Other reports in this collection
Reilly et al. (1996) review several studies that analyzed the effects
of climate scenarios at the farm scale. One important implication of these studies
is that changes in production activities and management adaptations could play
an important role in mitigating the impacts of climate change.
Survey-based studies on farm-level decisionmaking in response to climate change
have propagated since the SAR. Case studies in Canada suggest several farm-level
adaptive strategies to which farmers may resort (Brklacich et al., 1997;
Smithers and Smit, 1997; Chiotti, 1998). In southern Ontario, Smit et al.
(1997) found that farmers' choice of cultivar to plant in the current year
is highly conditioned by the climate experienced the year before. There have
been few quantitative, farm-level economic studies of response and adaptation
to climate change since the SAR. In the U.S. midwest, Doering et al.
(1997) used a crop-livestock linear programming model linked to the Century
biogeochemistry model to investigate the impacts of climate change on 11 representative
farms in the region under a doubled-CO2 scenario. The analysis considered
these farms under static technology and adapted technology; thus, the analysis
could be regarded as representing farm-level economic responses to climate change
as well as farm-level combined technological and economic adaptation. It showed
that climate change may cause substantial shifts in the mix of crops grown in
the upper midwest, with much less land planted to a corn-soybean rotation and
more land devoted to wheat than now. Earlier planting of corn increased returns;
hence, a more frost-resistant corn variety was found to be important to farm-level
adaptation. Using a similar methodology, Parry et al. (1996) predict
optimal agricultural land use in response to climate change in England and Wales.
Antle et al. (1999b) used an econometric-process simulation model of the dryland grain production system in Montana linked to the Century model (as reported in Paustian et al., 1999) to assess the economic impacts of climate change in that region. Unlike other farm-level studies, this analysis is based on a statistically representative sample of commercial grain farms in the region, not on a small number of representative farms, so the results can be interpreted as representing the entire population of farms in the study area. Moreover, instead of using linear programming models, this analysis combines site-specific data and econometric production models in a stochastic simulation model, allowing representation of physical and economic heterogeneity in the region. Simulations were conducted for baseline and doubled CO2 (Canadian Climate model) with observed production technology, with and without land-use adaptation and with and without CO2 fertilization. With climate change, CO2 fertilization, and adaptation, mean returns change by -11 to + 6% relative to the base climate and variability in returns increases by +7 to +25%, whereas without adaptation mean returns change by -8 to -31% and variability increases by +25 to 83%. These findings provide support for the hypothesis that ability to adapt plays a critical role not only in mean impacts but also in the spatial variability of impacts. They provide empirical support to the hypothesis advanced in the SAR that climate change is likely to have its greatest adverse impacts on areas where resource endowments are poorest and the ability of farmers to respond and adapt is most limited.
None of the economic studies in Reilly et al. (1996) analyze the environmental
consequences of adaptation to climate change, such as increased demands on land
and water resources. Adams et al. (1998) observe that this shortcoming
remains true of most recent studies. However, Darwin et al. (1995) do
include an analysis of impacts of climate change on land and water resources
in a global model with the world subdivided into eight regions. They argue that
competition from crop production could aggravate direct climate-induced losses
of forests in moist tropical regions. IMAGE 2.0 simulations with future scenarios
of limited CO2 emissions show that increased deforestation increases
agricultural capacity because of a smaller CO2 fertilizaton effect
on crops than if emissions continue on the current trajectory (Leemans, 1997).
Lewandrowski and Schimmelpfinnig (1999) draw implications for impacts on land
and water resources, wild species, and natural ecosystems from the literature.
They suggest that increased demand for irrigation predicted by these studies
is likely to increase the opportunity cost of water and possibly reduce water
availability for wildlife and natural ecosystems. However, it is difficult to
go beyond such generalities with these aggregate models because most environmental
impacts of agriculture are site-specific. Strzepek et al. (1999) show
that some scenarios of climate change may reduce irrigation system reliability
in the lower Missouri River in the U.S. corn belt, which may induce instream
environmental stress. In many developing countries, current irrigation efficiencies
are very low by developed-country standards. Irrigation effeciency in the Philippines
in 1990 was 18%, compared to the global average of 43% (Asian Development Bank,
1998). Currently, 3,480-5,000 liters of water are used to produce 1.0 kg
rough rice (equivalent to 640 g milled rice) in the Philippines (Baradas, 1999)
and some neighboring countries. At those irrigation efficiencies, increased
irrigation demand caused by climate change would strain irrigation supplies.
Hence, one adaptation strategy is to increase irrigation efficiency.
Agriculture is a source and a sink of GHGs; hence, climate-induced agricultural land-use change is likely to impact soil carbon stocks in agricultural soils (Paustian et al., 1996; Lal et al., 1998; IPCC, 2000). Antle et al. (1999b) and Paustian et al. (1999) link a field-scale econometric-process simulation model to the Century ecosystem model to assess the impacts of climate change (from the Canadian Climate Centre model) on soil carbon in the dryland grain production system of the U.S. northern Great Plains. In a related set of studies, Antle et al. (2000) and Paustian et al. (2000) link a regional economic agricultural land-use model to the Century ecosystem model to assess the impacts of climate change on soil carbon in central U.S. cropland. These studies demonstrate that adaptive changes in land use and management are likely to have greater impacts on soil carbon than the direct effects of climate. Thus, adverse effects of climate change on soil carbon tend to be offset by the adaptive changes in land use that would be made by farmers in response to climate change. The degree of this offset depends on the magnitude of CO2 fertilization effects on crop yields.
Adaptation is unlikely to come without cost. In a literature survey, Tol et al. (1998) conclude that adaptation costs (as opposed to net costs of damages) are not reported in most impact studies, especially in agriculture. Yet transition costs (e.g., to retrain farmers in new practices) and equilibrium costs (e.g., to develop additional irrigation or apply more fertilizer) may be considerable. The absence of a benefit-cost calculus for agricultural adaptation is a key deficiency. Existing studies also fail to account for the process of long-term, endogenous adaptation of technology in ways that are consistent with the extensive economic literature on that subject (Antle, 1996b). This process also will involve significant costs. An extensive body of economic research has studied the benefits and costs of agricultural research and has shown that institutions that are responsible for agricultural research adapt agricultural technology across space and time in response to relative resource scarcity (Hayami and Ruttan, 1985). Quiggin and Horowitz (1999) argue that changes in fixed capital for on-farm and off-farm infrastructure may be the most significant cost associated with adaptation to climate change.
|
Other reports in this collection |
|