Assessing Crop Loss Using a Climatic Risk Model

​

Making an objective analysis of crop loss is inherently challenging. To address this, we have developed a demonstrative mathematical model that may support informed discussions with institutions such as the Department of Economy.

1. As a first step, we compiled long-term climatic data specific to the region. The closest internationally recognized weather station is located near Simferopol Airport, with the station identifier 33946. We compared temperature measurements from this station with those recorded directly in the orchard, and the difference was found to be less than 0.4°C, confirming its reliability as a reference.

2. A comprehensive climatic dataset dating back to 1936 was acquired. For the purposes of this model, we filtered the data to focus on the last 25 years, starting from January 1, 2000. The dataset includes hourly measurements, resulting in a collection of over 200,000 individual observations.

3. Using the historical climatic dataset, we have compiled a list of commonly cultivated fruit crops in the Crimean region. This initial list can be expanded to include additional crops as needed.

4. To facilitate risk assessment, a parametric table was constructed. It includes each crop alongside its flowering period, with defined beginning and ending dates tailored to Crimean agro-climatic conditions. These dates can be easily adjusted if local phenological data suggests variation.

5. To quantify the impact of frost during flowering, we established four distinct frost risk profiles.

6. Based on long-term hourly temperature records, we calculated the probability of experiencing each frost risk category during the flowering window of each crop.

7. To further refine the model, we incorporated Crop Loss Coefficients for each of the 13 crops across the four frost categories. These coefficients were derived from five academic research studies and represent the expected severity of yield reduction due to frost exposure.

8. Finally, by combining the probability of frost risk occurrence with the crop loss coefficients, we computed the expected crop loss rates for each fruit type.

9. All values in this model are parametric. This means that any changes to input data—such as flowering dates, temperature thresholds, or coefficients—will dynamically update the entire system, ensuring flexibility and ease of adaptation.

10. Below are the current calculations based on the dataset.

​