New Remotely-Sensed Methods to Predict Crop Growth Stages
Understanding crop growth stages through timely and accurate data is key to agricultural system modeling as well as crop management practices, which rely heavily on accurately-timed farming methods coinciding with the stage of crop development. Furthermore, crops in different stages of growth have varied responses to environmental stresses caused by water resources, temperature conditions, and soil nutrient availability, making information on the current growth stage vital to farmers’ ability to properly schedule irrigation and fertilizer management methods for increased yields. Although satellite data has been used in the past to acquire vegetation phenology metrics such as green-up (increased vegetation), this information is different from the data on crop growth stages that are used to inform crop management and crop growth models.
Motivated by this data gap, scientists supported by the University of Maryland, NASA Harvest, and the University of Nebraska-Lincoln and led by Varaprasad Bandaru (UMD/NASA Harvest) teamed up to develop a new approach to estimating crop stages using remotely sensed satellite data, dubbed the “PhenoCrop” framework. In their recently-published paper, “PhenoCrop: An integrated satellite-based framework to estimate physiological growth stages of corn and soybeans,” the group of experts describes the PhenoCrop framework, implement the method within the state of Nebraska for 2012-2016, and test its performance in three experimental sites in eastern Nebraska. They also examined the spatial and temporal patterns of predicted growth stages for corn and soybean and validated these estimates using the experimental site data and state-level data from the United States Department of Agriculture’s National Agricultural Statistical Services (USDA NASS) Crop Progress Reports. Finally, in an effort to detail the variations using coarse and high spatial resolution satellite imagery, they compared data with estimates produced from MODIS 250m reflectance data.
In line with previous studies, findings demonstrated that the satellite data used in this research was able to provide insight into the onset dates of crop growth stages, reflecting both spatial and temporal differences in crop growth stage progressions. However, the authors note that when the data was downscaled, accuracy was improved and the data better captured small variations at the field scale and also highlight the fact that fused MODIS-Landsat data may be best for estimating individual crop growth stages in heterogeneous agricultural landscapes. Overall, the PhenoCrop framework showed excellent potential for enabling accurate estimates of various crop growth stages of both corn and soybean. This framework will be useful for improving crop characterization applied in crop modeling, thereby improving the accuracy of regional-scale crop condition simulations including yields and carbon fluxes. Additionally, using seasonal and sub-seasonal weather forecasts, the PhenoCrop method can provide information with sufficient lead time that can assist farmers in scheduling their crop management practices within a given season.
Read the full publication for more on the PhenoCrop framework and how it can be applied to in-season farming methods.