Irrigation demand modelling using the UKCP09 weather generator: lessons learned

Abstract

The determination of irrigation demand is typically based on crop modelling using a long historic record of local daily weather data. However, there are rarely adequate weather station records near to given sites; often any local records cover a limited number of years, are incomplete, costly or are of poor quality. This paper examines whether version 1 of the UKCP09 weather generator can provide a simpler and effective method of calculating irrigation demand with sufficient accuracy for regulatory and design purposes. The irrigation demands at seven sites distributed around England were modelled using the UKCP09 baseline climatology and compared with results modelled using daily observed weather records. For the design dry year used for irrigation planning, the weather generator replicated the observed conditions with reasonable accuracy. The weather generator was however less successful at replicating extreme dry years. These results are encouraging but also provide a note of caution for the use of these generated datasets for studying current irrigation demand and by implication for modelling future needs under climate change. The study also demonstrated a simple sub-sampling approach for reducing the processing demands if using the dataset in more complex models, though this would not remove any underlying error.