(Approx. 30 min reading incl. references)
Crop phenology is a critical component of agricultural systems, and its response to environmental changes can have significant impacts on crop yields, quality, and productivity, as well as on the broader ecosystem services that support agriculture, such as pollination, pest control, and nutrient cycling (Gao and Zhang 2021). As climate change and extreme weather events accelerate, the need to understand and monitor crop phenology has become increasingly critical, especially in the frame of global food security. However, traditional ground-based field observations are often limited in their ability to capture the scale and complexity of these changes. RS technology offers a powerful solution to this challenge, enabling the monitoring and analysis of crop phenology at regional and global scale (Gong et al. 2024). In this section few examples from recent studies are provided.
RS-based phenometrics has been applied in various studies to classify crop species and rotation types. For instance, Li et al. (2021) developed an improved flexible spatiotemporal data fusion (IFSDAF) model to conduct data fusion using MODIS and Landsat imagery and extract NDVI time series with both high spatial and temporal resolution. The study proposed a Random Forest (RF) model and a decision-rule-based model for mapping crop species and rotation types, achieving an overall accuracy of 90% and 89.7%, respectively. Another study leveraged phenology information of existing data inventories using Time-Weighted Dynamic Time Warping (TWDTW) to address the problem of automatic crop sample generation, achieving promising results for classes with reduced inter-classes similarity (Belgiu et al. 2021).
In addition to crop classification or rotation detection, phenometrics has been used to assess the impact of climate change on agricultural production. A study from 2012 found that globally, 27% of cereal crop areas have experienced changes in the length of the growing season since 1981, with the majority of these changes resulting in longer growing seasons (Brown et al. 2012). The study also investigated the correlation between the peak timing of the growing season and agricultural production statistics, finding that variations in the peak of the growing season have a strong effect on global food production in certain countries.
Other authors investigated how phenology-based time windows affect corn yield predictions, using machine learning algorithms and multi-source environmental data from the U.S. Corn Belt. The study found that the use of phenology-derived crop growth windows significantly improves the accuracy of yield prediction by approximately 10% compared to the fixed-season method (Pei et al. 2025). Dimov et al. (2022) evaluated the use of three different sets of object-based predictors for sugarcane yield estimation, achieving R² of up to 0.84 for the estimation of yield and up to 0.82 for the estimation of sugar quantity through RF regression. In the last example, the authors evaluated the application of RS data to improve agroecological farming strategies under unfavorable weather conditions and biotic stress events. Their findings revealed that intercropping is more effective for maize in Kenya than traditional monocropping, and they recommended the widespread adoption of this system (Liepa et al. 2025).