Intermountain GIS 2019

High temporal and spatial resolution satellite imagery provides a valuable resource to support the decision-making process in agricultural fields. It allows the monitoring of crops throughout the growing season as well as helping to predict crop yield. This research utilizes Planet and Sentinel-2A datasets to study potato fields in Lebanon and Idaho to build a yield prediction model. Spectral indices including NDVI, SAVI and NDWI show the crops’ response to changing weather conditions and indicate critical times where extra irrigation or nutrients are needed. Though the Planet data has a higher spatial resolution of 3 m compared to 10 m with Sentinel-2A, it is the daily temporal resolution that plays the major role in showing detailed information and changes during the growing season. Planet data with daily revisit provided a very rich and dense dataset and can compensate for cloudy days while Sentinel-2A has considerably fewer data points as it has a revisit time of 10 days. With Planet’s high temporal resolution, we can see critical changes that correspond to the weather conditions and growth stages of the potato plants. The spectral indices allowed us to determine the critical week for yield prediction for Russet and Norkotah varieties. Based on the critical growth stage for crops identified in Idaho, SAVI and slope predicted yield values in Lebanon. Despite the offset in the growing seasons between Idaho and Lebanon, the yield model was successful.