Agricultural insurance is crucial for transferring a significant portion of the risk due to unfavourable weather conditions outside the farm. Index-based insurances were proposed as an alternative to traditional products based on direct damage assessment because of their potential to reduce insurance costs while being unaffected by subjectivity during damage quantification. However, they may be affected by basis risk when indices are poorly related with the underlying biological processes and thus with actual yield losses. To overcome this limitation, we developed a new framework to derive indices characterized by a low basis risk, based (i) on the use of a complex biophysical crop model extended to account for the extreme weather events of interest and (ii) on crop- and region-specific meta-models derived from the input-output structure of the crop model. The procedure was applied to frost and drought-heat damage to barley, soft and durum wheat in two Italian regions. Meta-models resulted easy to understand for farmers and accurate in reproducing percentage yield losses, with mean normalized relative root mean square error equal to 27.0% and 39.0% for damage caused respectively by frost and by the combined effect of drought and heat. The procedure, successfully adopted in operational contexts for the injuries and crops considered in this study, is going to be extended to other weather events, crops and regions.

农业保险对于转移由于农场外部恶劣天气导致的大部分风险至关重要。基于指数的保险被提议作为基于直接损害评估的传统产品的替代方案，因为它们可以降低保险成本，同时不受损害量化过程中的主观影响。但是，当指数与基础生物过程之间的关联性不佳，从而与实际产量损失之间的关联性不佳时，它们可能会受到基础风险的影响。为了克服这一局限性，我们开发了一个新框架来导出具有低基础风险的指数，基于（i）使用复杂的生物物理农作物模型来扩展以解决感兴趣的极端天气事件，以及（ii）基于从农作物模型的投入产出结构得出的特定于作物和地区的元模型。该程序应用于意大利两个地区对大麦，软小麦和硬质小麦的霜冻和干旱热损害。元模型使农民易于理解，并能准确地再现产量损失百分比，因霜冻以及干旱和高温的综合影响，平均归一化相对均方根误差分别为27.0％和39.0％。该程序已在本研究中考虑的伤害和农作物的操作环境中成功采用，并将扩展到其他天气事件，农作物和地区。该程序应用于意大利两个地区对大麦，软小麦和硬质小麦的霜冻和干旱热损害。元模型可以使农民容易理解，并能准确地再现产量损失的百分比，因霜冻以及干旱和高温的综合影响，平均归一化相对均方根误差分别为27.0％和39.0％。该程序已在本研究中考虑的伤害和农作物的操作环境中成功采用，并将扩展到其他天气事件，农作物和地区。该程序应用于意大利两个地区对大麦，软小麦和硬质小麦的霜冻和干旱热损害。元模型可以使农民容易理解，并能准确地再现产量损失的百分比，因霜冻以及干旱和高温的综合影响，平均归一化相对均方根误差分别为27.0％和39.0％。该程序已在本研究中考虑的伤害和农作物的操作环境中成功采用，并将扩展到其他天气事件，农作物和地区。对于因霜冻以及干旱和高温的共同影响而造成的损坏，赔偿金为0％。该程序已在本研究中考虑的伤害和农作物的操作环境中成功采用，并将扩展到其他天气事件，农作物和地区。对于因霜冻以及干旱和高温的共同影响而造成的损坏，赔偿金为0％。该程序已在本研究中考虑的伤害和农作物的操作环境中成功采用，并将扩展到其他天气事件，农作物和地区。