Abstract

Accurate estimates of tree crop orchard age and historical crop area are important to develop yield prediction algorithms, and facilitate improving accuracy in ongoing crop forecasts. This is particularly relevant for the increasingly productive macadamia industry in Australia, where knowledge of tree age, as well as total planted area, are important predictors of productivity, and the area devoted to macadamia orchards is rapidly increasing. We developed a technique to aggregate more than 30 years of historical imagery, generate summary tables from the data, and search multiple combinations of parameters to find the most accurate planting year prediction algorithm. This made use of known planting dates of more than 90 macadamia blocks spread across multiple growing regions. The selected algorithm achieved a planting year mean absolute error of 1.7 years. The algorithm was then applied to all macadamia features in east Australia, as defined in an recent Australian tree crops map, to determine the area planted per year and the total cumulative area of macadamia orchards in Australia. The area estimates were refined by improving the resolution of the mapped macadamia features, by removing non-productive areas based on an optimal vegetation index threshold.

Highlights

  • The macadamia nut is indigenous to Australia, but is grown in many places including Hawaii, Central America, and South Africa [1]

  • There is no significant dip in the normalized difference vegetation index (NDVI) around the planting year, which appeared common for the Macksville region

  • This led us to reject the use of algorithms which search for dips in Normalized Difference Spectral Indices (NDSIs) time series, and to instead find an optimal algorithm that defines an NDSI threshold corresponding to the average NDSI of trees of a particular age

Read more

Summary

Introduction

The macadamia nut is indigenous to Australia, but is grown in many places including Hawaii, Central America, and South Africa [1]. It accounts for ~2% of global tree nut consumption. Crop yield prediction is important for growers and industry, as it guides marketing, finance and logistic decisions. Region-scale predictions rely on accurate estimates of crop productive area [2]. Predicting the yield of tree crops requires accurate estimates of tree age. Other applications for historical planted area data for tree crops are to facilitate forecasts of future productivity and water resource usage at a regional or national level [4,5]

Objectives
Methods
Results
Discussion
Conclusion
Full Text
Published version (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