Abstract
Crop extent and frequency maps are an important input to inform the debate around land value and competitive land uses, in particular between cropping and mining in the case of Queensland, Australia. Such spatial datasets are useful for supporting decisions on natural resource management, planning and policy. For the major broadacre cropping regions of Queensland, Australia, the complete Landsat Time Series (LTS) archive from 1987 to 2015 was used in a multi-temporal mapping approach, where spatial, spectral and temporal information were combined in multiple crop-modelling steps, supported by training data sampled across space and time for the classes Crop and No-Crop. Temporal information within summer and winter growing seasons were summarised for each year, and combined with various vegetation indices and band ratios computed from a pixel-based mid-season spectral synthetic image. All available temporal information was spatially aggregated to the scale of image segments in the mid-season synthetic image for each growing season and used to train a number of different predictive models for a Crop and No-Crop classification. Validation revealed that the predictive accuracy varied by growing season and region and a random forest classifier performed best, with κ = 0.88 to 0.91 for the summer growing season and κ = 0.91 to 0.97 for the winter growing season, and are thus suitable for mapping current and historic cropping activity.
Highlights
Agricultural production and productivity have important economic and environmental implications, most notably for present and future food security
We demonstrate the use of the full Landsat archive to dynamically create maps that describe cropping activity for each growing season and year in the archive, with the objective of being able to provide the information in an ongoing operational program, rather than as one-off static maps
The study area (Figure 1) is the Western Cropping Zone (WCZ) within the State of Queensland, Australia, which is based on a strategic cropping land map by the Queensland Department of Natural
Summary
Agricultural production and productivity have important economic and environmental implications, most notably for present and future food security. These implications are magnified by a steadily increasing global population, a changing climate and, with that, a greater frequency of extreme weather events. Urban expansion and mining interests, along with grazing and forestry, are the most common types of land uses that compete with cropping [6]. This competition may in the future be aggravated by an increasing demand for biofuel production [7]. Managing land resources to support the needs of the population is a mandate common to all levels of government
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
