Predicting Climate Fluctuations and Agricultural Impacts

The key weather variables for crop prediction are rainfall, temperature and solar radiation, with humidity and wind speed playing also a role. As Doblas-Reyes et al. (2006) explained, seasonal climate forecasts are able to provide insight into the future climate evolution on timescales of seasons and longer because slowly-evolving variability in the oceans significantly influences variations in weather statistics. The climate forecast community is now capable of providing an end-to-end multi-scale (in space and time) integrated prediction system that provides skilful, useful predictions of variables with socio-economic interest.

Seasonal forecasts can be produced using mathematical models of the climate system. A wide range of forecast methods, both empirical-statistical techniques and dynamical methods, are employed in climate forecasting at regional and national levels (WMO 2003). Operational empirical-statistical methods, based on statistical links between current observations and weather conditions some time in the future, include analysis of general circulation patterns; analogue methods; time series, correlation, discriminant and canonical correlation analyses; multiple linear regression; optimal climate normals; and analysis of climatic anomalies associated with ENSO events. Dynamical methods (used principally in major international climate prediction centers) are model-based, using atmospheric GCMs in a two-tiered prediction system, or dynamically coupled atmosphere-ocean GCMs. These dynamical forecast models - an extension of the numerical methods used to predict the weather a few days ahead -are based on systems of equations that predict the evolution of the global climate system in response to initial atmospheric conditions, and boundary forcing from the underlying ocean and land surfaces.

Doblas-Reyes et al. (2006) emphasize the importance of a fully probabilistic approach during all the stages of the forecasting process. Predictions of the climate system evolution in seasonal timescales suffer mainly from two sources of uncertainty: initial condition and structural model uncertainty (Doblas-Reyes et al. 2006). To address the first source of uncertainty, forecast models are run many times from slightly different initial conditions, consistent with the error to estimate the effect of this initial-condition uncertainty. One way to represent model uncertainty is to incorporate, within the ensemble, independently derived models, resulting in a multi-model ensemble system (Palmer et al. 2004). Hagedorn et al. (2005) and Doblas-Reyes et al. (2005)

showed that the DEMETER multi-model ensemble system, made up with 7 European coupled models, is intrinsically more useful and more skilful than forecasts from any one (e.g. national) model. "Forecast assimilation" deals with statistically combining multiple dynamical and statistical forecasts to maximize the content information (Stephenson et al. 2005).

For agriculture, climate forecasts must be interpreted in terms of production outcomes at the scale of decisions if farmers and other agricultural decision-makers are to benefit. Interest in linking seasonal climate forecasts from general circulation models (GCMs) with crop models is motivated by: (a) the need for information that is directly relevant to decisions, (b) use for ex ante assessment of potential benefits to enhance credibility and support targeting, and (c) support for fostering and guiding management responses to advance climate information (Hansen 2005).

At the time of the inaugural CLIMAG workshop in 1999, nearly all quantitative efforts to translate seasonal forecasts into agricultural terms and assess the value of management responses have used categorical indices of ENSO to select historic analog years as inputs to crop models. Interest in incorporating forecasts based on dynamic climate models were slowed by concerns about the difference in spatial and temporal scale of GCMs and crop models, and the dynamic, nonlinear, often nonmonotonic relationship between meteorological variables and crop response. However, the past five years have seen increasing interest and some methodological advances in using dynamic climate model output as input to process-level crop models, synthetic daily weather or daily climate model output to drive the crop model; statistical transfer functions trained on crop model predictions run with historic weather data; and variations on the analog method that include weather classification, hidden Markov models and probability-weighted historic analogs (Hansen et al. 2006).

Several avenues are likely to enhance the quality of forecasts of agricultural impacts of climate variations over the next five to ten years. First, dynamically coupling crop models within climate models will support refined two-way interaction between the atmosphere and agricultural land use. Second, remote sensing and proliferation of spatial environmental databases provide substantial opportunities to expand the use and enhance the quality and resolution of climate-based crop forecasts. Finally, climate-based crop forecasts will benefit from climate research in the emerging area of "weather within climate."

Renewable Energy Eco Friendly

Renewable Energy Eco Friendly

Renewable energy is energy that is generated from sunlight, rain, tides, geothermal heat and wind. These sources are naturally and constantly replenished, which is why they are deemed as renewable.

Get My Free Ebook


Post a comment