Forecasting grape yield with high spatial resolution optical remote sensing

Research output: Book chapter/Published conference paperConference paperpeer-review

15 Downloads (Pure)

Abstract

Quantitative descriptors of 52 Cabernet Sauvignon grapevine canopies from a single vineyard block in the Riverina (NSW, Australia) were derived from high spatial resolution (0.25 m) vegetation index images acquired on nine occasions over two growing seasons. Two canopy descriptors were derived from theimagery: one to describe grapevine canopy area and the other vegetation density. The descriptors were compared to yield data collected from the same vines at harvest in both years. The strength and type of the relationships varied with phenological stage. Significant (p < 0.05) correlations between canopy densityand yield collected in the same year were observed, which were negative for descriptors derived from imagery collected pre-veraison then positive for descriptors derived from imagery collected at veraison. In addition there was a strong negative correlation between canopy density from imagery acquired at flowering in year 1 and yield in year 2. The relative change in the area and density of the grapevine canopy between adjacent imaging missions was also calculated using the remotely sensed canopy descriptors. The relationship between the relative change in canopy density between adjacent imaging missions during year 1 and yield in year 2 produced the greatest magnitude correlation in the study. A best subsets regression analysis of the remotely sensed canopy descriptors produced a three variable model to estimate year 2 yield data with an adjusted-R2 of 0.69, suggesting forecasting grape yield from variability in vine canopies detected by high spatial resolution optical remote sensing is achievable.
Original languageEnglish
Title of host publicationFrutic Chile 2009
Subtitle of host publication8th Fruit, Nut and Vegetable Production Engineering Symposium
EditorsStanley Best
Place of PublicationChillan, Chile
PublisherINIA
Pages33-40
Number of pages8
Publication statusPublished - 2009
EventFruit, Nut and Vegetable Production Engineering Symposium - Concepcion, Chile
Duration: 05 Jan 200909 Jan 2009

Conference

ConferenceFruit, Nut and Vegetable Production Engineering Symposium
Period05/01/0909/01/09

Fingerprint

Dive into the research topics of 'Forecasting grape yield with high spatial resolution optical remote sensing'. Together they form a unique fingerprint.

Cite this