Improved seasonal rainfall prediction for grain growers using farm level data and novel modelling

作为澳大利亚重要的出口商品，谷物的成功生产在很大程度上取决于可靠的季节性预测。然而，高度多变的气候意味着，对于澳大利亚的2.5万名粮食种植者来说，目前的预测在空间和时间上缺乏细节。该项目将结合模糊分类和人工神经网络，利用1.7万名粮食种植者协会成员提供的高密度气候数据和气象局提供的海洋表面温度等气候驱动因素，开发一个本地详细的、不断更新的数据驱动的季节预报系统。根据观察到的数据进行验证后，预测将通过基于web的门户传递给用户。

Successful grain production, a key export commodity for Australia, depends heavily on reliable seasonal forecasts. However, the highly variable climate means that for Australia’s 25,000 grain growers current forecasts lack detail in space and time. Using a combination of fuzzy classification and artificial neural networks, this project will develop a locally detailed continuously updating data-driven seasonal forecast system using high density climate data from the 17,000 Grain Growers Association members and climate drivers such as sea surface temperature from the Bureau of Meteorology. After validation against observed data, the forecasts will be delivered via a web-based portal to users.