricultural pests and diseases 1991), Ireland (McWilliams 1991), UK (UK Department of the Environment 1991), and New Zealand (Salinger et Recent reviews echo the earlier IPCC conclusion that al. 1990; Martin et al. 1990). ecological conditions for insect growth and abundance are As these regional studies come forward, they highlight expected to improve overall through lengthening of the IPCC conclusion that impacts will vary greatly accord• breeding seasons and extension of ranges (Porter et al. ing to types of climate change and types of agriculture. 1991; Pimentel et al. 1992), but also suggest that drier Both beneficial and detrimental effects of climatic change conditions may decrease pest damage in some regions. are projected, although these will not be evenly distributed Porter et al. (1991) investigated the impact of one GCM over the world. In particular, it appears more certain now climate change scenario on the European com borer that cool and temperate climatic zones may benefit, but in (Ostrinia nubilalis) in Europe, finding northward shifts in the tropics a further increase in temperature will be potential distribution of up to 1220 km, and an additional undesirable. At mid and high latitudes, increased tempera• generation in nearly all regions where the insect is current• tures can benefit crops presently limited by Cold tempera• ly known to occur. tures and short growing seasons. In the tropics, crops may Pimentel et al. (1992) considered the effects of a 2°C be subjected to growing season temperatures higher than temperature rise in the US and Africa and found that, if optima, and continue to suffer from nutrient shortages, North America becomes warmer and drier, crop losses due which impair the potential to utilise the beneficial direct to plant diseases are expected to decline as much as 30 % effects of increased CO2 (Goudriaan and Unsworth 1990). below current levels. If Africa becomes warmer and Newly published agricultural impact studies continue wetter, crop losses to diseases will increase up to 133% to emphasise the uncertainties inherent in such research as above current levels for some crops. High percentages of long as regional climate change scenarios remain unreli• crop losses to pests are expected to be sustained in Africa able. Other key uncertainties are projections of future because effective pest control technologies are not exten• technology and adaptation strategies. Methodological sively in use, nor are they expected to improve appreciably improvements are slowly accruing in dealing with such in the future. US crop losses to weeds are estimated to rise factors as direct COj effects, pests and pathogens, increas• between 5% and 50% (depending on the crop), because of ing fuel costs and resource substitution, but many studies intensified competition from weeds, which are often better neglect one or more of these possibly key factors. Given adapted to arid conditions than are crops. Herbicidal these considerable caveats, researchers concur that regional controls tend to be less effective under hot/dry conditions impact studies should continue to be regarded as sensitivity than they are under the more cool/wet current conditions studies only. (Pimentel et al. 1990).

8.1 North America 8 Research on regional impacts 8.1.1 United States Note: the following section summarises recent research on both 'Potential effects on crop yields and livestock produc• Further results from the US EPA national study published tivity' and on 'Effects on regional and national produc• since the IPCC Impacts Assessment imply a range of tion,' determining if the research upholds previous IPCC possible outcomes for US agriculture, depending on the findings or indicates new conclusions. severity of climate change and the compensating effects of Another focus of substantial amounts of research CO2 on crop yields (Adams et al. 1990). Simulations activity since the first IPCC synthesis has been regional combining results from climate, crop and economic models assessment. Some of the regional assessments have been indicate that the role of the US in agricultiaral export conducted or are under way in the EC (Carter et al. 1991a, markets may change, regional patterns of agriculture are C.rV Agriculture and forestry 49 likely to shift, and demand for irrigation water is likely to agricultural chemicals. Irrigation did not appear to be a increase. viable solution to heat-stress-related maize yield losses in An integrated research project on the US midwest Texas and Oklahoma. region (Missouri, Iowa, Nebraska and Kansas) used the 1930s climate (1.1 °C increase in temperature) as a warmer and drier analog. Methodology included a crop simulation 8.1.2 Canada model and its use for evaluation of both available and induced adaptations; inclusion of technological change; Research on potential agricultural climate change impacts analysis of climate variability effects on crop yield varia• in Quebec indicates that for one GCM (Goddard Institute bility; and integration of impacts on multiple sectors for Space Studies (GISS)) climate change scenario, yields including agriculture, forestry, water resources, and of com, soybeans, potatoes, phaseolus beans and sorghum energy. The research found that impacts on agriculture would increase and yields of cereal and oilseed crops, ie overall would be small given adaptation (Rosenberg and wheat,. barley, oats, sunflowers and rapeseed would Crosson 1991; Easterling et al. 1991). Relating these decrease (Singh and Stewart 1991). Apple and grape results to the IPCC best estimate of 2.5°C warming production might be enhanced and the northern part of the implies agricultural losses of about 10% for equilibrium province would benefit most. However, even if climatic warming of doubling of COj equivalent (Cline 1992). conditions in Quebec Were to improve, limitations of soil However, at the margins of the region losses could be fertility would still effectively constrain the significant considerable with a shift in irrigation from west to east. expansion of the agricultural land base. Effective adjustments to the 1930s climate change scenario in the US niidwest were earlier planting combined with longer season varieties of annuals and shorter season 8.2 Central and South America varieties in pereimials, and the use of furrow dyking in warm season crops. Wheat benefited from reduction of Several conferences have echoed the IPCC projections of cold stress and earlier break in dormancy. possible negative impacts of global warming in Central and Other recent research that used a simple relationship South America (Fundacion Universo Veinteuno 1990; between yield and COj and temperature reports similar Universidad de Sao Paulo 1990). Liverman (1991) and yield declines for the US Great Plains (up to 42-45''N) Liverman and O'Brien (1991) have used GCM output to (Okamoto et al. 1991). In summer, wanning caused by in• project declines in moisture availability and maize yields at creases in atmospheric COj appears to be undesirable for several sites in Mexico, following on from earlier studies spring wheat, soybean and com even if positive effects of cited in the initial IPCC Impacts Assessment. In contrast, COj are taken into account. Several CO2 emissions rates a study of the implications for climate change and direct were tested. CO2 effects in the Caribbean region found that agriculture These studies support the IPCC conclusion that US and food supply should benefit, although more intense grain production would decline with global warming. hurricanes and higher storm surges could bring added Another simulation with a mechanistic crop growth model damage to low-lying coastal areas (Granger 1991). with a soil water component found that grain yields at one site in Illinois increased, or decreased only slightly with temperature increases, except when the climate change was 9 Europe assumed to result in severely decreased rainfall (Muchow and Sinclair 1991). 9.1 European Community For the southern US, research using a climate model, a crop growth model, and a field level pesticide transport Ongoing work for the European region, with a focus on mode] studied the impacts of climate change on various the EC countries, is using a combination of techniques aspects of maize production (Cooler 1990). The results including large-scale mapping of agricultural potential, suggest that substantial changes in agricultural production crop-climate simulation modelling and laboratory experi• and management practices may be needed to respond to the mentation, to evaluate the possible effects of climate climate changes expected to occur in this region. These change on agricultural and horticultural production. Early changes include the need for heat tolerance as the control- results for grain maize, sunflower and soybean, all of hng factor in the introduction of new varieties, decrease in which are constrained by low temperatures in northern use of existing pesticides because of excessive loss on Europe, indicated that northward shifts in crop suitability mtensively cropped, rapidly leaching soils and an increase of 200-335 km per 1°C warming could be expected, with in the potential risk of aquifer contiimination by long-lived 50 C.N Agriculture and forestry these estimates varying locally by a factor of two owing to 9.4 Ireland terrain and regional climate (Carter et al. 1991b). For different GCM estimates of an equilibrium climate The agricultural production potential of Ireland may be response to equivalent COj doubling, estimates of north• enhanced with climate change (McWilliams 1991). Produc• ward shifts in potential due to warming varied from 700 to tion options could increase, new crops could be cultivated, 1900 km. Rates of northward shift depend critically on the and the overall costs of agricultural production could likely rate of warming. Regional estimates by the GISS GCM of be less than at present. While cereal crops may have little transient climate response to low, middle and high emis• or no increase in yields, yields of other crops like sugar sions scenarios in Europe give shifts varying from 10 km beet and potatoes could increase by up to 20%. Crop per decade to well over 100 km per decade up to the 2050s regions could extend northward and new crops, such as (Carter et al. 1991a,b). maize, sunflower, and flax, may become viable over much Further mapping work is examining the thermal arid of the south. Vegetable crops and fruit crops could be moisture constraints on grain maize production at a grown over wider areas than is currently the case, because European scale. Crop yields are being estimated at individ• warming should bring a reduction in the frequency of late ual sites, including winter wheat (different EC countries spring frosts. and four different crop models), cauliflower (ten locations in northern Europe), and soybean (France). COj enrich• ment and temperature experiments with lettuce and carrot 9.5 United Kingdom are also under way (Parry et al. 1992a). In the UK, higher temperatures would decrease cereal crop yields, and increase yields of such crops as potatoes, sugar 9.2 Commonwealth of Independent States beet and forest trees (UK Department of the Environment 1991). Viability of grassland, animal production and Since the first IPCC Impacts Assessment, additional work forestry in the uplands would improve, but might require has been conducted in Russia based on outputs from considerable investment. New crops and tree species could several GCM-based 2xC02 climate experiments. These be introduced into the UK. Maize and sunflower might be indicate that the agricultural production level in Russia is grown for their grain yield as well as for fodder. Horti• expected to drop by 5-15% without direct physiological cultural crops may benefit from warmer winters and effects of CO2 on crop growth and water use (Sirotenko reduced glasshouse heating costs. However, soils are likely 1991). In contrast, predictions based on palaeoclimate to be seriously affected by climate change which encom• analogs show a significant rise in productivity in Russia for passes variations in moisture, because cultivation may a doubling of atmospheric COj (Sirotenko et al. 1990), become more difficult under drier conditions, and because especially when direct COj effects are taken into account soil shrinkage could lead to more cracking with conse• (Menzhulin 1992). A crop modelling investigation of quences for moisture balance, efficiency of fertilisers, winter and spring wheat yields with GCM scenarios herbicides and pesticides, and groundwater pollution. Land combined with direct CO2 effects has shown that winter use policy for agriculture and agro-forestry may need to be wheat may respond more favourably to climate change reviewed in light of climate change effects. than spring wheat (Menzhulin et al., 1992). A review published since the IPCC assessment summarises research on climate change agricultural 9.6 Rhine Basin impacts in Russia with similar conclusions to those cited by the original IPCC (MacCracken et al. 1990). Wolf and van Diepen (1991) estimated the effects of one climate change scenario ( + 3°C with seasonally varying precipitation changes) and doubled. levels of atmospheric 9.3 Hungary CO2 on crop production and agricultural land use in the Rhine basin. Using a crop growth model, they found that The Hungarian Meteorological Service of the Hungarian the climate change and COj scenario tested had generally Ministry for Environment and Regional Policy has pub• positive effects on simulated crop production in the Rhine lished several studies on climate change and its potential basin. Average grain production of winter wheat increased socioeconomic impacts. These volumes consider specific by 35%, average production of permanent grassland by potential effects on Hungarian agriculture (Emanuel and 53 %, while silage maize production decreased. Changes in Starosolszky 1990; Farago et al. 1990; Farago et al. land use induced by climate change projected on the basis 1991). of a literature review included an increase in area of C.rV Agriculture and forestry 51 permanent crops, especially for vineyards; decrease in area demand for irrigation increased 15%. The east coast of of permanent grassland; decrease in areas cultivated with Peninsular Malaysia may become too wet for rubber root crops, oats, rye, rape and turnip rape; and increase in cultivation. Additional adverse effects on agriculture were areas cultivated with grain maize, sunflower and soybean. estimated to occur in both Indonesia and Malaysia, and Thailand as well, as a result of land losses due to sea-level rise.. 10 Africa

Downing (1991) has presented a methodology for research 12 Australia/Oceania on the potential impact of climate change in Africa that focuses on vulnerability to hunger (or food security), 12.1 New Zealand combining sensitivity to climatic variations with the multiple dimensions of regional and household food In New Zealand, higher temperatures may result in security. Downing et al. (1992) assessed the relative substantial shifts in agricultural potential, with temperate vulnerability of food production and needs to arbitrary crops, especially cereals, being grown 200 km further climate change scenarios in several African countries. In south and 200 metres higher for each 1°C temperature in• Senegal, a 4°C warming would reduce crop yields by 30% crease. At the same time, the potential range of these crops by the middle of next century, which could leave an could contract southward if winter chilling requirements additional one to two million people without domestically are not fulfilled (Salinger et al. 1990). The main horticul• produced cereals. In Zimbabwe, under a 2°C warming, tural crops (apples and kiwifruit) face displacement south yields that are currently achieved in seven out of ten years of the present day main production areas. This could have would be reached in only two to four out of ten years. In serious implications for kiwifruit production, given the Kenya, climate change could increase total potential assumed climate change scenarios. Viticulture may expand national production, while exacerbating food shortages in south into the South Island with a change to red winegra- semi-arid areas. pes in northern areas. Citrus and subtropical production A analysis of current crop yields and potential crop mjght undergo expansion particularly in northern New yield changes due to changes in temperature, moisture, Zealand. Animal industries could show substantial changes UV-B radiation, COj, pests, improved technology and reflecting the change in the quantity and quality of seasonal losses to pests in Africa concluded that, if rainfall increas• pasture production and changes in pests and diseases. es 10%, crop yields will improve to a limited extent Dairying may decline in the far north, but could increase (Pimentel et al. 1992). Water shortages may well persist, in the far south with temperature increases, resulting in an however, as well as serious crop losses to pests. The overall increase in production. At the same time, sheep potential impacts of the greenhouse effect on Africa are a numbers might show an overall decline, except in southern subject of regional concern (Suliman 1990). New Zealand. Beef cattle numbers are projected to increase in many areas of New Zealand with climate change. 11 Asia

A review paper on the climatic impacts on agriculture in 13 Global assessments East Asia suggests that year-to-year fluctuation of rice yields may become bigger, even though average yield may The US Department of Agriculture has conducted studies be greater (Yoshino 1991). This is based on statistical of the sensitivity of world agriculture to potential climate analysis of historical rice yields in Japan, which suggests changes (Kane et al. 1991, 1992). Preliminary results of that the fluctuation of yields is affected not only by these studies were presented in the first IPCC Impacts climatic variability, but by step-wise development of Assessment; recent results continue to indicate that the technological and socioeconomic development as well. overall effect of moderate climate change on world and A climate change agricultural impacts study in South• domestic economies may be small as reduced production east Asia has been completed which used a climate change in some areas is balanced by gains in others. However, scenario developed from the GISS GCM and crop model• further accumulation of GHGs could intensify the climate ling techniques (Parry et al. 1992b). In Indonesia, average effects. nee, soybean, and maize yields are projected to decrease A recent global study of climate change and crop yield about 4%,. 10% and 50% respectively. In Malaysia, changes has estimated changes in yields of world crops in simulated rice yields decreased from 12% to 22%, while both major production areas and vulnerable regions caused 52 C.rV Agriculture and forestry by GCM climate change scenarios and the associated direct Regarding nutrient applications, Buol et al. (1990) effects of increasing levels of COj (Rosenzweig and project differing regional fertiliser use, as did the first Iglesias 1992). Yield changes were estimated by crop IPCC Impacts Assessment. In temperate countries where specialists in 18 individual countries using compatible crop crops are already heavily fertilised, there will probably be growth models and common climate change scenarios. no major changes in fertilisation practices, but alterations Adaptations to the possible climate changes were also in timing and application method (eg careful adjustment of tested at two levels of effort. Preliminary results appear to side-dress applications of nitrogen during vegetative crop amplify the IPCC conclusion that agriculture in lower growth) are expected with changes in temperature and latitude regions may be more vulnerable to the projected precipitafion regimes. Fertilisation rates may need to be changes, even with adaptation efforts, because simulated raised to utilise the beneficial direct effects of increased yield declines tend to be greater in these areas in contrast CO2 (Goudriaan and Unsworth 1991). In tropical countries, to less negative or even positive yield changes at mid and where currently recommended fertiliser application rates high latitudes. When the economic consequences of the are not always applied, augmentation of fertiliser applica• yield changes were projected in a world food trade model, tions will be needed in any case. results showed that cereal prices and the number of people Adaptation to more intense pest infestations by at risk of hunger could increase with climate change increasing application of chemical control may increase (Rosenzweig and Parry 1992). chemical loadings in agricultural regions and may not be efficacious in regions where effective pest control tech• nologies are not extensively in use nor expected to improve 14 Adaptation appreciably in the future. Therefore, high percentages of crop losses to pests may be sustained in such regions. While some recent reviews and studies have emphasised Research has continued on the process by which farmers the possibilities of adaptation to climate change especially adapt, taking into account that imperfect information on in developed countries (US National Academy of Sciences climate change will be translated into incomplete and 1991; Rosenberg and Crosson 1991), conference state• imprecise reactions by farmers (Yohe 1992). In this work, ments and studies in the last eighteen months also suggest a utility-based decision model is used to test how farmers that it is important to remember that adaptation may not be in the US midwest may decide to switch crops, or at least complete and that some adaptive measures may have their mix of crops in response to growing evidence that the detrimental impacts of their own (Second World Climate climate appears to be changing. Yohe suggests that Conference 1990; Smit 1991, 1992; Bazzaz and Fajer consistent integration of these imprecise decisions over 1992). For example, changes in planting schedules and larger systems could add a sense of transition to static readily available cultivars may be easily adopted to portraits of potential futures. minimise impacts on agricultural incomes, but modifying Kaiser et al. (1992) have similarly studied the process the types of crops, grown does not guarantee equal levels of economic adjustment in the US midwest to synthetic or nutritional quality of food production or equal profits climate scenarios consistent with GCM predictions. This for farmers. work finds that farmers in southern Minnesota can make Similarly, irrigation is beginning to be perceived as a some adjustments to mildly changing climate that may limited option for regions with increasing aridity caused by significantly moderate initial negative effects. This was due climate change. The integrated US midwest study found primarily to excellent water-holding capacity of the soil in that agriculture may be less able to compete for surface the region tested. water and that declining groundwater supplies would hasten Adaptation in agricultural policy is emerging as a the abandonment of irrigation in the western portions of necessary focus. Inertia and current policy structures may the region (Rosenberg and Crosson 1991). Excessive actually inhibit agricultural adaptation to climate change overdraft of groundwater, soil salinisation, waterlogging, and variability (Smit 1991, 1992). Another study suggests and potential increased demand from competing sectors that current US commodity programs may well slow may limit the viability of irrigation as an adaptation to adaptation to climate change by providing financial incen• climate change. However, studies continue to show that the tives for farmers to remain with existing cropping systems, need for irrigation is likely to increase with global warm• rather than to switch to more climatically appropriate ones ing, as was concluded in the first IPCC assessment. In the (Lewandrowski and Brazee 1992). US, irrigation has been predicted to increase for a variety of climate change scenarios with higher temperature (Peterson and Keller 1990). C.rV Agriculture and forestry 53

15 Gaps in research and monitoring or specific to different agricultural systems in different regions of the world should be encouraged. Critical research and monitoring gaps that remain to be • While the agricultural sector (especially in regard to fdled include: productivity) is probably among the most monitored of human activities, much could be done to analyse • Further experimental studies of temperature and COj current and historical data for climate and COj sensi• interactions over the full growing period on crops and tivity. These types of studies can investigate, for cultivars from diverse agricultural systems are needed. example, how crop production regions have shifted These studies should be designed to test a wide range geographically in the past, if they shifting now, and of levels of both factors, at varying water-stress whether shifts are climate-related. levels, and to quantify a full complement of crop physiological and phenological responses and their effects on both biomass productivity and grain produc• 16 Recommendations and conclusions tion. Studies should also be conducted on the interac• tions of CO2, high temperature, and other stresses Since the publication of the first IPCC Impacts Assess• such as elevated UV-B radiation and tropospheric ment, study of agricultural impacts of climate change has ozone. continued in many areas. The findings of these studies • Initial studies of changed climatic variability should be reinforce the conclusion of the IPCC Assessment Report amplified by many more research efforts. The ability that, globally, climate change should not compromise our to test changes in climate variance is a needed advance ability to produce sufficient food. These findings also in impact studies on agriculture, particularly for indicate that previous work may have overestimated investigation of the effect of changes in climatic negative, and underestimated positive, impacts due to variances on variability of crop yields, farm retums climate change. However, there is a need for improved and risks, regional production and key factors of regional climate models and the extension of GCM simula• agricultural stability. tions beyond the equivalent COj doubling point, as well as • The viability of a wide variety of adaptation options a need for projecting how our agricultural systems will be (both biophysical and socioeconomic) needs to be configured in the future. tested for differing agricultural environments and The available studies suggest, however, that the impact systems. Research should be expanded in identifying of global warming on agriculture may be significant, the range of adaptability that exists within current crop especially if warming is at the upper end of the range species, in tracking the expansion of crop production projected by Working Group I. Impacts of climate change into new environments, and in determining the physio• must be considered in the context of the necessity for logical mechanisms whereby limits to production in continuing increases in regional agricultural production as stressful environments have been reduced so that the the world's population grows from 5 billion to a projected climatic range of successful production has been 8.5 billion by 2025. The following points summarise the expanded. Research initiatives should continue to contributions that recent studies have made to the field of emphasise the processes by which farmers and institu• agricultural climate change impacts. tions (both public and private) adapt, taking into Present studies have focused attention on how transient account that imperfect information on climate change projections of agricultural technology should be estimated. will be translated into incomplete and imprecise For example, conventional plant breeding and biotechnol• reactions by farmers, agricultural organisations and ogy are each expected to improve yields on the order of government agencies. Both economic and environ• 1-2%/year, and such estimates should be factored into mental costs and benefits of proposed adaptations projections of future impacts of climate change. Evaluation should be evaluated. of a wide variety of adaptation options is needed for • A needed focus is a better understanding of the differing agricultural environments and systems. linkages between climate impacts and socioeconomic Many regional assessments have been conducted since structures and the critical thresholds where significant the IPCC, or are under way in the EC, Russia, the US change takes place. Regions suffering from food midwest and agricultural regions of Canada, Southeast poverty today need to be evaluated for vulnerability to Asia and Mexico. Additional national assessments have the potential for increased stress from climate change. been published by the Hungarian Ministry for Environment Much can be done to study ways to enhance resilience and Regional Policy, the Ireland Department of the of agricultural activities to climate variability. Devel• Environment, UK Department of the Environment, and the opment of improved analytical models either generic New Zealand Ministry for the Environment. Additional 54 CIV Agriculture and forestry

integrated regional impact studies linking biological, spheric C02-Plant Growth Relationships. US Dept. of physical and socioeconoinic factors should be,carried out Energy and US Dept. of Agriculture. DOE/ ER-0450T. wherever possible. 61 pp. As these regional and national studies come forward, Aveyard, J.M. 1988. Climatic Change Implications for Land they highlight the IPCC conclusion that impacts will vary Degradation in New South Wales. Dept. of Conservation and Land Management, Chatswood, New South Wales. greatly according to types of climate change and types of Bazzaz, F.A. and E.D. Fajer 1992. 'Plant life in a COj-rich agriculture. Beneficial and detrimental effects of climafic world', Scientific American. January 68-74. change will not be evenly distributed over the world. In Bouwman, A.F. (ed.) 1990. Soils and the Greenhouse Effect. particular, it appears more certain now that cool and John Wiley and Sons. 575 pp. temperate climatic zones may benefit, but in the tropics a Brklacich, M. and B. Smit 1992. 'Implications of changes in further increase in temperature will be undesirable. climatic averages and variability on food production op• Research continues to address the relative importance portunities in Ontario, Canada', Climatic Change 20(1): of direct and indirect effects of COj on future crop produc• 1-21. tion. Some scientists weigh the beneficial effects as the Buol, S.W., P.A. Sanchez, J.M. Kimble and S.B. Weed dominant factor, while some are sceptical that the large 1990. 'Predicted impact of climate warming on soil benefits seen in experimental settings will be seen in prpperties and use', in: Kimball, B.A., N.J. Rosenberg, farmers' fields under changing climate conditions. Com• L.H. Allen Jr. (eds.) Impact of Carbon dioxide, Trace Gases, and Oimate Change on Global Agriculture. ASA bined studies of direct and indirect effects of CO2 on crops Special Publication Number 53. American Society of have been intensified to provide clarification. A key point Agronomy, pp. 71-82. which is emerging is that the wanning effects of CO2 are Carter, T.R., M.L. Parry and J.R. Porter 1991a. 'Climatic projected to increase beyond doubling point, but direct change and fumre agroclimatic potential in Europe', Int. physiological effects may level off. The future is also J. Climatol., 11:251-269. projected to bring simultaneous multiple stresses, such as Carter, T.R., J.*R. Porter and M.L. Parry 1991b. 'Chmatic elevated UV-B radiation and increased tropospheric ozone, warming and crop potential in Europe: Prospects and un• in combination with high COj, and these may negate some certainties'. Global Environmental Change 1:291-312. of the beneficial physiological effects in some crops. Cline, W. 1992. The Economics of Global Warming. Institute Recent studies have reinforced the concern for in• for International Economics. Washington, D.C. creased incidence of both droughts and floods, and on Cohen, S.J. 1991. 'Regional impacts of projected global warming: A research proposal for the Mackenzie Basin', water resources for agricultural in general. Changes in in: Second Symposium on Global Change Studies (Pre• water resource availability (both in timing and quantity) prints). American Meteorological Society. Boston, pp. pose the greatest risks to agriculture resulting from climate • 74-77. change. Thus, water resource management for food and Cooter, E.J. 1990. 'The impact of climate change on continu• fibre production (in relation to water needs for other ous corn production in the southern U.S.A',. Climatic sectors as well) is a crucial task. Cliange 16:53-S2. Current reviews confirm the earlier conclusion of the Downing, T.E. 1991. 'Vulnerability to hunger and coping IPCC that ecological conditions for insect growth and with climate change in Africa', Global Environmental abundance are expected to improve overall. Change 1(5):(in press). Downing, T.E. et al. 1992. Implications of Climate Change for International Agriculture: Vulnerable Regions. Environmental Change Unit. University of Oxford. References Oxford, UK. (in press). Eamus, D. 1991. The interaction of rising COj and temper• Adams, R.M., C. Rosenzweig, R.M. Peart, J.T. Ritchie, ature with water use efficiency. Plant, Cell, and Envi• B.A. McCarl, J.D. Glyer, R.B. Curry, J.W. Jones, K.J. ronment 14:843-852. Boote and L.H. Allen, Jr. 1990. 'Global climate change Easterling, W.E., M. McKenney, N.J. Rosenberg and K. and US agriculmre', A'arwre 345:219-224. Lemon 1991. /I Farm-Level Simulation of the Effects of Akita, S. and D.N. Moss 1973. 'Photosynthetic responses to Qimate Qiange on Crop Productivity in the MINK Re• CO2 and light by maize and wheat leaves adjusted for gion. Resources for the Future. Dept. of Energy constant stomatal apertures', Crop Science 13:234-237. DOE/RL/01830T-H5. Washington, D.C. Allen, Jr., L.H. 1990. 'Plant responses to rising carbon Emanuel, A. and O. Starosolszky 1990. Role of the Qimate dioxide and potential interactions with air pollutants', J. and Climate Change in the Life of Hungary. Hungarian Environ. Qual. 19(l):15-34. Meteorological Service. Hungarian Ministry for Environ• Allen, S.G., S.B. Idso, B.A. Kimball, J.T. Baker, L.H. ment and Regional Policy. 42 pp. (In English). Allen, Jr., J.R. Mauney, J.W. Radin and M.G. Ander- Farago, T., Zs. Ivanyi and S. Szalai 1990. Climate Variabili• • son, April 1990. Effects of Air Temperature on Atmo• ty and Change (I): Causes. Processes, Regional Impacts C.JV Agriculture and forestry 55

with Special Emphasis on the Socio-economic Impacts Kane, S., J. Reilly and J. Tobey 1992. 'An empirical study and the Tasks Related to the International Cooperation. of the economic effects of climate change on world agri• Hungarian Meteorological Service. Hungarian Ministry culture', aimatic Change. 21:17-36. for Environment and Regional Policy. lOO pp. (In Hun• Lewandrowski, J.K. and R.J. Brazee 1992. 'Government garian with English summaries). farm programs and climate change: a first look', in: J. Farago, T., Zs. Ivanyi and S. Szalai 1991. Qimate Variabili• Reilly and M. Anderson (eds.). Economic Issues in ty and Change (II); Qianges in Composition of Atmos• Global Climate Orange: Agriculture, Forestry, and phere and in the Climatic Characteristics, Detection, Natural Resources. Westview Press. Boulder, CO. Modelling, Scenarios and Impacts of the Regional Chan• Liverman, D.M. 1992. 'The potential impacts of global ges. Hungarian Meteorological Service. Hungarian warming in Mexico: some preliminary results', in: J. Ministry for Environment and Regional Policy. 210 pp. Reilly and M. Anderson (eds.). Economic Issues in (In Hungarian with English summaries). Global Qimate Change: Agriculture, Forestry, and Fundacion Universe Veinteuno 1990. Memoria del Semin- Natural Resources. Westview Press. Boulder, CO. ario Internacional sobre Calentamiento Global: Una Liverman, D.M. and K. O'Brien 1991. 'The impacts of Vision Latinoamericana. Fundacion Universe Veinteuno. global warming in Mexico', Global Environmental Mexico D.F. Change. 1:351-364. Giorgi, F. and L.O. Mearns 1991. 'Approaches to the simu• MacCracken, M.C, M.l. Budyko, A.D. Hecht, and Y.A. lation of regional climate change: A review', Reviews of Izrael 1990. Prospects for Future Climate. A Special Geophysics 2%2):\9\-2\e. US/USSR Report on Climate and Climate Change. Goudriaan, J. and M.H. Unsworth 1990. 'Implications of Chapter 7. Impacts of Climate Change on Water Re• increasing carbon dioxide and climate change for sources and Agriculture, pp. 185-224. agricultural productivity and water resources'. In; Manabe, S. and K. Bryan, Jr. 1985. 'COj-induced change in Kimball, B.A., N.J Rosenberg, L.H. Allen Jr, (eds.) a coupled ocean-atmosphere model and its paleoclimatic Impact of Carbon dioxide, Trace Gases, and Climate implications', J. of Geophysical Research 90(C6): 11, Change on Global Agriculture. ASA Special Publication 689-11,707. Number 53. American Society of Agronomy, Inc. pp.' Martin, R.J., M.J. Salinger and W.M. Williams 1990. 71-82. 'Agricultural Industries', in: Climate Change: Impacts on Granger, O. 1991. 'Climate change interactions in the New Zealand. Ministry for the Environment. Wellington, -.^ Gvts.i&rC&nhh^&v',The Environmental Professional 13: pp. 165-173. 43-58. McWilliams, B.E. 1991. Climate Cliange: Studies on the Hendrey, G.H., K.F. Lewin and J. Nagy 1992. 'Free air Implications for Ireland. Department of the Environment. carbon dioxide enrichment: Development, progress, Dublin. results', Vegetatio. (in press). Mearns, L.O., C. Rosenzweig and R. Goldberg (1992). 'The Hesketh, J.D. and D.M. Aim 1992. 'Synthesis: Modelling effect of changes in interannual climatic variability on field-crop ecosystems and issues for the future', in: C.J. CERES-Wheat yields: sensitivity and 2xC02 GCM Pearson (ed.). Field Crop Ecosystems. Elsevier. Amster• studies', J. ofAg. and For, Meteor, (in press). dam, pp. 519-536. Menzhulin, G.V. 1992. 'The impact of expected climate Houston Advanced Research Center 1991. 77?^ Regions and changes on crop yields: Estimates for Europe, the USSR, Global Warming. Woodlands Conference Statement. and North America based on paleoanalogue scenarios', HARC. Houston, TX. in: J. Reilly and M. Anderson (eds.). Economic Issues in Jodha, N.S. 1989. 'Potential strategies for adapting to Global Climate Change: Agriculture, Forestry, and greenhouse warming: Perspectives from the developing Natural Resources. Westview Press. Boulder, CO. world', in: N.J. Rosenberg, W.E. Easterling, P.R. Menzhulin, G.V., L.A. Koval and A.L. Badenko 1992. Crosson, and J. Darmstadter (eds.). Greenhouse Warm• 'Potential effects of global warming and atmospheric CO^ ing: Abatement and Adaptation. Resources for the Future. on wheat production in the USSR', in: Rosenzweig, C. Washington, D.C. pp. 147-158. and A. Iglesias (eds.). Implications of climate change for Kaiser, H.M., S.J. Riha, D.G. Rossiter and D.S. Wilks international agriculture: Crop modeling study. US (1992). 'Agronomic and economic impacts of gradual Environmental Protection Agency. Washington, D.C. (in global warming: A preliminary analysis of midwestem press). crop farming', in: J. Reilly and M. Anderson (eds.). Miglietta, F., A. Raschi, R. Resti and G. Zipoli 1991. 'Dry Economic Issues in Global Climate Change: Agriculture, matter production, yield and photosynthesis of soybean Forestry, and Natural Resources. Westview Press. grown in natural COj enriched environment', Agronomy Boulder, CO. pp. 91-116. Abstracts. American Society of Agronomy. Madison, Kane, S., J. Reilly and J. Tobey 1991. Climate Change: WI. p. 22. Economic Implications for World Agriculture. US Depart• Muchow, R.C, and T.R. Sinclair 1991. 'Water deficit ment of Agriculture. Economic Research Service. AER-? effects on maize yields modeled under current and 647. 21 pp. "greenhouse" climates',/t^ron. y. 83:1052-1059. 56 CIV Agriculture and forestry

Okamoto, K., T. Ogiwara, T. Yoshizumi and Y. Watanabe Dept. of Energy DOE/RL/01830T-H5. Washington, 1991. 'Influence of the greenhouse effect on yields of D.C. 35 pp. wheat, soybean and corn in the United States for differ• Rosenzweig, C. and A. Iglesias (eds.). with B. Baer, W. ent energy scenarios', Qimatic Qiange 18:397-424. Baethgen, M. Brklacich, T.-Y. Chou, R.B. Curry, R. Parry, M.L., T.R. Carter, J.R. Porter, G.J. Kenny and P.A. Dellecolle, H.M. Eid, C.R. Escano, J. Jones, Z. Karim, Harrison 1992a. Qimate Change and Agricultural L. Koval, D. Liverman, G.V. Menzhulin, W.S. Meyer, Suitability in Europe. Environmental Change Unit. P. Muchena, A. Qureshi, G. Rao, J.T. Ritchie, O. Sala, University of Oxford. Report 1. Oxford, UK. H. Seino, O.J.F. de Siqueira, M.L.C. Tongyai and J. Parry, M.L., M. Blantran de Rozari, A.L. Chong and S. Zhiqing 1992. Implications of Qimate Change for In• Panich (eds.) 1992b. The Potential Socio-Economic ternational Agriculture: Crop Modeling Study. US En• Impacts of Climate Change in South-East Asia. United vironmental Protection Agency. Washington, D.C. (in Nations Environment Program. Nairobi. press). Patterson, D.T. and E.P. Flint 1990. 'Implications of Rosenzweig, C. and M.L. Parry 1992. Qimate Change and increasing carbon dioxide and climate change for plant World Food Supply. Environmental Change Unit. communities and competition in natural and managed University of Oxford. Oxford, UK. (in press). ecosystems', in: Kimball, B.A., N.J Rosenberg, L.H. Salinger, M.J., M.W. Williams, J.M. Williams and R.J. Allen Jr. (eds.) Impact of Carbon dioxide. Trace Cases, Martin 1990. 'Agricultural Resources', in: Qimate and Qimate Qiange on Global Agriculture. ASA Special Qiange: Impacts on New Zealand. Ministry for the Envi• Publication Number 53. American Society of Agronomy, ronment. Wellington, pp. 108-132. Inc. pp. 83-110. Scharpenseel, H.W., M. Schomaker and A. Ayoub (eds.) Peterson, D.F. and A. Keller 1990. 'Irrigation', in: P.E. 1990. Soils on a Warmer Earth. Elsevier Scientific Pub• Waggoner (ed.). Qimate Change and US Water Re• lishers. Amsterdam. (Proceedings of an International sources. Report of the American Association for the Workshop on Effects of Expected Climate Change on Advancement of Science Panel on Climatic Variability, Soil Processes in the Tropics and Subtropics, Nairobi, Climate change and the Planning and Management of US Kenya, February 1990). Water Resources. John Wiley & Sons. New York. pp. Second World Climate Conference 1990. Conference 269-306. Proceedings. World Meteorological Organization. Phillips, D.L., D. White and C.B. Johnson 1991. Qimate Geneva. Change and Soil Erosion in the United States. US Singh, B. and R.B. Stewart 1991. 'Potential impacts of a Environmental Protection Agency. Corvallis, OR. C02-induced climate change using the GISS scenario on Pimentel, D., N. Brown, F. Vecchio, V. La Capra, S. agriculture in Quebec, Canada', Agriculture, Ecosystems Hausman, O. Lee, A. Diaz, J. Williams, S. Cooper and and Environment 35:327-347. E. Newburger 1992. 'Ethical issues concerning global Sirotenko, O.D., A.A. Velichko, V.A. Dolgy-Trach and warming on food production', J. of Ag. and Environ. V.A. Klimanov 1990. 'Global warming and future agro- Ethics, (in press). climate resources of the Russian plains', Priroda Journal Pimentel, D., L. .McLaughlin, A. Zepp, B. Lakitan, T. 3:83-88. (In Russian). Kraus, P. Kleinman, F. Vancini, W.J. Roach, E. Graap, Sirotenko, O.D. 1991. 'Soviet CLIMATE-YIELD simulation W.S. Keeton and G. Selig 1990. 'Environmental and system'. Meteorology and Hydrology Journal 4:67-73. economic impacts of reducing US agricultural pesticide (In Russian). use', in: D. Pimentel (ed.). Handbook on Pest Manage• Smit, B. 1991. 'Potential future impacts of climate change on ment in Agriculture. CRC Press. Boca Raton, FL. the Great Plains', in: G. Wall (ed.). Symposium on the Pittock, A.B., A.M. Fowler and P.H. Whetton 1991. Impacts of Qimatic Change and Variability on the Great 'Probable changes in rainfall regimes due to the enhanced Plains. (Occ. Paper 12, Geography, Univ. of Waterloo, greenhouse effect'. International Hydrology & Water Re• Canada), pp. 81-91. sources Symposium. Perth, pp. 182-186 Smit, B. 1992. 'Decisions in agriculture in the face of Reynolds, J.F., D.W. Hilbert, J.-L. Chen, P.C. Harley, uncertainty: Role of climate impact assessment', in: P. P.R. Kemp and P.W. Leadley 1992. Modeling the Re• Dzikowski (compiler) Proceedings of a Symposium/- sponse of Plants and Ecosystems to Elevated CO, and Workshop on Changing Climate in Relation to Sustain• Qimate Change. US Department of Energy. DOE/ER- able AgricuUure. Alberta Agriculture, Edmonton, pp. 60490T-H1. Washington, D.C. 33-42. Rind, D., R. Goldberg, J. Hansen, C. Rosenzweig and R. Suliman, M. (ed.) 1990. Greenhouse Effect and Its Impact on Ruedy 1990. 'Potential evapotranspiration and the likeli• Africa. Institute for African Alternatives. London. 90 pp. hood of future drought', J. Geophys. Res. 95(D7): Sundquist, E.T. 1990. 'Longterm Aspects of Future Atmo• 9983-10,004. spheric COj and Sea-Level Changes', in: Roger R. Rosenberg, N.J. and P.R. Crosson 1991. Processes for Revelle et al., Sea-Level Change. National Research Identifying Regional Influences of and Responses to Council and National Academy Press. Washington, D.C. Increasing Atmospheric CO, and Qimate Change: the pp. 193-207. MINK Project, An Overview. Resources for the Future. C.rV Agriculture and forestry 57

Teramura, A.H., J.H. Sullivan and L.H. Ziska (1990). 'The Whetton, P. and B. Pittock 1991. 'Toward regional climate interaction of elevated UV-B radiation and CO2 on change scenarios: How far can we go'? Proceedings of productivity and photosynthetic characteristics in wheat, Global Climate Change: Its Mitigation through Improved rice, and soybean'. Plant Physiology 94:470-4 75. Production and Use of Energy. Los Alamos, New Mexi• United Kingdom Climate Change Impacts Review Group co. 1991. The Potential Effects of Climate Change in the Wilks, D.S. 1991. 'Adapting stochastic weather generation United Kingdom. HMSO. London, pp. 124. algorithms for climate change studies', Qimatic Change Universidad de Sao Paulo and Woods Hole Research Cen• (in press). ter 1990. Regional Conference on Global Warming and Wolf, J. and C.A. van Diepen 1991. Effects of Qimate Ch• Sustainable Development. Sao Paulo, Brazil. 18-20 June ange on Crop Production in the Rhine Basin. DLO-The 1990. Conference Statement. Winand Staring Centre. Rijkswaterstaat RIZA Wagenin- US National Academy of Sciences 1991. Report on Adap• gen. Lelystad, The Netherlands. tation to Climate Change. Committee on Science, Yohe, G.W. 1992. 'Imbedding dynamic responses with Engineering, and Public Policy. National Academy Press. imperfect information into static portraits of the regional Washington, D.C. impact of climate change', in: J. Reilly and M. Anderson US National Science Foundation 1992. Soil-Warming Ex• (eds.). Economic Issues in Global Qimate Change: periments in Global Change Research. Report of a Work• Agriculture, Forestry, and Natural Resources. Westview shop held in Woods Hole, Massachusetts, September Press. Boulder, CO. pp. 200-214. 27-28 1991. US National Science Foundation, Ecosystem Yoshino, M. 1991. 'Impact of climatic change on agriculture Studies Program. Washington, D.C. from the viewpoint of East Asia', in: K. Takasuchi and Vinnikov, K.Ya. and I.B. Yeserkepova 1991. 'Soil moisUire: M. Yoshino (eds.). The Global Environment. Springer- empirical data and model results', J. of Qimate 4:66-79. Verlag. Berlin, pp. 16-41. Wang, Y.P. and K.J. Hennessy 1991. 'Predicted greenhouse effect on wheat quality and production', Australian Grain. August-September, pp. 26-27. 1

II The managed forest and the forest sector

Scope of the issue fime, reforestafion is the only known post-combustion carbon abatement measure.' (Sanghi and Michael 1991). For the purposes of this report, the managed forest Various options for reforestation/afforestation explore the includes all forests in which some harvest takes place, with methods for the establishment of a variety of species and particular emphasis on forests where there is some degree the potential increase in productivity. (Schroeder and Ladd of management. The majority of these managed forests are 1991). The data are based on experiences in a developed still of natural origin. This classification excludes forest country, but the approach has potential application in many parks/reserves and wilderness areas. countries, with different species, economic and forest Projected changes to forest ecosystems suggest chang• sector infrastructures. es in the geographic distribution of forest species and Data from the IPCC Impacts Assessment demonstrated changes in the growth rates with the most significant the benefits associated with increasing levels of manage• changes likely in the temperate and boreal areas. New ment and recent reports have continued to emphasis the information available mdicates that the productivity of the need for precise costing of all options and the accounting forest sector due to increased greenhouse gas concen• of the multiple benefits of environmental, human, spiritual, trations and associated climate change may have been ethical and economic values associated with forests. imderestimated. However, the socioeconomic consequences There is an important and unpredictable, human com• of climate change for the forest sector could be significant ponent involved in the questions of impacts and subsequent in regions where economic and social welfare and develop• responses concerning species and vegetation conservation. ment is highly dependent on the forest sector. The global This is further compounded by the lack of change in the ability of the forest sector to cope with climate-induced sociological and economic sciences to calculate the multiple changes will depend on the near-term acquisition of benefits with any degree of certainty. For example, there knowledge through focused research and assessment and has been no improvement on 'the regional scale, of the on prudent, fimely and proactive forest management poli• science, either in climate change or in the putative re• cies and strategies. sponse since the publicafion of the IPCC Impacts Assess• ment (1990) although efforts in this field continue (Wood• ward and McKee 1991). Progress since the IPCC Impacts Assessment in 1990 The human population must be seen as both a part of the solution (ability to respond) and part.of the problem 1 Socioeconomic importance of forests (population growth). Within this context, Forestry Canada, as part of Canada's 'Green Plan' (Anon 1991), has begun Clearly, the dependence and the importance of forests to a community tree-planting program designed to encourage the national economies has not diminished. This was the planting of up to 325 million trees in urban and rural demonstrated in the previous assessment and the updates in areas across Canada. By the year 2000, this will result in Tables 1 and 2 serve to reinforce this statement. Likewise, the absorption of some 5.2 million tonnes of CO^, thus management strategies are urgently needed for the adapta• helping to offset futiare carbon emissions by including the tion, mitigation and protection of the forest sector. involvement of the public at large in this issue. In addition Countries must continue to address the socioeconomic to Canada's Green Plan, the US has the Global Releaf issues related to forests in their response to the impacts of Program of the American Forestry Association. global and climate change. For example, the Sao Paulo Recent public opinion polls indicate that, at least in Conference in 1990 stated clearly that 'The warming of the developed countries, people are willing to pay for carbon earth ia clear evidence that current patterns of economic sequestration via reforestation, although this will not activity are not sustainable.'. The conference identified the address the whole problem. The potential willingness of vulnerability of a number of patterns of activity for the society to exceed the costs associated with the desired level developing countries. of sequestration would 'make possible even larger refores• tation benefits.' (Sanghi and Michael 1991).

2 Reforestation/afforestation and more managed forests 3 Tropical deforestation

The need for reforestation/afforestation is a global issue The rate of deforestation has increased by 90 % during the that must involve all countries. 1980s, mainly by shifting cultivators, and is expected to A recent report from the USA states that 'Except for continue to increase during the foreseeable future (Myer CO2 scrubbing, which is not a proven technology at this 1991). Most efforts to reduce deforestation have con-

J CrV Agriculture and forestry 59 centrated on commercial enterprises, such as loggers/ the Bangkok Forestry Conference (June 1991) recom• ranchers, but they account for only 40% of all current mended that the nations of the world should be 'living '^'deforestation. off the interest and not drawing down the principal of Further work is needed to liberalise current economic the earth's ecological capital (and this should be theories to account for the human demands and behavioural accomplished by) integrating economics and the shifts within society. For example, current studies indicate environment into decision-making.'. This attitude, also that the greatest impact on atmospheric carbon can be stated at the Sao Paulo Conference (Sao Paulo 1990) realised by halting deforestation compared to promoting and elsewhere (Woodward and McKee 1991; Schroe- reforestation. However, the former solution may not be der and Ladd 1991) reflects a universal concem and realistic as 'the depletive pressure is likely to increase if the need for action. only through growing human numbers and growing human demands.' (Houghton 1991). In addition, the IPCC Impacts Forest health monitoring programs, including forest Assessment (1990) summarised the substantial potential climate observations above and within forests, need to be that exists to increase the growth performance of managed established to detect damage to forest trees and soils that tropical stands with rates of carbon sequestration that are may have been caused by anthropogenic causes by isolating more than twice the rates of managed forests in temperate damage not attributable to natural causes or management or boreal regions. practices (Hall and Addison 1991, Miller 1991). Integrated monitoring systems based on observed or proxy climate data sources should also include changes in vegetation and 4 Research and monitoring gaps soils attributable to other impacts on representative forest ecosystems. These health monitoring systems need to be The IPCC Impacts Assessment (1990) listed many items flexible and adaptable to allow for their transfer to other worthy of support in relation to forest ecosystems in countries withm the global community. The ability to close general. These were echoed in more detail in the Bangkok this research and monitoring gap can be immediately and Forestry Conference report (June 1991): 'There was a need highly successful given the demonstrated state of science for better information on the extent of forests, including and technology for electronic climate and biological fiirther effort to ensure that appropriate definitions were monitoring systems. - used for each forest type and objective.'. Other problems Forests are a most significant renewable resource of cited included the use of various nomenclature and the biomass, timber, and fibre. Due to their extent forests have failure to include tree crops associated with agriculture and a large effect on the global carbon cycle. They are at the agroforestry in discussions of forest responses. Numerous same time sinks, sources and reservoirs of carbon. The urgent data needs are becoming increasingly apparent, role of forests to the global carbon cycle is universally including (but not limited to): recognised as a sink for carbon, especially in the temperate and boreal regions (Sedjo 1992, Kauppi et al. 1992), but • the need for regional climate change impact studies forests also provide valuable benefits as resources in the (Graham et al. 1990, Overpeck et al. 1990, Smith et short term and in the long term to biodiversity. However, al. 1990). there needs to be an increased effort to quantify the global • an improved understanding of the genetics/physiology carbon estimates and the socioeconomic benefits. This of tree species. includes the development of analytical models that account • the identification of critical points within the carbon for the exchanges of carbon within the ecosystem and the cycle such as soil carbon contributions and the re• relative contribution of forests, under different assumptions sponse of plants to elevated CO2 and climate change such as species types, soil conditions and various land use (Reynolds et al. 1992). patterns etc. Robust estimates of the costs of carbon • a better understanding of the linkages between climate sequestered through afforestation are needed. These impacts and socioeconomic structures and the thresh• science-based models would have wide applicability in olds where change takes place. The Bangkok Techni• other countries as a shared resource and should not be cal Workshop to Explore Forestry Options (April restricted by proprietary rights. The universality of this 1991) stated explicitly that 'better information on the type of scientific transfer should be encouraged to facilitate cost effectiveness of the social and economic basis for changes in policy. different options for global forest management, and on quantifying the multiple roles of forests, was an urgent research priority'. All results since then would seem to confirm the urgency of this statement. For example, 60 CIV Agriculture and forestry

5 Recommendations/conclusions Forestry Canada, Science and Sustainable Development Directorate, Inform. Rep. DPC-X-34. 42pp. • The data on resources reinforces the vulnerability of Graham, R.L., M.G. Turner andV.H. Dale 1990. 'How the forest in all countries. Climate change will im• increasing COj and climate change affect forests'. Biosci• ence, 40(8):575-587. pinge on the social and economic sustainability of Houghton, R.A. 1991. 'Tropical deforestation and atmos• these forests. Results since the 1990 IPCC Impact pheric carbon dioxide'. Climate Change 19:99-118. Assessment reinforce the political and economic Kauppi, P.E., K. Mielikainen and K. Kuusela 1992. 'Bio• urgency in this sector. mass and carbon budget of European forests 1971 to • More reforestation and afforestation programs are 199Q\ Science, 256:70-14. recommended. The major goal is to sequester carbon n»CC 1990. Qimate Change: The IPCC Impacts Assessment. but other benefits must not be neglected. Policies must Edited by W.J.McG. Tegart, G.W. Sheldon and D.C. take a holistic view and consider the cumulative Griffiths, Intergovernmental Panel on Climate Change. benefits of forests. WMO and UNEP, Australia. • Research needs must centre on regional impacts Myers, N. 1991. 'Tropical forests: present status and future studies, process studies, and the establishment and outlook.' Qimate Change 19:3-32. definition of the socioeconomic linkages. Overpeck, J.T., D. Rind and R. Goldberg 1990. 'Climate induced changes in forest disturbance and vegetation'. • Forest monitoring and development of analytical Nature, 343(4):51-53. models are essential, and must be widely applicable, Miller, G. D. 1991. 'Forest health: where does it hurt'? acceptable and consistent for all countries. American Forests, 97:46-48. • Research leading up to the IPCC Impacts Assessment Reynolds, J.F., D.W. Hilbert, J.-L. Chen, P.C. Harley, and subsequent work, has amply demonstrated the P.R. Kemp, and P.W. Leadley 1992. Modeling the Re• continuing need for intensive and accurate technical sponse of Plants and Ecosystems to Elevated CO, and and financial cooperation to optimise the management Qimate Change. US Department of Energy. DOE/ER- of our global forest. This assumes that there are or 60490T-H1. Washington, D.C. will be programs in place that take advantage of local Sanghi, A.K. and K.S. Michael 1991. Carbon Sequestration and regional opportunities, particularly, and wherever Through Reforestation: Cost and Potential for New York. possible within institutional infrastructures. Air and Waste Management Association. Paper presented at the 84th Annual Meeting and Exhibition, Vancouver, B.C, Canada. 14pp. Sao Paulo 1991. Regional Conference on Global Warming References and Sustainable Development, Sao Paulo Conference, Organised by the University of Sao Paulo, Brazil, and Anon 1991. The State of Forestry in Canada: 1990 Report to The Woods Hole Research Centre, Woods Hole Mass., Parliament. Corporate and Public Affairs, Ottawa. 80pp. USA. Bangkok, April 1991. Technical Workshop to Explore Schroeder, P. and L. Ladd 1991. 'Slowing the increase of Forestry Options. Report prepared by H. Volz, W. atrnospheric carbon dioxide: a biological approach', Kriebitzsch and T.Schneider, Germany, IPCC Forestry Qimatic Qiange 19:283-290 AFOS Working Group. 39pp. Sedjo R. 1992. 'Temperate forest ecosystems in the global Bangkok, June 1991. Conference Statement. International carbon cycle', Ambio (in press) Conference on Global Warming and Sustainable Devel• Smith, W.H. 1990. 'The health of North American forests: opment: An Agenda for the 90s. Organised by the stress and risk assessment', J. of Forestry, 88:32-35. Thailand development research Institute, Bangkok, Woodward, F.I. and I.E. McKee 1991. 'Vegetation and Thailand, The Instimte for Research on Public Policy, climate'. Environment International 17:535-546. Ottawa, Canada and The Woods Hole Research Centre, Woods Hole Massachusetts, USA Bangkok, Thailand. Hall, J.P. and P. A. Addison 1991. Response to Air Pollution: ARNEWS Assesses the Health of Canada's Forests. C.rV Agriculture and forestry 61

TABLE lA. Summary of forest resources and forest management species (1987)

Canada USA Austria FRG Productive Forest Land (million ha.) 244 210 3.0 7.3 Impact of Forestry GNP on Economy 506 415 000 95 244

(SbillionCdn) Import 1.59 15.44 1.1 506 Export 16.87 8.73 2.48 4.98

Managed Species: Canada: Picea spp., Pinus spp., Abies balsamea, Populus USA: Picea spp., Pinus spp., Quercus, Carya, Acer, Fagus Betula, Populus, Abies, Larix, Pseudo Tsuga, Tsuga, Juglanis Austria: • Picea abies, Pinus sylvestris, Larix decidua, Abiel alba, Fagus sylvatica Fed. Rep. of Germany: Picea abies, Pinus sylvestris, Abies alba, Pseudotsuga menziesii, Larix decidua, Fagus sylvatica, Quercus robur

TABLE IB. Summary of forest resources and forest management species (1987)

Indonesia China India Brazil Productive Forest Land (million ha.) 45 122 45 350 Impact of Forestry GNP on Economy 85 260 246 250

($billion Cdn) Import 0.22 3.38 0.22 0 Export 2.00 0.68 0.2 1.22

Managed species: Indonesia: Meranti romin, Tectonia grandis, Pinus merkusii, Pterocarpus indicus. Eucalyptus deglupta. Acacia decurrens China: Dawn redwood, Cathaya, Golden Larch, Chinese swamp cypress, fokienia cedar, larches, Chinese fir, Korean pine India: Dipterocarpus spp., Shorea robusta, Cedrus deodara, Pinus roxburghii, Abies densa, Picea smithiana Brazil: Eucalyptus spp,, Tectona grandis, Khaya spp., Swietenia macrophylla 62 CIV Agriculture and forestry

TABLE IC. Summary of Forest Resources and Forest Management Species (1987)

New Kenya Zambia Finland Zealand Chile Productive Forest Land (million ha.) 1.1 4.1 18.2 2.8 4.7 Impact of Forestry GNP on Economy 1.7 2.1 • 96.9 23.2 16.4

(Sbillion Cdn) Import 0.02 0.09 0.4 0.16 0.16 Export 0.004 0 7.15 0.55 0.48

Managed species: Kenya: Ocotea spp., Myrica salicifolia,, Acapia labai. Acacia abyssinica, Podocarpus gracilior. Juniperus procera Zambia: Brachystegia spp., Jubernardia angolensis. Eucalyptus grandis, Pinus kesiya Finland: Picea abies, Pinus sylvestris, Bemla pendula New Zealand: Pinus radiata, Podocarpus totara, Rimu cupressinum, Kamaki, Rata Chile: Pinus radiata. Eucalyptus spp., Populus, Pinus

TABLE 2A. Roundwood production, imports, exports and their value (1989*)

Indonesia China India Brazil Roundwood production (thousand m') 175 730 274 590 269 451 255 455 Roundwood imports (thousand m') 0 13 383 902 26 Value (Cdn$000)** 0 1 435 171 104 260 6 574 Roundwood export(thousand m3) 1 131 21 76 46 Value (Cdn$000) 38 910 21 749 1 032 4 525

Kenya Zambia Finland New Zealand Roundwood production (thousand m') 35 650 12 204 46 262 10 557 Roundwood imports (thousand m') 0 0 6 784 5 Value (Cdn$000) 0 0 307 774 3 383 Roundwood exports (thousand m') 0 0 1 047 2 470 Value (Cdn$000) 0 0 97 825 138 254

TABLE 2B. Roundwood production, imports, exports and their value (1989*)

Chile Canada USA Austria FRG Roundwood production (thousand m') 16 864 176 976 533 168 16 086 35 332 Roundwood imports (thousand m') 0 4 268 3 471 4 832 3 909 value (Cdn$000) 0 223 542 149 221 298 007 445 675 Roundwood exports (thousand m') 4 679 5 093 28 232 1 722 5 489 value (Cdn$000) 188 362 424 400 3 348 651 93 118 366 181

* FAO Yearbook Forest Products 1989—PubHshed in 1991 ** Exchange Rate 1 CDN$ = 1.183432 US$ V Natural terrestrial ecosystems

Lead authors: R B Street (Canada) S M Semenov (USSR) C H D Magadza (Zimbabwe)

Contributors: T V Callaghan (UK) J I Holten (Norway) R Christ (Austria) L Mortsch (Canada) A W Diamond (Canada) O J Olaniran (Nigeria) G R Evans (USA) S D Sastrapradja (Indonesia) J T Everett (USA) R R Vaghjee (Mauritius) R S de Groot (The Netherlands) N Vygodskaya (USSR) A Fischlin (Switzerland) V Natural terrestrial ecosystems

1 Background • rare and mainly middle and high alpine species may be threatened with extinction (direcUy or indirectly) As identified in the IPCC Impacts Assessment (1990), by climate change; and natural terrestrial ecosystems could face significant envi• • migration barriers, both natural and anthropogenic, ronmental impacts as a result of the global increases in the could prevent or delay dispersal of more temperate atmospheric concentration of greenhouse gases and the plant and animal species into Scandinavia. associated climatic changes, as well as from the increasing UV-B radiation due to the atmospheric ozone depletion and The process of reduction of tropical forest on the African the combined effect of biospheric oxidants and acidic continent, first noted in the tropical rain forest areas of depositions. The rate of these changes will be the major Zaire and Uganda about 9000 BP (Livingston 1975) with factor determining the type and degree of impacts. The the arrival of humans, will continue unabated and even be type of disturbances resulting from climate change will be reinforced by global warming, resulting in an accelerated unique for each ecosystem and the community within, as reduction of tropical forest and the expansion of the arid each responds to the added pressure of a changing environ• zones of Africa (Magadza 1990). Projected cliinatic change ment. Adding to the complexity of assessing these changes is likely to lead to an encroachment of the Sahel syndrome is that they could occur as a series of 'jumps' rather than into the savannas. This region supports approximately 200 in a uniform manner across an area like a 'moving wave' million people and any adverse effects of climate warming (Holten and Carey 1991). could result in the worsening of already precarious produc• The social and economic consequences of these tion systems in the area. impacts will be significant, especially for those regions of The degradation of the savanna ecosystems has a the globe where societies and related economies are further implication for animals and migratory birds. Many dependent on natural terrestrial ecosystems for their social animal and bird species utilise savanna weflands and and economic welfare. Projected changes within natural shallow lakes, such as Lake Abiata in Ethiopia, Lake terrestrial ecosystems suggest that changes in the availabili• Turkana in northern Kenya and other rift valley lakes, the ty of food, fuel, medicine, fibre, construction materials Kafue flats of Zambia and, in West Africa, the flood plains and income are possible. of the Niger and Senegal rivers, Lake Chad and several other water bodies. Decreased rainfall or soil moisture as projected by several climate models for the African 2 Progress since the IPCC Impacts Assessment savanna ecosystems suggests that large-bodied animals, such as the elephant, hippopotamus and the large bovines, One of the major issues arising from the impacts of climate are liable to be affected more adversely than smaller change on terrestrial ecosystems is water availability. This species. Many migratory bird species utilise seasonal will be of increasing concern as populations increase and wetlands, locally known as dambos, vleis or bugas in the need for water for drinking and economic activity (eg eastern and southern Africa. Songbirds (eg common irrigation) increase concomitantly. Recent studies suggest whitethroat, sedge warbler) breeding in Europe and that while water use efficiency should be increased in a wintering in the Sahel have declined in years following higher COj atmosphere, the same amount of water could droughts in the Sahel, apparently owing to reduced habitat be used per unit area because of increased leaf area ratios and survival in the wintering grounds. Projected climate due to greater biomass produced through CO2 fertilisation. change of the order anticipated by several climate models Furthermore, any benefits accruing from the more efficient will probably result in the reduction of many such habitats use by plants of soil water would be lost if wasteful (Magadza 1990) and, therefore, could negatively impact on irrigation technologies, such as flood irrigation, persist. these animals and migratory birds. There is evidence (Holten and Carey 1991) suggesting Seasonal wetlands are also critical to waterfowl and that: other migratory birds in the Great Plains region of North America (Diamond 1990; Diamond and Brace 1991). • in boreal, alpine and arctic regions, mesic and hygric In sub-humid savannas, climate change projections plant communities will in general be more sensitive to suggest increased aridity. It is therefore likely that rivers climate change than the xeric communities; in these regions that show marked seasonality in flow • in temperate and southern to middle boreal regions in would have reduced flows during periods of drought. Europe, major floristic and faunistic changes may Furthermore, many shallow water bodies, such as Lake take place in some of the thermophilous deciduous Chad, could experience complete aridity during extended forests as a result of projected improved conditions for drought periods, or experience very high salinity levels. Dutch elm disease; Such episodes would result in eventual extinction of the C.V Natural terrestrial ecosystems 65 fish species in those water bodies. Since many river basins a mesohaline salt marsh on the Chesapeake Bay) exposed display some degree of endemicity, such species decima• to twice normal COj concentrations showed increased tion could lead to reduction in species diversity, where carbon accumulation of 88% and 40% respectively. only those types, such as the Claridae, which can readily Callaghan and Carlsson (in press) showed that high migrate between river basins, would have the potential to productivity is not necessarily commensurate with success• survive. ful reproduction and greater fitness. This is particularly Projected climate change may be one of the most true in short growing seasons (eg due to temperature important factors affecting fisheries now and in the next limitations in the tundra or aridity in the tropics) when few hundred years. The level of the impact could vary climate conditions within the growing season allow high widely and may depend on attributes of the species as well productivity, but the shortness of the growing season as on their regional specificity. Freshwater fish populations restricts successful reproductive development. Climate are contained within the bonds of their watershed and in change projections suggest that in tundra areas, the most cases carmot migrate, but must endure (Shuter and growing season may lengthen as a result of warming; Post 1990), adapt, suffer population declines or become however, in other areas, increased aridity may shorten the extinct (Coutant 1990) when their environment deterio• growing seasons. This suggests the importance of con• rates. sidering population dynamics of plants as well as produc• The added stresses to the freshwater ecosystems tivity. Such approaches also consider age-related processes accompanying projected climate changes can be expected, and show that long-lived plants (eg trees), and those in the short term, to reduce species number and genetic reproducing sporadically (eg many high-latitude and high- diversity within freshwater populations (Everett, in press). altitude plants), are threatened more by climate change Species that survive a changing environment will likely be than short-lived plants (eg annuals) because of the limita• the ones with adequate existing genetic variation to allow tions on adaptive responses within each generation. survival and selection of at least some individuals. Whether The report 'Arctic Interactions: Recommendations for the projected rate of climate change is within the capability an Arctic Component in the International Geosphere- of freshwater species to adapt genetically will depend on Biosphere Programme' (1988) stresses the point that the existing genetic variability and on the rapidity of estimates suggest that 200-400 GT (109 metric tons) of change (Mathews and Simmerman 1990). non-oxidised carbon is stored in circumpolar peatlands With projected climate change, profound impacts, both (mostly in the boreal zones of Canada, the USSR and beneficial and destructive, can be expeyted for the Fennoscandia). If a major part of this carbon were released distribution and productivity of valuable fisheries and the because of changes in surface temperature and moisture—a industries associated with them (Healey 1990; Hill and physical possibility under projected climate changes over Magnuson 1990; Johnson and Evans 1990). Among the a few decades or centi^ries-then northern regions could first fish populations responding to climate change will emit a substantial additional amount of carbon to the probably be those in streams (a high rate of heat transfer atmosphere (potential positive feedback). from the air) on the warm or cold margins of their species' In their draft map of ecoclimatic sensitivity regions, native ranges (Holmes 1990). With warming, a longer Holten and Carey (1991) have suggested that within growing season is expected which should lead to greater Norway, mountain and subalpine environments are the fish productivity where temperature is currently a limiting most sensitive to climate change, through climate-induced factor, such as in high latitude lakes and reservoirs changes in snow cover (depth and duration) and hydrology. (Schlesinger and McCombie 1983). Species whose current The arctic and flat river valley regions appear to be as distribution is confined to cool waters may be eliminated sensitive as the alpine and subalpine environments. from parts of their range, leading to reduced fisheries, in spite of higher growth rates of the species that can tolerate warmer water. 3 Uncertainties and knowledge gaps Photosynthesis in plants grown at elevated concentra• tions of carbon dioxide is often reported to be reduced Uncertainties and gaps in the knowledge base exist in after a few days or weeks exposure. This reduction is often terms of our understanding of the environmental impacts attributed to acclimation of carbon metabolism. Some for natural terrestrial ecosystems and associated socioeco• studies of photosynthesis response to elevated CO2, nomic consequences of climate changes. These uncertain• however, show either an increase or no reduction in ties and gaps exist as a result of deficiencies in our photosynthesis capacity. For example, data collected over theoretical knowledge of: four growing seasons on two monospecific stands: one dominated by a C3 sedge and the other by a C4 grass (in 66 C.V Natural terrestrial ecosystems

• Fundamental ecological process: effects of a changing has been estimated that 25-60% of the present heritage set of interactive environmental parameters on plant sites and reserves could be unfit for the current species/ net primary productivity and water use, ecosystem communities they protect under projected changes in carbon and nutrient budgets, vegetation structure, climate. This suggests that areas of adaptation will be species composition, forage quality, plant-animal and required and that management strategies (ie in relation to vegetation-soil interactions; population dynamics of dessication and pest/disease control), adjustment of policies major species and vegetation types as well as of pests on selection/siting of protected areas and the development and noxious species; and migration rates and mecha• of specific stijdies and protective measures, should be nisms for species dispersal and settlement including required for 'transferral' of threatened plants and animals the impacts of barriers and corridors. to suitable habitats. • Linkages between climate and atmospheric chemistry The correlative relationships between climate and plant changes (ie elevated levels of atmospheric pollutants and animal species need to be experimentally evaluated to such as SOj, NOx etc and atmospheric fertilisers) on establish causal relationships and the relative impacts of the one hand and population and ecosystem dynamics changes in climate compared to other controlling factors. (including implications for competitiveness, reproduc• This comparative analysis is essential as there are other tion and productivity) on the other, can provide an factors that can and are contributing to ecosystem disap• assessment of the sensitivity (including buffering pearance and degradation (eg urbanisation, atmospheric capacity, plasticity and resilience) of key species, pollutants, land conversion, agricultural land management communities or ecosystems. practices, water course alterations). This requires that • Linkages between ecosystem dynamics and human fundamental ecological research be done in parallel with social and economic development and welfare under a research on socioeconomic responses (eg development of changing climate can provide an assessment of the impact models that bridge the gap between environmental vulnerability of social and economic systems to impacts and social and economic consequences) as ecologi• changes in natural terrestrial ecosystems. cal response may well be determined as much by socioeco• nomic reactions to climate change as by climate change The significance of the environmental impacts and socio• itself economic consequences as a result of climate change (IPCC Impacts Assessment 1990) and the need to develop viable and effective response strategies warrants the 4 Addressing the uncertainties and knowledge gaps allocation of sufficient resources to tackle these deficien• cies through climate impacts research, management, To reduce these uncertainties and gaps in our knowledge, awareness raising, national and international cooperation integrated national, regional and global programs and etc. networks of research and systematic observations are Within many developing countries, most of the climate required. These should include the improvement and change impacts activities (and for the environment in further develpment of the global network of research and general) are orchestrated by outside agencies with Uttle integrated systematic observation sites (eg Terrestrial input from local scientific/technical experts. The capability Ecosystem Monitoring and Assessment (TEMA) of UNEP, of developing countries of taking long-term views and to Global Change and Terrestrial Ecosystem or using previ• initiate programs on climate change impacts is hampered ous IBP research sites) with the aim of examining the by lack of the necessary institutional framework. In linkages and relationships essential to defining environ• addition, the monitoring networks within many developing mental impacts and identifying where and what action is countries are either too sparsely distributed or unservice• necessary. Also essential is the need to increase the skills able to allow for a worthwhile assessment of the impacts of human resources and level of supportive funding of, and responses to, climate change at the regional level. available for participation in global programs and net• The most urgent need is for fundamental research on works, especially for developing countries. the sensitivity and likely response of key physical and Integrated, representative, long-term systematic biological constituents of critical ecosystems (including observation programs including biophysical, hydrological those locally defmed) to climatic variability and change. and meteorological parameters on a regional, national and This research should also include wildlife and freshwater global basis, as well as coincident economic and social fish populations for which currently there is very little (including health) data are a necessary building block information available. towards reducing the uncertainties. These observation There is also a need for more attention on the implica• programs should combine relatively coarse data based on tions of climate change for heritage sites and reserves. It remote sensing with ground-based sampling over transects C.V Natural terrestrial ecosystems 67

on carefully chosen terrain covering the latitudinal/ A comprehensive observation program designed to altitudinal extent of selected ecotones and ecotypes. improve understanding of fundamental ecological processes Systematic observations of population dynamics and should include two elements: ecosystem responses should support and be supported by field experimentation and modelling. Specific attention • broad-scale, low intensity, randomly or regularly dis• should be paid to vulnerable and critical (including those persed sampling units (transects, plots, or points)to locally defined) habitats, typical ecotonal areas (eg Holten detect regional scale changes in the distribution and and Carey 1991) and species along steep climate gradients abundance of biota and distribution and integrity of including margins and isolated populations, and coastal and ecological zones; and island ecosystems as well as concentrating in transects • site-specific integrated monitoring projects located in across ecotones (desert-steppe/forest, savanna-steppe/ sensitive ecosystems. These intensively monitored sites forest, alpine timberline, forest-steppe and forest-tundra). would also contribute to the understanding of These data should be archived in such a manner that they cause/effect relationships between climate change, could be readily accessible. Heritage sites and reserves are biota and ecosystems that is necessary to interpret the excellent benchmarks against which to monitor the impacts changes detected through the element noted above. of climate change and, therefore, should be considered as a priority when establishing a network of benchmark sites. Peat dynamics is a key element in the prediction of future Monitoring of more sensitive ecosystems and species climate change. Studies of thermal and hydric regimes of should be given priority since they would provide early peatlands should have high priority for research and indication of impact of climatic change. systematic observation. In order to quantify the emission Remote sensing (satellite and radar, and the use of and storage of carbon by northern peat-rich ecosystems, GIS) can be extremely useful for some aspects of monitor• refinement is needed of their present extent and carbon ing ecosystem status and dynamics (eg geographic boun• content. daries of vegetation zones). Ground-based or field obser• Long-term ecological research sites should be present vation, however, is also essential as a means of ground- in each major ecological zone. Where possible, use should tmthing the remotely sensed data and for monitoring the be made of existing sites where long-term information is subtleties in ecosystem and species responses which cannot already available. Strengthening the international agree• be measured remotely. ments on ecosystem monitoring such as the US, USSR, Considerable progress could be made in a relatively Canada ecological transect in the northern latitudes should short time by assembling and analysing existing data on be promoted as a key component of the required research interactions among climate, ecosystem and species dynam• and monitoring program. Other monitoring programs, ics, and social and economic parameters with the objective including the Man and the Biosphere (MAB) Reserves and of identifying likely sensitivities and responses. The initial UNEP/TEMA's Global network, should be instituted focus should be on ecosystems and regions considered through appropriate agreements to develp a global stand• most likely to be affected in the short term. ardised dataset. These long-term ecological research and monitoring sites would provide opportunities for intensive process-oriented research in addition to monitoring of 5 Fundamental ecological processes effects. Research efforts should include model development An essential step in climate impacts assessments is to and assessment (both correlative and dynamic) and suppor• establish a consistent baseline of resources at risk through tive controlled environment experiments and field experi• assembling relevant inventories of species and ecosystems ments. Field manipulation techniques, at different levels on a national, regional and global basis. This will require within the ecosystem (ie at species, population or plant development and acceptance of consistent protocols for conununity level) are an essential component in addressmg identiiying and making inventories of the boundaries, the uncertainties in understanding ecosystem processes and extent and dynamics of plant and animal populations functions. (including disease and pest infestations) supported by field More field programs should focus on natural terrestrial 8nd laboratory manuals. Existing protocols should be ecosystem process research (eg what are the key climatic, evaluated for their sensitivity to the identification of the hydrologic, geomorphological, biological and chemical impacts of climate change, including whether or not they processes that occur within these ecosystems and their provide observation across current ecosystem boundaries interactions in determining ecosystem responses). Research 3nd include vulnerable species/communities. topics to be addressed at the intra-specific level include physiology, phenology, regeneration and reproduction. 68 C.y Natural terrestrial ecosystems dispersal and migration. At the inter-specific level, 7 Identifying social and economic consequences research should focus on changes in plant-animal interac• tion (eg pollination), predation and competition, decompo• To identify and assess the social and economic conse• sition, and pest and disease dynamics. quences of climate change there is a need for directed With regard to entire life communities and vegetation monitoring and research programs at the national, regional complexes, research questions to be considered include and global levels. These programs should be multi-disci• changes in surface covering, vegetation structure, species plinary, involving not only physical and biological scien• composition, biomass, diversity and succession patterns. tists, but also representatives from the social sciences, Experimental work on specific interactions or specific including economics. National and international research variables are also required to understand cause and effect funding should be capable of recognising and supporting because of the complex interactions between environmental this type of research. variables in nature. The International Tundra Experiment These research programs should explicitly address (ITEX) being developed for arctic studies, which includes impacts and consequences in the context of changes in non- basic systematic observations and environmental manipula• climatic factors (eg population changes, economic growth, tions is one example of the type of program required. land use changes, lifestyle changes). Also included is the need to develop more realistic projections of supply, demand and use of natural resources 50 to 100 years from 6 Identifying the linkages with climate change and now, and to develop methods for characterising non- greenhouse gas concentrations economic values of natural terrestrial ecosystems, such as scenic vistas, recreation, wildlife habitat and biodiversity. A focus for systematic observations of natural terrestrial Of particular interest when undertaking this type of ecosystems should be the identification of a set of wildlife research is information on the combined effect of both and vegetation indicator species which could be used for direct and indirect impacts on species and ecosystems. detecting climate-induced ecosystem changes and to assist Examining the impact of a particular variable on a static in quantifying vegetation, wildlife and freshwater fish environment and society, although interesting, is limited in sensitivity to climate and possible responses to climate its applicability. It would be more realistic to examine the change. Rather than trying to predict what will happen impacts and consequences with fixed societal and environ• when climate changes, focus should be placed on assessing mental changes or, better still, with a responsive society the sensitivity of species for climate change (thresholds/ and environment. These types of stiadies would lead to the limits) in order to select indicator species which may be identification of thresholds within both the natural terrestri• used to detect effects of climate change 'in the field'. In al ecosystems and associated social economic structures. selecting indicator species consideration should be given to Essential to these programs is the need to combine species which are expected to expand their range of biological monitoring that can identify the ecological distribution (horizontally and/or vertically), species which consequences of climate change with coincident data and are expected to retreat or become extinct, and species information on associated social and economic systems. As which have a wide distribution with many ecotypes which a first step, existing databases comprising coincident may increase or decrease in abundance within the present physical and biological data, as well as social and econom• range of distribution (Ketner and de Groot, in press). ic information, should be identified. Assembling these Once identified, the distribution of indicator species databases will provide a means of undertaking the required should be systematically mapped using a grid system that impacts research and assist in refining the definition of the lends itself to various modelling approaches, including required research and monitoring programs. computer-based models. To facilitate implementation of research and monitor• Related research should focus on identifying the ing programs directed at identifying and assessing social cause/effect relationships between climate, ecosystems and economic consequences of climate change, national processes (eg biophysical, biochemical, geochemical and regional teams of experts representing the physical, processes), functions (eg habitat, water quantity and biological and social scientific and economics communities quality), distribution and productivity and associated should be assembled to undertake impact studies in an wildlife and freshwater resources. Such research programs integrated fashion. This Can be accomplished through should be interdisciplinary by nature and not only assess strengthening existing programs at the global and regional the impacts of direct changes in climate, but also the levels. At the global level, the World Climate Programme effects of associated changes in soils, water regimes, and and in particular, its World Climate Impacts and Response landscape structure and functions. Strategies Programme should be encouraged to play a more active role in stimulating and facilitating regional assess- C. V Natural terrestrial ecosystems 69 ments. The estabUshment of regional programs such as the K.M. Peterson and D. LinviU (eds) Oimate change: the Landscape-Ecological Impact of Climate Change (LICC) biological implications. NATO/ARW Symposium, program for Europe should be encouraged. Clemson, South Carolina. NATO Advanced Studies Series. Coutant, C.C. 1990. 'Temperature-oxygen habitat for freshwater and coastal striped bass in a changing 8 Conclusions climate,' Transactions of the American Fisheries Society 119:240-253. Promotion of the high and irreplaceable values of natural Diamond, A.W. 1990. 'Wildlife as biological indicators for terrestrial ecosystems is essential in gaining public support assessing impacts of climate change,' in: G. Wall and M for sustaining them in a changing climate. Action is also Sanderson (eds.). Climate Change impUcationsfor water necessary to increase public awareness and to encourage and ecological resources. Tit^t. Geog Pub. Series, Occ. support in their defence, as well as for enhancing the Pap No. 11, University of Waterloo, pp. 225-233 transcending importance of climate change impact studies Diamond, A.W. and R. K. Brace. 1991. 'Climate and and the assessment of their consequences and of the waterfowl populations,' in: G. Wall (ed.) Symposium on response options. the impacts of climatic change arul variability on the Great Plains. Dept. Geog.Pub. Series, Occ. Pap No. 12, Managers, decision makers and local people need to University of Waterloo, pp. 245-257 be aware that policies and practices should be flexible to Everett, J.T., in press. 'Impacts of CUmate Change on accommodate the implications of climate change and the Living Aquatic Resources of the World,' prepared for: dynamics of this change. World Fisheries Congress, Athens, May 1992. In the near term, the identification of sensitive species, Healey,M.C. 1990. 'Implications of climate change for conununities and ecosystems should be one of the most fishery management policy,' Transactions of the important tasks undertaken. This sensitivity should be American Fisheries Society 119: 366-373. defined not only from an environmental perspective, but Hill, D.K. and J.J. Magnuson 1990. 'Potential effects of should also consider social and economic implications. As global climate warming on the growth and prey consump• far as possible, the scale at which sensitivities are defined tion of Great Lakes fish,' Transactions of the American should be consistent with that at which decision can be Fisheries Society 119:265-275. made. Holmes, J.A. 1990. 'Sea lamprey as an early responder to climate change in the Great Lakes basin,' Transactions of On the basis of these identified sensitivities, appropri• the American Fisheries Society 119: 292-300. ate response strategies should be developed and implement• Holten, J.L and P.D. Carey 1991. 'Responses of climate ed for those species, communities and ecosystems which change on natural terrestrial ecosystems in Norway,' face the most deleterious impacts (eg semi-arid and arid Draft report. Norwegian Institute for Nature Research. regions and tundra regions). Humans must be prepared to n*CC Impacts Assessment 1990 intervene where vital ecosystems or species are in jeopar• Johnson, T.B. and D.O. Evans 1990. 'Size-dependent winter dy. In developing and implementing response strategies, mortality of young-of-the-year white perch: climatic consideration should be given to a wide range of response warming and invasion of the Laurendan Great Lakes,' options, the specific local environmental; social and Transactions of the American Fisheries Society 119: 301- economic circumstances and the degree of uncertainties 313. and risks. Ketner, P. and R.S. de Groot, in press. 'The use of Consideration should also be given to reducing major bioindicators to detect climate change in N.W. European Ecotones,' Paper submitted to: Ecological Applications, human-caused stresses such as pollution and those resulting based on a manuscript prepared for the SCOPE Work• from the impacts of other human activities such as logging shop 'Ecotones in a changing environment: a workshop and the grazing of domestic animals at subsistence level. on ecotones and global change', April 26-29 1991, These stresses may originate outside the boundaries of the Michigan, USA affected ecosystem but often, once they are reduced, the Koshida, G. and L. Mortsch 1991. Climate change and water elasticity of that ecosystem increases, possibly decreasing level impacts on wetlands: a bibliography. Downsview: the impacts of climate change. Canadian Climate Centre, Atmospheric Environment Service, CLI-1-91, 50p. Livingston, D.A. 1975. 'The late quaternary climatic changes References in Africa,' Annual review of ecology and systematics. 6: 249-2. Bardecki, M. 1991. 'Wetlands and Climate Change: A Magadza, C.H.D. 1990. 'Some possible impacts of climate Speculative Review,' Can. Water Res. J.. 16(l):9-22. change on African Ecosystems,' in: Proceedings of the Callaghan, T.V. and Carlsson, B.A. in press. 'Plant popula• Second World Qimate Conference pp 385-390. tion dynamics in a changing climate at high latitudes,' in: 70 C.V Natural terrestrial ecosystems ^

Mathews, W.J. and E.G. Zimmerman 1990. 'Potential Schlesinger, D.A. and A.M. McCombie 1983. 'An evalu• effects of global warming on native fishes of the southern- ation of climatic morphoedaphie, and effort data as great plains and the southwest,' Fisheries 15, No.6: predictors of yields from Ontario sport fisheries,' Ontario 26-32. , Fisheries Technical Report Series No. 10. Ministry of Mortsch, L. 1990. 'Climatic change and variability: the Natural Resources, Ontario, Canada. effects of an altered water regime on the Great Lakes Shuter, B.J. and J.R. Post 1990. 'Climate, population coastal wetlands,' in: Climate change: implications for viability, and the Zoogeography of temperate fishes,' water and ecological resources, G. Wall and M. Transaction of American Fisheries Society 119:314-336. Sanderson, eds. University of Waterloo, Dept. of Ge• Wiersma, G.B. 1991. Feasibility Study of a Global Network^ ography Publication Series, Occasional Paper No., 11, p. of Integrated Monitoring Sites. Idaho National Engineer• 217-224. ing laboratory, Idaho Falls, Idaho. Sept. Osterreichische Akademie der Wissenschaften 1991. Wiersma, G.B. 1991. Scientific Advisory Committee on AnthropogeneKlimaanderungen:Moglicheauswirkungen Terrestrial Ecosystem Monitoring (SACTEMA), Work• auf osterreich—mogliche massnahmen in Osterreich ing Group, Evaluation. UNEP/GEMS. Sept. VI Hydrology and water resources

Co-chairmen and lead authors: E Stakhiv (USA) HLins(USA) I Shiklomanov (Russia) VI Hydrology and water resources

1 Introduction at present a full 'systems view' of climate and the hydrological cycle has yet to emerge .... Climate Since the IPCC Climate Change Impacts Assessment report models have yet to be developed which account for the (1990b), a substantial number of studies have been under• full hydrological cycle and its interactions with the taken that deal with the hydrologic effects and water atmosphere, oceans and land. resources management impacts of global warming. (Chahme 1992). Although the body of information is growing rapidly, espe• cially the proliferation of case studies of hydrologic Water resources management, by its nature, is intend• impacts, few new insights have been offered by these stud• ed and has served to ameliorate the extremes in climate ies, especially for water resources management impacts. variability. It stands to reason, then, that managed water• This is due largely to the reality that the analyses are based sheds, ie those under some form of water control and on General Circulation Model (GCM) outputs which sim• distribution, are apt to experience less dramatic water re• ply do not provide the information required for such sources impacts than those without such water management analyses, nor are the results compatible with one another systems even though the natural hydrologic response may for the five GCMs most frequently cited (US National be similar. However, specific regional changes in precipi• Academy of Sciences 1992). Overall, the recent studies tation and runoff are as yet impossible to predict with any have confirmed the findings and conclusions of the 1990 degree of confidence (US National Academy of Sciences Impacts Assessment, albeit with nuances and refinements 1992; Chahine 1992). For hydrologic studies, precipitation that were not explicitly considered then. is the variable of greatest interest, but GCM simulations of Among the difficulties in trying to interpret and precipitation for present climate conditions are notoriously compare the findings was that a great majority of studies poor (Wood et al. 1992). Much attention has been focused were conducted without regard to conventional impact on decreases in mean annual runoff, but some areas may assessment and evaluation principles, while also ignoring well experience increased runoff, at least during the IPCC global warming scenarios. This diminished their traditional seasonal peak flood period. Most model results direct value in trying to understand the hydrologic response do seem to forecast increased spring runoff for all the sensitivity, not to mention the comparative value in major rivers, although a comparative analysis by Kuhl and examining cross-regional impacts of global change on the Miller (1992) and Miller and Russell (1992) suggest the vulnerability of comparably managed water resources. The reason may be that at least when considering the GISS impacts of climate change on watershed response and the GCM, that model uniformly overestimates mean annual vulnerability of managed systems presents a greater degree streamflow and peak spring runoff when compared to of evaluation difficulty; the few studies that have attempted historical records. to deal with these issues should be viewed in the context of Thus far, the numerous impact studies are largely an evolving yet tenuous methodology. suggestive in nature, serving to prompt planners and policy Even though evidence for global warming and its long- analysts to search for worthwhile ways by which to range impacts is becoming more compelling with each new account explicitiy for and deal with potential changes in report, like it or not, there is still considerable disagree• average runoff and a greater variability in climate in the ment about the foreseeable regional hydrologic effects of foreseeable future. Water resources impacts are consid• global warming, not to mention the more complex socio• ered to be among the most significant consequences of economic and environmental impacts associated with water climate change, along with ecological impacts, affecting resources management. The reality is that generic global especially the all-important agricultural sector. Hence, a warming scenarios do not lend themselves to simple cause- degree of educated guessing is warranted as a way of and-effect extrapolations..There are countless simultaneous better preparing for the uncertainties that lie ahead, at least positive and negative feedback mechanisms to consider just in terms of preparing robust national strategies for coping to account for hydrologic processes alone. Water re• with uncertainty. Potential effects in the agricultural, sources management impacts add an additional level of forestry and urban sectors, although of justifiable concern complexity. to warrant separate inquiries, are actiially derivative prob• lems stemming from the fundamental effects of climate, The hydrological cycle has emerged as the central change on water availability (quantity, quality, distribution element in studies of climate change and timing); and we must not forget that while the long- term implications of global warming have focused on water yet deficits and droughts, there is reason to believe that climate variability could also increase considerably, so that C.VI Hydrology and water resources 73 societies must contend with large uncertainties at both ends The inability to forecast precipitation, runoff and of the hydrologic spectrum — floods as well as droughts. evapotranspiration accurately should not be used as an Since the first assessment of the IPCG (1990a,b,c), argument to negate the application of GCMs to river basm scientists participating in the Impacts Assessment Working level assessments, nor to diminish the value of undertaking Group II realised that producing an estimate for water re• hydrologic sensitivity analyses of the various postulated sources availability and other conditions extant within a climate change scenarios. Rather, the caveats are intended climate that may exist 50 or 100 years hence, could be to be cautionary in tenor - as a way of recognising that misleading at best, and completely imrealistic at worst. much too little is 'known' with certainty and spurious The reason is that the first step of analysing the cumulative conclusions are drawn too often from information that is, consequences of global warming requires that the physical at best, suggestive in nature. Towards this end, the focus changes of hydrologic effects associated with climate of future IPCC efforts should be to develop a more uniform change be differentiated from 'normal' climate variability methodology for impact assessments as part of a continuing as derived from historic records. Second, future water effort of targeting regional river basin and country studies resources management impacts (Water use, availability, in order to obtain a reasonably comparable basis for distribution, operation and maintenance etc) should be evaluation. viewed only as an increment (or decrement) to the water Hence, this update and assessment of recently avail• supply and use trends that are likely to occur under the able information is presented in the spirit of simply 'baseline' or 'without' climate change scenario. The compiling the latest information, without a rigorous indicators of water use that reflect the baseline condition, comparative analysis and evaluation, for it is simply too projected over a period of 50 to 100 years will, in all difficult to judge the relative merit of various inquiries Ukelihood, be dominated by population growth, economic inasmuch as a formal and consistent evaluative framework growth, and patterns of land use and settlemerit. These is yet to be formulated. There are no accepted standards socioeconomic changes would severely stress existing for analysis. However, a number of principles and water management systems even under a benign climate preliminary guidelines for assessing impacts of climate scenario. change were developed by an expert group assembled by Clearly, climate change could exacerbate an increas• the IPCC (1992d), but they have yet to be applied. ingly complex resource management problem. Yet, the expectation is that, despite all the large uncertainties that infuse this issue, water resources management opportuni• 2 Overview of new information ties for adapting to climate change are perhaps better understood and offer more realistic promise to mitigate This supplement to the first IPCC Impacts Assessment re• whatever adverse consequences that may materialise than port summarises recent materials (1990-1992) that were our collective ability to predict, understand and manage solicited and received from many countries (Australia, Bel• fiiture socioeconomic changes. This theme had been gium, Venezuela, Canada, Indonesia, Israel, New Zealand, emphasised first in the IPCC (1990c) Response Strategies Norway, Peru, Switzerland, Poland, Romania, Senegal, report and in subsequent reports by the US National USSR, USA, Uruguay, Finland, France, Chile, Japan), Research Council (NRG 1991), the US National Academy international organisations (WMO, UNESCO, IIASA, of Sciences (NAS 1992), American Association for the Ad• Commission on Mekong). Proceedings of international vancement of Science (Waggoner (ed.) 1990), and a con• and national symposia and meetings on this problem held ference of practising US federal water resources agencies during the period 1991-92 in Japan,Denmark, Australia, and water utility managers (Ballentine and Stakhiv (eds) Iceland, Uruguay, USA and Indonesia are also included, 1992). as are selected papers from conferences held on the subject The task of predicting regional hydrologic and water and supplemented by published literature. management consequences of climate change is a formid• The studies and analyses that were made available able one. At present, we can offer Only a semblance of subsequent to the IPCC Impacts Assessment which consensus on insights about how certain managed and assessed the hydrologic and water resources management uncontrolled watersheds could respond under a variety of consequences of climate change reflected the range of ap• climate change scenarios. Using the GCMs currently proaches that have been pursued in the past by scientists, available, it is not possible even to replicate accurately hydrologists and water resources engineers. These include: current average climate conditions, much less forecast water balance methods; use of GCMs to obtain estimates regional climatic conditions over annual and decadal time of changes in climatic and hydrologic characteristics; use scales (Lins et al. 1990; Kuhl and Miller 1992; Chahine of deterministic hydrologic models; statistical analysis of 1992). long-term records; use of palaeo-analogues; and a combi- 74 C. VI Hydrology and water resources nation of the aforementioned approaches. It is recognised conceptual models of arid and semiarid basins suggest that since the publication of the first IPCC Impacts Asess- that streamflow is less sensitive to climate change than ment, many studies have been initiated that rely on a previous statistical models have indicated. . . comparison of GCMs but, with few exceptions, the final results were not yet available for this report. Similarly, Furthermore, they conceded that changes in runoff to be many recent publications delve into better understanding of expected under the selected range of temperature changes historical climate changes as an implicit analog or baseline (up to ±4°C) would: for extrapolating future changes (eg studies of the El Nino phenomenon or the Sahelian drought). These studies were not be statistically different from the historical record not included as they did not explicitly deal with evidence unless precipitation changed by more than 10%. This of hydrologic changes induced by anthropogenic global confirmed a pathbreaking study by Klemes (1985) that warming. Most studies focused on understanding the had long been neglected by many analysts. sensitivity of hydrologic response to climatic perturbations. Far fewer studies attempted to take on the more compli• A critical aspect of water resources management is an cated problem of directly analysing the relationship understanding of the statistical properties of climate and between hydrologic watershed sensitivity and the vulnera• weather-driven events, such as floods and droughts. For bility of managed water systems. it is the extremes of precipitation and runoff that influence Methodological approaches varied widely, so it would design criteria and management strategies. Little research be inappropriate to infer that the studies portray distin• has been done linking climate change with frequency of guishable patterns of regional or geographic response, flood or drought, although several earlier attempts were despite the fact that the recent suite of water-resources- made to place the implications of climate change and related climate studies cover a much larger and diverse variability within the context of water resources system geographic area. It is impossible to. draw'any specific in• yield reliability using methods based on stochastic hydro- sights or conclusions about regional, river basin or water• logic analysis (Lettenmaier and Surges 1978; Matalas and shed-level impacts because of wide differences among the Fiering 1977; Nemec and Schaake 1982; Klemes 1985; GCMs and different analytical approaches that were used Fiering and Rogers 1989). A recent study by Wood etal. in these studies. In other words, the impacts will depend (1992) laid out a procedure for assessing the impact of on the characteristics of the river basin, the GCM selected, climate change on the flood frequency distribution for a and on the particular analytical method pursued. Given the river basin and tested the methods for CO, doubling large uncertamties associated with GCM outputs, it is scenarios produced by the GFDL GCM. First, the authors premature to use them for predictive purposes. At best, note that the GFDL GCM failed to reproduce the historical they may have a heuristic value for setting the stage for the climate characteristics for all seasons, uniformly underes• next phase of detailed hydrologic impact studies. timating precipitation for the two watersheds in the State Nevertheless, there are several recent noteworthy of Washington in north-westem US In fact, the differ• studies that shed some new light on the difficulties of ences between historical precipitation and the present GCM conducting hydrologic sensitivity studies and potential climate simulations were larger than the differences water management responses, and which serve as caution• predicted between the GCM IxCO, scenario (present) and ary reminders of a few principles and insights that may the IxCO, scenario. Hence, the authors caution that while serve future endeavours. For example, Kuhl and Miller the effect of GCM model biases on their approach may be (1992) compared the GISS GCM hindcasts with observed diminished by making relative comparisons, one could not runoff of the 20 largest rivers in the world and found be entirely certain because the models are highly nonlinear. systematic overestimation of the mean annual runoff. Notwithstanding the caveats, the authors concluded on the - Furthermore, for several of the rivers (Mississippi, Amur, basis of their test that stochastic transfer models offer a Colorado, Indus, Murray) the model's spring maximum quantitative means of linking large-scale circulation condi• runoff was significantly larger than observed, and the tions, as inferred from GCMs to local surface meteorolog• minimum summer and autumn runoff was too low. ical conditions, and can provide information for evaluating Nash and Gleick (1991) studied the hydrologic re• climate change impacts on flood frequencies. sponse of the Colorado River (US) and compared their re• A study conducted for a large semi-arid agricultural sults, using a well-calibrated water balance accounting area in the US, the Missouri-Iowa-Nebraska-Kansas model, with several previous statistical studies for the same (MINK) area, was important for several innovative river. The authors concluded that the previous studies approaches and implications for water management (DOE overestimated the decreases in runoff for the various 1991). The study approach was to postulate a recurrence scenarios involving increasing temperatures and that: of the most severe drought of record (1931-40) under C.VI Hydrology and water resources 75 conditions of a doubled COj climate in the year 2030, both that exceed, by a factor of two or more, the rates that with and without adjustments (such as water conservation, would be yielded by use of the more realistic consistent pricing, technological innovation etc). Detailed sectoral temperature. This explains why most GCMs have pro• water uses, demands and projections were developed. For duced excessively low values of summer soil moisture and the MINK area, irrigation withdrawals were projected to raises further questions conceming the results of studies of increase by 14% and consumptive uses by 23% without soil-moisture changes induced by an increase in greenhouse employing rational adjustments. Crop yields would decline gases. from 10 to 25% in the absence of adjustments, and plant Finally, in one of the few available studies of the requirements for irrigation water would increase by about vulnerability of managed urban water resources systems, 20%. However, as a compensating factor, it is expected Kirshen and Fennessey (1992) analysed the potential that the direct effects of increasing CO2 concentration from impacts of climate change on the water supply of the 350 to 450 ppm would reduce irrigation requirements by Boston metropolitan area, which represents a large, well- 10% (Easterling et al. 1991). Wollock and Homberger managed urban water supply system. The study included (1991) examined changes in assumptions about the effects a hydrologic analysis of various climate change scenarios, of plant stomatal resistance changes due to the CO2 enrich• comparing the outputs of several GCM models (GISS, ment effect on the sensitivity of catchment runoff. They GFDL, UKMO and OSU) and an associated analysis of found that lack of good information led to large variability water management impacts, including financial and eco• in the projections of runoff response. Most importantly, nomic Consequences and potential responses. The authors perhaps, is that despite all the uncertainties, the adverse re• showed that the GISS and GFDL models generally predict• gional economic impacts of such a drastic drought condi• ed decreased watershed yield whereas the OSU and tion was found to be small, in the order of several percent• UKMO GCM scenarios indicated increases in runoff and age points reduction in total regional production (DOE increases in system safe yield. As a result of these 1991, p27). contradictory outcomes, the Massachusetts Water Re• A somewhat different approach to/ analysing the sources Authority (MWRA) is taking a 'wait and see' hydrologic effects of climate change was undertaken by response similar to that of other urban US water utility McCabe et al. (1990). The authors examined the response managers (Schwarz andDillard 1990) because they believe of the Thomthwaite moisture index to climate change that the impacts of climate change are too uncertain arui scenarios generated by three GCMs (GISS, GFDL, OSU). the costs of action are too high to warrant such actions. Their reasoning was that the Thomthwaite moisture index However, the MWRA has engaged in an aggressive is a useful and early measurable indicator of water supply program of cost-effective demand management and water (precipitation) relative to its demand (potential evapotran• conservation actions that, in effect, represent the founda• spiration). Results showed that all three GCMs (steady tion of a sensible anticipatory adaptive strategy. Similar state CO2 doubling) predicted drier conditions for most of institutional responses are expected in most urban metro• the US, although the amount of decrease depends on the politan regions, whether the cause is growth in water GCM climate scenario selected. Results also suggested demand or in water shortages caused by droughts or global that changes in the moisture index are related mainly to climate change (Rhodes, Miller and MacDonnell 1992; changes in the mean annual potential evapotranspiration as Frederick 1992; Rogers 1992; and Stakhiv 1992). a result in changes in the mean annual temperature, rather than to changes in the mean annual precipitation. The greatest decrease in the index is expected to occur 3 Case study summaries in the northem US, especially in the Great Lakes region, but only small decreases in the moisture index will occur Studies have been carried out for the river basins and in the already dry areas of the south-west US. It is regions located in various natural climatic zones (tem• important to note at this point that the parameterisation of perate, high and middle latitudes, arid and semi-arid and continental evaporation in many atmospheric general circu• humid tropical) as well as by using palaeoclimatic scen• lation models is demonstrably inconsistent with observed arios for the continents of the Earth as a whole. conditions. In the turbulent transfer relation for potential Temperate zones are most comprehensively presented in evaporation, GCMs use the modelled actual temperature to the studies on European countries for the river basins evaluate the saturated surface humidity, whereas the where most of the annual runoff is formed as a result of consistent temperature is the one reflecting cooling by the snowmelt in spring. hypothetical potential evaporation. Milly (1992) demon• In Finland, a COj doubling scenario generated by the strated that in GCM experiments where soil moisture is GISS GCM model was used to estimate changes in runoff limited the models produce rates of potential evaporation for 12 watersheds of 600 km^ to 33 500 km^ Temperature 76 C.VI Hydrology and water resources increases of 2 to 6°C and monthly precipitation increases show that there are no statistically discemable changes in of 10 to 30 mm were produced by the GISS model. The annual river runoff in the basins of Volga, Dnieper, Don, results showed 20 to 50 % increases in mean annual runoff Ural, Zapadnaya, Dvina and Neman with a l''C warming. for individual watersheds, and a more stable interannual There is an expectation of increased annual runoff (by 5 to monthly runoff pattern for most watersheds (Vehvilainen 10%) on the rivers of the Caucusus and West and East and Lohvansuu 1991). Maximum spring discharges were Siberia. Somewhat larger increases are expected in the foimd to decrease considerably (up to 55 %) and minimum mountain rivers in Central Asia. However, the estimates monthly flows to rise dramatically (up to 100 to 300%). obtained for mountainous regions are very crude because Hydrologic sensitivity studies in Norway relied on a of great uncertainty of palaeoclimatic scenarios. comparison of two assumed (non-GCM) scenarios for CO, With a 2°C global warming, annual runoff in the doubling by the year 2030. According to the first scen• major river basins of the former USSR increases more ario, winter temperature would rise by 3.0 to 3.2°C, significantly. In particular, in the Volga basin annual summer by 1.5 to 2.0°C and armual precipitation by 5 to runoff increases by 5 to 7%; in the Dnieper, Don, 15%. According to the second scenario, by 3.5 to 5.0°C Dniester basins as well as in the Baltic River basins by 20 and 15 to 20%, respectively. The changes in runoff for to 25%; in Siberia and the Far East by 10 to 20%. This seven regions in Norway were calculated by usmg an trend also occurs in the mountain rivers in the Caucasus existing simplified conceptual hydrological model for a 30- and Central Asia. year historic period and the two climate scenarios As to changes in monthly and seasonal runoff, the (Saelthun et al. 1990). The results show that annual runoff most, detailed reconstructions have been made for the three increased somewhat (up to 10 to 15%) in the mountainous largest rivers of the European part of Russia: the Volga, regions and in the regions with large annual precipitation. Dnieper and Don. Proxy records indicate considerable In low forested areas, it decreased slightly. Seasonal potential change in seasonal runoff Winter runoff increas• runoff changes significantly: spring floods decrease; winter es dramatically and spring runoff decreases significantly runoff increases substantially, while summer runoff (Shiklomanov and Lins 1991), ostensibly in response to a decreases. The frequency of floods increases in autumn shift in the snowmelt season from spring to late winter as and winter (Saelthun et al. 1990). temperature rises. Kaczmarek and Krasuski (1991) obtained very similar For most of Canada, global warming would increase results for one of Poland's largest rivers, Varty (area is annual river runoff and its intra-annual distribution except 54 530 km^, using CO, doubling scenarios generated by for the Great Lakes basin, where according to all the GCM the GFDL model coupled with a monthly water balance scenarios, an increase in air temperature and decrease in hydrological model. With global warming the annual river runoff are most probable. This could result not only runoff of this river changes, but slightly. However, winter in a changed water balance and lake level, but also in runoff rises (21%) and summer runoff decreases (24%). serious economic and ecological consequences (Hartmann Similar results have been obtained for small and medium- 1990; Croley 1991; Quinn 1992). sized watersheds in Belgium and Switzerland. Studies The impacts of climate change on groundwater has were conducted by Bultot et al. (1991) and Gellens received little attention, although a recent evaluation of (1991a; 1991b) who applied GCM scenarios with the Australia's major aquifers represents an awakening to the conceptual hydrological model of the Royal Meteorological importance that groundwater recharge and management has Institute of Belgium with a diurnal time interval. For in the overall scheme for water resources management. example, comprehensive studies of an experimental basin Ghassemi et al. (1991) have noted that precipitation trends in Switzerland reveal relatively insignificant changes in have changed markedly in this century over Australia and annual runoff with increasing winter runoff by 10% and they expect trends and patterns to continue with global decreasing the summer one by 11% (Bultot et al. 1991) warming. Overall, they conclude that for a large area of For the nations of the former USSR, a set of improved Australia, especially the populated south-eastern quadrant, scenarios of future climate with a 1°C and 2°C global where groundwater is already overdrafted, groundwater warming based on palaeoclimatic analogues was developed recharge will increase. The shallow aquifers in the arid in 1991 by a group of climatologists guided by Professor and semi-arid zones of Australia (eg the Amadeus basin) M I Budyko. These scenarios were used to obtain ap• will also benefit from increased precipitation and recharge. proximate estimates of potential changes in annual and However, global warming will likely be detrimental for a seasonal runoff of main river watersheds. Methodo• number of major regional aquifers in south-western logically, these computations are based on water balance Australia, in the Perth basin and Murray basin, because of methods (for average annual runoff) and hydrological mod• decreased rainfall. els of river watersheds with 10-day time intervals. They C.VI Hydrology and water resources 77

Considering the qualitative results of the various tion increases up to 5%. With this scenario, it is unlikely studies of global warming effects on annual and seasonal that runoff will increase because evapotranspiration will re• river runoff in the northern and temperate zone of the sult in a net reduction in runoff. Northern Hemisphere, it would appear that in the regions In wet tropical regions, the assessment of global where a considerable part of annual runoff is formed wanning effects on water resources has been done for two during the spring flood, the intra-year runoff distribution river basins of Venezuela (Andressen and Rincon 1991); appears to be more sensitive to air temperature changes for the basin of the Uruguay River in Uruguay (Tucci and than to annual precipitation. In this connection, projec• Damiani 1991; the Lower Mekong River basin (Mekong tions of variations of seasonal runoff in these regions could Secretariat 1990); the Ellagawa catchment area in the cen• be more reliable than of the absolute magnitude of annual tral part of Sri Lanka (Nophadol and Hemantha 1992) and runoff because the GCM temperature forecasts are thought Indonesia (Rozari et al. 1990). In these studies, the to be more reliable than precipitation. Yet, research by scenarios used are those inherent in the selected GCM, Karl and Riebsame (1989) contradict that impression by with CO, doubling and hydrological implications estimated showing that hydrologic (runoff) variability is almost by using a variety of conceptual, deterministic simulation entirely controlled by precipitation variability. models of river basins. Estimates of global warming effects on hydrology in In the Sri Lanka study, Nophadol and Hemantha arid and semi-arid regions have been made for the southern (1992) chose to compare three GCMs (GISS, UKMO and area of the European part of the former USSR, Australia, GFDL) with respect to their ability to replicate historic some regions of the US, South America and Africa temperature and precipitation conditions. They concluded including Sahel (Nash and Gleick 1991; Urbiztondo et al. that for Sri Lanka the GFDL was least compatible and 1991; Australian Bureau of Meteorology 1991; Sirculon chose to continue analysis with the GISS and UKMO 1990; Vannitsem and Demaree 1991). The most recent models. The authors concluded that under a doubling of case studies provide mixed and often contradictory implica• CO2, the mean monthly and daily precipitation would in• tions, most likely stemming from the diverse methods crease during the peak rainy season, although no sig• applied and the divergence among GCM scenarios. nificant change in mean armual precipitation would occur. However, there are indications that the earlier conclusions The peak precipitation period would shift from April-July about the sensitivity of watersheds in arid and semi-arid to September-November. However, low flows would regions to even small changes in climatic characteristics increase, as would severity of droughts. may be overstated (Nash and Gleick 1991). Nevertheless, The study by Rozari et al. (1990) relied solely on the it is apparent that annual runoff appears to be more GISS GCM transifionalCOj doubling scenario. The sensitive to changes in precipitation than in air temperature authors selected this model because it generated rainfall for arid areas. ' results that best reflected the situation in Indonesia, Analysing the hydrological implications of global according to their judgment. The study analysed the warming in arid regions, the problem of the Sahelian zone overall impacts on hydrologic sensitivity for Indonesia and should be emphasised, as a severe drought of the past then focused on three catchment areas in Java. Based on decades caused disastrous consequences in many countries. a general water balance analysis for the three watersheds, The comprehensive analysis of climate change impacts on the analysis showed that, in general, the periods of water water resources of Sahel is presented in a study by deficits are shortened by at least a month, with significant Sirculon (1990) who emphasises a strong sensitivity of increases on monthly runoff for all months in all three river systems to changes in climatic characteristics and watersheds. An associated study (Rozari 1991) analyses in consequent vulnerability of the water management system. greater detail the agricultural impacts of increased precipi• As for the future of water resources in Sahel, there is a tafion and runoff citing increased erosion rates, reduction large amount of uncertainty because of contradictory data of soil productivity with consequent reduction in crop on future climate obtained by GCM computations and from yields of up to 9 %. Although the three watersheds are palaeoclimatic analogs. Palaeoclimatic analog information regulated by large reservoirs for flood control, hydro• suggests that global warming should lead to a large electric power irrigation, municipal and industrial water increase in precipitation in this region with relatively small supply and inland water fisheries, no analyses on water increases in temperature, resulting in a considerable management impacts were conducted. increase in runoff However, according to the outputs of Andressen and Rincon, in a personal communication three types of GCMs (IPCC 1990a) which show that the (1991), provide preliminary results of a study of climate air temperature in the Sahelian region rises by 1 to 2°C, change on the hydrology of three drainage basins in with a doubling of CO,, the models show winter precipi• western Venezuela, ranging from 1000 to 25 000 km^. A tation decreases of 5 to 10%, whereas summer precipita• relatively simple (6-parameter) monthly hydrologic simula- 78 C. VI Hydrology and water resources

tion model was used in conjunction with a Thomthwaite scenarios (exponential and arithmetic growth of COj). For water-balance model. Three different (non-GCM gener• all the hydrologic stations analysed, all the climate scenario' ated) climatic scenarios were developed: forecasts showed longer dry periods and wet periods. Also, the pattems of monthly rainfall generally shifted one (1) a low sensitivity scenario with no precipitation change to two months earlier in the monsoon season, along with and a +2°C increase; the time of occurrence of the driest and wettest months. (2) a medium sensitivity scenario with a 20 % precipitation Generally, rainfall increased significantly m the wet increase and a +3°C increase; and months, while the decrease in the dry months was smaller, (3) a high sensitivity scenario with a 40% precipitation although the duration of rainfall-deficient months in• increase and +4°C increase. creased. Reservoir operations would have to change to accommodate the greater variability in peak flows and low- Based on these hypothesised scenarios, the authors con• flow periods. Hydro-electric power production will be clude that the streamflow peaks will not shift; that average affected, but irrigation releases would not be affected. The annual streamflow will increase for the low and medium authors concluded that new reservoir operation rules would scenarios, but decrease slightly for the high sensitivity mitigate many of the potential adverse effects. scenario. In the study of the Uruguay River (Tucci and Damiani In perhaps the best existing suite of comparative water 1991), which was part of the triad of studies conducted for management studies available, the US Environmental the US Environmental Agency, the authors generally Protection Agency (1990) initiated a comparison of five followed the same approach as for the Mekong River. complex river basin management schemes of the La They compared three GCMs (GISS, GFDL, UKMO) with Plata/Uruguay, Mekong, Indus, Zambezi and Nile Rivers. both equilibrium and transient COj- doubling scenarios. Only preliminary analyses of the Mekong, Uruguay and They found that all the models underestimate total historic Zambezi Rivers were available for this assessment. A rainfall, but that the GISS model performed the best in similar effort, undertaken by the United National Environ• terms of replicating modem historic conditions. A hydro- ment Programme (Parry, Magalhaes and Ninh (eds) 1991) logic model was used to estimate flows in the Uruguay compared impacts for Brazil, Indonesia, Malaysia and River from the precipitation generated by the GCMs. The Thailand, although the analysis for Brazil examined only predicted flows indicated an 11.7 % decrease with the GISS the impacts of current climate variability and drought on model; a 21.5% increase with the GFDL and a 6.4% agriculture. Overall, the analysis for Southeast Asia was decrease with the UKMO model. The GISS transient conducted by comparing three GCMs (GISS, GFDL, and model also showed an 11.7% reduction in mean flows. OSU). COj-doubling analyses for Indonesia were focused All the GCM models predicted lower dry-season flows on agricultural impacts, showing higher levels of precip• for the Uruguay River, but the models differed widely on itation, which increased the area of potential agriculture, the inter-annual variation of flows and on the magnitude of but also increased soil erosion and loss of productivity. In peak flows. The study did focus on water management Malaysia, there did not appear to be a significant change impacts, as the basin is developed and managed for hydro• in the seasonal pattern in rainfall. Rice yields were electric power, flood control, irrigation and water supply. expected to decrease because of an increase in temperature . Much more hydro-electric power is planned for the basin, and decrease in solar radiation due to increased cloudiness. hence the impacts on hydro-electric production was Results of COj-doubling for Thailand (GISS model) shows considered very important. However, the different models a warming of 3 to 6°C, and a significant reduction in generated contradictory results. For example, the UKMO precipitation. Northern Thailand would be drier in most and GISS models predicted a decrease in hydro-electric months, which would be a benefit to agriculture in that power output of 2.5 to 4.75%, whereas the GFDL model area. Overall, the methods used were not comparable, and showed an increase of 17.3%. Three of the four model the results presented in these studies were inconsistent and scenarios predict smaller and fewer floods, resulting in highly variable among the three countries, as they were flood damage avoidance benefits. However, reduction in conducted by different groups. low flows and lengthening of low flow duration was The Mekong River is among the largest rivers in the forecast, showing decreases of 3 to 18%. This implies a world, with almost three-quarters of the watershed lying greater frequency of failure and water rationing, especially within the tropical zone; its climate is controlled largely by for irrigation withdrawals. seasonal monsoons. In a study conducted by the Mekong Analysis of the impacts of climate change on water Secretariat (1990), equilibrium scenario outputs of three resources management of the Zambezi River was con• GCMs (GISS, GFDL, UKMO) were compared for water ducted by Urbiztondo et al. (1991) for the US Environ• resources impacts analysis, along with two transient mental Protection Agency (1990). Preliminary results of C.VI Hydrology and water resources 79 a rainfall runoff simulation analysis and comparison of maintenance of existing systems; and those that modify the GCM scenarios generated by the GFDL, UKMO and GISS operational capacity of existing systems. There are many models showed that runoff based on GFDL and GISS uncertainties faced by water managers and planners, not steady-state CO^ doubling was consistently lower than the the least of which is future population and demand. base scenario, representing current climate. The UKMO Hydrologic uncertainty, ironically, may be among the less scenario projects an increase in river flow. The GISS important considerations. Transient A and Transient B scenarios show decadal varia• Sheer (1991), Fiering (1992), and Stakhiv (1992) tions in precipitation, with decades of increases and de• make similar points in that planning water resources creases. The Middle Zambezi River is regulated by the systems involves making many interrelated design and Kariba Reservoir, a large multipurpose storage dam. A operating decisions under uncertainty. These decisions are highly simplified reservoir optimisation model was used to made continuously and focus on the extremes and variabili• examine the impacts that GCM-generated inflows would ty of the present climate. Various levels of buffering have on hydro-electric power generation.' Kariba capacity, redundancy and resilience are built into individual Reservoir generates up to 70% of Zimbabwe's electricity, design criteria and components of a water management so that any pronounced reduction in runoff or changed project and its operation within a larger system as an seasonality Of flows would have a major adverse economic inherent hedge against an unforeseeable event. Hence, impact on that country. However, the simulation and com• managed water systems, while not entirely fail-safe, are parison of various GCM scenarios, both stationary and designed to be robust enough to guarantee a high degree of transient, showed that annual target energy demands, based reliability for rare or extreme events within the historic on a plant load factor of 65 % and an efficiency of 95 %, record. These rare events may become more frequent as would be met under virtually all scenarios. Nevertheless, part of a potential shift in the climate mean and variability. the impacts, during dry years would be severe, as is the But planners and designers are constantly adjusting their situation under present climate conditions, design criteria and, in effect, are anticipating subtle chang• es in the statistical properties of the evolving record of precipitation and streamflow. In addition, new and more 4 Water resources management impacts sophisticated techniques are employed to optimise the capacity, operating rules and economic efficiency of the Global warming is expected to catalyse changes in water upgraded water management systems. mapagement in many regions of the world. The quality Furthermore, water resources management is a highly and quantity of groundwater; the structure and character of dynamic and adaptive endeavour in the sense that new water consumption might alter; the conflicts and contradic• technologies, economic instruments, institutional and legal tions among individual water users and consumers would changes, and demand management measures are constantly heighten; and design and operational changes would need being introduced, tested and employed as water scarcity to be re-examined. All these issues have been treated, to becomes a chronic problem (Frederick 1992). Hence, the varymg degrees, for several river, basins with prescribed various strategies outlined in the Resource Use and climate scenarios both in the Northern and Southern Hemi• Management Subgroup of the IPCC Response Strategies spheres. Working Group III (1990c) are largely in place or are Climate uncertainty and water management was the . being implemented in most of the developed countries. topic of two recent conferences in the United States (NRC For a number of existing water management systems, 1991; Ballentine and Stakhiv (eds) 1992). Rogers (1992) comprehensive estimates have been developed, as examples addressed the role that uncertainty of climate change of potential climate change effects on water consumption information played in decision making related to water and use, socioeconomic and ecological consequences of management issues. Rogers essentially concluded that global warming and changes in hydrological characterist• even if the GCM models were scientifically well grounded ics. Few comprehensive and detailed water management and their prediction considered perfect, the information studies (as opposed to hydrologic sensitivity analysis) have provided is largely peripheral to practical engineering been conducted, employing the full range and hierarchy of decisions. The reason is that hydrology and availability of models required for such an undertaking. Among the water is only one among several important factors in water better studies are those of Kirshen and Fennessey (1992); management decisions, which include economic, political, Lettenmaier and Gan (1990); Lettenmaier et al. (1990); sociological, technological and demographic consideration. Lettenmaier and Sheer (1991) and Fiering and Rogers There are three categories of decisions in water resources (1989). planning and management: plaiming and design that goes Two studies (Lettenmaier and Gan 1990; Lettenmaier into new investments; those dealing with the operation and and Sheer 1991) explicitiy modelled large managed river 80 C.VI Hydrology and water resources basins in California, wherein hydrologic simulation models 5 Conclusions were coupled with snowmelt and soil moisture accounting Since the publication of the IPCC First Assessment Report, models developed by the US National Weather Service. several studies on impacts of climate change on hydrology The studies used GCM outputs and stochastically created and water resources have been conducted. Unfortunately, 100 years of rainfall and, temperature records from 30 there is not yet adequate information on regions affected years of historic data. Snow accumulation, ablation and by aridity and desertification and an effort should be runoff were the primary variables of concern. The undertaken to fill that gap. The new studies expanded the response of all four catchments to changed climate condi• geographic scope of the original surveys, while confirming tions were dominated by temperature-related changes in many previous conclusions; but few new insights were snowmelt and were relatively unaffected by GCM-predict- offered on hydrologic sensitivities and vulnerability of ed rainfall changes. Reservoir system performance for the existing water resources management systems. Sacramento-San Joaquin River basin was analysed using the hydrologic information. While the average annual The principal conclusions suggested by the new studies runoff did not change markedly, the shift in seasonal are: runoff to the winter months had a large impact on the reliability of water deliveries to the various water supply • Significant progress has been made in hydrologic users and on the flood control capabilities of the system. sensitivity analyses in developed countries, yet large Two conceptual studies for sizing the optimal capacity gaps exist in the information base regarding the and developing ideal operating rules for the design of implications of climate change for less developed hypothetical reservoirs under climate uncertainty were nations; undertaken by Fiering and Rogers (1989) and Lettenmaier • Comparative sensitivity analyses that rely on existing et al. (1990). Both studies attempted better to understand GCMs offer generic insights regarding the physical the vulnerability, resilience and robustness of reservoirs hydrologic effects and water resources management and managed systems imder climate uncertainty as a way impacts, but the differences in the outputs of the of improving the design of new systems and operation of GCMs coupled with large differences in hydrologic existing systems. Essentially both studies confirmed an sensitivity analyses makes it difficult to offer region- intuitive notion, that water supply performance, ie the specific impact assessments; minimisation of shortages or failures (or maximisation of • Temporal streamflow characteristics in virtually all reliability) is controlled more by reservoir storage capacity regions exhibited greater variability and amplification than by any variation in operating policy. Lettenmaier of extremes, with larger flood volumes and peak flows and Gan (1990) varied reservoir capacity from 0.25 to as well as increased flow episodes and a shift in the 0.50 of mean annual runoff. Fiering and Rogers (1989) turning of the seasonal runoff; showed that reservoir reliability did not change much with • The higher the degree of water control, regulation and changes in the climate-induced variation in runoff if the management of sectoral water demands, the smaller reservoir capacity was larger than 0.5 of the mean armual the anticipated adverse effects of global warming. runoff Conversely, unregulated hydrologic systems are more As a whole, the studies on water resources and water vulnerable to potential hydrologic alterations. management effects of climate change allow the corrobo• The principal recommendations are: ration of the following inference: under the condition of global warming and a great uncertainty of local climate • Increased variability of floods and droughts will variability, large water management systems are capable of require a re-examination of engineering design as• greater flexibility in adapting to changes in timing and sumptions, operating rules, system optimisation and magnitude of runoff. In this connection, the watersheds contingency planning for existing and planned water with a large degree of control over river runoff under management systems; global warming would have considerable advantages for • More studies on hydrologic sensitivity and water re• redistributing the available water supply and flood control source management vulnerability need to be directed compared to the regions with natural unregulated runoff towards arid and semi-arid regions and small island regimes. states. • A uniform approach to tlie climate change hydrologic sensitivity analyses needs to be developed for com• parability of results. C. VI Hydrology and water resources 81

References mate Change and Water Resources Management, US Army Instimte for Water Resources, Fort Belvoir, VA. Andressen, R. and B. Rincon 1991. 'Preliminary Assessment Gellens, D. 1991(a). 'Impact of a COj-Induced Climatic of the Impacts of Climatic Changes on the Hydrological Change on River Flow Variability in Three Rivers in Regimes of Three Basins in Western Venezuela'. Center Belgium,' Draft Paper, Royal Meteorologic Institute of for Tropical Climatology, University of Los Andes, Belgium. Venezuela, (Personal Communication). Gellens, D. 1991(b). 'Sensitivity of die Streamflow and of the Australian Bureau of Meteorology 1991. 37!^ Impacts of Sizing of Flood and Low Flow Control Reservoirs to the Climate Change on Hydrology and Water Resources. 2XCO2 Climate Change and Related Hypotheses: Study of Climate Change Impacts Workshop, June, 1991, Two Catchments in Belgium,' 17 pp. Melbourne, 76 pp. Ghassemi, F., G. Jacobsen and A.J. Jakeman 1991. Major Ballentine, T. and E.Z. Stakhiv (eds) 1992. Climate Change Australian Aquifers: Potential Climatic" Change Impacts. and Water Resources Management. Proceedings of a US IWRA Water International, 16(1) 38-44. Interagency Conference on Climate Change and Water Hartmann, H.C. 1990. 'Climate Change Impacts on Resources Management, US Army Institute for Water Laurentian Great Lakes Levels,' Qinmtic Change, 17, Resources, Fort Belvoir, VA. 49-67. Bultot F., D. Gellens, M. Spreafico and B. Schadler 1991. Intergovernmental Panel on Climate Change 1990(a). 'Repercussions of a CO2 Doubling on the Water Balance: Climate Change: The IPCC Scientific Assessment, J.T. A Case Study in Switzerland,' Draft Report of the Swiss Houghton, G.J. Jenkins and J.J. Ephramms (eds), Cam• National Hydrological and Geological Survey, Berne, bridge University Press, 365 pp. Switzerland. Intergovernmental Panel on Clunate Change 1990(b). Chahine, M.T. 1992. 'The Hydrological Cycle and Its Qimate Change: The IPCC Impacts Assessment, W.J. Influence on Climate'. Nature, (359), 373-380. McG. Tegart, G.W. Sheldon and D.C. Griffidi (eds), . Croley, T.E. 1990. 'Laurentian Great Lakes Double CO^ Australian Government Publishing Service, Canberra. Climate Change Hydrologicallmpacts'. Climatic Change, Intergovernmental Panel on Climate Change 1990(c). The IPCC Response Strategies, Response Strategies Working Department of Energy 1991. 'Process for Identifying Group in. World Meteorological Organization and the Regional Influence of and Responses to Increasing United Nations Environment Program, 270 pp. Atmospheric CO2 and Climate Change—the MINK Pro• Intergovernmental Panel on Climate Change 1990(d). ject: An Overview,' Report TR052A. Preliminary Guidelines for Assessing Impacts for Qimate Easterling, W.E., M. McKenney, N.J. Rosenberg and K. Change. T.R. Carter, M.L. Parry, S. Nishioka, and H. Leman 1991. 'A Farm Level Simulation of the Effects of Harasawa (eds). WMO/UNEP, 28 pp. Climate Change on Crop Productivity in the MINK Kaczmarek Z., and D. Krasuski 1991. 'Sensitivity of Water Region,' Department of Energy, Working Paper 11-13. Balance to Qimate Change and Variability,' IIASA Fiering, M.B. 1992. 'Runoff to Management', in: T. Working Paper. Ballentine and E.Z. Stakhiv (eds). Climate Change and Karl, T.R. and W.E. Riebsame 1989. 'The Impact of Water Resources Management, Proc. US Interagency Decadal Fluctuations in Mean Precipitation and Tempera• Conference on Climate Change and Water Resources \ ture on Runoff: A Sensitivity Study Over the United Mgmt., US Army Institute for Water Resources, Fort States', Qimatic Change, (15) 423-447. Belvoir, VA. Kirshen, P.H., andN.M. Fennessey 1992. 'PotentialImpacts Fiering, M.B. and P. Rogers 1989. 'Climate Change and of Climate Change Upon the Water Supply of the Boston Water Resources Planning Under Uncertainty', Draft Metropolitan Area', Draft Report to US Environmental Report to the US Army Engineer Institute for Water Protection Agency, 66 pp. Resources, Fort Belvoir, VA. Klemes, V. 1985. 'Sensitivity of Water Resource Systems to Fiering, M.B., and P. Rogers 1990. Climate Change and Climate Variations', World Chmate Applications Pro• Water-Resource Planning, in: Y.Y. Haimes and E.Z. gramme Publication No. WCP-98, World Meteorological Stakhiv (eds), Proc. Fourth Engineering Foundation Organization, Geneva, 17 pp. (and 5 appendices). Conference on Risk-Based Decisionmaking in Water Kuhl, S.C. and J.R. Miller 1992. 'Seasonal River Runoff Resources, American Society of Civil Engineers, New Calculated from a Global Atmospheric Model,' Water York, pp91-97. Resources Research, 28(8) 2029-2039. Frederick, K. 1992. 'Adaptive Responses to Climate Change: Lettenmaier, D.P. and S.J. Burges 1978. 'Climate Change: Demand Management', in: T. Ballentine and E.Z. Detection and Its Impact on Hydrologic Design'. Water Stakhiv (eds), Qimate Change arul Water Resources Resources Research, 14(4), 679-687. Management, Proc. US Interagency Conference on Cli• Lettenmaier, D.P., K.L. Brettman and L. Vail 1990. 'Robustness of a Multiple-Use Reservoir to Seasonal '1

82 C.VI Hydrology and water resources

Runoff Shifts. Associated with Climatic Change', Rhodes, S.K., K.A. Miller and L.J. MacDonnell 1992. pp91-97, in: Y.Y. Haimes and E.Z. Stakhiv (eds), Risk- 'Institutional Response to Chmate Change: Water Provid• Based Decisionmaking in Water Resources. Proc. er Organizations in the Denver Metropolitan Region,' Engineering Found. Conf., ASCE, New York, NY. Water Resources Research, 2S(l)n-l%. Lettenmaier, D.P. and T.Y. Gan 1990. 'Hydrologic Sensi• Rogers, P. 1992. 'What Water Managers and Planners Need tivities of the Sacramento-San Joaquin River Basin of to Know About Climate Chaiige and Water Resources California, to Global Warming'. Water Resources Res., Management,' in: T. Ballentine and E.Z. Stakhiv (eds), 26(1) 69-86. Qimate Change and Water Resources Management, Lettenmaier, D.P. and D.P. Sheer 1991. 'Climatic Sensitivi• Proc, US Interagency Conf on Climate Change and ty of California Water Resources'. ASCE J. Water Re• Water Resources Management, US Army Institute for sources Planning and Mgmt., 117(1) 108-125. Water Resources. Lins, H.F., LA. Shiklomanov and E. Stakhiv 1990. Impacts Rozari, M.B. 1991. 'Study on Socioeconomic Impacts and on Hydrology and Water Resources, Jaeger, J. and H.L. Policy Response Resulting from Climate Change: A Case Ferguson (eds), Qimate Oiange: Science, Impacts and Study in Three Districts of West Java's Northern Coastal Policy, Proc. Second World Climate Conference, Plain', Presented at Seminar on Global Climate Change - CambridgeUniversity Press, pp. 87-97. National Response Strategies, 8-9 Nov. 1991, Jakarta, Matalas, N.C. and M.B. Fiering 1977. Studies in Geophys• Indonesia. ics, in: Qimate, Qimate Qiange and Water Supply. Rozari, M.B. etal. 1990. 'Socioeconomiclmpacts of Climate National Academy Press, Washington, DC. Change: Indonesian Report', Report Submitted to United Mekong Secretariat 1990. 'A Study, on Impacts of Climate Nations Environment Programme, 53 pp. Change on Water Resources in the Lower Mekong Saelthun, N.R. et al. 1990. 'Climate Change Impact on Basin', Prepared by University of Colorado, Boulder, Norwegian Water Resources,' Norwegian Water Re• CO for US Environmental Protection Agency, 134. sources and Energy Administration, v. 42, 41 pp. McCabe, G.J., D.M. Wollock, L.E. Hay and M.A. Ayers Schwarz, H.E. and L-A. Dillard 1990. 'Urban Water,' in: 1990. 'Effects of Climatic Change on the Thornthwaite P.E. Waggoner (ed.), Oimate Change and US Water Re• Moisture Index', Water Resources Bulletin, 26(4) 633- sources, pp341-366. J, Wiley & Sons, New York. 643, J. American Water Resources Association. Sheer, D. 1991. 'Creative Water Management', in: 'Manag• Miller, J.R., and G.C. Russell 1992. 'The Impact of Global ing Water Resources in the West Under Conditions of Warming on River Runoff, Geopliys. Res., 97, 2757- Climate Uncertainty'. Proc. Colloquium of the National 2764. Research Council. National Academy Press, Washington, Milly, P.C.D. 1992. 'Potential Evaporation and Soil Moisture DC, pp290-295. in General Circulation Models'. J. Qimate 5(209-226) Shiklomanov, I.A. and H. Lins 1991. 'Climate Change Nash, L.L. and P.H. Gleick 1991. 'Sensitivity of Streamflow Impact on Hydrology and Water Management'. in the Colorado Basin to Climatic Changes', Hydrology, Meteorologia and Gidrologia. No.4, pp51-66, 125, 221-241. Sirculon, I. 1990. 'Impact Possible des Changement Qlima- Nemec, J. and J. Schaake 1982. 'Sensitivity of Water Re• tiques a Venir Sur les Resources in Eau des Regione source Systems to Climate Variation'. Hydrological Aridos et Semiarides,' WMO/TD-N, 380. Sciences, 27, pp. 327-343. Stakhiv, E.Z. 1992. 'Water Resources Planning and Manage- - Nophadol, I., and E.J. Hemiantha 1992. 'Impact Assessment ment Under Climate Uncertainty', in: T. Ballentine and of Global Warming on Rainfall-Runoff Characteristics in E.Z. Stakhiv (eds), Qimate Change and Water Resources a Tropical Region (Sri Lanka),' in: R. Herrmann (ed.) Management. Proc, US Interagency Conference on AWR Symposium Managing Water Resources During Climatic Change and Water Resources Management, US Global Change, pp547-556. Army Institute for Water Resources. Parry, M., A.R. Magalhaes and N.H. Ninh (eds) 1991. 'The Tucci, C.E.M. and A. Damiani 1991. 'International SUidies Potential Socioeconomic Effects of Climate Change; A on Climate Change Impacts: Uruguay River Basin', Summary of Three Regional Assessments', United Na• Presented in Proceedings of a Workshop on Analysis of tions Environment Programme. Potential Climate Changes in the Uruguay River Basin, Quinn, P.M. 1992. 'The Sensitivity of Water Resources Hydrologic Publication 25, Institute of Hydraulic Re• Management to Climate Change: A Great Lakes Case search, Federal University of Rio Grande do Sul, Brazil- Smdy.' in: T. Ballantine and E.Z. Stakhiv (eds), Qimate Urbiztondo, B.I., W.R. Rose and P. Restrepo 1991. 'Poten• Change and Water Resources Management, Proc, US tial Climate Change Impacts in die Middle Zambezi Interagency Conf on Climate Change and Water Re• River Basin. Report', Prepared for US Environmental sources Management, US Army Institute for Water Protection Agency, University of Colorado, Boulder, Resources. CO, 33 pp.

J C.VI Hydrology and water resources 83

US Environmental Protection Agency 1990. 'Progress Vehvilainen, R. and T. Lohvansuu 1991. 'The Effects of Reports in International Studies of Climate Change Climate Change in Discharges and Snowcover in Fin• Impacts'. Draft Report to the US EPA, University of land,' Hydrological Sciences Journ., 36(2)4, ppl09-121. Colorado, Waggoner, P,E. (ed.) 1990. Climate Change and US Water US National Research Council 1991. 'Managing Water Resources, Report of the American Association for the Resources in the West Under Conditions of Climate Advancement of Science. J. Wiley & Sons, New York, Uncertainty'. Proc. of a Colloquium. National Academy 496 pp. Press, Washington, DC, 344 pp. Wollock, D.M. and G.M. Hornberger 1991. 'Hydrological US National Academy of Sciences 1992. Policy Implications Effects of Changes in Levels of Atmospheric Carbon of Greenhouse Warming—Mitigation, Adaptation and the Dioxide', i. Forecasting, 10:105-116. Science Base. National Academy Press, Washington, Wood, E.F., J.P. Hughes and D.P. Lettenmaier 1992. 'An DC, 918 pp. Approach for Assessing the Sensitivity of Flood to Vannitsem, S., and G.R. Demaree 1991. 'Detection et Regional Climate Change'. Presented at MIT Sympo• Modelication des Secheresses au Sahel: Proposition d'une sium, The World at Risk: Natural Hazards and Climate Nouvelle Methogologique,' IRMB, Bruxelles. Change. Massachusetts Institute of Technology.