Wool producers with remote control: new tools for whole of property management

Summary

Pasture management in the pastoral zone is crucial to long-term sustainability of woolgrowing enterprises, but the large size of properties and scarcity of labour make it difficult for pastoralists to know the condition of the pasture base across the whole property. As well, like all agricultural producers, pastoralists are under increasing scrutiny to ‘prove’ whether they are managing their natural resources well. The ability of satellite imagery to provide frequent and whole-of-property information that can be easily archived over many years means that satellites are potentially useful for pastoral management. However, to date, impediments in processing and interpreting satellite imagery have rendered it beyond the reach of most pastoralists.

Gary Bastin from CSIRO Sustainable Ecosystems in Alice Springs, led a team who worked with eight pastoral families in northern South Australia to devise a prototype for presenting satellite information in a useful format. Three products in particular seem to show the most potential, namely:

(a) maps of cover at various times for the property (or particular paddocks) scaled between the historically lowest and highest values of cover for the area of interest;

(b) maps of vegetation cover summarised as time traces of average cover across land types within paddocks; and

(c) cover values showing changes in categories (such as low, medium, high) over time hence allowing woolgrowers to manage against preset cover targets.

The project team would welcome enquiries from potential collaborators who are interested in taking the technology forward.

Aims

The project had two main aims. Firstly, Gary and his team (Dr Craig James and Vanessa Chewings from CSIRO and Amanda Brook from the South Australian Department of Water, Land & Biodiversity Conservation) aimed to work with the participating families both in workshops and on their individual stations, to get some insight into the way pastoralists make their decisions and to determine the manner in which satellite imagery might be usefully presented to the families. The second aim was to develop a prototype means of processing and displaying satellite information in ways helpful to pastoralists in making management decisions. The team used all satellite resources at its disposal but particularly focussed on applying the MODIS technology because it shows the most potential as a basis for a satellite-based information service.

The project area covered nine stations (eight families) in the north east of South Australia. The stations cover an area of 8,213 km2, are a mix of family and company-owned, and run predominantly merinos with cattle also present on some stations. The country varies from native pine and black oak woodlands on foothills of the Flinders Ranges in the far west to predominantly chenopod shrublands with areas of mixed acacia shrublands and alluvial floodouts further east.

Outcomes

During initial workshops with woolgrowers in late March 2005 to introduce the project, participants were sceptical that they would use information derived from satellite data to make decisions about their current grazing management practices (e.g., number of sheep in each paddock and the time over which these flock sizes were maintained in the paddock). However all producers saw value in satellite-derived maps of cover as a way of complementing their current methods used to monitor their land and vegetation. Most woolgrowers also said that they would appreciate current information being placed against historical context (“knowing where I am now compared to where I have come from”). The expanding archive of satellite imagery (dating back more than 30 years) means that this request was achievable. Encouragingly, all woolgrower families were keen to participate in the project. They requested that that the project used the highest spatial resolution MODIS imagery available (~250-m pixel size), provide as much historical context as possible and in some cases, woolgrowers nominated focus paddocks for particular investigation. A more complete list of woolgrower views on useful applications appears on p. 14 of Gary’s technical report (see the Want to know more? section below).

Subsequently, the project team met with individual woolgrower families in September, October and December 2005, where they provided historical Landsat imagery and freeware image-viewing software. The primary purpose of the station visits was to tailor images of cover (based on the “visible red” band of MODIS imagery) to the particular interests and requirements of each family. The team provided updates of vegetation cover to each family (December 2005 and March 2006) based on the archive of MODIS imagery (113 image dates from late 2000 to 17th January 2006). The MODIS images used were NASA 16-day cloud-free composites.

The “visible red” band of MODIS imagery was scaled to an index of fractional cover for each station (or focus paddock) where zero equalled the lowest cover in the area of interest and one represented the maximum cover (excluding areas of dense timber cover). This index reasonably indicates cover regardless of whether the vegetation is green (actively growing) or “hayed off” (dry). (It is possible for satellite images to provide information on the ‘greenness’ of plant cover and this is relevant on large stations for better judging the growth response following large or scattered storms. However ‘greenness’ is less relevant for managing feed supply in the pastoral zone where pasture is generally hayed-off for most of the year).

The index represents spatial differences in cover reasonably well but is affected by the presence of moisture, either on the soil surface or in the atmosphere, which cause falsely high cover values most frequently in winter. Gary recommends that for now, these anomalous images be removed from the time sequence of cover images. As well, the technology has some limitations, including inability to distinguish perennial plants from annuals (though inferences can be made about this characteristic depending on the time of year the image is taken).

From experiences gained through working with woolgrowers during station visits and during workshops at the end of the project, Gary and his team gleaned three product-types that are probably useful for assisting cover management. These are:

Maps of cover at various times for the property (or particular paddocks) scaled between the historically lowest and highest values of cover for the paddock of interest, the property or the region (in this case, the project area). The first two types maximise the information for individual managers while the third option is more useful for showing how management of cover on a property is tracking relative to the region (i.e., a form of regional benchmarking).
Maps of vegetation cover can be summarised as time traces of average cover across land types within paddocks. These time traces can be viewed either as a continuous record for the duration of MODIS imagery or as a year-on-year record.
Cover values can be presented to show changes in categories (such as low, medium, high) over time hence allowing woolgrowers to manage against preset cover targets.
The Project team have developed the technology to a prototype stage. They have prototype presentation layouts and processes as well as a wealth of knowledge about woolgrower needs and would welcome enquiries from potential collaborators who are interested in taking the technology forward.

Background

The pastoral zone is characterised by huge properties and scarcity of labour. Consequently, it is difficult for pastoralists to see all parts of the property quickly. Therefore, there is the potential for mismatching stocking rates to available fodder, both in terms of overgrazing and missing opportunities where forage is abundant. Pasture management in the pastoral zone is perhaps more critical than in higher rainfall country, in so far as there are less opportunity to ameliorate damage to pastures from overgrazing. Potentially, poor decisions in the pastoral zone can destroy the pasture base almost indefinitely.

Techniques that allow pastoralists to ‘see’ their whole property, frequently and close to real time have potential to improve their ability to better match stock numbers to available feed. Satellite imagery has the potential to offer this ‘bird’s eye view’ at frequent intervals close to real time. In addition to stocking rate and environmental management considerations, satellite imagery has the potential to corroborate a pastoralist’s claim to good management because the imagery provides a valuable and independent historical record. This may become more important as governments and markets increase their scrutiny of the way producers manage the environment.

Satellite imagery is not new. The Landsat satellite has been providing imagery for over 30 years. However, until recently, only government agencies and specialists have had ready access and ability to interpret satellite data. The MODIS satellite has been providing imagery since 2000 – not as long as Landsat - but the MODIS imagery has the advantage of being available at practically no charge (in its raw form) and provides images at least as frequently as Landsat. However, its resolution is not as fine as Landsat or some other systems (eg: Ikonos and Quickbird).

However, even MODIS imagery requires processing and interpretation before it can provide meaningful information that might be applied to paddock scale management. For the non-specialist user, relatively simple and robust tools (mainly software) will be needed in this regard.

Publications and Resources

None listed

Citation

Land & Water Australia. 2007. Wool producers with remote control: new tools for whole of property management . [Online] (Updated June 10th, 2008)
Available at: http://lwa.gov.au/node/4 [Accessed Wednesday 29th of February 2012 08:20:56 PM ].