Fine-grained multi-factor hail damage modelling

Abstract

A fine-grained multi-factor estimation of crop-hail damage is required to progress from manual inspection of crops post-event to automated assessment and accurate forecasting of the expected impact on agricultural areas. Such automated processes will enable more accurate claims processing, improve customer satisfaction, and reduce insurance losses. This paper demonstrates the value of Gaussian Processes for the construction of such a multi-factor model of crop-hail damage. This is underpinned by a survey of public datasets, and a description of the target dataset to support an operational crop-hail damage model.