Australia-Victoria

Overview

Our ultimate goal is to build the digital infrastructure necessary for the Australian grains industry to map, monitor, manage and forecast production across Australia. We rely on climate data, crop models, and optical and SAR satellite data which we combine with various sources of field data. We use the JECAM site in Victoria to test and benchmark our methods before porting them to regional and national scales.

Project Objectives

Mapping Agricultural Areas

Operational Implementation Plan

We are testing several machine learning methods to classify crops using optical images and SAR data. Last season, a particular point of interest was the identication of  We have been ramping up to apply our local and regional crop type mapping algorithms to national scale. While we have the capabilities to process the satellite imagery, representative field data is the main drawback preventing us to achieve it. We have also led and brought to completion the second cross site experiment on the effect on roadside sampling for cropland mapping.

Scale

JECAM site/regions/states

Forecasting Agricultural Variables

Operational Implementation Plan

We have developed a series of satellite-driven methods for yield estimation based on MODIS which can be applied in-season to forecast grain yield. To complement this approach, we also rely on crop modelling to provide yield estimates early in the season. Future research will include machine learning of crop yield for national scale assessment.

Agricultural Variables (large scale)
  • Yield
  • Crop rotation

Site Description

Landscape TopographyGently undulating landscape
Typical Field SizePaddocks are often as large as 250 ha.
Climatic ZoneThe climate is temperate with winter rainfall and dry summers. The average growing season rainfall (April to October) is 250mm.
Major Crops and Calendars

Other leguminous crops (Normal):
Calendar: March - December
Typical Rotation: There are at least three common rotations: 1. Wheat barley pulse (lentils or faba beans) 2. Wheat canola pulse 3. Wheat barley pasture canola

Grassland and Meadows (Early,Normal,Late):
Calendar: April - December
Typical Rotation: There are at least three common rotations: 1. Wheat barley pulse (lentils or faba beans) 2. Wheat canola pulse 3. Wheat barley pasture canola

Lentils (Normal):
Calendar: April - December
Typical Rotation: There are at least three common rotations: 1. Wheat barley pulse (lentils or faba beans) 2. Wheat canola pulse 3. Wheat barley pasture canola

Canola (Normal):
Calendar: April - December
Typical Rotation: There are at least three common rotations: 1. Wheat barley pulse (lentils or faba beans) 2. Wheat canola pulse 3. Wheat barley pasture canola

Oats (Normal):
Calendar: April - December
Typical Rotation: There are at least three common rotations: 1. Wheat barley pulse (lentils or faba beans) 2. Wheat canola pulse 3. Wheat barley pasture canola

Winter Barley (Early,Normal,Late):
Calendar: April - December
Typical Rotation: There are at least three common rotations: 1. Wheat barley pulse (lentils or faba beans) 2. Wheat canola pulse 3. Wheat barley pasture canola

Winter Wheat (Early,Normal,Late):
Calendar: April - December
Typical Rotation: There are at least three common rotations: 1. Wheat barley pulse (lentils or faba beans) 2. Wheat canola pulse 3. Wheat barley pasture canola

Soil Type & Texture

Organic:

  • Humic

Inorganic:

  • Loamy Sand
  • Sandy Loam
  • Clay
Soil Drainage Class["Well drained","Imperfectly drained"]
Irrigation Infrastructure["Center pivot irrigation","Not Frequent"]
Other Site Details

Mainly rainfed, mechanized and intensive cropping systems.

In Situ Observations

Crop type

  • Crop Type(s): Winter Wheat, Winter Barley, Oats, Canola, Lupins, Lentils, Chick peas
  • Collection Protocol:

    Roadside sampling on main and secondary roads. Surveys are conducted once in the season  usually flowering. We also source some crop labels from growers and consultants. Paddock boundaries are manually digitized using very high resolution and in-season Sentinel-2 images.

  • Frequency: Once per season, around flowering

Yield

  • Crop Type(s): Winter Wheat, Canola, Winter Barley
  • Collection Protocol:

    Ad hoc collection of yield maps from growers or consultants.

  • Frequency: seasonally

EO Data

Optical Data Requirements

  • Approximate Start Date of Acquisition: 01/04
  • Approximate End Date of Acquisition: 31/12
  • Spatial resolution: Medium resolution (20-60m)High resolution (5-20m)
  • temporal_frequency: WeeklyBiweekly
  • Level of Expertise: IntermediateAdvance
  • Latency of Data Delivery: 2-5 days
  • Challanges: Data storage and processing

    • SAR Data Requirements

  • Approximate Start Date of Acquisition: 01/04
  • Approximate End Date of Acquisition: 31/12
  • Spatial resolution: Medium resolution (20-60m)
  • temporal_frequency: Biweekly
  • Frequency: C
  • Incidence Angle: Steep
  • Polarisation:
  • Level of Processing: Level 2Level 3
  • Orbit: Any
  • Level of Expertise: IntermediateAdvance
  • Latency of Data Delivery: 2-5 days
  • Challanges: Processing

    • Passive Microwave Data Requirements

    Thermal Data Requirements

    Results

    Documents and Files

    Links to paper

    Project Reports

    Study Team

    Team Leader

  • Name: Franz Waldner
  • Affiliation: CSRIO Agriculture & Food
  • Affiliation Webpage: https://www.csiro.au/
  • Position: PostDoc
  • Email: franz.waldner@csiro.au
  • Personal Webpage: https://people.csiro.au/w/f/franz-waldner
  • Phone number:
  • Postal Address:

    • Other Team Members

  • Name: Zvi Hochman
  • Affiliation: CSIRO Agriculture & Food
  • Affiliation Webpage: https://www.csiro.au/
  • Position: Senior Principal Research Scientist, Agricultural Systems
  • Email: zvi.hochman@csiro.au
  • Personal Webpage: https://people.csiro.au/h/z/zvi-hochman
  • Role: Co-Lead