Predicting walnut (Juglansspp.) crop yield using meteorological and airborne pollen data
Crop yield determines economy by influencing prices on the trade market, and so accurate forecasts of the yield are important for planning various aspects of agricultural production. The main aim of this study is to construct a model for predicting walnut yield in an important walnut production area (the region of Novi Sad in Northern Serbia). Relationships between the amount of walnuts produced annually (2000–2011) and abiotic (e.g. meteorological) and biotic (e.g. airborne pollen data) factors were examined using Pearson correlation analysis. Walnut yield data were then entered into linear regression models with variables that had the highest correlations. The models were constructed using 10 years of data, and tested using 2 years of data not included in constructing the model. This paper has shown that walnut yield is greatly dependent on weather conditions, particularly during fertilisation and seed growth, but the amount of available airborne pollen also plays an important role. The introduction of the seasonal pollen index, as a proxy for the amount of pollen available for fertilisation, improved the performance of models predicting walnut yield.