Abstract
Machine-learning algorithms used for modelling olive-tree phenology generally and largely rely on temperature data. In this study, we developed a prediction model on the basis of climate data and geophysical information. Remote measurements of weather conditions, terrain slope, and surface spectral reflectance were considered for this purpose. The accuracy of the temperature data worsened when replacing weather-station measurements with remote-sensing records, though the addition of more complete environmental data resulted in an efficient prediction model of olive-tree phenology. Filtering and embedded feature-selection techniques were employed to analyze the impact of variables on olive-tree phenology prediction, facilitating the inclusion of measurable information in decision support frameworks for the sustainable management of olive-tree systems.
Highlights
Introduction for Predicting Olive PhenologyMonitoring the timing of phenological events is key to understanding the effects of climate change on agroecosystems [1], and scheduling appropriate control strategies.Phenological information can be derived through the direct ground observation of crop development, though this is a time-consuming and labor-intensive process [2]
The use of multitemporal data has proven useful for monitoring and characterizing spatial and temporal patterns of crop cover changes, with improved accuracy compared to that of a single date [8]. Both normalized difference vegetation index (NDVI) and enhanced vegetation index (EVI) are important indicators to estimate the start of the growing season, SOS [9,10,11], or other phenological metrics [10]
Olive-tree phenology observation data were obtained from the monitoring network established in Tuscany (Italy), consisting of 21, 19, and 18 olive-tree orchards in 2008, 2009, and 2010, respectively
Summary
Introduction for Predicting Olive PhenologyMonitoring the timing of phenological events is key to understanding the effects of climate change on agroecosystems [1], and scheduling appropriate control strategies.Phenological information can be derived through the direct ground observation of crop development, though this is a time-consuming and labor-intensive process [2]. Satellite time series can be conveniently used to determine vegetation indices and monitor phenological phases on the regional level [3] Vegetation indices, such as the normalized difference vegetation index (NDVI) and the enhanced vegetation index (EVI), were used to infer crop growth, and changes in vegetation dynamics and land cover [4,5]. The use of multitemporal data has proven useful for monitoring and characterizing spatial and temporal patterns of crop cover changes, with improved accuracy compared to that of a single date [8] Both NDVI and EVI are important indicators to estimate the start of the growing season, SOS [9,10,11], or other phenological metrics [10]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
