PRELIMINARY CONCERNS ABOUT AGRONOMIC INTERPRETATION OF NDVI TIME SERIES FROM SENTINEL-2 DATA: PHENOLOGY AND THERMAL EFFICIENCY OF WINTER WHEAT IN PIEMONTE (NW ITALY)

Abstract. TELECER project is supported through Rural Development Programme regional action of EU CAP and is aimed at providing Precision Agriculture–devoted services for cereals monitoring in the Piemonte Region (NW-Italy) context. In this work authors explored some general and preliminary issues mainly aimed at demonstrating and formalizing those evident relationships existing between NDVI image time series and the main ordinary agronomic parameters, with special focus on phenology and thermal efficiency of crops as related to Growing Degrees Day (GDD). Winter wheat was investigated and relationships calibrated at field level, making possible to spatially characterise environmental and management effects. Two different analysis were achieved: (i) one aimed at mapping crop phenological metrics, as derivable from NDVI S2 time series; (ii) one aimed at locally modelling relationship linking GDD and NDVI to somehow test the thermal efficiency of crops in the different parts of the study area. The first analysis showed that the end of season appears to be the most constant phenological metric in the study area possibly demonstrating a time concentration of harvest operations in the area. Differently, the peak of season and the start of season metrics showed to be largely varying in the study, thus suggesting to be stronger predictors: (i) of crop development; (ii) of the effects induced by local agronomical practices. Several base temperatures were used to compute correspondent GDD. These were tested against NDVI and modelled by a parabolic model at field level. Model coefficients distribution were analysed and mapped the correspondent agronomic interpretation suggested.