Objectives:

To develop methods of predicting crop yields on seasonal timescales, and of assessing the impacts of natural and anthropogenic climate change on crop productivity.

Work description:

UREADMY with UREADAG: Crop models for maize and peanuts, staple to the agriculture of Africa and India, will be driven by reanalyses and seasonal prediction ensembles. The sensitivity of these crops to extreme events will be studied. The specific effects of heat stress, in particular, may well be far more important in current semi-arid climates than is realised, and are likely to become increasingly important globally in future climates with warmer temperatures. The skill of seasonal forecasts in providing the necessary crop sensitive parameters will be assessed along with identification of potential shortcomings in the meteorological and crop numerical models. The ultimate aim is the development of an integrated system which applies probabilistic methods for the interpretation of seasonal forecasts in terms of crop development and yields.

CIRAD: In collaboration with the AGRHYMET Regional Center (Niamey, Niger), software has been developed to predict crop yields 1-2 months before harvesting. Using deterministic water balance models coupled to empirical models based on statistical correlations between water consumption and yield, 70 to 90% of the cereals' yield variability in Sahelian regions can be predicted. This technique is spatialized with METEOSAT real time processing (for rainfall estimation) and a Sahelian low resolution GIS for soil, climate, and crop parameters. This approach will be extended to take into account the seasonal rainfall prediction from numerical techniques. The accuracy of the predicted crop yields will be assessed with the aim to design new and real time recommendations to farmers.

The Met. Office: Using results from climate simulations driven with future CO2 emissions and land use scenarios, and which include a carbon cycle model, ecosystem changes (e.g. in gross and net primary productivity, plant and soil respiration, and plant functional type coverage) will be analysed. The relative sensitivity of the results to the impacts of CO2 increases vs. land use changes (e.g.deforestation) on climate and the remaining natural vegetation will be assessed.

Deliverables:

D3101: Identification of the important meteorological parameters for crop models and assessment of the potential shortcomings in seasonal prediction and crop models.

D3102: Progress towards the development of an integrated system which applies probabilistic methods for the interpretation of seasonal forecasts in terms of crop development and yields, as well as recommendations to farmers.

D3103: Assessment of the impacts of future CO2 emissions and land use scenarios on ecosystem changes, particularly crop productivity.

Milestones:

Year 1: Report on the changes in tropical vegetation due to increasing CO2. Test crop models with reanalysis data.

Year 2: Perform numerical experimentation on land use changes. Develop probabilistic methods for the interpretation of seasonal forecasts in terms of crop development and yields, including millet and maize.

Year 3: Report on progress towards the development of a combined seasonal weather and crop productivity forecasting system

Year 3: Report on relative importance of land use change and CO2 increase for tropical climate change and the implications for crop productivity.