Simulating Adaptive Management using Impact Models in a Risk Framework


Nina Pirttioja

Suomen ympäristökeskus


Model estimates of future crop development and yield are uncertain because key parameters, assumptions about crop management and projections of future climate can all vary across a range of values. This study presents an approach for undertaking a simultaneous analysis of parameter uncertainty, sensitivity to climate and probabilistic climate projections, which can be used to offer insights into future risks of yield shortfall and potential adaptation options.


The approach involves combining joint probability density functions (pdfs) for future time periods, describing uncertainties in ensemble climate model projections of temperature and precipitation change, with impact response surfaces (IRSs). Each IRS is constructed based on a sensitivity study of modelled crop yield as a function of temperature and precipitation relative to the present-day, and assuming a scenario of future CO2 concentration. Uncertainties in yields are represented by constructing IRSs for multiple combinations of plausible crop and management parameter values applied to a crop model.


The risk of yield shortfall is estimated by defining a threshold yield level, superimposing the climate pdf on a given IRS and integrating over the combinations of climate change that generate yields below the threshold. Using this procedure for the same IRS over different time periods produces an estimate of evolving risk attributable to uncertainties in future climate. Repeating the exercise across multiple IRSs, generates an ensemble of risk estimates encapsulating both climate and crop model uncertainties.


The method is applied to yields of barley in south-west Finland, simulated using the WOFOST crop model for a locally grown cultivar. Crop yields and climate change uncertainties are presented for 10-year intervals into the future assuming an SRES A1B scenario. Some options for adapting cropping to changing climate are investigated by matching tested parameter combinations to simulations using parameters appropriate for alternative crop cultivars grown outside Finland.