Integrated modelling of farm adaptation to climate change in East Anglia, UK: Scaling and farmer decision making

We argue that an effective representation of farmer adaptation to climate change can only be achieved by integrating different modelling approaches that are spatially and temporally multi-scale and dynamic. We introduce a modelling framework that addresses these issues. To illustrate our approach, we simulate a group of farms representing a catchment in the East Anglian region of England, with and without water trading (spatially multi-scale). Different modelling approaches are used to represent weather (year-to-year variation) and climate effects (predicted crop yields for the 2020s and 2050s). Dynamics are introduced by (i) allowing the outcome of previous years’ decisions to inform decision making in subsequent years and (ii) by modelling large investment decisions (buildings and irrigation), as longer term commitments. Variability and uncertainty are captured by running the framework multiple times with inputs drawn at random. Example results demonstrate that farm-level allocation of abstraction licences constrains water abstraction at the catchment level—with water trading summer water abstraction increased from 66.4% (compared to 62.2% with no trading) for the baseline period, through 75.6% (65.9%), for the 2020s, to 93.2% (72.5%) for the 2050s. Hence, modelling the catchment as a single unit would have over-estimated water abstraction. Under the model assumptions investment in winter abstraction capacity was not justified. Due to their high value, irrigation levels were maintained on potatoes (Solanum tuberosum L.), resulting in relatively stable crop areas over time; in contrast, for the lower value sugar beet (Beta vulgaris L.) yields and areas were more variable due to their dependence on variable weather and surplus water availability. There was no major shift from native cropping to exotic crops (oilseed rape [Brassica napus L.] to sunflowers [Helianthus annuus L.]) in either the 2020s or 2050s. We conclude that omitting farm-level constraints in regional models may overestimate the degree of adaptation possible and underestimate the negative effects of climate change.